Влияние сенсорной среды на исходы состояния здоровья госпитализированных пациентов
Abstract
Background
Hospital environments have recently received renewed interest, with considerable investments into building and renovating healthcare estates. Understanding the effectiveness of environmental interventions is important for resource utilisation and providing quality care.
Objectives
To assess the effect of hospital environments on adult patient health‐related outcomes.
Search methods
We searched: the Cochrane Central Register of Controlled Trials (last searched January 2006); MEDLINE (1902 to December 2006); EMBASE (January 1980 to February 2006); 14 other databases covering health, psychology, and the built environment; reference lists; and organisation websites. This review is currently being updated (MEDLINE last search October 2010), see Studies awaiting classification.
Selection criteria
Randomised and non‐randomised controlled trials, controlled before‐and‐after studies, and interrupted times series of environmental interventions in adult hospital patients reporting health‐related outcomes.
Data collection and analysis
Two review authors independently undertook data extraction and 'Risk of bias' assessment. We contacted authors to obtain missing information. For continuous variables, we calculated a mean difference (MD) or standardized mean difference (SMD), and 95% confidence intervals (CI) for each study. For dichotomous variables, we calculated a risk ratio (RR) with 95% confidence intervals (95% CI). When appropriate, we used a random‐effects model of meta‐analysis. Heterogeneity was explored qualitatively and quantitatively based on risk of bias, case mix, hospital visit characteristics, and country of study.
Main results
Overall, 102 studies have been included in this review. Interventions explored were: 'positive distracters', to include aromas (two studies), audiovisual distractions (five studies), decoration (one study), and music (85 studies); interventions to reduce environmental stressors through physical changes, to include air quality (three studies), bedroom type (one study), flooring (two studies), furniture and furnishings (one study), lighting (one study), and temperature (one study); and multifaceted interventions (two studies). We did not find any studies meeting the inclusion criteria to evaluate: art, access to nature for example, through hospital gardens, atriums, flowers, and plants, ceilings, interventions to reduce hospital noise, patient controls, technologies, way‐finding aids, or the provision of windows. Overall, it appears that music may improve patient‐reported outcomes such as anxiety; however, the benefit for physiological outcomes, and medication consumption has less support. There are few studies to support or refute the implementation of physical changes, and except for air quality, the included studies demonstrated that physical changes to the hospital environment at least did no harm.
Authors' conclusions
Music may improve patient‐reported outcomes in certain circumstances, so support for this relatively inexpensive intervention may be justified. For some environmental interventions, well designed research studies have yet to take place.
PICOs
Резюме на простом языке
Влияние сенсорной среды на исходы состояния здоровья госпитализированных пациентов
Больничная среда (звуки, картины, ароматы, дизайн, качество воздуха, мебель, архитектура и планировка), может оказать влияние на здоровье пациентов, находящихся в больнице. Целью этого обзора является обобщение лучших доступных доказательств по больничной среде, для того, чтобы помочь людям, участвующим в разработке больничной среды, принимать решения, которые будут полезными для здоровья пациентов.
Обзор выявил 102 соответствующих исследования, 85 из которых были по использованию музыки в больнице. Другими рассмотренными аспектами среды были: ароматы (два исследования), аудиовизуальные отвлечения (пять исследований), декорирование (одно исследование), качество воздуха (три исследования), тип палаты (одно исследование), полы (два исследования), мебель и меблировка (одно исследование), освещение (одно исследование), температура (одно исследование), а также различные изменения в дизайне (два исследования). Ни одно из исследований, отвечавших критериям включения, не оценивало: искусство, доступ к природе, например, больничные сады, атриумы, цветы и растения, потолки, вмешательства по сокращению больничного шума, способы контроля пациента, технологии, помощь в нахождении пути в больнице, или наличие окон.
В целом, кажется, что музыка в больнице может помочь улучшить исходы по сообщениям пациентов, такие как тревога; однако, существует меньше доказательств в поддержку использования музыки для влияния на физиологические исходы (такие как, снижение частоты сердечных сокращений и кровяного давления) и для уменьшения использования лекарств. Что касается других аспектов больничной среды, нет большого числа хорошо спланированных исследований, которые помогут с принятием дизайнерских решений на основе доказательств. Исследования, которые были включены в этот обзор показывают, что физические изменения, предпринятые для «улучшения» больничной среды, в целом не приносят вреда.
Authors' conclusions
Background
The International Academy for Design and Health held its 7th World Congress in 2011, highlighting the growing multinational interest in environmental design that promotes health. A reform on a global scale is underway to make hospitals much more than places to go and receive treatment; hospitals are now being conceptualised as places that have the potential to positively impact health through being restorative, healing environments (Dilani 2001). The World Health Organization (WHO) highlights the important role of the environment in the health process, and calls for action towards creating supportive environments ('Ottowa' 2004). In accordance with this movement, international organisations such as The Centre for Health Design (CHD) have been established ('CHD' 2009). CHD's mission is 'to transform healthcare environments for a healthier, safer world through design research, education and advocacy'.
Other initiatives have been stimulated by the US‐based Society for the Arts in Healthcare ('SAH' 2009), which attempts to advance arts as an integral component in health care, and the UK‐based Medical Architecture Research Unit (Etheridge 2008), which provides research, consultancy, and training, with a vision to "explore the interface between health service organisational culture and the built environment response". Research into, and implementation of 'healing' healthcare environments is also being carried out in Japan (Cooper 2002; Takayanagi 2004), and across Europe (Pelikan 2001).
In the Middle East, various major investment projects have been supporting a rapid growth in the hospital sector (News 2004) and the hospital industry in the United States has been going through a major building boom with billions more dollars being spent on replacing or renovating old facilities (Babwin 2002). In the United Kingdom, the Private Finance Initiative (PFI) has prompted a renewed interest in hospital design, investing billions of pounds into the biggest new hospital building programme in the history of the NHS (Milburn 2001). Investments such as these have provided an opportunity for hospitals to be considered as 'therapeutic environments' (Gesler 2004), spurring on initiatives such as the UK‐based scheme in which the Kings Fund and the Department of Health offer grants to health authorities to enhance the environment ('Kings Fund' 2009).
Some argue that this expenditure is a waste of resources (Lipley 2001), whilst others have highlighted the lack of 'evidence‐based practice' when it comes to hospital design (Frumkin 2003). Schemes set up to enhance the hospital environment are sometimes not evaluated on the grounds that it is too logistically complex (Comer 1982), or by simply stating that the effects are obvious (Parker 2000). There is some empirical evidence in support of creating better environments in care facilities however, and researchers are finding that changes to the physical environment can positively influence patient outcomes (Devlin 2003; Rubin 1996). In an invitation‐only conference entitled 'Designing the 21st Century Hospital: serving patients and staff', held by the Robert Wood Johnson Foundation ('Webcast' 2004), Craig Zimring and Roger Ulrich referred to a literature review they had conducted, finding over 600 studies on the effects of the hospital environment on patients, families and staff (Ulrich 2004a, later updated: Ulrich 2008). In their presentation of the research, Zimring and Ulrich emphasised the evidence mainly in terms of quantity and consistency of findings. Following the presentations, delegates discussed how worthwhile applying quality criteria to the current evidence would be, such as in a Cochrane systematic review.
Description of the intervention
The environment can be altered in endless ways. Some environmental changes may be detrimental to health (Thompson 2000), further exemplifying the need for using evidence‐based design. As a framework for assessing the evidence on hospital design, the literature could be considered in terms of whether it regards: (1) aspects that are added to the environment as positive distracters (such as art and music); and (2) aspects of the environment that are changed to reduce stressors (such as light, noise or air quality).
In this review, we are interested in elements of the sensory environment; that is, aspects of the hospital surroundings that can be seen, touched, smelt, or heard (such as the building design and layout, decor, furniture and furnishings, air quality and aromas/odours, and noises/sounds). In this review we utilise the phrase 'sensory environment' to define the hospital characteristics under study; in the literature, other terms are utilised such as 'healing environments', 'supportive environments', and 'health‐promoting environments'. We have opted for a less directive phrase as we are interested in determining which environmental factors have positive, negative, and neutral effects on the health of individuals.
How the intervention might work
Sensory environments have been advocated on the basis of their perceived ability to reduce anxiety, lower blood pressure, improve postoperative outcomes, reduce the need for pain medication, and shorten the hospital stay (Ulrich 1992); good design has also been implicated to improve quality of sleep (Hewitt 2002; Marberry 2002), reduce hospital acquired infections (Ulrich 2004b), and improve staff retention and well‐being (Neuberger 2003; Marberry 2002; Gross 1998). Theories underpinning these effects are wide‐ranging and stem from a variety of perspectives, for example: Attention Restorative Theory (a 'functionalist‐evolutionary' model; Kaplan 1989); a 'psycho‐evolutionary' model (Ulrich 1983) and the Biophilia Hypothesis (Ulrich 1993a); Henry's model of neuro‐endocrine responses (Parsons 1991, which incorporates the ideas of the 'Fight or Flight' response to environmental stressors and Selye's model of stress and disease, 'General Adaptation Syndrome'); the Intake‐Reject Hypothesis (Lacey 1974), a controversial hypothesis which has implications for distraction therapies; the 'Gate Control Theory' (Melzack 1965) and the 'Neuromatrix Theory' of pain (Melzack 1999); and the Broaden‐and‐Build model (Fredrickson 2000), which offers a premise for creating environments which help cultivate positive emotions. We will not go into a full explanation and debate of all these theories here, but suffice is to say that although there are some disparities between these explanatory models, and some sit controversially within their fields, they complement each other on the general principles that removing environmental stressors, and using the environment to calm, distract, and elicit positive emotions may have positive implications for health.
Why it is important to do this review
Clearly this is a broad and complex area of study, with the 'environment' being considered as an intervention influencing health‐related outcomes; nevertheless, it is imperative that the evidence‐base for sensory environments is assessed systematically, to ensure that patients are provided with the best possible opportunity to recover and that the system remains cost‐effective.
Objectives
To assess the effect of the sensory hospital environment on adult patient health‐related outcomes.
Methods
Criteria for considering studies for this review
Types of studies
In this area of research it may often be very difficult to conduct a Randomised Controlled Trial (RCT) due to the nature of the intervention and logistical complexities. Therefore, this review included a variety of study designs: RCTs, (non‐randomised) Controlled Clinical Trials (CCTs), Controlled Before and After studies (CBAs), and Interrupted Time Series (ITS). To be included in the review, a CBA had to have two intervention sites and two concurrent control sites with outcomes measured both before and after the intervention was implemented. An ITS had to have at least three data collection points before the intervention, and at least a further three data collection points after the intervention. All studies must have been conducted prospectively.
Types of participants
The review included adults attending hospital as in‐patients, day hospital or out‐patients. Studies were included if over 90% of the participants were over 18. Both elective and non‐elective patients were included in the review. We have included all diagnoses including psychiatric patients (Gross 1998).
Where possible, studies were characterised by the type of hospital visit, where 90% of participants was considered as a cut‐off point for inclusion in a specific category. In‐patients were classed as those that required a hospital bed and required an overnight stay for tests or surgery; day‐patients were patients that required a hospital bed for specialised observation or health care for a limited number of hours of the day, but did not need to stay overnight; out‐patients were people who were referred to see a hospital consultant for a specialist opinion or examination and did not require a hospital bed.
Types of interventions
The review incorporated studies that investigated any aspect(s) of the sensory environment. Interventions were those that altered the environment by one or a mixture of the following ways.
(1) Providing positive distracters to complement the treatment already being administered. Positive distracters are elements of the sensory environment; they do not include therapies (such as bright light therapy), which are received instead of orthodox treatments. Patients could be offered a choice of distraction but we did not include instances when patients were actively involved in creating a distraction (e.g. creating a work of art). Positive distracters included:
-
aromas/scents (different aromas/patient choice of aroma versus none);
-
viewing artwork (comparing different styles/patient choice of art versus no art);
-
viewing performance art (versus none);
-
audiovisual distractions, such as television/video (absent versus present/differences between content/patient choice of content);
-
decoration (colour of walls etc.);
-
music (versus no music, different styles of music, or other environmental comparison);
-
access to nature, for example via atriums, gardens, window views, or indoor planting (versus no access to nature, other views or urban retreats).
(2) Reducing environmental stressors by implementing physical changes. We have not included changes that are made to policy (e.g. ensuring multi‐bed rooms are unisex). Physical changes to the sensory environment include:
-
noise‐reducing aids (e.g. sound‐absorbing ceiling tiles versus regular);
-
way‐finding aids (e.g. colour‐zoned areas, landmarks);
-
patient controls (e.g. access to lighting and ventilation controls);
-
lighting (e.g. natural versus fluorescent);
-
people/privacy (e.g. open versus closed wards, decentralised nurse stations).
(3) Multi‐faceted interventions:
Some studies manipulate many variables and as such span across the above three broad categories of environmental interventions (such as when a whole ward is redesigned; Leather 2003). We have included these studies in the review provided they were not confounded with non‐environmental changes, such as changes to policy.
We excluded studies from the review if the intervention was not clearly defined (to the extent that it could be replicated). In order to meet this requirement, studies must have, where applicable, either:
-
provided the manufacturing details of the intervention being assessed, if appropriate;
-
provided pictures or diagrams, if appropriate;
-
provided a detailed description of the objective properties of the intervention, (e.g. an intervention of colour change needs to describe the specific hue; simply stating 'blue' is not specific enough);
-
provided a contact from which more detailed information could be sought.
All studies were carried out in a hospital setting. A hospital was defined (Ward 2008) as a health facility that:
-
provides communal care where there is an expectation that this care is time limited;
-
provides overnight accommodation;
-
provides nursing and personal care;
-
provides for people with illness and disability.
This definition includes hospices.
Studies may have been conducted in any area of the hospital grounds, such as general wards, specialist wards (e.g. Intensive Care Units), waiting rooms, common areas, and gardens.
Types of outcome measures
This review included all validated health‐related patient outcomes reported in the research. Outcomes of interest included validated measures of: anxiety; pain; length of hospital stay; patients' satisfaction; quality of sleep; aggression and mood; physiological outcomes; medication utilisation; hospital‐acquired infections; and mortality. We included a broad range of outcomes since the environment may affect many aspects of a patient's physical and psychological health and different interventions may be applicable to some outcomes and not others. When summarising the results of studies, we have reported up to five relevant outcomes for each comparison and grouped the remaining reported outcomes for that intervention under a heading "other outcomes".
Search methods for identification of studies
Electronic searches
We searched databases covering the fields of health, medicine, psychology and architecture. To identify possible studies, a strategy for MEDLINE was developed using relevant MeSH terms and text words (dates searched 1902 to December 2006; Appendix 1). This strategy was adapted for other databases searched. These included: the Cochrane Central Register of Controlled Trials (CENTRAL; last searched January 2006; Appendix 2); EMBASE (January 1980 to February 2006; Appendix 3); Royal College of Nursing/British Nursing Index (BNI; January 1985 to August 2005; Appendix 4); PsycINFO (January 1806 to December 2006; Appendix 5); Construction and Building Abstracts (CBA; January 1985 to August 2005; Appendix 6); Royal Institute of British Architects (RIBA) library online catalogue (last searched December 2005; Appendix 7); InformeDesign (last searched January 2005; Appendix 8); NHS Estates Knowledge and Information Portal (Architecture in Healthcare Database; complete database searched November 2004); Avery Index to Architectural Periodicals (January 1996 to December 2001; Appendix 9); Cumulative Index to Nursing and Allied Health Literature (CINAHL; January 1982 to August 2005; Appendix 10); Web of Science (January 1970 to January 2006; Appendix 11); Applied Social Sciences Index and Abstracts (ASSIA; January 1987 to December 2004; Appendix 12); UK National Research Register (last searched February 2006; Appendix 13); Architecture Publication Index (January 1978 to March 2002; Appendix 14); Turning Research Into Practice (TRIP) database plus (last searched January 2006; Appendix 15); and Zetoc (The British Library's Electronic Table of Contents; last searched January 2006; Appendix 16). This review is currently being updated (MEDLINE last search October 2010; Appendix 17), see Studies awaiting classification.
Searching other resources
We reviewed reference lists of relevant articles, sourced grey literature from relevant organisations' web pages (e.g. Centre for Health Design, NHS Estates and Facilities Division), and contacted researchers for further information and other potential studies.
Data collection and analysis
Selection of studies
One review author (AD) conducted the initial search (Trials Search Co‐ordinator 'JM' conducted the EMBASE search). The initial search of databases retrieved 78,480 'hits'; these were screened for relevance and the majority of deleted records were double‐checked by a second review author. After screening, 15,140 titles were recorded on the main review database. Two review authors (AD/RS/DW/TD/EG) independently screened the studies obtained from the initial search. Each title was rated as 'hit' (maybe eligible), 'unsure' (probably not eligible) or 'reject' (not to be assessed further). Any disagreement with regard to eligibility was resolved through a third review author and discussion where necessary. We obtained full‐text (English and non‐English) papers for the 'hits' and abstracts for the 'unsures'. All abstracts (2029 total) were assessed independently by two review authors (AD/RS/DW/TD/EG) and rated as 'hit', 'unsure', or 'reject'. We then obtained full‐text papers (524 total) for 'hits' and 'unsures' and these were assessed for inclusion by at least two review authors. We discussed full‐text papers rated as 'unsure' in group meetings. A final corpus of 102 studies was selected for inclusion in the review (see Figure 1 for a flow diagram of the study process).
Study flow diagram.
Data extraction and management
Two review authors undertook data extraction independently, using a modified version of the EPOC data collection checklist (AD/DG/HM/RS/DW/TD/EG). Any disagreements were resolved through discussion between review authors.
Assessment of risk of bias in included studies
Retrieved studies were independently evaluated for risk of bias by two review authors (AD/DG/HM/RS/DW/TD/EG), using the criteria described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2008). For studies where there was difference of opinion a consensus was reached through discussion between review authors.
Data synthesis
The results section is structured alphabetically by intervention (and control) with multi‐intervention studies (e.g. whole ward redesign) being grouped as a separate category. Where appropriate, we have summarised the results of each intervention against different types of control (e.g. other form of environment, standard care) separately; this is because the effectiveness of interventions will vary depending on the comparison, and it may not be appropriate to combine all different types of control group together. Outcomes for individual interventions are looked at in turn and the heterogeneity of the studies explored.
For continuous variables, we calculated a mean difference (MD) for identical measures, or standardized mean difference (SMD), where different techniques were used to measure the same outcome domain, with 95% confidence intervals (CI) for each study. For dichotomous variables, we calculated a risk ratio (RR) with 95% CI. We used sensitivity analyses to explore the influence of risk of bias ('similarities at baseline'; 'sequence generation'; 'concealment of allocation'; 'completeness of data'; and 'blinding of healthcare personnel'), and the influence of deciding to include individual studies that were ambiguous as to whether they met the inclusion criteria.
Where statistical analyses were inappropriate or unfeasible, a discursive account of the results is presented with supporting tables. When it was appropriate to combine the studies, we used a random‐effects model of meta‐analysis. We have presented continuous data that were reported using medians and ranges in tables only.
Subgroup analysis and investigation of heterogeneity
We identified the presence of statistical heterogeneity by visually examining the forest plots, and using the I2 test for heterogeneity (where it was considered that: 0% to 40% might not be important; 30% to 60% may represent moderate heterogeneity; 50% to 90% may represent substantial heterogeneity; 75% to 100% represents considerable heterogeneity). The importance of the observed value of I2 was evaluated in conjunction with (i) magnitude and direction of effects and (ii) strength of evidence for heterogeneity (i.e. P value from the Chi2 test). We explored heterogeneity qualitatively (based on the characteristics listed below), and with subgroup analyses (where appropriate).
Heterogeneity was explored based on:
-
case mix (reason for hospitalisation; psychiatric/non‐psychiatric);
-
hospital visit characteristics (in‐patient/out‐patient/day‐patient; area of hospital studied);
-
geography (countries in which studies were undertaken).
To aid interpretation, we have presented the findings for anxiety from studies on music in subgroups (this decision was made post‐hoc and is based on the following rationale); Given the temporal dependence of anxiety as a transitional state (Spielberger 1983), the studies have been grouped according to the methodological criteria of when the musical intervention was provided, i.e. according to whether music was provided in the waiting period prior to a medical procedure (and hence the outcome of anxiety was obtained after the music intervention but before the medical procedure), or if music was provided throughout a medical procedure (anxiety was measured after both the music and medical procedure), or music was provided in the post‐operative period, or in an intensive care environment. We have displayed Chi2 tests assessing subgroup differences where appropriate (i.e. when the data presented in each subgroup is independent).
Results
Description of studies
Overall, 102 studies have been included in this review; one study was published twice on the same population but with different outcomes reported (Barnason 1995/1996) and two studies (Barnason 1995/1996 ; Lembo 1998) explored more than one type of intervention. Environmental interventions explored were: those that provided positive distracters to complement healthcare treatment already being administered, to include aromas (two studies), audiovisual distractions (five studies), decoration (one study), music (85 studies); those that reduced environmental stressors by implementing physical changes, to include studies on air quality (three studies), bedroom type (one study), flooring (two studies), furniture and furnishings (one study), lighting (one study), and temperature (one study);and multifaceted interventions (two studies). No studies meeting the inclusion criteria were found to evaluate: art, access to nature for example through hospital gardens, atriums, flowers, and plants, ceilings, interventions to reduce hospital noise, patient controls, technologies, way‐finding aids, or the provision of windows.
Studies awaiting assessment
We have two studies published in Korean, which we have not yet been able to assess due to translation difficulties (Hur 2005; Son 2006), one ongoing study (Characteristics of ongoing studies), and a further 66 studies awaiting assessment which are part of an ongoing update of this review (Studies awaiting classification).
Risk of bias in included studies
A summary of risk of bias judgements for all studies can be seen in Figure 2. A narrative description for each intervention type is given below.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Effects of interventions
Providing positive distracters
Aromas
Description of studies on aromas:
One RCT (Graham 2003) and one CCT (Holmes 2002) have been included on the use of aromas in the hospital (Table 1). These studies investigated the use of aromas in 328 patients overall; in Graham 2003 it is unclear how many participants there are per group, and Holmes 2002 is a cross‐over trial of 15 psychiatric patients. The overall mean age was 65.64 years old, with 169 males and 159 females included in the studies. Studies were conducted in Australia and England. Patient groups assessed were those undergoing radiotherapy treatment, and psychiatric in‐patients in a psychogeriatric ward.
| Study ID | Methods | Participants | Interventions | Outcomes |
| RCT; 3 parallel groups. | DESCRIPTION: 313 patients undergoing radiotherapy, in Australia. | FRAGRANT PLACEBO: Patients were administered the carrier oil with low‐grade essential oils. The carrier oil was sweet almond cold‐pressed pure vegetable oil. The low‐grade fractionated oils (lavender, bergamot, and cedarwood) were of unknown purity (supplied by Naturistics, Hornsby, Australia). These fractionated oils were diluted with the carrier oil in a ratio of 1:2. | ANXIETY, DEPRESSION, and FATIGUE: Measured via the Hospital Anxiety and Depression Scale (HADS), and the Somatic and Psychological Health Report (SPHERE), which is composed of the General Health Questionnaire (GHQ) and Symptoms of Fatigue and Anergia (SOFA) scales. | |
| CCT; Cross‐over trial, 2 conditions. | DESCRIPTION: 15 psychiatric inpatients in the communal area of a long‐stay hospital psychogeriatric ward for patients with behavioural problems, in England. | LAVENDER: The communal area of the unit was diffused with a standard concentration of lavender oil (2%), using three aroma‐streams for a period of two hours between the period of 4pm and 6pm. | AGITATION: Measured on the 16‐point Pittsburgh Agitation Scale by a blinded observer for the final hour of each two hour study period. Outcomes are presented as median scores for each patient in each condition. |
SD: standard deviation
Fragrances were administered via a necklace with a plastic‐backed paper bib (Graham 2003) and an aroma‐stream in the communal area (Holmes 2002). One study had three comparison groups (Graham 2003). Fragrances evaluated were: low‐grade fractionated oils (combination of lavender, bergamot, and cedarwood, diluted with a carrier oil), 100% pure essential oil (combination of lavender, bergamot, and cedarwood), and lavender oil (2%). Control conditions were: sweet almond cold‐pressed pure vegetable oil only; and water.
Outcomes assessed were: anxiety, depression, fatigue, and agitation.
We have tabulated 15 excluded studies (Table 2).
| Study ID | Reason for exclusion |
| Intervention | |
| Study design | |
| Study design | |
| Study design | |
| Setting and population | |
| Editorial | |
| Study design | |
| Outcomes | |
| Study design | |
| Data unsuitable for cross‐over study | |
| Commentary | |
| Setting | |
| Setting | |
| Aromas administered via nasal cannula, judged to be too invasive to constitute an 'environmental' intervention. | |
| Intervention; outcome not validated |
Risk of bias in included studies on aromas:
One RCT and one CCT were included on hospital aromas. Methods of group assignment were via telephone contact to a data management centre (Graham 2003; adequacy unclear), and alternate days (Holmes 2002). Concealment of allocation is adequate in Graham 2003. Blinding (of patients, physicians, and outcome assessment) was attempted in Graham 2003. In this study 9% of patients in the non‐fragrant placebo group believed they had received pure essential oil, 25% in the fragrant placebo group believed they had received pure essential oil, and 24% in the pure essential oil group believed they had received the pure essential oil. Holmes 2002 blinded the outcome assessor to study group (through the use of nose callipers), although in this study it was not feasible to blind patients to the scent on the ward. Completeness of outcome data was satisfactory in the studies (> 80% complete). Graham 2003 did not report withdrawals and drop‐outs, and in Holmes 2002 there were no withdrawals. It is unclear whether or not there is selective outcome reporting in either study. Protection against contamination could not be achieved in Holmes 2002 as this was a cross‐over trial, however, for the other study, this was not a problem.
Findings from studies on aromas:
Anxiety
One study on patients undergoing radiotherapy treatment, measured the outcome anxiety on the Hospital Anxiety and Depression (HAD) scale (Graham 2003). Graham 2003 reports a multivariate analysis, in which there were significantly fewer anxious patients (cases were classified as anxious when scoring > 7) in the non‐fragrant placebo group (13%), than in both the essential oil group (25%), odds ratio (OR) = 2.6 (95% confidence interval (CI) 1.1 to 6.1), and fragrant placebo group (23%), OR = 2.8 (95% CI 1.1 to 6.7).
Other outcomes
Graham 2003 found no strong evidence of effects for depression, fatigue, and general health (data insufficient for extraction) between any of the three groups.
Findings from non‐randomised studies:
Holmes 2002 investigated agitation using the Pittsburgh Agitation Scale (PAS) in a psychogeriatric ward communal area (N = 15, cross‐over study) and reports a significant effect (Wilcoxon Signed‐Ranks test P value 0.016) in favour of the lavender oil aroma‐stream group (median PAS score = 3, range = 1 to 7) versus diffused water (median PAS score = 4, range 3 to 7).
Art
Description of studies on art:
No studies have been included on the use of art in hospital. We have tabulated 16 excluded studies (Table 3).
| Study ID | Reason for exclusion |
| Qualitative report | |
| Descriptive article | |
| Participants | |
| Intervention interactive | |
| Qualitative report | |
| News article | |
| Qualitative | |
| Descriptive article | |
| Narrative article | |
| Commentary | |
| No group comparisons presented, unable to obtain further details from authors | |
| Study design | |
| Study design (part of same study as Staricoff 2001) | |
| Conference abstract, unable to obtain further details from author | |
| Setting | |
| Setting |
Audiovisual distractions
Description of studies on audiovisual distractions:
Five RCTs were conducted on audiovisual distractions (Table 4; NB. two articles report on Barnason 1995/1996; although the articles report differently on the choices of audiovisual distractions made available to the intervention group, the patient demographics and study designs are identical). These studies result in a total sample of 387 participants (Audiovisual group = 144, Control group = 243). Included patients had a mean age of 53.92 years old (range = 18 to 90), with 231 males and 156 females. Four studies were conducted in the USA and one was conducted in China. Three studies were carried out during endoscopy interventions, one was conducted during dressing changes for burns, and one during the post‐operative period.
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| RCT; 3 parallel groups. | DESCRIPTION: 96 in‐patients in the cardiovascular ICU and progressive care units having undergone elective coronary artery bypass grafting, in USA. | MUSIC GROUP: Choice of 5 tapes: 'Country Western Instrumental' or 'Fresh Aire' by Mannheim Steamroller, 'Winter into Spring' by George Winston, or 'Prelude' or 'Comfort Zone' by Steven Halpern. Played via headphones for 30 minutes. | Barnason 1995 reports: | Data extracted for state anxiety (STAI measure). Pain data are extracted for the end values on the VRS. | |
| RCT; 2 parallel groups. | DESCRIPTION: 80 in‐patients and out‐patients undergoing FB in Mariland, USA. | BEDSCAPES GROUP: Countryside river scene with associated sounds of nature played through headphones. Intervention available before, during, and after FB procedure. The scene was mounted by the bedside in the recovery area and on the ceiling in the procedure room. | ANXIETY: State anxiety via STAI; | SDs for anxiety have been estimated from P value of a t‐test. | |
| RCT; 3 parallel groups. | DESCRIPTION: 157 day‐patients undergoing colonoscopy in an Endoscopy Suite in Hong Kong, China. | VISUAL DISTRACTION: Eyetreck system (olympus) with preset home made movie (mainly scenic views), patient wears earphones but with no sound. | PAIN: scored using 10 cm VAS; | ||
| RCT; 3 parallel groups. | DESCRIPTION: 37 patients undergoing flexible sigmoidoscopy in Calfornia, USA. | AUDIOVISUAL: Virtual‐i glasses, personal display system showing an ocean shoreline with corresponding sounds (via headphones) | DISCOMFORT: Measured via VAS which asked patients to rate their level of abdominal discomfort from faint to severly intense. | Data for discomfort entered as pain scores. | |
| RCT; 2 parallel groups. | DESCRIPTION: 17 in‐patients undergoing burns care treatment and dressings change in a Burn Special Care Unit, Cincinnati, USA. | AUDIOVISUAL GROUP: "Muralvision" (Muralvision Studios, Inc., Eugene, Ore.)‐ on a bedside television, video programmes composed of scenic beauty (ocean, desert, forest, flowers, waterfalls, and wildlife) with accompanying music. | PAIN: Measured via the McGill Pain Questionnaire, with the Pain Rating Index and Present Pain Intensity scales. | For data extraction in the review, change scores from baseline were calculated and associated estimated standard deviations, using the F statistics provided. |
SD: standard deviation; STAI: State Trait Anxiety Inventory; VAS: visual analogue scale
Audiovisual interventions were all nature‐based, with Diette 2003 consisting of a static "photomural" of nature and the remaining being video‐based (Lee 2004a and Lembo 1998 used video eye glasses, and Barnason 1995/1996 and Miller 1992 used a bedside television). Diette 2003 and Lembo 1998 played nature sounds which corresponded to the visual distraction, and Barnason 1995/1996, Lee 2004a, and Miller 1992 played accompanying music. Three studies had more than one control group. Control groups consisted of: standard care alone (N = 4; Diette 2003; Lee 2004a; Lembo 1998; Miller 1992), visual distraction alone (N = 1; Lee 2004a), audio distraction alone (N = 2; Barnason 1995/1996; Lembo 1998), and scheduled rest (N = 1; Barnason 1995/1996).
Outcomes investigated were: Anxiety (N = 4), patient‐reported pain (N = 5), heart rate (N = 1), blood pressure (N = 1), sedation medication requirement (N = 1), sleep (N = 1), satisfaction (N = 1), hypotensive episodes (N = 1), oxygen desaturation (N = 1), recovery time (N =1), anger (N =1), stress (N = 1), and fatigue (N = 1).
We have tabulated 16 excluded studies on audiovisual distractions (Table 5).
| Study ID | Reason for exclusion |
| Intervention | |
| Study design | |
| Study design | |
| Intervention‐ interactive virtual reality | |
| Intervention‐ interactive virtual reality | |
| Intervention‐ patient education video | |
| Inadequate information provided (Ulrich study) | |
| Intervention interactive | |
| Unclear validity of outcomes | |
| Intervention interactive | |
| Intervention interactive | |
| Intervention‐ snoezelen | |
| Population, < 90% over 18 years old. | |
| Setting | |
| Setting | |
| Participants not adults |
Below, we summarise findings for the following comparisons:
-
Audiovisual distraction versus audio distraction (music)
-
Audiovisual distraction versus scheduled rest
-
Audiovisual distraction versus standard care alone
-
Audiovisual distraction versus visual distraction
-
Visual distraction versus standard care alone
Risk of bias in included studies on audiovisual distractions:
Five RCTs were included on audiovisual distractions. For Lembo 1998 and Miller 1992, the method of randomisation is unclear; Lee 2004a allocated patients via a computer‐generated list; Barnason 1995/1996 drew lots; and Diette 2003 allocated patients according to clinic day (which was randomised to intervention and control). Of the five included studies, it is unclear if concealment of allocation was used in four studies, and it was not used in one (Barnason 1995/1996 ). Lee 2004a reports blinding of recovery nurses (who assessed some outcomes) to patient allocation, and blinding of the endoscopists to two patient groups (but not to the standard care group). In the remaining studies blinding of healthcare personnel was not possible. Only Lee 2004a was judged to address incomplete outcome data, and the others remain unclear. Lee 2004a reported withdrawals and drop‐outs (eight patients; unclear from which groups), the reasons for which appear unrelated to the interventions, and Barnason 1995/1996 reports that all participants completed the study. Barnason 1995/1996 and Diette 2003 have been judged to be at risk of selective outcome reporting, whilst the remaining studies are unclear. All studies offered protection against contamination.
Findings from studies on audiovisual distraction versus audio distraction (music):
Anxiety
Two studies (Barnason 1995/1996; Lembo 1998) reported on the outcome anxiety (audiovisual group = 42, audio group = 45). Although both studies showed no strong evidence that the intervention had an effect (both individually and when combined: standardised mean difference (SMD) −0.24, 95% CI −1.05 to 0.56, P Value = 0.55), when combined they demonstrate moderate statistical heterogeneity (I2 = 65.6%, Chi2 = 2.91, df = 1, P value = 0.09; Analysis 1.1).
Exploring the heterogeneity of studies on anxiety:
With only two studies, it is difficult to explore the reasons for heterogeneity in terms of methodological and clinical differences. Both studies are subject to risk of bias; neither study had adequate allocation concealment and in Barnason 1995/1996 allocation concealment was not used. Additionally, Lembo 1998 was a very small study with no power calculation. Both studies were conducted in the USA and used video visual images. Barnason 1995/1996 was conducted on post‐operative patients and Lembo 1998 was conducted on patients undergoing an endoscopic procedure.
Pain
Two studies (Barnason 1995/1996; Lembo 1998) reported on the outcome pain (audiovisual group = 42, audio group = 45). With one small study (N = 25) in favour of audiovisual distraction (Lembo 1998: SMD −1.72, 95% CI −2.66 to −0.78), and the other (N = 62) showing no strong evidence that the intervention had an effect (Barnason 1995/1996: SMD 0.20, 95% CI −0.30 to 0.70), the studies combine with considerable statistical heterogeneity (I2 = 92%; Analysis 1.2), so we have not pooled these studies in a meta‐analysis.
Exploring the heterogeneity of studies on pain:
The comparison of Barnason 1995/1996 with Lembo 1998 has already been assessed for the outcome anxiety above.
Heart rate
Studies with insufficient data for extraction:
One study (Barnason 1995/1996; audiovisual N = 29, music N = 33) collected data on heart rate and reported that there was no strong evidence that the intervention had an effect.
Blood pressure
Studies with insufficient data for extraction:
One study (Barnason 1995/1996; audiovisual N = 29, music N = 33) collected data on blood pressure and reported that there was no strong evidence that the intervention had an effect.
Other outcomes
Lembo 1998 additionally reported on anger and fatigue (data insufficient for extraction). This study (audiovisual group = 13, audio group = 12) found the audiovisual group to have significantly lower anger scores (mean difference (MD) −0.40 points on a 10‐point visual analogue scale (VAS), 95% CI −0.68 to −0.12, P value = 0.005) and no difference in fatigue ratings. Barnason 1995/1996 (audiovisual = 29, music = 33) collected data on sleep quality to find no strong evidence that the intervention had an effect (MD 0.40 on the Richards‐Campbell Sleep Questionnaire, 95% CI −0.71 to 1.51, P value = 0.48).
Findings from studies on audiovisual distraction versus scheduled rest:
Anxiety
One study (Barnason 1995/1996) reported on anxiety (audiovisual group N = 29, scheduled rest N = 34) and found no strong evidence for an intervention effect (MD −1.60 points on the State Trait Anxiety Inventory (STAI), 95% CI −8.74 to 5.54, P value = 0.66).
Pain
One study (Barnason 1995/1996) reported on pain (audiovisual group N = 29, scheduled rest group N = 34) and found no strong evidence for an intervention effect (MD 0.15 points on a 10 point verbal rating scale (VRS), 95% CI −0.82 to 1.12, P value = 0.76).
Heart rate
Studies with insufficient data for extraction:
One study (Barnason 1995/1996; audiovisual N = 29, scheduled rest N = 34) collected data on heart rate and found no strong evidence for an intervention effect.
Blood pressure
Studies with insufficient data for extraction:
One study (Barnason 1995/1996; audiovisual N = 29, scheduled rest N = 34) collected data on blood pressure and found no strong evidence for an intervention effect.
Other outcomes
One study (Barnason 1995/1996 reported on sleep quality (audiovisual group N = 29, scheduled rest group N = 34) and found scores to significantly favour the audiovisual group (MD 1.57 on the Richards‐Campbell Sleep Questionnaire, 95% CI 0.47 to 2.67, P value = 0.005).
Findings from studies on audiovisual distraction versus standard care alone:
Anxiety
Three studies reported on anxiety (audiovisual group = 63, standard care group = 59). The largest of these studies (Diette 2003; N = 80) showed no strong evidence that the intervention had an effect, whilst the smaller two studies (Lembo 1998; Miller 1992) have significant findings favouring audiovisual distraction. When combined these studies show considerable statistical heterogeneity (I2 = 93%; Analysis 2.1) so we have not pooled these studies in a meta‐analysis.
Exploring the heterogeneity of studies on anxiety:
With few studies it is difficult to explore reasons for heterogeneity. All three studies are subject to risk of bias with unclear allocation concealment. Other than the size and quality of the individual studies, a further explanation for the differences in findings could be that the intervention Diette 2003 utilised was a static picture, whereas the other studies utilised video. All studies were conducted in the USA, with two (Diette 2003; Lembo 1998) being conducted on endoscopy patients and the other being conducted on patients undergoing burns dressing changes.
Pain
Three studies (Lee 2004a; Lembo 1998; Miller 1992) reported sufficient information for extraction on pain (Analysis 2.2). Although all three studies were significantly in favour of audiovisual distraction for pain relief, there is considerable heterogeneity between study effect estimates (I2 = 93%), therefore we have not pooled these studies in a meta‐analysis.
Exploring the heterogeneity of studies on pain:
With few studies it is difficult to explore reasons for heterogeneity. All three studies are subject to risk of bias, with unclear allocation concealment. All three studies used a video audiovisual distraction. Two were conducted in patients undergoing an endoscopic procedure and one (Miller 1992) was conducted during burns dressing changes. Two were conducted in the USA and one (Lee 2004a) was conducted in China.
Studies with insufficient data for extraction:
Diette 2003 also reported on pain scores; this study, which used static pictures also found a significant effect for pain in favour of patients who received an audiovisual distraction.
Sedation medication requirements
Lee 2004a reported on propofol requirement in patients undergoing an endoscopy procedure in China. This study found that those who received an audiovisual distraction (N = 52) required significantly less sedation medication than those who received standard care (N = 53), MD −0.37 mg/kg (95% CI −0.58 to −0.16, P value = 0.0005).
Other outcomes
Other health‐related outcomes reported were recovery time (Lee 2004a), oxygen desaturation episodes (Lee 2004a), hypotensive episodes (Lee 2004a), satisfaction (Lee 2004a), anger (Lembo 1998), and fatigue (Lembo 1998). Outcomes (each from just one study) favouring audiovisual distraction were anger (MD −2.20 cm on a 10 cm VAS, 95% CI −2.63 to −1.77, P value < 0.00001) and satisfaction (MD 2.30 cm on a 10 cm VAS, 95% CI 1.28 to 3.32, P value < 0.00001). There is no strong evidence of an intervention effect for recovery time, oxygen desaturation episodes, hypotensive episodes, or fatigue.
Findings from studies on audiovisual distraction versus visual distraction:
Pain
Lee 2004a found that audiovisual distraction (N = 52) is significantly better than visual distraction alone (N = 52) at reducing patient‐reported pain (MD −1.10 cm on a 10 cm VAS, 95% CI −2.01 to −0.19, P value = 0.02).
Sedation medication requirements
Lee 2004a found that audiovisual distraction (N = 52) is significantly better than visual distraction alone (N = 52) at reducing sedation medication requirements (MD −0.36 mg/kg, 95% CI −0.62 to −0.10, P value = 0.006).
Other outcomes
Lee 2004a also reported on recovery time, satisfaction, episodes of oxygen desaturation, and episodes of hypertension. For these outcomes there is no strong evidence that audiovisual distraction was more effective than visual distraction alone.
Findings from studies on visual distraction versus standard care alone:
Pain
Lee 2004a found no significant difference for patient‐reported pain (MD −0.80 cm on a 10 cm VAS, 95% CI −1.68 to 0.08, P value = 0.07) between a visual distraction group (N = 52) and standard care (N = 53).
Sedation medication requirements
Lee 2004a found no significant difference for sedation medication (propofol) requirements (MD −0.01 mg/kg, 95% CI −0.28 to 0.26, P value = 0.94) between a visual distraction group (N = 52) and standard care (N = 53).
Other outcomes
Lee 2004a found that patients allocated to a visual distraction group (N = 52) were significantly more satisfied (MD 2.10 cm on a 10 cm VAS, 95% CI 1.08 to 3.12, P value < 0.0001) than those allocated to standard care (N = 53).
Decoration
Description of studies on decoration:
One RCT on hospital décor has been included in the review (Table 6; Edge 2003). In this study there were 39 participants overall; 13 patients were assigned to beige rooms, 10 to purple, nine to green, and seven to orange. In Edge 2003 participants' ages ranged from 26 to 89 years old (average unknown), with 20 males and 19 females. The study was conducted in the USA on patients in a cardiac care unit.
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| RCT; 4 parallel groups. | DESCRIPTION: 39 in‐patients (10 post‐operative for cardiac surgery, 29 undergoing cardiac observations) in a cardiac care unit in Florida, USA. | BEIGE: Walls remained original colour of beige (similar to Sherwin Williams colour SW6658) in four rooms. | ANXIETY: Measured via STAI on day of discharge (after 2 to 5 days in hospital). Presented as mean (SD). No significant differences reported. | Partients in this study were not approached for informed consent until Day 3 of the study. |
STAI: State Trait Anxiety Inventory
The intervention colours were painted on to one wall at the foot of the patient beds, and were colour co‐ordinated with the rest of the room (e.g. with the colours in the curtains). Colour descriptors were: beige (original room colour, similar to Sherwin Williams colour SW6658), purple (SW6556), green (SW6451), and orange (SW6346).
Health‐related outcomes assessed were: anxiety, pain medication requests, and length of stay.
Twelve excluded studies have been tabulated on decoration (Table 7).
| Study ID | Reason for exclusion |
| Outcomes not validated | |
| Qualitative report | |
| Setting | |
| Study design | |
| Study design | |
| Participants | |
| Descriptive article | |
| Study design | |
| Descriptive article | |
| Outcomes | |
| Counfounding | |
| Staffing confound |
Risk of bias in included studies on decoration:
In the one RCT (Edge 2003) on decoration, participants were randomly assigned by administrative staff, but the method is unclear, and concealment of allocation was not used. Blinding of group allocation was not possible. Length of stay and pain medication requests were obtained from patient records, and subjective anxiety assessment was not blinded. The study is small, but of the consenting participants completeness of data was achieved. The author does give a description of withdrawals, drop‐outs, and non‐consenting patients (participants were approached for consent after allocation). Overall 11 patients were not included in the study for reasons seemingly unrelated to the intervention (Table 6), and one of whom was withdrawn by the researcher because she felt the patient was falsely answering the questions on anxiety (because the patient feared that high anxiety would affect her length of stay). This study is at risk of selective outcome reporting. It is likely that there was adequate protection against contamination.
Findings from studies on decoration:
Anxiety
Studies with insufficient data for extraction:
Edge 2003 found no strong evidence of an effect between groups for the outcome anxiety.
Pain medication requests
Studies with insufficient data for extraction:
Edge 2003 found no strong evidence of an effect between groups for the outcome pain medication requests.
Length of stay
Studies with insufficient data for extraction:
Edge 2003 found no strong evidence of an effect between groups for the outcome length of stay.
Music
Description of studies on music:
The 85 included RCTs investigating music (Table 8), resulted in a total sample of 6061 patients. Including two cross‐over trials (of 24 and 20 participants respectively; Davis‐Rollans 1987; Wong 2001). There were 2980 patients allocated to music, and 3124 allocated to a control (leaving one unknown case due to poor reporting in Taylor 1998). Patient characteristics were not reported in all of the studies; however, based on the 69 studies that reported information on mean age (details for 1108 participants remain unknown), the mean population age was 53.82 years old (range 14 to 99 years old; NB. > 90% of participants were 18 or older). Based on the 71 studies that reported information on gender, there were 2874 males and 2642 females included in the review (545 unknown cases). Studies were conducted in 16 different countries, predominantly the USA (N = 43) and China (N = 10). Five studies were conducted in each of Sweden and Canada, four in Germany, three in each of England and Taiwan, two in each of Australia, Japan, Turkey, and India, and one in each of Spain, Austria, Thailand, Slovenia, and Poland (NB. one study was carried out in two countries). The use of music was investigated in patients waiting for medical procedures (N = 13), undergoing endoscopic examinations (N = 12), undergoing percutaneous or surgical medical interventions (N = 34), undergoing non‐invasive medical procedures (N = 6), during labor (N = 1), post‐surgery (N = 7), in coronary care or intensive care environments (N = 10), or in ward environments (N = 2).
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| RCT; 2 parallel groups. | DESCRIPTION: 40 day‐patients undergoing ophthalmic surgery in New York, USA. | MUSIC GROUP: Patient choice of 22 types of music (e.g. soft hits, classical guitar, chamber music, folk music, popular singers from 1940's and 1950's), played via headphones throughout pre‐operative, surgical, and post‐operative periods. | PHYSIOLOGICAL: HR and BP measured via Propaq Monitor (Protocol Systems, Inc., Beaverton, OR) every 5 minutes during pre‐operative, surgical, and post‐operative period. Averages for the last three recorded measures within each time period were used for analysis in the paper. For purposes of review, data is extracted for the mean post‐operative scores only. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 118 out‐patients undergoing colonoscopy in a Digestive Endoscopy Unit in Spain. | MUSIC GROUP: Series of classical tracks (e.g. Bach, Grieg, Mozart, Delibe, Faure, and Mendelssohn) played via headphones during procedure. | ANXIETY: State anxiety measure pre and post procedure using the STAI; Reported as post ‐ pre difference with 95% CI. | There were no significant differences between groups regarding abnormal events. This data has not been extracted for the review. | |
| RCT; 3 parallel groups. | DESCRIPTION: 83 in‐patients undergoing cardiac catheterization in Germany. | MUSIC GROUP: Music was played via headphones, which were worn half on so patients could still hear the medical personnel, during the cardiac catheterization. A music therapist was present only to control the volume. Music played was "Entspannung" [relaxation] by Markus Rummel, composed specially for relaxation. | ANXIETY: Measured via the STAI before and after cardiac catheterization. Unclear whether post measurements were taken on the following day after cardiac catheterization. | ||
| RCT; 3 parallel groups. | DESCRIPTION: 90 patients undergoing urological surgery with spinal anaesthesia and PCS in Connecticut (USA) and Beirut (Lebonon). | MUSIC GROUP: Patients brought own music from home. | PROPOFOL REQUIREMENTS: Recorded as mg/kg/min and % of patients not using any propofol. Unclear if data presented are the SDs, and if the data presented as mg/kg/min is based on the total N or % of patients who used propofol. | Data not extracted for meta‐analysis. | |
| RCT; 2 parallel groups. | DESCRIPTION: 107 patients undergoing diagnostic coronary angiography or a percutaneous intervention procedure, in Ontario, Canada. | MUSIC GROUP: patient selected music (classical, soft rock, relaxation, country, own/other) played via headphones before, during, and after procedure, continued as the patient desired. | ANXIETY: State Anxiety via STAI pre and post procedure; | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 3 parallel groups. | DESCRIPTION: 96 in‐patients in the cardiovascular ICU and progressive care units having undergone elective coronary artery bypass grafting, in USA. | MUSIC GROUP: Choice of 5 tapes: 'Country Western Instrumental' or 'Fresh Aire' by Mannheim Steamroller, 'Winter into Spring' by George Winston, or 'Prelude' or 'Comfort Zone' by Steven Halpern. Played via headphones for 30 minutes. | Barnason 1995 reports: | Data extracted for state anxiety (STAI measure). See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 20 patients undergoing regional anaesthesia in North Carolina, USA. | NATURE TAPES: choice of sounds of birds, the ocean, a lagoon, or deeply resonant chimes, played for the duration of the surgery. | STATE ANXIETY: STAI administered pre‐operatively and one hour post‐operatively to calculate the change score. | Scores given for state anxiety are outside of the normal range for this questionnaire (20‐80). Method of calculating scores is not described. | |
| RCT; 3 parallel groups. | DESCRIPTION: 95 in‐patients undergoing coronary artery bypass (n = 92) or valvular heart surgery (n = 3) in Ohio, USA. | TAPED THERAPEUTIC SUGGESTIONS: excluded from review. | POST‐OPERATIVE STAY (days); Data extracted for analyses. | See Cepeda 2006 for details on music for pain relief. | |
| RCT: 3 parallel groups. | DESCRIPTION: 156 in‐patients undergoing chest tube removal after open heart surgery in Virginia, USA. | MUSIC GROUP: Patients preselected music they would prefer to hear from a library of 10 pre‐recorded music cassettes with no lyrics. Cassettes were produced by students in a music therapy programme under the supervision of a music therapist. Patients listened to the music via headphones for 10 minutes prior to and then during chest tube removal. | PAIN INTENSITY: Measured using a 10 cm NRS at 3 time points: (1) 10 minutes prior to chest tube removal, (2) immediately after chest tube removal, (3) 15 minutes after chest tube removal. | Report states that 18 participants had missing physiological data. Unclear from which groups these belonged, so for purposes of data extraction, it has been assumed that 6 participants were missing from each group. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 170 pre‐operative patients to undergo vascular angiography in California, USA. | MUSIC GROUP: Selection of 5 categories (classical, jazz, rock, country western, easy listening), played via headphones for 15 minutes prior to angiography. Patients could continue to listen to music during the angiography after collection of outcomes. | ANXIETY: Measured via STAI before and after 15 minute study period. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 140 in‐patients in the cardiac units with intravascular sheaths or an intra‐aortic balloon pump (IABP) in place, in USA. | MUSIC GROUP: 30 minutes of music through headphones. Mixture of symphonic music and nature sounds selected by the researchers. Same music played to all those in the music group. | PAIN PERCEPTION: measured via 10 mm VAS pre‐ and post‐intervention. | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 193 day‐patients undergoing lithotripsy for renal stones (America). | MUSIC GROUP: Music of type preferred by patient played via headphones, starting 10 minutes prior to procedure and continuing until 10 minutes after lithotripsy is complete. Patients wore additional ear protectors to protect patients from the noise of the lithotriptor. | PAIN INTENSITY: rated verbally every 5 minutes throughout the lithotripsy on a NRS (0‐10). | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 220 female out‐patients undergoing colposcopy in China. | MUSIC GROUP: CD compilation of slow‐rhythm music (instrumental ballad). Patients could choose to listen to any song(s) within the compilation. Music played through speakers during the examination. Prior to onset of study women attending the clinic were surveyed on musical preferences to inform compilation disc. | ANXIETY: State anxiety via Chinese version of STAI measured pre and post colposcopy; | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 43 in‐patients undergoing application of a C‐clamp (which applies pressure to stop bleeding when sheaths are removed after percutaneous coronary interventions), in an ICU, Hong Kong, China. | MUSIC GROUP: 3 choices of soft, slow music without lyrics (slow rhythmic songs, Chinese slow rhythmic music, Western slow rhythmic music). Music played via headphones during the application of the C‐clamp (approximately 45 minutes). | VITAL SIGNS: BP, HR, RR, and oxygen saturation, recorded at baseline, 15, 30, and 45 minutes via a bedside monitor. | Outcomes extracted for end time‐point only. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 64 patients undergoing cesarean section in Taiwan. | MUSIC GROUP: Choice of western classical, new age, or Chinese religious music, played via headphones. Participants listened to music for at least 30 minutes from start of anaesthesia to end of surgery. Volume low enough to allow mutual conversation with the researcher. | ANXIETY: measured via a 10 cm VAS, researcher filled it in after asking participant to indicate how they were feeling. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 20 in‐patients receiving mechanical ventilation in private patient room areas in critical care units in Midwest USA. | MUSIC GROUP: Selection of classical music played for 30 minutes via headphones. Patients instructed to close eyes and concentrate on the music. | MOOD: Short‐form POMS. | No SDs are presented for the post‐intervention physiological outcomes. Significant differences were found for HR and RR in favour of music. Other variables did not differ significantly. | |
| RCT; 2 parallel groups. | DESCRIPTION: 54 in‐patients receiving mechanical ventilation in one of 4 Intensive Care Units, USA. | MUSIC GROUP: choice of non‐lyric tapes 60‐80 bpm, classical, new age, country western, religious, and easy listening (30 mins). | ANXIETY: State anxiety via short form STAI (6 items); | Due to the lack of clarity over withdrawals and drop‐outs, it is unclear how many people were analysed in each group for the outcome anxiety. The degrees of freedom stated (49) suggests 51 observations were made, and we have assumed that N = 25 in the music group and N = 26 in the control group, based on the descriptions given for 3 of the withdrawals in the paper. | |
| RCT; 2 parallel groups. | DESCRIPTION: 64 out‐patients undergoing flexible sigmoidoscopy (FS) in Midwest USA. | MUSIC GROUP: choice of music (classical, country‐western, new‐age, easy listening, pop, rock, religious, jazz, era‐specific, motion picture soundtracks), played via headphones during FS procedure. Patients instructed to concentrate on music and that the investigator would meet with them afterwards to discuss their experiences. | ANXIETY: State anxiety measured via STAI | ||
| RCT; 2 parallel groups. | DESCRIPTION: 68 pre‐operative patients undergoing extracorporeal shock wave lithotripsy (ESWL) in Taiwan. | MUSIC GROUP: listened to natural music via headphones for 5 minutes prior to ESWL. | BLOOD PRESSURE: Method of data collection unclear. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 60 in‐patients and out‐patients undergoing flexible fibreoptic bronchoscopy (FB) in California, USA. | MUSIC GROUP: "Relax" (Expansion Records, Manchester, UK) consisting of piano improvisations (60bpm), played via headphones during FB procedure. | ANXIETY: State and Trait anxiety measure via STAI. | ||
| RCT; 3 parallel groups. | DESCRIPTION: 180 pre‐operative day patients scheduled for day surgery in Australia. | MUSIC GROUP: Choice of classical, jazz, country & western, new age, easy‐listening, and "other" mostly by contemporary artists. Music played for 30 minutes via headphones. | ANXIETY: State anxiety measured via STAI. | 95% Confidence Intervals are presented in the paper based on logarithmically transformed scores. Standard deviations were calculated by first back‐translating data to a log scale before utilising the CIs to estimate the SDs. Data entered into the review analysis are expressed as the natural log. | |
| RCT; 4 parallel groups. | DESCRIPTION: 121 patients undergoing elective cataract extraction in Canada. | MUSIC GROUP: Classical music accompanied by soothing sounds of nature, played via headphones intra‐operatively. | ANXIETY: Measured via STAI before and after surgery. | Data reporting unclear. Cannot extract SDs. | |
| RCT; 3 parallel groups. | DESCIPTION: 90 pre‐operative in‐patients waiting for dental restoration surgery under general anaesthesia, or orthopedic surgery, in Germany. | MUSIC GROUP: listened to 45 minutes of music pre‐operatively. Choice of music arranged the evening before surgery. 11 patients chose classical, 19 chose pop music. | ANXIETY: Measure via STAI and a tick‐box anxiety scale tailored for the clinic. There was a significant decrease in state anxiety from pre‐treatment to post‐treatment in the music group (change score = 2.2). There was no significant change in anxiety for the 'no medication' control group (change score = 0.633). | Not enough information for data extraction. | |
| RCT; Cross‐over study. | DESCRIPTION: 24 in‐patients with MI (N = 12) and other cardiac conditions (N=12) in a Critical Care Unit, Ontario, Canada. | MUSIC: 3 pieces each lasting approximately 12 minutes played in randomised Latin square design via headphones. a) Symphony no.6 by Beethoven; b) Eine Kleine Nachtmusik by Mozart; and c) The Moldau by Smetana. | PSYCHOLOGIC: Questionnaire not fully validated (content validity only). Not extracted for review. | Significant order effects for HR variability (P = 0.03) and heart rhythm (ectopy). | |
| RCT; 3 parallel groups. | DESCRIPTION: 143 out‐patients (93 included in present review) undergoing screening mammography in a clinic of a tertiary care hospital in Massachusetts, USA. | RELAXATION GROUP: excluded from review, as taped instructions constitute a psychological therapy. | ANXIETY: recorded before and after study period using the STAI, and at the end of the study with a Likert scale (1‐10) asking to rate level of anxiety felt during the procedure. | Investigator identified a possible floor effect. See Cepeda 2006 for details on music for pain relief (this study is not yet included in Cepeda 2006). | |
| RCT; 3 parallel groups. | DESCRIPTION: 56 in‐patients with unstable angina pectoris or acute MI at coronary care unit in Australia. | MUSIC GROUP: Received two or three 30‐minute sessions of light classical music (Bonny,Music Rx) played via headphones. | ANXIETY: Measured at pre and post test with three psychologic scales, 1) STAI, 2) Hospital Anxiety and Depression Scale, and 3) Linear Analogue Anxiety Scale (VAS). STAI scores extracted for review. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 33 in‐patients undergoing bone marrow transplant chemotherapy, in Columbus, USA. | MUSIC GROUP: Listened to 45 minute recording of self‐selected music via headphones at 6, 9, and 12 hours after the start of each infusion as an adjunct to antiemetic therapy. A variety of music selections was available and patients were encouraged to bring their favourite music. | NAUSEA: Measured on a VAS in the pictorial form of a thermometer (questionable validity). | Data not in sufficient detail for extraction. | |
| RCT; 2 parallel groups. | DESCRIPTION: 11 patients undergoing range‐of‐motion exercises as part of acute care rehabilitation for burns, Virginia, USA. | MUSIC GROUP: Choice of 6 cassette tapes (Lifescapes series) that met the guidelines for music in medical settings. Music was played during the range‐of‐motion exercises. | PAIN: Measured via VAS before and after rehabilitation exercises. There was a statistically significant increase in pain from pre‐treatment to post‐treatment in both groups (P = 0.04). There was no significant difference between groups (P = 0.38). | Data not in sufficient detail for extraction. See Cepeda 2006 for details on music for pain relief. | |
| RCT (post‐test only); 3 parallel groups. | DESCRIPTION: 15 pre‐operative patients in the waiting room for elective same‐day surgery in Pittsburgh, USA. | HUMOUR GROUP: excluded from review. | ANXIETY: Measured via a VAS after the intervention period. | ||
| RCT (post‐test only); 3 parallel groups. | DESCRIPTION: 46 pre‐operative patients scheduled for same‐day surgery in Pittsburgh, USA. | HUMOUR GROUP: excluded from review. | ANXIETY: Measure post‐intervention via VAS. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 50 pre‐operative patients to undergo septorhinoplastic surgery in Turkey. | MUSIC GROUP: Patients brought own CD or tape from home, those who forgot were provided with a 'suitable replacement', played via headphones during pre‐operative period, as patients were being sedated. | "modified" OAA/S: not included in review. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 93 in‐patients scheduled for laparoscope surgery, in Beijing, China. | MUSIC GROUP: Choice of 6 types of music (Western classical; light; pop; folk; folksong; opera) with 30 minutes listening time. Patients listened to music via headphones 1‐2 hours before their operation, whilst lying in bed. | ANXIETY: Measured via STAI before and after intervention period. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 78 patients undergoing colonoscopy in Kerala, India. | MUSIC GROUP: Choice of 6 tapes played via headphones during colonoscopy. Selection of: popular film songs (based on carnatic classical ragas), classical music, devotional songs, folk songs, soft instrumental music, bioacoustics (soft instrumental music with nature sounds). | SEDATION: Dose of midazolam (2 mg given on demand). | Data reported as median and range. See Cepeda 2006 for details on music for pain relief (this study is not yet included in Cepeda 2006). | |
| RCT; 2 parallel groups. | DESCRIPTION: 198 out‐patients awaiting gastrointestinal procedures (colonoscopy or esophagogastroduodenoscopy [EGD]) in California, USA. | MUSIC GROUP: Patient selected music (classical, rock, jazz, country western, easy listening) for 15 minutes prior to medical procedure. Were allowed to continue listening to music after outcome measures were taken. | ANXIETY: state anxiety measured via STAI. | This study has questionable clinical relevance: "To avoid introducing more anxiety with the explanation of the [gastrointestinal] procedure, the consent process for the [gastrointestinal] procedure was delayed until after the patients completed the music study". | |
| RCT; 3 parallel groups. | DESCRIPTION: 60 in‐patients who have undergone surgery with general anaesthesia, in a PACU in Iowa, USA. | MUSIC GROUP: Choice of three instrumental tapes: Calm classical (e.g. Bach, Debussy, Pachelbel), Stimulative classical (e.g. Strauss, Tschaikovsky), and calm popular music (e.g. piano solos by George Winston, guitar solos by William Ackerman and Steve Halpern). Patients decided which music they would like to listen to in the PACU in the pre‐operative visit. Played via headphones until discharge from PACU. | PAIN: Measured via 10cm VAS every 15 minutes while in the PACU. | Data insufficient for extraction. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 55 day‐patients undergoing laparoscopic sterilization or laparoscopic tubal dyeing as part of a fertility programme in Sweden. | MUSIC GROUP: Panpipe music played via headphones for 30 minutes pre‐operatively and 30 minutes post operatively. | PHYSIOLOGICAL MEASURES: RR, BP, and HR, were measured by an attending nurse blinded to group allocation, before and after each 30 minute session (pre‐ and post‐surgery). | Sensitivity analyses conducted using data extracted for pre‐operative and post‐operative scores for anxiety, RR, BP, and HR. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 3 parallel groups. | DESCRIPTION: 110 in‐patients and out‐patients undergoing IV catheter insertion in Southwestern public and private hospitals, USA. | SALINE GROUP: excluded from review as confounding non‐environmental intervention. | PAIN INTENSITY: measured via 100mm VAS. | See Cepeda 2006 for details on music for pain relief. Insertion difficulty is not included in the present review as a health‐related outcome. | |
| RCT; 3 parallel groups. | DESCRIPTION: 76 in‐patients undergoing surgery with general anaesthetic, in Sidcup, UK. | MUSIC GROUP: Classical music ('Adagio' Karajan, Deutsche Grammophon, 445 282‐4). Played via headphones during surgery. | FENTANYL REQUIIREMENTS: This served as an indication of the adequacy of nociception control provided during the operation. Fentanyl (ųg) was given intravenously if intra‐operative BP or HR increased by 20% or more above baseline values for more than 5 minutes. | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 34 out‐patients undergoing urologic procedures using spinal anaesthesia, USA. | MUSIC GROUP: All patients requested to bring their favourite CD to hospital on morning of surgery. A suitable substitution was provided to those who did not have access to their favourite CD. Music played via occlusive headphones intraoperatively. | BP (mm Hg), HR (bpm), PROPOFOL REQUIREMENTS (mg/min): recorded every 10 minutes for duration of surgery; | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 43 patients undergoing lithotripsy, USA. | MUSIC GROUP: Patients brought own music from home to listen to via occlusive headphones during surgery. | BP (mm Hg), HR (bpm), ALFENTANIL REQUIREMENTS: recorded every 15 mins. | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 3 parallel groups. | DESCRIPTION: 163 in‐patients undergoing hysterectomy in Vienna, Austria. | MUSIC GROUP: Patient choice of classical (Bach), Entertainment ("Musik zum Träumen" [Music for Dreaming]), or Relaxation (Oliver Shanti "Rainbow Way"). Played via headphones for 45 minutes beginning at the start of the operation (at the abdominal incision). | PAIN: Pain intensity measured via 10 cm VAS once a day for 5 days. There were no differences between groups. | Where means and SDs are presented (for length of stay), there are missing data from 23 participants so it is unclear how many people are in each group. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 104 patients undergoing gastrointestinal endoscopy in India. | MUSIC GROUP: Classical Indian instrumental music played for 10 minutes prior to procedure and then throughout the procedure. | PHYSIOLOGICAL OUTCOMES: BP, HR, and RR measured at beginning of the consultation and at the end of the procedure. Methods unclear. Data presented as change scores. | ||
| RCT; 3 parallel groups. | DESCRIPTION: 58 patients undergoing noxious medical procedures (e.g. tissue biopsy, vascular port placement) at an oncology clinic in Midwestern USA. | MUSIC GROUP: Patient choice from selection of CDs (pop, rock, easy listening, classical, religious hymns, jazz or blues, country), listened via headphones prior to and during procedure. | PAIN: Pain intensity measured via a NRS prior to treatment, during treatment (retrospectively), and post treatment. | Values provided are adjusted scores (from analysis of covariance), adjusted for baseline scores, medications, and gender. Numbers reported as SDs in the text although too small so assumed to be the standard errors, which reflect the non‐significant findings described. See Cepeda 2006 for details on music for pain relief | |
| RCT; 3 parallel groups. | DESCRIPTION: 165 out‐patients undergoing elective colonoscopy in Hong Kong, China. | MUSIC ALONE: excluded from review as no appropriate control (i.e. music with no PCS). | PAIN SCORE: 10 mm VAS; | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 64 in‐patients on mechanical ventilation in an Intensive Care Unit in Hong Kong, China. | MUSIC GROUP: Choice of Chinese classical music, religious music (Buddhist and Christian), Western classical, natural sounds with slow beats. Played via headphones. | ANXIETY: State anxiety using Chinese STAI ‐short version (6 items). Participants responded to questions by holding up corresponding number of fingers. | ||
| RCT; 3 parallel groups. | DESCRIPTION: 37 patients undergoing flexible sigmoidoscopy in Calfornia, USA. | AUDIOVISUAL: Virtual‐i glasses, personal display system showing an ocean shoreline with corresponding sounds (via headphones) | DISCOMFORT: Measured via VAS which asked patients to rate their level of abdominal discomfort from faint to severely intense. | Data extracted for review on anxiety and anger. Arousal and attention not considered health‐related outcomes. | |
| RCT; 2 parallel groups. | DESCRIPTION: 50 in‐patients and out‐patients undergoing non‐oncologic surgery under spinal anaesthesia in Canada. | MUSIC GROUP: Choice of pop, jazz, classical, and new age played via non‐occlusive headset. Patients received the music prior to, during and after surgery. | PCS: Amount of midazolam consumed during the perioperative period was recorded. Data extracted for total midazolam consumed (mg). | Estimated anxiety data extracted. | |
| RCT; 2 parallel groups. | DESCRIPTION: 35 in‐patients in one of 5 Intensive Care Units having had an acute myocardial infarction, Midwestern USA. | MUSIC GROUP: 3 sessions of music listening on 1st, 2nd, and 3rd day (or 2nd‐4th day) of hospitalisation. Music session consisted of listening to 3 pieces of music each session: Bach's 'Largo', Beethoven's 'Largo', and Dubussy's 'Prelude to the Afternoon Faun'. Each session lasted approximately 22 minutes. | STATE ANXIETY: measured with the STAI at two time points: (1) during the first 48 hours of admission, and (2) on the 3rd or 4th day of hospitalisation. For the music group, these measurements were taken prior to the first music session and at the end of the 3rd music session. | ||
| RCT; 3 parallel groups. | DESCRIPTION: 45 patients undergoing femoral angiography in Boston, USA. | RELAXATION TAPE: Progressive muscle relaxation. Excluded from review. | STATE ANXIETY: Measured via STAI immediately pre‐ and post‐procedure. | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 44 post‐operative in‐patients in an Orthopaedic Department, Japan. | MUSIC GROUP: Choice of: Western classical music, Gagaku (Japanese traditional court music), Noh songs, or Enka. All participants choose Enka (a melodramatic and representative genre of Japanese popular songs, usually about sad aspects of life, irrecoverable destiny, and desertion by a lover, sung with a slow tempo), which is popular among elderly Japanese people. There were 10 Enka CDs to choose from. Music played via headphones for 20 minutes whilst lying in bed. | PAIN: Measured via 10 cm VAS and the Wong/Baker Faces Pain Rating Scale (includes 6 categories of facial expressions ranging from '0, a happy smiling face' and '5, a tearful face'). | See Cepeda 2006 for details on music for pain relief (study not yet included in Cepeda 2006). | |
| RCT; 4 parallel groups. | DESCRIPTION: 52 in‐patients undergoing various surgical procedures in USA. | MASSAGE THERAPY: excluded from review. | STATE ANXIETY: measured via the short‐form STAI‐6 in the PACU (after surgical procedure). | See Cepeda 2006 for details on music for pain relief (study not included in Cepeda 2006). | |
| RCT; 2 parallel groups. | DESCRIPTION: 38 emergency department admissions undergoing laceration repair in Pittsburgh, USA. | MUSIC GROUP: Listened to music via headset. Choice of 50 styles and artists to choose from. | ANXIETY: State anxiety measured via STAI before and after laceration repair. SDs presented in graphical format. Estimated figures extracted for purposes of review. | SDs for pain and anxiety are estimates from graphs. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 30 patients undergoing surgery with general anaesthesia, Canada. | MUSIC GROUP: Selected a CD preoperatively from a choice of 4: classical, jazz, new‐age, popular piano music. Selected listening volume. Anesthesiologist started CD once patients were anaesthetised. Music played via headphones. | Repeated measures observations (4 time points)‐ T1: immediately after arterial line insertion, T2: 5 min after peritoneal incision, T3: at skin closer, T4: 30 min after arrival in the recovery area. | Data extracted for end time points only. See Cepeda 2006 for details on music for pain relief | |
| RCT; 2 parallel groups. | DESCRIPTION: 28 post‐operative in‐patients who have undergone elective abdominal hysterectomy, USA. | MUSIC GROUP: Four easy listening selections (out of an original 10) were selected by nursing graduate students to be harmonious, pleasant, and calming. Patients listened to music for 10 minutes via headphones and were requested to close eyes on the first and second post‐operative day. | PAIN: measured via VAS on first and second post‐operative day (pre and post 10 minute intervention period). | Outcomes extracted for 2nd post‐operative day as only 6 and 5 outcome measures were obtained for the intervention and control group on the 1st post‐operative day (all 28 participants completed the 2nd day of testing). See Cepeda 2006 for details on music for pain relief. | |
| RCT; 3 parallel groups. | DESCRIPTION: 90 (58 included in present review) in‐patients undergoing elective hysterectomy in Sweden. | MUSIC GROUP: relaxing music accompanied by sea waves (ref: Uneståhl L‐E. Avslappningmusik. Träningsprogram för kropp och själ. [Relaxation music. Training programme for body and soul]. Örebro, Sweden Veje International AB: 1970). Played through headphones from time of skin incision to time of wound closure. | PAIN INTENSITY: 10‐point VAS used every hour for the first 24 post‐operative hours, and every 3 hours after that until the patient felt no pain. Data presented as group mean score of patients' median score. | Data extracted for responses given on the day of surgery. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 3 parallel groups. | DESCRIPTION: 151 day patients undergoing surgery of varicose veins or inguinal hernia repair under general anaesthesia, in Sweden. | INTRA‐OPERATIVE MUSIC GROUP: Music played via headphones from end of induction of anaesthesia until wound dressing via headphones. Music was soft instrumental comprising 7 melodies of a new‐age synthesizer. Post‐operatively, in the PACU, patients were exposed for 1 hour to a blank (silent) disc via headphones. | PAIN INTENSITY: Post‐operatively on a NRS (0 to10), every 30 minutes for 2 hours in the PACU. | Anxiety and fatigue not presented in enough detail. Findings on night sleep not reported at all. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 3 parallel groups. | DESCRIPTION: 182 (125 included in present review) day patients undergoing surgery for varicose veins or inguinal hernia repair under general anaesthesia, from two hospitals in Sweden. | MUSIC GROUP: Soft classical music via headphones which allow conversation to take place. Played on auto‐reverse from the time of arrival at the PACU until the patient wanted to stop listening. | PAIN INTENSITY: taken every 30 minutes until patient reports a pain level of ≤ 3 on a VAS (0 to 10). Data presented as the mean of median scores in 120 minutes. | There was a significant difference in the length of time patients listened to their allocated tape. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 3 parallel groups. | DESCRIPTION: 75 day patients undergoing surgery of open Lichtenstein inguinal hernia repair, in Sweden. | INTRA‐OPERATIVE MUSIC GROUP: exposed to new‐age synthesizer music during operation, and a sham (silent) CD for 1 hour post‐operatively in the PACU. | PAIN: assessed by NRS (0 to 10), 30 minutes before anaesthesia and 1 hour after admission to PACU. | Data extracted for post‐operative music group versus control group. See Cepeda 2006 for details on music for pain relief (study not yet included in Cepeda 2006). | |
| RCT; 2 parallel groups. | DESCRIPTION: 36 in‐patients with bronchial asthma, in Poland. | MUSIC GROUP: Underwent 10 day rehabilitation programme, including 45 minutes each day of: exercise of breath control, correction of respiratory pattern, training of diaphragm and additional respiratory muscles, plus an additional 15 minutes of music listening (C.M. Weber "Adagio"; J.S. Bach "Air on a G‐string"; V.A. Mozart "Andante z Divertimenta D‐dur"). | ANXIETY: Measured via the STAI at baseline and after the 10th session. | Data not provided in sufficient detail for extraction. | |
| RCT; 3 parallel groups. | DESCRIPTION: 104 day‐patients scheduled for elective surgical procedures (gynaecology, general surgery, urology) in a Day Surgery Unit in England. | MUSIC + BINAURAL BEAT: taken to a quiet environment pre‐operatively and asked to listen to music with their eyes closed; 30 minute soundtrack (Holosync Solution, 'Awakening Prologue', Centerpointe Research Institute, Beaverton, OR) which produces binaural beats (through two similar pure tones being presented separately to each other). | STATE ANXIETY: measured via the STAI before and after intervention period. Results converted to percentages instead of presenting STAI score. | It is unclear if > 90% of participants are 18 years old or over. | |
| RCT; 2 parallel groups. | DESCRIPTION: 50 out‐patients undergoing flexible sigmoidoscopy in Maryland, USA. | MUSIC GROUP: Choice of 20 tapes (classical/country‐western/popular/rhythm and blues/gospel) played via headphones throughout the procedure. | ANXIETY: Measured via STAI. An analysis of variance identified a significant difference between groups (P < 0.002): | No SDs reported. Insufficient data for extraction. | |
| RCT; 2 parallel groups. | DESCRIPTION: 110 in‐patients giving birth for the first time (to a singleton fetus), in Southern Thailand. | MUSIC: patient choice of western music without lyrics, including synthesizer, harp, piano, orchestra, and jazz (60‐80 bpm). Patients listened to their choice via headphones for the first 3 hours of the active phase of labour, starting when cervical dilation was 3 or 4 cm with uterine contraction of 30 to 60 seconds. Women could stop listening to music for 10 minutes if they wished. | PAIN SENSATION: measured using 100 mm VAS at four time points‐ once at the start of the study before the treatment period, and then every hour for 3 hours. | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 119 patients undergoing diagnostic endoscopy in Germany. | MUSIC GROUP: Played Radio Arabella (105.2 MHz) throughout the procedure. This is a well‐known regional broadcasting company in Munich, and plays various trends of pop, rock, soul, and 'Deutsche Schlager' music, specialising in 'oldies'. | SEDATION & ANALGESIA: Number of patients requiring midazolam and pethidine, and amount administered. It is unclear from the report if sedation is an outcome measure or baseline characteristic. | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 30 patients undergoing cerebral angiography in Hannover, Germany. | MUSIC GROUP: Choice of nine tapes (international pop, German pop, oldies, meditation, rock, techno, instrumental, classic, and traditional). Played via stereo speakers throughout the angiogram; patients could adjust the volume. | ANXIETY: Measured via STAI the evening before the angiography and during the angiography. | Data not sufficient for extraction. | |
| RCT; 2 parallel groups. | DESCRIPTION: 86 post‐operative in‐patients who have undergone cardiac surgery, in cardiovascular units, Midwest USA. | MUSIC GROUP: Participants advised to clear minds and allow muscles to relax. They were given a choice of easy listening, classical, or jazz music, played via headphones, for 20 minutes whilst participant remained in bed. Music qualities were: no dramatic changes, consonance, instrumental, 60 to 70 bpm. The environment was made conducive to rest. | HR: Recorded on bedside monitor. No differences between groups. | Paper only reports data for the first 3 sessions due to missing data. SDs obtained from authors (unpublished). Data extracted for morning session of the first post‐operative day. See Cepeda 2006 for details on music for pain relief (study not yet included in Cepeda 2006). | |
| RCT; 2 parallel groups. | DESCRIPTION: 42 male out‐patients undergoing radiation therapy in a Veterans Affairs Hospital in Southeastern USA. | MUSIC GROUP: Patient choice of rock and roll, big band, country and western, classical, easy listening, Spanish, religious. Four to six tapes were available in each category. Each patient could choose one category for the duration of the study. Patients listened via headphones to the music before and during the radiation simulation appointment, and during daily radiation treatments for the duration of course of therapy. | STATE ANXIETY: measured using the STAI at 5 time points: (1) time of evaluation; (2) post simulation appointment; (3) end of first week of treatment; (4) end of third week of treatment; and (5) end of the fifth week or end of radiation therapy. | Data extracted for end time point. | |
| RCT; 2 parallel groups. | DESCRIPTION: 32 out‐patients undergoing colonoscopy, USA. | MUSIC GROUP: Patient choice of classical, jazz, pop rock, easy listening, played via headphones throughout pre‐sedation and procedure. | SEDATION REQUIREMENTS (mg). Amount of Versed and Demerol administered was recorded (not a primary outcome). | State anxiety data extracted from graph in article, bars taken as standard errors and converted into SD. | |
| RCT; 2 parallel groups. | DESCRIPTION: 17 in‐patients with HIV/AIDS staying in the HIV/AIDS ward, London, UK. | MUSIC GROUP: Attended a live music concert played in the public area on the hospital ground floor. Unclear what type of music was played. | LEVELS OF CD4 AND CD8 LYMPHOCYTES (cells/mm3): measured within one hour before and after the study period. | Data insufficient detail. | |
| RCT; 2 parallel groups. | DESCRIPTION: 120 in‐patients undergoing surgery with epidural anaesthesia in Taiwan. | MUSIC GROUP: Walkman music via headphones. Choice of 5 types (Manderin pop song, local Taiwanese folksong, Western, Classical, Buddhist hymn). Music played throughout operation. | Subjective feelings of anxiety and sedation: Method of assessment unclear (data EXCLUDED from review). | ||
| RCT; 3 parallel groups. | DESCRIPTION: 61 in‐patients who have had an elective abdominal hysterectomy using general anaesthesia in Arizona, USA. | MUSIC GROUP: patient choice brought from home or from selection provided (classical; jazz; light rock; country; rock and roll; easy listening; gospel), played via headphones. | PAIN INTENSITY: Two measures used: a 9‐inch VAS, and a 10‐point NRS. Measures were taken every 15 minutes for the duration of PACU stay (unspecified). Results of VAS are reported as a mean value for each group; results of NRS are reported as mean rating at 1 hour and at discharge from PACU. | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 3 parallel groups. | DESCRIPTION: 30 pre‐operative in‐patients scheduled to undergo cardiac catheterization (CC) in Hong Kong, China. | SENSORY INFORMATION: excluded from review. | STATE ANXIETY: measured via STAI before and after intervention period; | Outcomes extracted for T1 ‐ after the study intervention but before cardiac catheterization, as the study has been categorised in the "pre‐procedure" subgroup. | |
| RCT; 2 parallel groups. | DESCRIPTION: 200 patients undergoing flexible bronchoscopy in Slovenia. | MUSIC GROUP: Easy listening and relaxation ambient music selected by the investigator, started immediately after the beginning of the procedure and stopped when the procedure is over. | PHYSIOLOGICAL: HR and BP recorded before and after the procedure, method of data collection not described. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 59 in‐patients undergoing elective laparoscopic cholecystectomy in an operating theatre in Japan. | NATURE SOUNDS: Patients were played their choice of nature sounds via headphones throughout operation, after induction of anaesthesia until the last suture of surgery. | PHYSIOLOGICAL MEASURES: BP and HR was recorded through 'non‐invasive' methods using the Philips patient monitoring system for anaesthesia [unpublished]. Outcomes were measured pre‐anaesthesia, at the start of surgery, at gallbladder removal, at the end of surgery, at extubation, at the end of anaesthesia, and in the PACU. | Data insufficient detail for extraction. Authors unable to provide means and SDs for physiological measures. | |
| RCT; 2 parallel groups. | DESCRIPTION: 60 in‐patients undergoing coronary artery bypass graft or vascular surgery, in the OR and ICU, Florida, USA. | MUSIC GROUP: Choice of 6 discs played during surgery and post‐operatively. Family encouraged to bring in additional music choices post‐operatively. Music selection from Prescriptive Music Inc., ('Clarity'‐ melodies from classical motion pictures; 'Timeless'‐ heartfelt originals, 'Towards'‐ piano improvisation, 'Interlude'‐ piano music by Mozart, 'Universe'‐ synthesized compositions, 'Essence'‐ cello and piano). | ANXIETY: Measured via the STAI on the 3rd post‐operative day. | ||
| RCT; 3 parallel groups. | DESCRIPTION: 61 post‐operative in‐patients following open heart surgery undergoing 30 minutes chair rest, in the surgical intensive care unit of a rural Midwestern hospital, USA. | MUSIC GROUP: Instructed to listen and follow music and allow it to distract and relax. Played through headphones. 6 choices: synthesizer, harp, piano, orchestra ,slow jazz, flute. 30 second excerpts provided for choice. Phone unplugged, blinds closed, lights dimmed and door closed. Do not disturb sign placed on door. | ANXIETY: Anxiety about chair rest measured via 100 mm VAS. | Change scores extracted for anxiety. | |
| RCT; 2 parallel groups. | DESCRIPTION: 93 pre‐surgical out‐patients, to undergo elective surgery (ear‐nose‐throat; orthopedics; plastics; or other general minor surgery), in USA. | MUSIC GROUP: Music played for 30 minutes via headphones prior to surgery in a hospital isolation room. Participants brought their own choice of music from home. | ANXIETY: Measured via STAI before and after experiment. | Outcomes are expressed as a percentage of the baseline score (mean and SD). | |
| RCT; 2 parallel groups. | DESCRIPTION: 40 in‐patients in acute care unit for acute myocardial infarction in Midwest USA. | MUSIC GROUP: listened to 4 adagios selected by the investigator for 25 minutes. Primarily string composition, low‐pitched, simple and direct musical rhythm, tempo approximately 60 bpm. Played via headphones. | ANXIETY: state anxiety measured via STAI. | ||
| RCT; 3 parallel groups. | DESCRIPTION: 45 in‐patients who had an acute MI, in private rooms in an ICU in Midwest USA. | MUSIC GROUP: Experimenter selected classical music, played via headphones. Asked to assume a comfortable position in bed, lights lowered, telephones unplugged, curtains drawn, doors closed, advised to clear mind and let muscles relax. 20 minutes. | PHYSIOLOGICAL: HR (chart extraction), RR (auscultation with a stethoscope for 30 seconds), BP (non‐invasive automatic oscillometric BP cuff), measured pre, immediately post and at 1 and 2 hours post the intervention period. Heart rate variability determined using power spectral analysis and fast Fourier transform from 3 hours of continuous electrocardiographic data (30 minutes pre to 2 hours post intervention period). | Data presented as mean and SE. SEs converted into SDs for purposes of review. | |
| RCT; 2 parallel groups. | DESCRIPTION: 50 day patients scheduled for elective same‐day surgery of gynaecological procedures (e.g. exploratory laparoscopies, laparoscopic tubal ligation, ovarian cysts excision, and intrauterine device removal), in New Jersey, USA. | MUSIC GROUP: Choice of classical, country, jazz, popular, or show music, played via headphones. | ANXIETY: State and trait anxiety via STAI on entry and exit of holding area. Results reported as Mean +/‐ SEM. | Anxiety data presented as mean and SEs. SEs converted into SDs for purposes of review. | |
| RCT; cross‐over study (2 allocation groups). | DESCRIPTION: 20 ventilator‐dependent in‐patients in an ICU in Hong Kong, China. | MUSIC GROUP: Choice of 7 cassettes (Chinese folk song; Chinese instrumental music; Chinese music with western instruments; Buddhist music; Western classic; Western movie music; piano music). Participants instructed to close eyes and focus on the flow of the music. | STATE ANXIETY: short‐form STAI, patients held up appropriate number of fingers in response to questions. Taken before and after each intervention. | This cross‐over study has not reported patient‐specific differences between the two intervention measurements. Therefore, the data has been extracted as if it is independent groups (as there was not enough data to calculate a correlation coefficient). | |
| RCT; 2 parallel groups. | DESCRIPTION: 39 patients undergoing eye operations with anaesthetics and analgesics in China. | MUSIC GROUP: Given a choice of 3 musical styles (pop, light, classical) with each type including 4 songs. Patients were asked to listen to the music 2 to 3 times the night before the operation and relax (not clear if patients were in‐patients). Patients then listened to the music throughout their operation. | ANXIETY: STAI presented as change scores. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 66 pre‐operative surgical patients in the OR holding area in Hong Kong, China. | MUSIC GROUP: Slow music played via headphones for 20 minutes pre‐operatively. Participants had a choice of 3 tapes‐ Chinese instrumental, Western instrumental, or Western and Chinese slow songs. | ANXIETY: State anxiety measured via C‐STAI pre‐ and post‐intervention period. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 110 in‐patients undergoing abdominal hysterectomy in China. | MUSIC GROUP: Patient choice of music they felt to be calming and comforting, played via headphones intraoperatively, from 6 minutes before skin incision and 1 minute before the loading dose of anaesthesia to wound closure | PHYSIOLOGICAL: BP, HR, Bispectral Index recorded at 10‐minute intervals during surgery. Results reported as baseline and mean score during surgery (during target sedation period). | See Cepeda 2006 for details on music for pain relief (study not yet included in Cepeda 2006). | |
| RCT; 3 parallel groups. | DESCRIPTION: 75 in‐patients with suspected myocardial infarction in a coronary care unit in Midwest USA. | MUSIC GROUP: 30 minutes of self‐selected music choice of instrumental tapes: Halpern relaxation tape, classical music, country and western. Played via headphones. | ANXIETY: state anxiety measured via STAI | ||
| RCT; 2 parallel groups. | DESCRIPTION: 40 in‐patients with chronic cancer pain in acute care, Midwestern USA. | MUSIC GROUP: Choice of 10 relaxing instrumental tapes. Participants without a preference were given a Halpern antifrantic tape. The researcher suggested to the participant that the music would help them relax and reduce their pain. Participants listened to the music for 30 minutes via headphones. Participants lay on their beds and the lights were dimmed. | PAIN: Measured before and after 30 minute test period via McGill Pain Questionnaire (Pain Rating Index, Number of Words Chosen, and Present Pain Index). Additionally, pain intensity was measured via a 10 mm [sic] VAS. | See Cepeda 2006 for details on music for pain relief. |
BP: blood pressure; HR: heart rate; PACU: post‐operative care unit; POMS: Profile of Mood States; RR: respiration rate; SD: standard deviation; STAI: State Trait Anxiety Inventory; VAS: visual analogue scale; VRS: verbal rating scale;
Typically the music intervention consisted of the patient choosing from a selection of music compilations offered by the researcher (N = 46). Some studies asked the patient to bring music from home to listen to (N = 7), whilst others offered a fixed condition chosen by the researcher (N = 32), with three of these being sounds of nature as opposed to music. Studies utilised a variety of comparison groups (with 11 studies utilising two comparison groups): standard care alone (N = 52); blank tape/headphones only (N = 25); white noise (N = 4); pre‐recorded operating room noise (N = 4); restful environment (N = 9); and audiovisual distraction (N = 2). Given that the type of comparison may influence the size or direction of treatment effect, each comparison is dealt with separately in the review.
Studies investigated the use of music on: patient‐reported anxiety (N = 52), patient‐reported pain (N = 32, see the Cochrane review Cepeda 2006, which has overlapping scope), blood pressure (N = 43), heart rate (N = 45), respiration rate (N = 20), pain‐medication requirements (N = 23, see also Cepeda 2006), anxiolytic medication requirements (N = 5), peripheral skin temperature (N = 4), skin conductance (N = 2), oxygen saturation (N = 5), requirement of oxygen supplementation (N = 1), lung function (N = 1), blood flow (N = 1), heart rate variability (N = 1), bispectral index (N = 3), stress hormones (N = 6), mood (N = 4), depression (N = 1), sleep quality (N = 2), headache (N = 1), fatigue (N = 4), urinary problems (N = 1), well‐being (N = 2), nausea (N = 5), uncertainty (N = 1), satisfaction (N = 5), length of stay (N = 8), time to mobilisation (N = 1), induction time of sedation (N = 1), activities of daily living (N = 1), and interleukins (N = 1).
We do not report on music for pain in this review since it has already been covered elsewhere (Cepeda 2006).
We have tabulated 138 excluded studies on music (Table 9;, Table 10, Table 11), including 24 CCTs which were excluded following review by our contact editor (Table 9; see section on 'Differences between protocol and review').
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| CCT; 2 parallel groups; | DESCRIPTION: 42 pre‐operative patients scheduled for ambulatory surgery in a Midwestern city hospital, USA. | MUSIC GROUP: Choice of 20 tapes (classical, environmental, new age, country‐western, general easy listening), played via headphones whilst resting in recliner chairs. Intervention lasted 15‐30 minutes depending on how long patient had left before surgery. | PHYSIOLOGICAL MEASURES: measured using "standard noninvasive technology" before and after the intervention period. | ||
| CCT; 2 parallel groups; | DESCRIPTION: 301 outpatients and inpatients undergoing colonoscopy or esophagogastroduodenoscopy (EGD) in Switzerland. | MUSIC GROUP: Patients could choose between "light" and "classical" music which was played in the background during the examination. | GENERAL EVALUATION: Not extracted for review. | Paper reports outcomes as medians. Authors provided means and standard deviations on request for purposes of review. | |
| CCT; 2 parallel groups; | DESCRIPTION: 44 pre‐operative in‐patients in the holding area before total joint replacement, in Florida, USA. | MUSIC GROUP: Listened to music via headphones for a minimum of 30 minutes in the pre‐operative holding area. Music consisted of music‐enhanced nature sounds (sea, thunder, rainstorms, wind, and waterfalls). | The paper provides descriptive statistics only. | ||
| CCT; 2 parallel groups; | DESCRIPTION: 49 out‐patients undergoing bronchoscopy, USA. | MUSIC GROUP: Played new wave music 'Reflections of Passion' by Yanni, via headphones for the duration of bronchoscopy procedure. | PHYSIOLOGICAL: Methods of obtaining data unclear. There were no significant differences between groups on any of the physiological parameters. | ||
| CCT; 2 parallel groups; | DESCRIPTION: 158 inpatients with cardiovascular and respiratory illnesses participating in a therapeutic exercise programme, in the Republic of Uzbekistan. | MUSIC GROUP: Were played classical and pop music during the two week exercise programme. | PULSE: Assessed increase in pulse rate during exercise in the first and second weeks of the exercise programme. Method of data collection unclear. In week one, the music group worked harder, as demonstrated by increased pulse. | Authors conclude that cardiovascular patients should use music from the 1st week of exercise, and respiratory patients should use music from the 2nd week of exercise for better adaptation. | |
| CCT; 2 parallel groups; | DESCRIPTION: 24 pre‐operative day surgery patients waiting for surgery under general anaesthesia, in Texas, USA. | MUSIC GROUP: Choice of "easy listening" music selections recorded by the medical staff played via headphones for 20 minutes pre‐operatively. | ANXIETY: Measured via the STAI and a VAS. There was no significant difference between groups on either measure. STAI: | ||
| CCT; 2 parallel groups; | DESCRIPTION: 65 in‐patients undergoing rehabilitation (physiotherapy, balneotherapy, re‐education, and physical exercise) for lower back pain, France. | MUSIC GROUP: Music provided for first 4 days of 12 day hospitalisation. Music played for 20 minutes in the afternoon after physical therapy via headphones in a silent room. Patients were given a choice of music, with each choice arranged to have progressive relaxation with re‐awakening period at the end. | PAIN: Measured via VAS at baseline, day 5, and day 12, plus immediately pre and post therapy sessions. Day 5 outcomes did not significantly differ between groups: | ||
| CCT; 2 parallel groups; | DESCRIPTION: 101 in‐patients and out‐patients waiting for cardiac catheterizations in a cardiac telemetry unit, USA. | MUSIC GROUP: Listened to 20 minutes of 'Trance‐Zendance' by Halpern. Played via headphones prior to cardiac catheterization. | ANXIETY: Measured via STAI pre‐ and post‐ 20‐minute intervention period. There was a significant difference in anxiety scores in favour of the music group: | ||
| CCT; 2 parallel groups; | DESCRIPTION: 20 pre‐operative patients scheduled for breast biopsy in a holding area in Kentucky, USA. | MUSIC GROUP: Choice from selection of "new age" music listened to via headphones for 20 minutes pre‐operatively. No other music types were offered and patients' preferences were not solicited. | ANXIETY: Measured via the STAI pre‐ and post‐ 20‐minute study period. | ||
| CCT; 2 parallel groups; | DESCRIPTION: 34 in‐patients undergoing elective lumbar microdiscectomy procedures consented to participate (although only 10 analysed), Kentucky, USA. | MUSIC GROUP: Participants had a choice of 3 cassettes (country, instrumental, classical), with music 60‐80 bpm, played via headphones at a volume pre‐selected by participants. Music began 30 minutes before the end of the surgery and continued without interruption for one hour in PACU. | No data reported in paper. The paper reports there were no differences between the two groups on any of the outcome measures. | ||
| CCT; 2 parallel groups; | DESCRIPTION: 33 out‐patients awaiting arthroscopic procedures in Philadelphia, USA. | MUSIC GROUP: 20 minutes of classical music (tape 3 of Music Rx, developed by Bonny), played via audiocassette played placed 1 foot away during the waiting period. | BP: Measured via a Dinamap monitor before and after the 20 minute study period. There were no significant differences between the groups. Standard deviations are not provided. | ||
| CCT; 2 parallel groups; | DESCRIPTION: 113 pre‐operative day patients undergoing cytoscopy, cauterisation, or endoscopy, in China (Hong Kong).NUMBERS: Music group = 58, Control group = 55.AGE, mean (SD): Music group = 50.0 (15.5), Control group = 51.9 (14.4) years old.GENDER (male/female): Music group = 31/27, Control group = 27/28.ETHNICITY: not described.INCLUSION CRITERIA: 18 years or older; undertaking noninvasive day procedures with regional or local anaesthetic.EXCLUSION CRITERIA: Cognitive disability; hearing impairment; received preoperative sedatives; received a colon preparation; pre‐existing co‐morbid illness; did not have sufficient time to participate. | MUSIC GROUP: Choice of eastern and western style easy listening music and Chinese pop music (10 CDs and 10 mini‐discs) played via headphones in reclining chairs for 20‐40 minutes pre‐operatively.CONTROL GROUP: Undertook usual pre‐procedural relaxing activities (e.g. reading, watching TV) in the waiting room. | STATE ANXIETY: Measured via STAI pre‐ and post‐intervention period. Only the music group had a significant drop in anxiety. Post‐intervention scores:Music group = 42.5 (5.7) Control group = 46.4 (6.5) PHYSIOLOGICAL OUTCOMES: Measured via "standard non‐invasive instruments" pre‐ and post‐intervention period. There were no significant differences between groups.BP scores: Music group: Systolic BP = 124.2 (21.1), Diastolic BP = 70.0 (10.8) Control group: Systolic BP = 129.4 (25.6), Diastolic BP = 72.0 (11.1) PULSE scores: Music group = 71.1 (10.4) Control group = 70.1 (8.6) RR:Music group = 16.6 (1.0) Control group = 16.7 (1.0) | ||
| CCT; 2 groups (non‐parallel), intervention group was cross‐over (2 types of music played in same order for all participants), control group may have been recruited post‐hoc (UNCLEAR). | DESCRIPTION: 45 patients with disease of the lungs or chest being treated at the Department of Chest Surgery in Zakopane, Poland. | MUSIC GROUP: Cross‐over study. Music was played via headphones. Patients were played Debussy's 3rd part of Bergamasque Suite (Clair de Lune) on the piano (soothing music), and then Bartok's Wonderful Mandarin (nerve racking/exciting music). First, there was a 3‐minute rest in the recumbent position then, before, in between, and after each musical piece was a 3‐minute pause. | Parameters were measured at 5 time points in the music group: 1) Following 3‐min rest; 2) Following soothing music; 3) Following 3‐min pause; 4) Following exciting music; 5) Following 3‐min pause. Parameters were measured at 4‐time points in the control group (to follow same time scale as the music group)‐ a final 5th measurement was not taken for the control group. | Data not sufficient for extraction. No SDs reported. Paper reports significance as P < 0.1. | |
| CCT; 2 groups (non‐parallel), intervention group was cross‐over (2 types of music played in same order for all participants), control group may have been recruited post‐hoc (UNCLEAR). | DESCRIPTION: 45 patients with disease of the lungs or chest being treated at the Department of Chest Surgery in Zakopane, Poland. | MUSIC GROUP: Cross‐over study. Music was played via headphones. Patients were played Debussy's 3rd part of Bergamasque Suite (Clair de Lune) on the piano (soothing music), and then Bartok's finale of the Wonderful Mandarin (nerve racking/exciting music). First there was a 3‐minute rest in the recumbent position, then a 3‐minute pause before the first musical piece. After the first piece of music, there was a 4‐minute pause before the second musical piece was played. | Parameters of airway resistance were measured at three time points in both groups: 1) Following 3‐min rest; 2) Following relaxing music; 3) Following exciting music. | Data not sufficient for extraction. | |
| CCT; 2 parallel groups; | DESCRIPTION: 80 patients undergoing minor surgery in a day‐surgery ward in China. | MUSIC GROUP: Choice of 3 types of music with slow rhythms: classical music (concertos and sonatas), contemporary popular music (e.g. "The heart will go on"), Chinese popular music (e.g. "Night plane"). Music had 45 minutes running time and played via headphones for duration of surgery. | ANXIETY: measured via C‐STAI post‐surgery. Patients were asked to fill out the questionnaire by thinking retrospectively over procedure. | ||
| CCT; 2 parallel groups; | DESCRIPTION: 17 day‐patients undergoing arthroscopic surgery, USA. | MUSIC GROUP: Choice of 4 musical tapes (classical tapes from 'Music Rx', Bonny; popular tapes from 'Music Rx'; New Age tape by Steven Halpern called "Dawn"; easy listening selections assembled by investigator). Music played via headset and auto‐reverse cassette player. From administration of pre‐operative medication to PACU. Participants were told to restart the music if they desired after their return to the ambulatory surgery unit. | ANXIETY: state anxiety measured via STAI approximately 2 hours post‐operatively. Non‐significant. | SDs are estimated from the t‐value. | |
| CCT; 2 parallel groups; | DESCRIPTION: 63 patients undergoing dialysis, Florida, USA. | MUSIC GROUP: Choice of classical, pop, rock, jazz, country/western, gospel, easy listening, swing, and bluegrass played via headphones during dialysis treatment. Music was played for 1 hour beginning 30 minutes after the onset of dialysis treatment, there then was an hour of no music, followed by another hour of music. | BP: Measured after each hour of dialysis treatment. BP was recorded daily from each patient's chart for a 2‐week baseline period and a 3‐week treatment period. The two groups did not significantly differ on systolic and diastolic BP readings for onset through final readings during the treatment period. | Data not sufficient for extraction. | |
| CCT; 2 parallel groups; | DESCRIPTION: 88 pregnant women attending a high‐risk antenatal clinic, London, UK. | MUSIC GROUP: One or two musicians playing in one corner of the waiting room with the chairs arranged in a semi‐circle around them. Harp, clarinet, or guitar were preferred, violin and cello not welcomed. Does not explain in depth what was played. | BP: Obtained by the clinician (method unclear) at the beginning of the consultation (after the waiting room experience). There was no significant difference between the groups. | This is one of a series of studies. Two more studies were conducted in the antenatal clinic that do not need inclusion criteria for the review due to the study design (before and after). | |
| CCT; 2 parallel groups; | DESCRIPTION: 9 in‐patients waiting for elective surgery in a theatre holding area, China. | MUSIC GROUP: Choice of: slow rhythmical songs; Chinese slow rhythmical music; Western slow rhythmical music. Played via headphones for 20 minutes pre‐operatively. | BP: Measured via a calibrated Dinamap BP monitor before and after 20 minute study period. No significant differences between groups. | ||
| CCT; 3 parallel groups; | DESCRIPTION: 76 patients presenting to an emergency department with minor musculoskeletal trauma, in Midwest USA. | MUSIC GROUP: Provided with a Walkman tape player and choice of music (classical, country, rock, pop, and jazz), or were allowed to listen to the radio if they preferred. Patients also received standard care. | PAIN: Pain intensity measured at 0, 30, and 60 minutes via 10‐point NRS. There was no statistical differences between groups at any time interval. All groups showed significant improvement from baseline. | No SDs provided. | |
| CCT; 2 parallel groups; | DESCRIPTION: 57 post‐operative in‐patients who have undergone elective nasal surgery. | MUSIC GROUP: Choice of Chinese and Western various music types and patients encouraged to bring music of their own choice. Listened for 30 minutes on four occasions: post‐operatively (T1), again 4 hours later (T2), on the first post‐operative day at 8am (T3) and again at noon (T4). | PAIN: Measured via VRS at baseline and after each intervention session. Music group gave significantly lower pain ratings at all four time points. T4 data: | ||
| CCT; 2 parallel groups; | DESCRIPTION: 60 post‐operative in‐patients after coronary artery bypass graft in a progressive care area, USA. | SOUNDS GROUP: Marsona Sound Conditioner (providing white noise in the form of rain, ocean waves, or a waterfall). 56/60 patients chose ocean sounds. Sounds were played at the bedside throughout the night (switched on between 20.30 and 21.00 hours). Sounds played for 3 nights. | SLEEP: Pre‐test evaluated for patients on their usual sleep at home, data collected on admission, prior to surgery. Assessed on the fourth day post‐transfer for the quality of sleep the previous night. Measured via the Richards‐Campbell Sleep Questionnaire, which includes VAS for sleep depth, latency to sleep onset, awakening, return to sleep, quality of sleep, and a total sleep score. No SDs reported. | ||
| CCT; 2 parallel groups; | DESCRIPTION: 50 patients undergoing surgery with local anaesthetic, Poland. | MUSIC GROUP: Played music before during and after surgery. Patients chose music from a selection of instrumental tapes (music was soft, bright, avoiding high‐pitched sounds, classical). | ANXIETY: Measured via STAI the day before and directly after surgery in the recovery room. | Data not in sufficient detail for extraction. | |
| CCT; 2 parallel groups; | DESCRIPTION: 57 in‐patients undergoing surgery with local anaesthetic, Japan. | MUSIC GROUP: Were played a set piece (by Elgar) from when they entered the theatre to the end of surgery. | ANXIETY: Measured via STAI before and after surgery. Participants were requested to think back to how they felt during surgery in the post‐treatment questionnaire. | ||
| CCT; 3 parallel groups; | DESCRIPTION: 30 pre‐operative patients waiting for transurethral resection of the prostate, in a theatre holding area, Hong Kong, China. | MUSIC GROUP: 20 minutes of slow rhythm soft music played pre‐operatively via headphones. Choice of three tapes that had been judged by a panel of 3 musicians (slow rhythm songs, Chinese slow rhythm music, Western slow rhythm music). No nurse presence. | PHYSIOLOGICAL: BP and HR recorded on an automated BP monitor (Dinamap 1846‐SX) before and after 20 minute study period. Paper reports no significant differences between groups (Kruskal‐Wallis test) |
BP: blood pressure; HR: heart rate; RR: respiration rate; STAI: State Trait Anxiety Inventory; VAS: visual analogue scale
| Study ID | Reason for exclusion |
| No data presented‐ summary paper | |
| Validity of outcomes | |
| Setting | |
| Validity of outcomes | |
| Outcomes not reported for relevant groups; data collection methods unclear. | |
| Study design | |
| Intervention provided outside of hospital | |
| Study design | |
| Setting and population | |
| Intervention | |
| Intervention | |
| Intervention‐ relaxation technique | |
| Intervention | |
| Questionable validity of outcome, relevant data not presented | |
| Setting | |
| Setting | |
| Qualitative report | |
| Outcome measure | |
| Outcomes | |
| Setting | |
| Intervention | |
| Setting | |
| Intervention not well defined | |
| Intervention provided during education programme | |
| Study design | |
| Music therapist confound; group differences in timing of data collection | |
| Intervention included coaching of participants | |
| Setting not hospital | |
| Study design | |
| Study design | |
| Intervention interactive music therapy | |
| Interventions not suitable for inclusion | |
| Invention group provided 20mins coaching. | |
| Intervention group provided reinforcement and training‐ bias | |
| Intervention group provided coaching on relaxing | |
| Secondary analysis of previous study | |
| Secondary analysis of previous study | |
| Secondary analysis of previous study | |
| Setting; qualitative | |
| Qualitative | |
| Relaxation (psychological) technique use with intervention | |
| Outcomes not health‐related | |
| Intervention‐ taught techniques | |
| Outcomes | |
| Case study | |
| Intervention not well defined | |
| Intervention not well defined | |
| Policy confound (restraints use) | |
| Policy confound (restraints use) | |
| Policy confound (restraints use) | |
| Outcome measure | |
| Outcome measure | |
| Study design | |
| Conference abstract‐ not enough detail | |
| Study design | |
| Intervention not well defined | |
| Data unsuitable for cross‐over study | |
| Setting | |
| Setting | |
| Intervention not well defined | |
| Validity of outcomes | |
| Study design‐ selection of participants by matching, different wards assigned to different conditions. | |
| Unable to clarify discrepancies in data with author | |
| Setting | |
| Setting; Duplicate | |
| Intervention began before admission | |
| Unclear methods and data | |
| Study design | |
| Intervention not well described (CCT) | |
| Intervention not well described (CCT) | |
| Outcomes not validated/reliable | |
| Study design | |
| Lack of information | |
| Confounding | |
| Intervention not well described | |
| Intervention not well described | |
| Intervention not well described, no data presented (CCT) | |
| Setting |
| Study ID | Reason for exclusion |
| Study design | |
| Study design | |
| Setting | |
| Inappropriate methods and analysis | |
| Outcomes not patient‐related | |
| Intervention | |
| Patients allocated retrospectively to music or control, depending on whether or not they had noticed or heard any music playing (unpublished information). | |
| Intervention not well described | |
| Data collection tool not validated | |
| Outcomes not validated | |
| Policy change‐ staff asked to remain quiet on intervention days | |
| Therapy as treatment | |
| Setting | |
| Overview‐ insufficient detail | |
| Music not well described; cross‐over trial with and with‐out vibration | |
| Setting | |
| Study design | |
| Study design | |
| Study design | |
| Study design | |
| Intervention not well defined | |
| Outcomes not validated | |
| Study design | |
| Setting | |
| Intervention | |
| Outcomes not validated | |
| Study design | |
| Study design; outcomes | |
| Insufficient information (intervention and data) | |
| Study design | |
| Study design | |
| Test (not clinical) situation, healthy controls | |
| Setting | |
| Intervention not well defined; validity of outcomes unclear | |
| Inappropriate control‐ drugs | |
| Duplicate study (Lee 2005) |
Below, we summarise findings for the following comparisons.
-
Music versus blank tape/headphones alone
-
Music versus pre‐recorded operating room noise
-
Music versus scheduled rest
-
Music versus standard care alone
-
Music versus white noise
Risk of bias in included studies on music:
Of the 85 RCTs, method of randomisation was unclear in 42 studies. The remaining studies allocated participants to groups via a computer‐generated sequence (N = 19), a table of random numbers (N = 11), drawing lots (N = 9), or coin flips (N = 4). Reporting of allocation concealment was poor with only 11 of 85 studies reporting adequate allocation concealment. Five RCTs had inadequate allocation concealment and 69 studies remain unclear. Blinding to group allocation was achieved in some studies that utilised a control condition involving headphones. Blinding of healthcare personnel and/or the data collector was reported in 20 studies, and double‐blinding of patients and staff was achieved in four studies which investigated the use of music during surgery with general anaesthesia. For studies comparing music to standard care, blinding to group allocation was not possible. Blinded assessment of patient‐reported outcomes was generally not possible for music interventions. Other outcomes, such as physiological measures, were reported as automated in 12 studies. Blinded assessment of at least one outcome was achieved in 16 studies. For 43 studies, no blinding or automation of any outcome measure was achieved, and in six studies it was unclear whether any of the outcomes were blinded or automated. The remaining studies had outcomes rated as a mixture of 'not done' and 'unclear' for blinded assessment.
Completeness of data (i.e. outcomes obtained for > 80% of participants) was also poorly reported, with 49 studies scoring as 'done' on this measure (31 were unclear). Five RCTs had incomplete (< 80%) data for at least one outcome measure. Reporting of withdrawals and drop‐outs was not done in the majority of studies (N = 47). In the 38 studies which did report withdrawals and drop‐outs, attrition ranged from zero to 33 participants per study, amounting to 235 withdrawals. Only 22 studies specified from which groups participants withdrew, and overall in these instances the music and control groups had similar attrition (43 and 49 participants respectively; missing information on 143 withdrawals). When studies are weighted according to final sample size, four RCTs arise as outliers in the number of withdrawals (Broscious 1999; Korunka 1992; Phumdoung 2003; Twiss 2006). Korunka 1992 withdrew 23 participants due to missing data (unclear from which groups), and reasons for the withdrawals (N = 92) from the other three studies are largely unrelated to the intervention itself.
Protection against contamination did not appear to be a problem for most studies (N = 80). Two RCTs were cross‐over designs so this could not be achieved. Three RCTs had unclear protection against contamination.
Findings from studies on music versus blank tape/headphones alone:
Anxiety
Eleven RCTs reported sufficient data for extraction on the outcome anxiety (with a total sample of 891 participants: 455 in the music group, and 436 in the control group; Analysis 3.1). Four studies investigated music in the pre‐procedure period (Cooke 2005; Guo 2005; Ikonomidou 2004; Wang 2002), four during a medical procedure (Andrada 2004; Colt 1999; Domar 2005; Mandle 1990), three post‐operatively (Ikonomidou 2004; Nilsson 2003b; Nilsson 2005), and one in the intensive care unit (Lee 2005). Ikonomidou 2004 investigated the use of music in the same patients both pre‐operatively and post‐operatively. Andrada 2004 reports change scores and as such it is inappropriate to combine the data of this study in a meta‐analyses where we are using the standardised mean difference (SMD).
Studies conducted in the pre‐procedure period were fairly homogenous (I2 = 0%) and were in favour of the use of music as compared with headphones only (SMD −0.82, 95% CI −1.03 to −0.60, P value < 0.00001). This translates into a mean difference of approximately −9.58 points (95% CI −12.04 to −7.02) on the State Trait Anxiety Inventory (assuming an SD of 11.64; Millar 1995), approximately −23.2 mm (95% CI −29.2 to −17.0) on a 100 mm VAS (assuming an SD of 28.35; Millar 1995), or approximately −3.8 points (95% CI −4.7 to −2.7) on the Hospital Anxiety and Depression scale (assuming an SD of 4.59; Millar 1995).
Studies conducted during medical procedures were also fairly homogenous (I2 = 26%), however, once combined these studies (with 91 and 92 participants in the music and headphones groups respectively) show no strong evidence of an effect when comparing music with headphones alone (SMD −0.12, 95% CI −0.47 to 0.23, P value = 0.49). Not included in this estimate is Andrada 2004 (with 63 and 55 participants in the music and headphones groups respectively); this study showed a significant difference in change scores (MD −5.08, 95% CI −9.04 to −1.12). If Andrada 2004 were to be included in the analysis, it is likely therefore to increase the heterogeneity and modify the effect estimate somewhat in favour of the music group.
Studies conducted during the post‐operative period showed no strong evidence of an effect on average, however these studies, which straddle the line of no effect, have substantial statistical heterogeneity (I2 = 66%). The one study conducted in the intensive care unit also showed no strong evidence of an effect between groups (SMD 0.17, 95% CI −0.32 to 0.66, P value = 0.50).
Due to the observed differences between subgroups, we have not combined all studies in a meta‐analysis, which would reveal substantial heterogeneity (I2 = 74% or 76%, depending on whether the pre‐procedure or post‐procedure data from Ikonomidou 2004 are used).
Exploring the heterogeneity of studies on anxiety:
Conducting a sensitivity analyses to remove studies with higher risk of bias (removing those with unclear or inadequate allocation concealment) leaves just two studies (Cooke 2005; Lee 2005), one of which has findings in favour of music (Cooke 2005: SMD −0.86, 95% CI −1.24 to −0.49) and the other found no strong evidence for an intervention effect (Lee 2005: SMD 0.17, 95% CI −0.32 to 0.66).
Grouping the studies by reason for hospitalisation (seven studies = surgery, two studies = endoscopy, one study = non‐invasive intervention, and one study = ICU), or geographical location (four studies = USA, three studies = Sweden, two studies = China, one study = Spain,and one study = Australia) does not help explain the heterogeneity. All studies were conducted on non‐psychiatric populations.
A post‐hoc analysis shows that four studies (Cooke 2005; Domar 2005; Guo 2005; Lee 2005) provided patients with a choice of music to listen to (I2 = 81.2%), one asked patients to bring music from home (Wang 2002), and the remaining six (five without Andrada 2004) provided set pieces of music to listen to (I2 = 64%, 52%, or 62%, depending on whether the pre‐procedure, post‐procedure, or no data from Ikonomidou 2004 are included). Grouping the studies in this way does little to explain the heterogeneity.
Studies with insufficient data for extraction:
One other RCT also investigated anxiety (Nilsson 2003a); this study was conducted in 151 day patients undergoing surgery with general anaesthetic. This study found no strong evidence of an effect between groups (which were either played music intra‐operatively only, post‐operatively only, or not at all) for anxiety. This finding is in concordance with the studies outlined above that were also conducted in the post‐operative period or during surgery.
Heart rate
Eight RCTs reported sufficient data for extraction on the outcome heart rate (Analysis 3.2). Davis‐Rollans 1987 is a cross‐over study (with 24 participants) presenting individual patient data. The standard deviations for this study have been adjusted to account for the correlation co‐efficient. When combined these studies include 519 participants (music group = 276, control group = 267) with consistent findings (I2 = 0%), showing no strong evidence of an intervention effect (MD 0.40 bpm, 95% CI −1.02 to 1.82, P value = 0.58). A sensitivity analysis conducted without Davis‐Rollans 1987 (495 participants, music group = 252, control group = 243), does little to alter the overall results (I2 = 0%, MD −0.04 bpm, 95% CI −1.86 to 1.95, P value = 0.96).
Studies with insufficient data for extraction:
Four other RCTs also investigated heart rate (Chlan 1995; Heitz 1992; Mandle 1990; Tsuchiya 2003). These studies included 149 participants (music group = 74, control group = 75). Chlan 1995 reported that the findings were significantly in favour of the music group; however this was a small study with only 11 participants in the music group and nine in the control group. Tsuchiya 2003 also reported a significant difference in favour of the music group at one time point (at extubation), however five other time points (start of surgery, at gallbladder removal, at the end of surgery, at the end of anaesthesia, and in the post‐operative care unit (PACU) demonstrated no strong evidence of an effect. The remaining two studies showed no strong evidence of an effect between groups. These findings typically support the findings of the RCTs reported above.
Blood pressure
Eight RCTs reported sufficient data for extraction on the outcome blood pressure (Analysis 3.3). Of these studies, seven reported systolic blood pressure (music group = 271, control group = 262 participants), six reported diastolic blood pressure (music group = 242, control group = 236 participants), and one reported arterial blood pressure (15 participants per group). The studies are statistically homogenous (I2 = 0%), and demonstrate no strong evidence that music (when compared with headphones only/blank tape) can help reduce systolic blood pressure (MD −0.40 mm Hg, 95% CI −2.48 to 1.67, P value = 0.70), diastolic blood pressure (MD −0.35 mm Hg, 95% CI −2.08 to 1.39, P value = 0.69), or arterial blood pressure (MD 4.00 mm Hg, 95% CI −5.33 to 13.33, P value = 0.40).
Studies with insufficient data for extraction:
Four other RCTs also investigated blood pressure (Chlan 1995; Heitz 1992; Mandle 1990; Tsuchiya 2003). These studies all found no strong evidence of an effect between groups, apart from Tsuchiya 2003 who found that the control group had significantly higher blood pressure at extubation than the music group. In Tsuchiya 2003, five other time points (as described in the findings for heart rate) did not demonstrate any evidence of an effect. These findings are in concordance with the RCTs described above.
Respiration rate
Two RCTs (Lee 2005; Ikonomidou 2004) reported sufficient data for extraction on the outcome respiration rate (Analysis 3.4). Combined, these studies investigated music on 119 participants (music group = 61, control group = 58). With low heterogeneity (I2 = 13.0%), these studies are in favour of music (MD −1.72 breaths/min, 95% CI −3.00 to −0.44, P value = 0.008). However, there are multiple ways to extract and analyse the data presented in these studies (see Analysis 3.4), and given the concerns of risk of bias and the data presented in Ikonomidou 2004 (see Table 8), it can be concluded that there is no strong evidence for an effect of music on respiration rate when compared to blank tape/headphones alone.
Studies with insufficient data for extraction:
Three other RCTs also investigated respiration rate (Chlan 1995; Davis‐Rollans 1987; Heitz 1992). One small study, Chlan 1995 (with 11 participants in the music group and nine in the control group), reports significant findings in favour of music. Heitz 1992 (N = 40), and Davis‐Rollans 1987 (cross‐over of 24 participants) reported no strong evidence for an effect. These findings further contribute to the conclusion reported above that there is currently no strong evidence to support the use of music for reducing respiration rate when compared to a blank tape or headphones alone.
Anxiolytic medication requirements
One RCT reported sufficient data for extraction on the outcome anxiolytic medication requirements (Mandle 1990). In this study of 30 patients (music group = 14, control group = 16), there was no strong evidence for an effect between groups on diazepam consumption (MD −0.50 mg, 95% CI −2.08 to 1.08, P value = 0.53).
Studies with insufficient data for extraction:
One other RCT also investigated anxiolytic medication requirements (Harikumar 2006). This study of 78 participants (music group = 38, control group = 40) reports that the control group received significantly more midazolam than the music group.
Other outcomes
Some studies reported on other outcomes (Table 12). For this set of outcomes, significant differences were found in favour of music for well‐being and induction time of sedation. No strong evidence of an intervention effect was found for abnormal events, activities of daily living, airway pressure, bispectral index, fatigue, headache, interleukins, mood, nausea, skin conductance, stress hormones, and urinary problems. Mixed evidence was found for heart rate variability characteristics, length of stay, oxygen saturation, and satisfaction. It should be noted that for many of these outcomes, only one or two studies are reported and in cases there is a high risk of bias.
| Outcome | Detailed RCTs (N) | Participants (N) | Heterogeneity (%) | Results | Other RCTs | Participants (N) | Findings |
| Oxygen saturation | N = 1 | Total = 50 | N/A | MD 1.60, 95% CI 0.05 to 3.15, P value = 0.04 | N = 1 | Total = 20 | No significant difference between groups. |
| Airway pressure | None | N/A | N/A | N/A | N = 1 | Total = 20 | No significant differences between groups. |
| Skin conductance | Total = 186 | I2 = 81.7% (one study significant, one study non‐significant) | SMD 0.13, 95% CI −0.55 to 0.80, P value = 0.71 (Analysis 3.5). | None | N/A | N/A | |
| Heart rate variability (RR intervals, low and high frequency bands, total power, low/high frequency ratio) | None | N/A | N/A | N/A | N = 1 | Total = 68 | Heart rate variability data (high frequency power, logarithm of low frequency, and high/low frequency ratio) showed significant positive changes in the music group but not control group. Other variables (RR intervals, low and high frequency bands, low frequency power, and logarithm of high frequency) were non‐significant. |
| Bispectral index (mean, time to reach BIS 60) | N = 2 | Total = 160 | N/A (studies reported different outcomes) | Bispectral index: MD 1.00, 95% CI −1.27 to 3.27, P value = 0.39; | None | N/A | N/A |
| Stress hormones (Cortisol, epinephrine, norepinephrine, adrenocorticotropic hormone‐ ACTH) | N = 3 | Total = 216 | Cortisol I2 = 51.8% (3 studies) | Cortisol: SMD −0.32, 95% CI 0.73 to 0.09, P value = 0.13; | N = 1 | Total = 75 | No differences in cortisol levels at any time between groups. Change scores at 2 hours post‐operatively were significantly greater in the post‐operative music group than control. Blood glucose levels did not differ between groups at any time. |
| Mood | N = 1 | Total = 20 | N/A | MD −8.50, 95% CI −18.55 to 1.55, P value = 0.10 | None | N/A | N/A |
| Abnormal events (hypoxaemia, hypotension, hypertension, bradycardia, tachycardia, respiratory depression, pruritis) | N = 1 | Total = 118 | N/A | Cardio‐respiratory incidents = 0; | N = 1 | Total = 193 | Vomiting (Intra‐operatively/PACU): Music = 0/0%, Control = 2.2/0% Unclear missing data as presented in % values. |
| Headache | N = 1 | Total = 115 | N/A | MD 0.00, 95% CI −0.15 to 0.15, P value = 1.00 | None | N/A | N/A |
| Fatigue | N = 1 | Total = 115 | N/A | MD −0.30, 95% CI −0.78 to 0.18, P value 0.22 | N = 1 | Total = 151 | There were no significant differences between groups. |
| Urinary problems | N = 1 | Total = 115 | N/A | MD −0.10, 95% CI −0.40 to 0.20, P value = 0.51 | None | N/A | N/A |
| Well‐being | N = 1 | Total = 115 | N/A | MD 0.30, 95% CI 0.02 to 0.58, P value = 0.03 in favour of music group. | None | N/A | N/A |
| Nausea | N = 2 | Total = 308 | N/A (different methods of measurement) | MD −0.20, 95% CI −0.50 to 0.10, P value = 0.19 | N = 1 | Total = 151 | There were no significant differences between groups. |
| Satisfaction | N = 1 | Total = 110 | N/A | MD 1.60, 95% CI 1.29 to 1.91, P value < 0.00001 | N = 2 | Total = 344 | There were no significant differences between groups. |
| Length of stay | N = 1 | Total = 61 | N/A | MD 0.00, 95% CI −0.99 to 0.99, P value = 1.00 | N = 2 | Total = 118 | Harikumar 2006 reports that recovery time was significantly longer (difference in medians = 10 minutes) in the control group. Heitz 1992 reports no significant differences between groups. |
| Activities of daily living | N = 1 | Total = 61 | N/A | MD −0.30, 95% CI −2.63 to 2.03, P value 0.80 | None | N/A | N/A |
| Serum interleukins (IL‐6) | N = 1 | Total = 110 | N/A | MD −7.40, 95% CI −22.61 to 7.81, P value = 0.34 | None | N/A | N/A |
| Induction time of sedation (minutes) | N = 1 | Total 110 | N/A | MD −6.00, 95% CI −10.49 to −1.51, P value = 0.009 in favour of music. | None | N/A | N/A |
CI: confidence interval; MD: mean difference; SMD: standardised mean difference; PACU: post‐operative care unit; RR: respiration rate
Findings from studies on music versus pre‐recorded operating room noise:
Anxiety
Studies with insufficient data for extraction:
One RCT investigated anxiety (Cruise 1997; N = 62) and found no strong evidence for an intervention effect.
Heart rate
Studies with insufficient data for extraction:
One RCT investigated heart rate (Cruise 1997; N = 62) to find no strong evidence for an intervention effect.
Blood pressure
Studies with insufficient data for extraction:
One RCT investigated blood pressure (Cruise 1997; N = 62). This study found no strong evidence for an intervention effect on diastolic blood pressure. However, significant differences were observed for systolic blood pressure; in this study systolic blood pressure was unexpectedly increased in the music group (but not significantly in the operating room noise group) immediately after retrobulbar block, and at 15 and 30 minutes after the retrobulbar block the music group had higher systolic blood pressure than the operating room noise group (which decreased over the course of the surgery).
Respiration rate
Studies with insufficient data for extraction:
One RCT investigated respiration rate (Cruise 1997; N = 62) to find no strong evidence for an intervention effect.
Other outcomes
Other outcomes reported for the comparison music versus operating theatre noise were: length of stay (Ayoub 2005; Korunka 1992); time to mobilisation (Nilsson 2001); nausea (Nilsson 2001); fatigue (Nilsson 2001); and well‐being (Nilsson 2001). The two studies that reported length of stay have differing findings (Korunka 1992 reports results significantly in favour of music, whereas Ayoub 2005 reports no strong evidence for an effect), however these studies can not be combined in a meta‐analysis due to insufficient data for extraction. Nilsson 2001 found that those who received music were faster to mobilise to a sitting position than the control group however there was no difference between groups in the time it took to stand and walk. For the patient‐reported outcomes (fatigue, nausea, and well‐being) Nilsson 2001 found no strong evidence of an effect.
Findings from studies on music versus scheduled rest:
Anxiety
Eight RCTs (Barnason 1995/1996; Chlan 1998; Elliot 1994; Sendelbach 2006; Voss 2004; White 1992; White 1999; Wong 2001) investigated music versus scheduled rest for anxiety (Analysis 4.1). These studies (music =189, control = 187), which individually are either non‐significant or in favour of music, have substantial heterogeneity (I2 = 82%), and therefore we have not pooled the data.
Exploring the heterogeneity of studies on anxiety:
One study (Barnason 1995/1996) did not use allocation concealment, however removing this study in a sensitivity analysis does little to reduce the heterogeneity (I2 = 80%). Studies were all conducted in either intensive, critical, or coronary care units (in Voss 2004 patients were undergoing chair rest in a Surgical Intensive Care Unit), and on non‐psychiatric populations. Six of the studies were conducted in the USA (I2 = 71%) and the other two were conducted in Australia (Elliot 1994) and China (Wong 2001).
A post‐hoc subgroup analysis of the three studies which played patients in the intervention group set pieces of music (I2 = 58%; Elliot 1994; White 1992; White 1999), and the remaining studies which offered patients a choice of music from a selection (I2 = 84%; Barnason 1995/1996; Chlan 1998; Sendelbach 2006; Voss 2004; Wong 2001), does little to explain the heterogeneity.
Heart rate
Four studies (Elliot 1994; Sendelbach 2006; White 1992; White 1999) reported sufficient data for extraction on heart rate (Analysis 4.2). These studies demonstrated little statistical heterogeneity (I2 = 0%), and overall (music group = 94, control group = 86) found no strong evidence of an effect between groups (MD −2.76 bpm, 95% CI −6.65 to 1.13, P value = 0.16).
Studies with insufficient data for extraction:
Two other RCTs also reported on heart rate (Barnason 1995/1996; Chlan 1998). Barnason 1995/1996 (music group = 33, control group = 34) found no strong evidence of an effect between groups, and Chlan 1998 (music group = 27, control group = 27) reports a significant effect in favour of music for heart rate reduction.
Blood pressure
Three studies (Elliot 1994; Sendelbach 2006; White 1999) reported systolic blood pressure (music group = 74, control group = 66), two studies (Elliot 1994; Sendelbach 2006) reported diastolic blood pressure (music group = 59, control group = 51), and one study (Wong 2001) reported arterial blood pressure (music group = 20, control group = 20). It is possible that this outcome is subject to selective outcome reporting (as it could be expected that studies would have collected both systolic and diastolic blood pressure data). For each outcome, studies demonstrated little statistical heterogeneity (I2 = 0%; Analysis 4.3). Combined findings for systolic blood pressure showed no strong evidence of an effect (MD −1.51 mm Hg, 95% CI −6.65 to 3.63, P value = 0.56), but for diastolic blood pressure the findings were favour of music (MD −5.29 mm Hg, 95% CI −8.78 to −1.79, P value = 0.003). The one study that assessed arterial blood pressure showed no strong evidence of an effect (MD −4.75 mm Hg, 95% CI −13.98 to 4.48, P value = 0.31).
Studies with insufficient data for extraction:
One study (Barnason 1995/1996; music group = 33, control group = 34) also assessed blood pressure and found no strong evidence of an effect.
Respiration rate
Three RCTs (White 1992; White 1999; Wong 2001) measuring respiration rate (music group = 55, control group = 55), when combined (I2 = 0%) found a significant difference (MD −2.04 breaths/min, 95% CI −3.43 to −0.66, P value = 0.004) in favour of the music group (Analysis 4.4).
Studies with insufficient data for extraction:
Chlan 1998 also reported on respiration rate, and this study (with 27 patients in each group) also reported a significant effect in favour of the music group for respiration rate over the course of the experiment.
Other outcomes
One study also reported on sleep quality (Barnason 1995/1996; music group = 33, control group = 34) and found no strong evidence of an intervention effect (MD 1.17 points on the Richards‐Campbell Sleep Questionnaire, 95% CI 0.00 to 2.34, P value = 0.05).
Findings from studies on music versus standard care alone:
Anxiety
Twenty‐nine RCTs reporting sufficient data for extraction (with a total sample size of 1812 participants: music group = 916, control group = 896), investigated the use of music versus standard care alone for the reduction of anxiety (Analysis 5.1). Nine studies provided music in the waiting period prior to a medical procedure (Buffum 2006; Cooke 2005; Gaberson 1991; Gaberson 1995; Hayes 2003; Padmanabhan 2005; Taylor‐Piliae 2002; Winter 1994; Yung 2003), 15 studies provided the music throughout a medical procedure (Argstatter 2006; Bally 2003; Binnings 1987; Chan 2003; Chang 2005; Chlan 2000; Kwekkeboom 2003; Lembo 1998; Lepage 2001; McRee 2003; Mennegazzi 1991; Smith 2001; Smolen 2002; Voss 2004; Yang 2003), and five studies investigated the use of music in the post‐operative period or in an intensive care environment (Lueders Bolwerk 1990; Mullooly 1998; Twiss 2006; White 1999; Zimmerman 1988).
In a random‐effects analysis, all three subgroups showed average treatment effects in favour of music (pre‐procedural SMD −0.37, 95% CI −0.62 to −0.12; procedural SMD −0.69, 95% CI −1.02 to −0.36; ICU/post‐operative SMD −0.58, 95% CI −1.01 to −0.15), as did the subgroups combined (SMD −0.55, 95% CI −0.74 to −0.36). Individual studies generally point in favour of music, or show no difference between groups. However, as indicated by the Chi2 statistics in the analysis, heterogeneity is present; the subgroups contain moderate to substantial heterogeneity (pre‐procedural anxiety I2 = 59.3%, procedural anxiety I2 = 79.6%, ICU/post‐operative anxiety I2 = 54.5%), as do all the studies combined (total I2 = 72.0%). A funnel plot of the studies indicates there may be some evidence of publication bias, with a lack of small studies with small treatment effects (Figure 3).
Funnel plot of comparison: 5 Music versus standard care, outcome: 5.1 Anxiety.
Exploring the heterogeneity of studies on anxiety:
Removing all studies with unclear concealment of allocation from the analysis, as well as Voss 2004 in which there were significant differences at baseline for the outcome anxiety, leaves 5 studies (Bally 2003; Chan 2003; Cooke 2005; Smith 2001; Twiss 2006), which combine with moderate heterogeneity (I2 = 56.5%), to estimate a slightly lower, average treatment effect (SMD) of −0.46 (95% CI −0.74 to −0.19) in favour of music.
Sub‐grouping studies into reason for hospitalisation (surgery or percutaneous interventions (n = 20); endoscopic procedures (n = 5; Chan 2003; Chlan 2000; Hayes 2003; Lembo 1998; Smolen 2002); coronary care (n = 3; Lueders Bolwerk 1990; White 1999; Zimmerman 1988); and radiation therapy (n = 1; Smith 2001), yields largely heterogeneous groupings (I2 = 73.2%, 79.9%, and 73.7%, respectively), so does not appear an appropriate variable to explain the heterogeneity between studies. All studies included in this comparison were carried out on non‐psychiatric populations. The majority of studies were conducted in the USA (n = 19), four in China (Chan 2003; Taylor‐Piliae 2002; Yang 2003; Yung 2003), two in Canada (Bally 2003; Lepage 2001), one in each of Taiwan (Chang 2005), Germany (Argstatter 2006), England (Padmanabhan 2005), and Australia (Cooke 2005). The two Canadian studies showed no strong evidence of a treatment effect (SMD −0.16, 95% CI −0.48 to 0.15, I2 = 0%), whereas studies from other locations achieved mixed results, favouring music overall but with substantial heterogeneity. It is likely therefore that location has little to do with the size of treatment effects for this comparison.
A post‐hoc subgroup analysis of all studies according to whether patients were provided with a choice of music or were provided with a set piece(s) selected by the researcher, suggests that providing patients with set pieces increases the average effect size. Nine studies (Argstatter 2006; Binnings 1987; Gaberson 1991; Gaberson 1995; Lembo 1998; Lueders Bolwerk 1990; McRee 2003; Padmanabhan 2005; White 1999) with some heterogeneity (I2 = 49.9%) provided set pieces (SMD −0.83, 95% CI −1.19 to −0.47, P value < 0.00001), and the remaining 20 studies (I2 = 74.8%) offered patients a choice of music (SMD −0.45, 95% CI −0.66 to −0.23, P value < 0.00001). The five studies listed above with adequate allocation concealment however, all offered patients a choice of music from the researchers' selection, which may confound the observation made here.
Studies with insufficient data for extraction:
A further five RCTs investigated music versus standard care for anxiety. One study conducted during the waiting period before a medical procedure favoured music (Daub 1988), in line with the findings above. Four studies investigated the use of music during a procedure (non‐invasive/rehabilitation = Ferguson 2004 and Nowobilski 2005; endoscopic = Palakanis 1994; percutaneous = Schneider 2001). Two of these studies (with 36 and 30 participants respectively) found no differences between groups (Nowobilski 2005; Schneider 2001). One study with 50 participants found significant differences in favour of music (Palakanis 1994), and one also favoured music (Ferguson 2004; with 11 participants) although the control group were also more anxious at baseline. These mixed results reflect the heterogeneity observed in the meta‐analysis above.
Heart rate
Twenty‐one RCTs reporting sufficient data for extraction (with a total sample size of 1653 participants: 838 in the intervention group, and 815 in the control) investigated the use of music versus standard care on heart rate. The studies combine with substantial heterogeneity (I2 = 60%; Analysis 5.2). In a random‐effects analysis, the average treatment effect is estimated to be −2.72 bpm (95% CI −4.70 to −0.74, P value = 0.007).
Exploring the heterogeneity of studies on heart rate:
Whereas for most studies allocation concealment is unclear, one study (Allen 2001) definitely did not use allocation concealment. Removing this study in a sensitivity analysis removes some of the heterogeneity (resulting in I2 = 52.1%) and reduces the overall effect size somewhat (MD −2.25 bpm, 95% CI −4.14 to −0.36, P value = 0.02). One study (White 1999) included a restful environment for the intervention group. Removing this study from the analysis, so all remaining studies are of music alone, increases the heterogeneity (I2 = 60.9%).
Sub‐grouping studies by type of procedure: pre‐procedure (N = 4), non‐invasive (N = 1), endoscopic (N = 2), percutaneous/surgical (N = 11), post‐operative (N = 1), and in care unit after myocardial infarction (N = 2), resolves some of the heterogeneity issue. The four pre‐procedure studies (Hayes 2003; Taylor‐Piliae 2002; Winter 1994; Yung 2003) fairly consistently (I2 = 26.7%) conclude that there is no strong evidence for an effect of music for reducing heart rate during this period (MD −1.07 bpm, 95% CI −4.44 to 2.30, P value = 0.53). The one study conducted during non‐invasive procedures (lithotripsy; Koch 1998b) also found no strong evidence of an effect. The two studies (Kotwal 1998; Triller 2006) conducted during endoscopic procedures had mixed findings (I2 = 60.6%), as did the 11 studies conducted during percutaneous interventions (I2 = 73.1%). One study on post‐operative patients (Masuda 2005, MD 0.20 bpm, 95% CI −8.66 to 9.06), and two studies (White 1999; Zimmerman 1988) conducted on patients recovering from myocardial infarction (I2 = 6.9%, MD −3.65 bpm, 95% CI −11.76 to 4.47) also found no strong evidence of an effect. The heterogeneity largely exists amongst the studies carried out during invasive procedures (endoscopic and percutaneous), with other procedures finding no strong evidence of an effect.
Studies were mostly conducted in the USA (N = 12), whilst some were conducted in China (N = 3), Taiwan (N = 2), India (N = 1), Japan (N =1), Slovenia (N = 1), and Germany (N = 1). Although there is some heterogeneity amongst the 12 USA studies (I2 = 53.0%), most of this can be explained by Allen 2001; the only study that did not use allocation concealment. Removing Allen 2001 from this subgroup reduces the heterogeneity (to I2 = 2.1%) and results in a mean difference of −0.97 bpm (95% CI −2.65 to 0.70, P value = 0.26), showing no strong evidence of an effect. The three studies conducted in China (Chan 2006; Taylor‐Piliae 2002; Yung 2003) show mixed findings (I2 = 82.8%), whereas the two studies conducted in Taiwan (Chang 2005; Tang 1993) were consistently (I2 = 0%) in favour of music (MD −7.09, 95% CI −10.13 to −4.06). Studies from India (Kotwal 1998), Japan (Masuda 2005), and Germany (Argstatter 2006) showed no strong evidence of an effect, whereas the study conducted in Slovenia (Triller 2006) did favour music.
A post‐hoc exploration can group studies according to whether patients were provided a choice of music (N = 13), asked to bring music from home (N = 2; Koch 1998a; Koch 1998b), or provided with set pieces (N = 6; Argstatter 2006; Cadigan 2001; Kotwal 1998; McRee 2003; Triller 2006; White 1999). The set piece subgroup has little statistical heterogeneity (I2 = 0%), and overall shows no strong evidence of an effect (MD −2.03 bpm, 95% CI −4.25 to 0.19, P value = 0.07). Moderate heterogeneity exists (I2 = 50.9%) between the two studies which asked patients to bring music from home. Together these two studies also show no strong evidence of an effect (MD 1.70 bpm, 95% CI −5.15 to 8.55, P value = 0.63). And, although favouring music for the reduction of heart rate (MD −3.54 bpm, 95% CI −6.25 to −0.84, P value = 0.01), the patient choice group has substantial statistical heterogeneity (I2 = 67.5%).
Studies with insufficient data for extraction:
Seven other RCTs also collected data on heart rate. Two studies investigated music during endoscopic procedures (Palakanis 1994; Smolen 2002), and with study sample sizes of 50 and 32 patients respectively, were in favour of music. Three studies (of 107, 50, and 30 participants) were conducted during percutaneous interventions (Bally 2003; Lepage 2001; Schneider 2001) and showed no strong evidence of an effect. One study conducted in the post‐operative period (Heitz 1992, with 60 participants) found no strong evidence of an effect, in line with the findings above. And one study was conducted during rehabilitative exercises (Ferguson 2004); patients in this study were undergoing different forms of exercise and rehabilitation (passive and active), making it unclear as to the appropriate direction of findings (i.e. if the aim of using music is for patients to work harder, then an increase in heart rate would be appropriate, but if the aim was to decrease stress, then a decrease in heart rate would be appropriate). This study of only 11 participants found no strong evidence of an effect.
Blood pressure
Eighteen RCTs reported sufficient information for assessment of the outcome systolic blood pressure (Allen 2001; Argstatter 2006; Broscious 1999; Buffum 2006; Cadigan 2001; Chan 2006; Chang 2005; Hayes 2003; Koch 1998a; Koch 1998b; Kotwal 1998; Masuda 2005; McRee 2003; Mennegazzi 1991; Triller 2006; White 1999; Winter 1994; Zimmerman 1988), 17 reported diastolic blood pressure (Allen 2001; Argstatter 2006; Broscious 1999; Buffum 2006; Cadigan 2001; Chan 2006; Chang 2005; Hayes 2003; Koch 1998a; Koch 1998b; Kotwal 1998; Masuda 2005; McRee 2003; Mennegazzi 1991; Triller 2006; Winter 1994; Zimmerman 1988), and two reported arterial blood pressure (Tang 1993; Yung 2003) (Analysis 5.3). Again, this outcome is subject to the risk of selective outcome reporting. The total sample size of the studies that measured systolic blood pressure was 1437 (730 in the intervention group and 707 in the control), with studies combining to show substantial statistical heterogeneity (I2 = 67%). All but one (White 1999) of these studies also reported diastolic blood pressure (overall N = 1407; intervention group = 715, control group = 692). For this outcome there was less statistical heterogeneity (I2 = 50%), and the overall findings showed no strong evidence of an effect (MD −0.97 mm Hg, 95% CI −2.58 to 0.63, P value = 0.23). There were 186 participants in the two studies that assessed arterial blood pressure (93 in the intervention group and 93 in the control group), and these studies were consistently (I2 = 0%) in favour of the music group for reduction of blood pressure (MD −9.86 mm Hg, 95% CI −12.06 to −7.65, P value < 0.00001).
Exploring the heterogeneity of studies on blood pressure:
In one study allocation concealment was definitely not used (Allen 2001). This study had the largest treatment effect in favour of music of all the studies that reported on systolic and diastolic blood pressure. Removing this study in a sensitivity analysis reduces the heterogeneity between studies (systolic blood pressure I2 = 55%, diastolic blood pressure I2 = 42%), and the combined effects continue to demonstrate no strong evidence of an effect (systolic blood pressure: MD −0.93 mm Hg, 95% CI −3.90 to 2.04, P value = 0.54; diastolic blood pressure: MD −0.66 mm Hg, 95% CI −2.20 to 0.87, P value = 0.40).
Ten studies were conducted on patients undergoing surgical/percutaneous procedures (Allen 2001; Argstatter 2006; Broscious 1999; Cadigan 2001; Chan 2006; Chang 2005; Koch 1998a; McRee 2003; Mennegazzi 1991; Tang 1993), two were conducted on patients undergoing endoscopic procedures (Kotwal 1998; Triller 2006), four were conducted during the waiting period before a medical procedure (Buffum 2006; Hayes 2003; Winter 1994; Yung 2003), two were conducted on patients in coronary or intensive care units (White 1999; Zimmerman 1988), one was conducted during a non‐invasive medical procedure (Koch 1998b), and one was conducted on post‐operative patients (Masuda 2005). Moderate heterogeneity exists in the nine studies conducted during percutaneous/surgical procedures measuring systolic and diastolic blood pressure (Systolic I2 = 74%; Diastolic I2 = 51%), attributable largely to Allen 2001. Overall, these studies show no strong evidence of an effect (Systolic MD −2.57 mm Hg, 95% CI −8.36 to 3.21, P value = 0.38; Diastolic MD −0.31 mm Hg, 95% CI −2.83 to 2.22, P value = 0.81). The two studies conducted during endoscopic procedures had little statistical heterogeneity (systolic and diastolic I2 = 0%) and were in favour of music for reducing systolic and diastolic blood pressure (Systolic MD −7.44 mm Hg, 95% CI −11.18 to −3.69, P value = 0.0001; Diastolic MD −5.44 mm Hg, 95% CI −7.88 to −3.00, P value < 0.0001). The studies in patients receiving music during the waiting period showed no strong evidence of an effect for systolic or diastolic blood pressure (Systolic MD 2.53 mm Hg, 95% CI −3.09 to 8.15, P value = 0.38 [I2 = 51%]; Diastolic MD −0.39 mm Hg, 95% CI −2.25 to 1.47, P value = 0.68 [I2 = 0%]), as did the studies conducted in coronary care units (Systolic MD −3.40 mm Hg, 95% CI −11.14 to 4.35, P value = 0.39 [I2 = 0%]; Diastolic MD 0.80 mm Hg, 95% CI −6.15 to 7.75, P value = 0.82 [Zimmerman 1988]). The studies conducted during non‐invasive procedures and during the post‐operative period also showed no strong evidence of an effect for systolic and diastolic blood pressure. Of the studies that measured arterial blood pressure (both significantly in favour of music), one was conducted on percutaneous/surgical patients and one was conducted during the waiting period.
Twelve of the studies that measured blood pressure were conducted in the USA (Allen 2001; Broscious 1999; Buffum 2006; Cadigan 2001; Hayes 2003; Koch 1998a; Koch 1998b; McRee 2003; Mennegazzi 1991; White 1999; Winter 1994; Zimmerman 1988). Two were conducted in China (Chan 2006; Yung 2003), two in Taiwan (Chang 2005; Tang 1993), and one in each of Germany (Argstatter 2006), India (Kotwal 1998), Japan (Masuda 2005), and Slovenia (Triller 2006). For the outcome systolic blood pressure there is substantial heterogeneity between the studies conducted in the USA (I2 = 73%), largely attributable to Allen 2001. For the studies that measured arterial blood pressure with positive findings, one was conducted in China, and one was conducted in Taiwan.
One study (White 1999), which reported on systolic blood pressure, incorporated a restful environment for those in the music group. Removing this study in a sensitivity analysis (so all remaining studies compare music alone to standard care), does little to change the statistical heterogeneity (I2 = 69%). Twelve studies offered participants a selection of music to choose from (Allen 2001; Broscious 1999; Buffum 2006; Chan 2006; Chang 2005; Hayes 2003; Masuda 2005; Mennegazzi 1991; Tang 1993; Winter 1994; Yung 2003; Zimmerman 1988), two requested patients to bring their own music from home (Koch 1998a; Koch 1998b), and six played participants set pieces selected by the investigators (Argstatter 2006; Cadigan 2001; Kotwal 1998; McRee 2003; Triller 2006; White 1999. For the outcome systolic blood pressure, grouping studies in this way does little to explain the heterogeneity (choice from selection: systolic I2 = 71%, diastolic I2 = 24%; music from home: systolic I2 = 0%, diastolic I2 = 0%; set pieces: systolic I2 = 52%, diastolic I2 = 66%). The two studies which requested participants to bring their own music from home produced consistent results showing no strong evidence of effect (systolic MD 2.24 mm Hg, 95% CI ‐5.33 to 9.81, P value = 0.56; diastolic MD 1.17 mm Hg, 95% CI −3.46 to 5.80, P value = 0.62), however, these studies were conducted by the same investigators and are likely to have many other similarities.
Studies with insufficient data for extraction:
Seven other RCTs collected data on blood pressure (Bally 2003; Ferguson 2004; Heitz 1992; Lepage 2001; Palakanis 1994; Schneider 2001; Smolen 2002). Within these studies, sample sizes range from 11 to 107 participants (mean = 48.57, SD = 30.50, median = 50). There is plausible risk of bias in these studies that raises some doubt about the results; only two studies were judged as having adequate allocation concealment (Bally 2003; Palakanis 1994). Studies were based in the USA (Ferguson 2004; Heitz 1992; Palakanis 1994; Smolen 2002), Canada (Bally 2003; Lepage 2001), and Germany (Schneider 2001). All studies offered patients a choice of music from a selection. Findings from these studies generally mirror those of the RCTs described above. Three studies were conducted during percutaneous/surgical interventions (Bally 2003; Lepage 2001; Schneider 2001), with one (Schneider 2001) reporting a significant drop in blood pressure from pre‐ to post‐treatment in the music group and no change in the control group (the other two studies showed no strong evidence of an effect). Two studies were conducted on patients undergoing endoscopic examinations (Palakanis 1994; Smolen 2002), and both reported significant findings favouring the music group. Ferguson 2004 was conducted on patients undergoing non‐invasive "range‐of‐motion" exercises with non‐significant findings. And Heitz 1992 was conducted in the post‐operative period, reporting no difference between groups.
Respiration rate
Nine RCTs reported sufficient information for data extraction on respiration rate. When combined these studies had an overall sample of 644 patients (322 in each group), and the study findings were heterogeneous (I2 = 80%; Analysis 5.4), so have not been pooled.
Exploring the heterogeneity of studies on respiration rate:
The quality of studies was mixed, with only one (Taylor‐Piliae 2002) reporting adequate concealment of allocation.
Five studies (Cadigan 2001; Chan 2006; Chang 2005; Mennegazzi 1991; Yung 2003) were conducted on patients undergoing percutaneous or surgical procedures (I2 = 78%), two studies (Buffum 2006; Taylor‐Piliae 2002) were conducted during the waiting period (I2 = 25%; MD −0.05 breaths per minute, 95% CI −1.24 to 1.13, P value = 0.93), one study (Kotwal 1998) was conducted during endoscopic procedures (favouring music), and one study (White 1999) was conducted in a coronary care unit (non‐significant findings).
Four studies were conducted in the USA (I2 = 64%), three in China (I2 = 85%), one in Taiwan, and one in India; grouping studies in this way does little to explain the heterogeneity.
One study (White 1999) incorporated a restful environment for the music group. Removing this study in a sensitivity analysis so that all remaining studies compare music alone to standard care does little to change the findings (I2 = 82%). Three studies exposed participants to set pieces of music (Cadigan 2001; Kotwal 1998; White 1999) and the remaining six studies offered participants a choice of music from a selection. Sub‐grouping studies in this way leaves 68% heterogeneity in the patient choice group (MD −0.60 breaths per minute, 95% CI −1.57 to 0.36, P value = 0.22) and 82% in the set pieces group. Much of the statistical heterogeneity in this set of studies is contributed by the significant findings of Chan 2006 and Kotwal 1998, however none of the clinical or methodological differences explored above explain these findings.
Studies with insufficient data for extraction:
Three other RCTs (Ferguson 2004; Heitz 1992; Lepage 2001) also collected data on respiration rate. Sample sizes were 11, 60 (with two control groups), and 50, respectively. Within these studies there is plausible bias which raises some doubt about the results. None of the studies had adequate allocation concealment. One study (Heitz 1992) reported that healthcare professionals were partially blinded to group allocation. This study had a third 'headphones only' group so nurses were not aware if those wearing headphones were actually receiving music, but they could recognise who was allocated to the standard care alone group. No studies reported a power calculation. None of the studies had clear completeness of dataset. Ferguson 2004 had different baseline characteristics between groups. None of the studies reported on withdrawals and drop‐outs. The studies (conducted during non‐invasive procedures, percutaneous interventions, and post‐operatively) all offered patients a choice of music from a selection and showed no difference between groups.
Anxiolytic medication requirements
Three RCTs reported sufficient data for extraction on the use of anxiolytic medications requirements (Analysis 5.5). These studies investigated 201 participants (music group = 100, control group = 101) and show considerable statistical heterogeneity (I2 = 91%) and for this reason have not been pooled. Two of the studies (Lepage 2001; Smolen 2002) report findings in favour of music, and one larger study reports no difference between groups (Schiemann 2002).
Exploring the heterogeneity of studies on anxiolytic medication requirements:
In all three studies allocation concealment is unclear, blinding of group allocation is not done, power calculations are not reported, and withdrawals and drop‐outs are not reported. All studies offer adequate protection against contamination and report that groups were similar at baseline. The two studies with positive outcomes (Lepage 2001; Smolen 2002) have unclear completeness of data, whereas the study with non‐significant findings (Schiemann 2002) did report completeness of dataset. Schiemann 2002 and Smolen 2002 were conducted on patients undergoing endoscopic procedures, and Lepage 2001 was conducted in patients undergoing percutaneous interventions. The studies with positive results were conducted in the USA and Canada, and study with non‐significant findings was conducted in Germany. The two studies with positive outcomes (Lepage 2001; Smolen 2002) offered patients in the intervention group a choice of music (I2 = 0%, MD −1.55 mg, 95% CI −2.10 to −1.00, P value < 0.00001). The study with non‐significant findings (Schiemann 2002, MD 0.03 mg, 95% CI −0.32 to 0.38) played patients in the intervention group set pieces of music. It is unclear in Schiemann 2002 if midazolam consumption is an outcome measure or baseline characteristic; this may provide an alternative explanation for the non‐significant findings in this study. From these observations, it is difficult to draw conclusions as to the explanation for the statistical heterogeneity.
Other outcomes
A number of studies reported on additional outcomes which we have summarised in Table 13. For this set of outcomes, apart from fatigue and uncertainty, significant differences in favour of music were found for most patient‐reported outcomes (mood, anger, depression, nausea, and satisfaction). Groups did not differ however, on most physiological outcomes (skin temperature, oxygen saturation, blood flow, bispectral index, lung function, requirement for oxygen supplementation). Findings for the outcome cortisol are mixed, and there was no strong evidence of an effect for other stress hormones (prolactin and catecholamines). Significant differences in favour of the music group were found for intubation time and length of stay (although another study at higher risk of bias, with insufficient data for extraction, was non‐significant). It should be noted that few studies reported on each of these outcomes (for 10 of the 16 outcomes only one study is included), and in some cases the risk of bias in studies is high.
| Outcome | Detailed RCTs (N) | Participants (N) | Heterogeneity (%) | Results | Other RCTs | Participants (N) | Findings | Comments |
| Skin temperature | N = 4 | Total = 298 | I2 = 0% | SMD 0.15, 95% CI −0.08 to 0.37, P value = 0.21; | None | N/A | N/A | |
| Oxygen saturation | N = 3 | Total = 150 | I2 = 79%; | MD −0.71% | None | N/A | N/A | |
| Requirement for oxygen supplementation | N = 1 | Total = 119 | N/A | OR 0.49, | None | N/A | N/A | |
| Blood flow characteristics | N = 1 | Total = 44 | N/A | Blood flow: MD −2.40 ml/min/100g, 95% CI −7.45 to 2.65, P value = 0.35; | None | N/A | N/A | |
| Bispectral index | N = 1 | Total = 39 | N/A | MD 0.22 BIS value, | None | N/A | N/A | It is unclear if this study meets the review inclusion criteria. Music group may have received intervention prior to coming to hospital. |
| Lung Function: (dyspnoea, tidal volume, minute ventilation, oxygen consumption, airway resistance). | None | N/A | N/A | N/A | N = 1 | Total = 36 | No significant differences. | |
| Stress hormones (Cortisol, Prolactin). | N = 1 | Total = 26 | N/A | Cortisol: MD 7.29, 95% CI −7.37 to 21.95, P value = 0.33; | N = 1 | Total = 30 | Cortisol significantly increased in the control group and remained unchanged in the music group. Catecholamines were non‐significant. | |
| Mood | N = 2 | Total = 170 | I2 = 0% | MD −1.18, | None | N/A | N/A | |
| Anger | N = 1 | Total = 24 | N/A | MD −1.80, | None | N/A | N/A | |
| Depression | N = 1 | Total = 39 | N/A | MD −3.29, | None | N/A | N/A | It is unclear if this study meets the review inclusion criteria. Music group may have received intervention prior to coming to hospital. |
| Fatigue | None | N/A | N/A | N/A | N= 1 | Total = 24 | No significant difference between groups. | |
| Uncertainty | N = 1 | Total = 30 | N/A | MD −3.53, 95% CI −12.15 to 5.09, P value = | None | N/A | N/A | |
| Satisfaction | N = 2 | Total = 174 | I2 = 0% | MD 0.46, 95% CI 0.16 to 0.76, | None | N/A | N/A | |
| Nausea | None | N/A | N/A | N/A | N = 1 | Total = 33 | The paper reports that the music group had significantly less nausea and vomiting than the control group (Mann‐Whitney U test, P < 0.017). | |
| Length of stay | N = 2 | Total = 153 | I2 = 0% | MD −6.00 minutes, 95% CI −10.72 to −1.28, P value = 0.01 in favour of music group (Analysis 5.10). | N = 1 | Total = 60 | Findings were non‐significant. | |
| Intubation time | N = 1 | Total = 60 | N/A | MD −200.20 minutes, 95% CI −391.03 to −9.37, P value = 0.04 in favour of music group. | None | N/A | N/A |
CI: confidence interval; MD: mean difference; SMD: standardised mean difference
Findings from studies on music versus white noise:
Anxiety
One RCT reported sufficient data for extraction on anxiety (Zimmerman 1988). In this study of 50 participants (music group = 25, control group = 25), there was no strong evidence of an effect (MD 0.90 points on the state anxiety scale of the STAI, 95% CI −5.78 to 7.58, P value = 0.79).
Studies with insufficient data for extraction:
One other RCT (music group = 32, control group = 29) also investigated anxiety (Cruise 1997). This study also found no strong evidence of an effect.
Heart rate
Two RCTs reported sufficient data for extraction on the outcome heart rate (Broscious 1999; Zimmerman 1988; Analysis 6.1). When combined, these studies (music group = 89, control group = 55) with statistical homogeneity (I2 = 0%), show no strong evidence of an effect (MD 4.67 bpm, 95% CI −0.76 to 10.10, P value = 0.09).
Studies with insufficient data for extraction:
One other RCT (music group = 32, control group = 29) also investigated heart rate (Cruise 1997) and also found no strong evidence of an effect.
Blood pressure
Two RCTs reported sufficient data for extraction on the outcome blood pressure (Analysis 6.2), both of which reported systolic and diastolic blood pressure (music group = 89, control group = 55). These studies were homogenous for the outcome systolic blood pressure (I2 = 0%), finding no strong evidence of an effect (MD −1.80 mm Hg, 95% CI −8.59 to 5.00, P value = 0.60). Although the findings for diastolic blood pressure also showed no difference between groups, the two studies show considerable statistical heterogeneity (I2 = 78%).
Studies with insufficient data for extraction:
One other RCT (music group = 32, control group = 29) also investigated blood pressure (Cruise 1997) and found no difference in both systolic and diastolic blood pressure between the music and white noise groups.
Respiration rate
Studies with insufficient data for extraction:
One RCT investigated respiration rate (Cruise 1997). This study (music group = 32, control group = 29) found no differences between groups.
Other outcomes
Other outcomes investigated were skin temperature (Zimmerman 1988; N = 50) and length of stay (Ayoub 2005; N = 62). For both these outcomes the findings showed no strong evidence of an effect.
Access to nature
Description of studies on providing access to nature:
There are no studies on provision of access to nature included in the review; We have tabulated five excluded studies (Table 14).
| Study ID | Reason for exclusion |
| Setting | |
| Commentary | |
| Outcomes | |
| Study design | |
| Study design‐ retrospective matched pairs |
Reducing environmental stressors by implementing physical changes
Air quality
Description of studies on air quality:
We included one RCT (Lohner 1979) and two CCTs (Engelhart 2003; Whyte 1969) on hospital air quality (Table 15). Engelhart 2003 reports numbers in terms of patient bed‐days (not number of patients), and in this study there were 6000 bed‐days (1200 bed‐days in the intervention group and 4800 in the control group). In the remaining studies there were 1771 participants, 824 in the intervention groups, and 947 in the control groups. Based on the data of 1771 patients reported in two of the studies (Lohner 1979; Whyte 1969; and assuming that reported "average" age is the mean), patients were on average 38.2 years old (ranging from at least 13 to at least 74 years old). Gender is not reported in Engelhart 2003. In the remaining two studies there were 815 males and 956 females included. Studies were conducted in Germany, Belgium, and Scotland. Patient groups included those in haematology‐oncology units, those undergoing treatment for acute leukaemia or bone marrow aplasia, and post‐operative patients in a surgical unit (urological surgery/general surgery).
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| CCT; 2 parallel groups. | DESCRIPTION: In‐patients, predominantly with leukaemia, multiple myeloma, and malignant non‐Hodgkins lymphoma, admitted to the haematology‐oncology unit in Bonn, Germany. | AIR FILTRATION: 3 rooms (single or double) were fitted with portable air filtration units (NSA model 7100A/B Environment Air System, National Safety Association Ltd., Memphis, USA). These units have a 95% filtering capacity for particles > 0.3 micrometers. They have two settings (high/low). Flow rate: High = 168, Low = 112 cubic m/hr. | INVASIVE ASPERGILLOSIS: Counts of cases (confirmed and suspected) were achieved via ward liaison, targeted chart reviews, and consultation with medical staff. | Compliance to the air filtration units was low; they were generally run on low due to the noise generated, and sometimes stopped during the night. | |
| RCT; 2 parallel groups. | DESCRIPTION: 45 inpatients undergoing treatment for acute leukaemia or bone marrow aplasia, in Belgium. | ISOLATION: Patients were isolated in a laminar air flow room. The isolation unit consisted of a normal bed enclosed in a plastic tent. The area available for the patient was about 5 sq metres. Air was pumped vertically through high‐efficiency filters positioned on the ceiling. All procedures on the patient were done through plastic gloves on the sides of the tent. All items entering the unit were either gas or steam sterilized, passing through locks irradiated by ultraviolet light. The tent was only opened when absolutely necessary. Any person approaching the isolated patients wore sterile boots, gloves, gown, and mask. Patients remained in isolation until either the bone marrow showed haematologic remission or myeloid proliferation and maturation. The study period ended when the patient was no longer in isolation. | BACTERIAL INFECTIONS: Fungal cultures of stools were performed at least once a week. Data is reported as a count of the number of patients in which suppression of bacterial growth from stool cultures was achieved. Cultures were also obtained from the nose, gingiva, throat, ear, and axilla. Results report the number of days with infection (per 1000 days with severe neutropenia). | ||
| CCT; 2 parallel groups. | DESCRIPTION: 1726 in‐patients residing in the wards of a surgical unit (urological surgery/general surgery), in Scotland, UK. | OPEN WARDS: Two wards (one male, one female) with natural ventilation (and no mechanical ventilation). The male ward had partitioned 4‐6 bedded bays, housing a total of 28 beds, and female ward was completely open plan with 29 beds. | STAPHYLOCOCCUS AUREUS (Staph. aureus): Nasal acquisition of Staph. aureus was monitored via swabs taken on admission, and then every Monday and Thursday thereafter. Data reported as the number of acquisitions, and the rate per 100 patient‐week. The number of tetracycline‐resistant strains of Staph. aureus and delayed acquisitions was also determined. |
Intervention groups included: isolation curtains with high‐efficiency particulate air (HEPA) filtration (Lohner 1979); single or double rooms with portable air filtration units (Engelhart 2003); and a closed ward with air conditioning, partially HEPA filtered (Whyte 1969). Comparison groups were with conventional airflow, which included mechanical and natural ventilation, and rooms of varying sizes.
Outcomes assessed were: invasive aspergillosis, mortality, remission, bacterial infections, and wound sepsis.
We have tabulated 31 excluded studies on air quality (Table 16).
| Study ID | Reason for exclusion |
| Study design | |
| Setting | |
| Outcomes | |
| Study design (no control) | |
| Study design (matched pairs) | |
| Policy confound (sterile versus clean attire) | |
| Study design (computer modelling) | |
| Confounding | |
| Patients‐ mix of adults and children, data inseparable | |
| Preliminary report of Rodriguez 1978; policy confound | |
| Sham operations | |
| One patient health‐related outcome reported (surgical site infections), but not enough detail on how it was assessed or if it was noted on an 'ad hoc' basis. | |
| Not patient outcomes | |
| Study design | |
| Not patient outcomes; test scenario | |
| Not patient outcomes; test scenario | |
| Study design | |
| Some relevant subgroups within study however cannot separate data (number of re‐operations in each subgroup). Additionally, there are discrepancies in reporting with Lidwell 1984 with 185 patients switching subgroups. | |
| Same study as Lidwell 1982. Non‐transparant reporting of findings‐ data not usable. Discrepancies between Lidwell 1982 and Lidwell 1984. | |
| Population‐ age | |
| Preliminary report of Lidwell 1982 and Lidwell 1984 | |
| Study design | |
| Retrospective study; population (age) | |
| Policy (decontamination) changes | |
| Mock setting | |
| Policy confounding | |
| Policy confounding | |
| Study design | |
| Study design | |
| Not patient outcomes | |
| Policy confounding |
Below, we summarise findings for the following comparisons.
-
Air conditioned closed ward versus naturally ventilated open ward
-
Laminar airflow (High Efficiency Particulate Air filter) versus conventional airflow
-
Portable air filtration unit versus standard care
Risk of bias in included studies on air quality:
We included one RCT (Lohner 1979) and two CCTs (Engelhart 2003; Whyte 1969) on air quality. Method of sequence generation and allocation concealment was unclear in two studies (Engelhart 2003; Lohner 1979), and allocation was by day of the week in Whyte 1969. Blinding of group allocation was not done in any study, although it is unclear if outcome assessments were automated or blinded in Lohner 1979 and Whyte 1969. Completeness of outcome data is unclear throughout. No studies reported withdrawals and drop‐outs. Protection against contamination was adequate in all studies. The reliability of outcome measures (case counts of invasive aspergillosis) is unclear in Engelhart 2003.
Findings from studies on air quality. Air conditioned closed ward versus naturally ventilated open ward:
Infection
Whyte 1969 (air conditioned wards = 800, naturally ventilated open wards = 929) found no strong evidence of an effect for number of septic episodes, presence of resistant bacteria, and acquisitions of Staphylococcus aureus.
Findings from studies on air quality. Laminar airflow (High Efficiency Particulate Air filter) versus conventional airflow:
Infection
Studies with insufficient data for extraction:
Lohner 1979 reports that the number of infections and cases of septicaemia were not different between groups. Lohner 1979 reports that days of infection and treatment with antibiotics were significantly more frequent (P value < 0.05) in the conventional airflow group (however the table contradicts the text, and it is unclear how the counts were standardised as the number of study days varied between groups).
Mortality
Lohner 1979 reported mortality data (laminar airflow group = 7/24, conventional airflow group = 5/21), showing no strong evidence of an effect (risk ratio (RR) 1.23, 95% CI 0.46 to 3.29, P value = 0.69).
Remission
Lohner 1979 reported "complete remission" data (laminar airflow group = 5/24, conventional airflow group = 3/21), showing no strong evidence of an effect (RR 1.46 , 95% CI 0.40 to 5.38, P value = 0.57).
Adverse events
Lohner 1979 reported that 10 patients were removed from the laminar airflow room due to "poor psychologic tolerance of confinement after an average confinement period of 18.6 days (range 15 to 23 days)". One further patient with schizophrenia was removed after 59 days in the laminar airflow room.
Findings from studies on air quality. Portable air filtration unit versus standard care:
Infection
The overall incidence of invasive aspergillosis between groups in Engelhart 2003 showed no strong evidence of an effect (portable air filtration unit = zero infections per 1200 patient‐days, standard care five infections per 4800 patient‐days, P value = 0.33).
Bedroom type
Description of studies on bedroom types:
One CCT (Lidwell 1971) has been included (Table 17), which assesses different bedroom types (e.g. open versus closed rooms). Lidwell 1971 reported data in patient‐weeks (not number of patients), and 3327 patient‐weeks were assessed (open rooms = 2750 patient‐weeks, closed single rooms = 577 patient‐weeks). Participant age is described to an extent in Lidwell 1971 (57% of patients were under 60 years old, and 43% were over 60 years old), but no details are given on participant gender. The study was conducted in the UK on in‐patients in two medical wards.The open rooms were partitioned into four‐bed bays (six bays on each of two wards), each with three proper walls and a fourth low dividing wall opening on to the corridor. The closed rooms were single‐bed rooms (five rooms on each of two wards) based on the same two wards as the open rooms.The outcome 'acquisition of Staphylococcus aureus' was assessed. We have tabulated 42 excluded studies (Table 18) on bedroom type.
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| CCT; 2 parallel groups. | DESCRIPTION: 3327 patient‐weeks were analysed. Participants were in‐patients on one of two medical wards in the UK. | OPEN WARD: Two identical wards each containing 6 four‐bed rooms, which opened up on to the corridor. The four‐bed rooms each had three proper walls with the fourth side being open to the corridor except for low dividing walls (3 ft.) on each side of the entrance. | PATIENT NASAL AQUISITION OF STAPHYLOCOCCUS AUREUS: A nasal swab was taken from patients as soon as possible after arriving on the ward (within 3 days). Subsequently a swab was taken from each patient on a set day of the week. | Ratio of nasal carriage rates in single rooms to 4‐bed bays: |
| Study ID | Reason for exclusion |
| Review article | |
| Participants (age) | |
| Cross‐sectional survey and qualitative interviews | |
| Confounding | |
| Retrospective study | |
| Policy changes | |
| Preliminary report of Rodriguez 1978; policy confound | |
| Healthy participants received intervention | |
| Study design | |
| Study design | |
| Study design | |
| Study design, outcomes | |
| Study design | |
| Outcome measure | |
| Participants | |
| Study design | |
| Study design; confounding | |
| Intervention | |
| Study design | |
| Confounding | |
| Study 1: inappropriate control; Study 2: intervention | |
| Study design | |
| Study design | |
| Study design; unclear intervention | |
| Participants; Study design | |
| Setting | |
| Study design | |
| Policy changes | |
| Patients‐ 38% <10 years old | |
| Study design | |
| Intervention (non‐environmental changes) | |
| Policy confounding | |
| Policy confounding | |
| Policy change | |
| Study design | |
| Study design | |
| Policy changes | |
| Study design | |
| Study design | |
| Study design‐ no control | |
| Validity of outcomes | |
| Policy confounding |
Risk of bias in included studies on bedroom types:
The one included CCT (Lidwell 1971) had unclear methods of sequence generation and allocation concealment was not done. Blinding of outcome assessors was unclear. Completeness of dataset was unclear. Withdrawals and drop‐outs were not described. There were insufficient data provided for extraction giving rise to a risk of reporting bias. It is unclear if there was protection against contamination.
Findings from studies on bedroom types:
Infection
Non‐randomised studies and studies with insufficient data for extraction:
Lidwell 1971 found that patients in single rooms (opening off the general ward area and not mechanically ventilated) acquired strains of Staphylococcus aureus from other patients at almost the same rate (23.9/1000 weeks) as patients in divided four‐bed bays (24.1/1000 weeks).
Ceilings
Description of studies on ceilings:
There are no studies on ceilings included in the review. We have tabulated two excluded studies (Table 19), which we excluded due to not meeting our study design and participant inclusion criteria.
| Study ID | Reason for exclusion |
| Participants (students) | |
| Study design |
Flooring
Description of studies on flooring:
We included two RCTs, including one cross‐over trial of 58 patients, on flooring in hospitals (Table 20). These two studies included 112 elderly adults exposed to carpeted floor (N = 86) and vinyl floor (N = 84). The mean age of participants was 79.39 years old. The gender of participants in the cross‐over trial (Willmott 1986) is not described; however there were 10 males and 44 females in the other trial (Donald 2000). Both studies were conducted in England. One study was carried out in an elderly care rehabilitation ward (Donald 2000), and the other study states "elderly hospital in‐patients" (Willmott 1986).
| Study ID | Methods | Participants | Interventions | Outcomes |
| RCT; 2 x 2 design (2 x flooring, 2 x therapy). | DESCRIPTION: 54 in‐patients in an elderly care rehabilitation ward, in England. | LINOLEUM FLOOR: Patient admitted to one of two four‐bed bays with latex vinyl square tile flooring. | NUMBER OF FALLERS: obtained from the accident report forms. | |
| RCT, cross‐over design, 2 conditions. | DESCRIPTION: 58 elderly hospital patients in England. | CARPETED CORRIDOR: Each patient walked along the corridor towards a staff member standing 15 m ahead. Chalk marks were drawn at an interval of 10 m. | GAIT SPEED and STEP LENGTH: The number of steps taken was counted from the time the patient crossed the first chalk line until the distal line was crossed by the leading forefoot, and elapsed time was measured by a stopwatch. |
Carpeted floors were heavy‐duty Flotex® 200, with no pile (Donald 2000) and in Willmott 1986 it was described as a carpeted corridor. Vinyl floors were latex vinyl square tiles in Donald 2000, and a reflective vinyl tiled corridor in Willmott 1986.
Outcomes assessed were: number of fallers, number of falls, independence, length of stay, gait speed, and step length.
We have tabulated seven excluded studies on flooring (Table 21), and one ongoing study (Characteristics of ongoing studies).
| Study ID | Reason for exclusion |
| Paediatric hospital | |
| Study design; outcomes | |
| Qualitative | |
| Study design | |
| Study design | |
| Study design | |
| Qualitative evaluation |
Risk of bias in included studies on flooring:
Included in the review was one parallel RCT (Donald 2000), with randomisation conducted via envelopes stratified by patient risk of falling (using a scale designed by the researcher), and one cross‐over RCT (Willmott 1986) with an unclear method of sequence generation. Donald 2000 was a factorial design including two types of physical therapy as well as the two flooring types. Concealment of allocation was unclear in both studies. Blinding of group allocation in both studies was not possible. Completeness of dataset was achieved in Donald 2000 for the outcomes of interest in this review, but it is unclear in Willmott 1986. Willmott 1986 does not describe any withdrawals and drop‐outs. Donald 2000 describes 22 withdrawals and drop‐outs by allocation group (Table 20), which were mostly (N = 13) due to non‐compliance with the therapy related outcome measures (not included in review), as well as death (N = 6), and patient transfer (N = 3). Both studies are at risk of selective outcome reporting bias. Neither study offered protection against contamination.
Findings from studies on flooring:
Number of fallers and falls
Donald 2000 found that the number of people who fell on carpeted floor (N = 7 of 28 people, fell altogether 10 times) was greater than on linoleum floor (N = 1 person of 26, fell once) but this showed no strong evidence of an effect.
Independence
Donald 2000 found no strong evidence of an effect between carpeted flooring (mean improvement = 1.1, SD 3.4) and linoleum flooring (mean improvement = 2.9, SD 3.3) groups in Barthel score (MD 1.80, reported P value = 0.08).
Length of stay
Studies with insufficient data for extraction:
Donald 2000 reported no strong evidence of an effect between linoleum flooring and carpeted flooring groups on length of stay.
Gait speed
Studies with insufficient data for extraction:
In a cross‐over trial of 58 patients, Willmott 1986 found that gait speed was significantly faster on carpet than vinyl (P value < 0.0005).
Step length
Studies with insufficient data for extraction:
In a cross‐over trial of 58 patients, Willmott 1986 found that step length was significantly greater on carpet than vinyl (P value < 0.0005).
Furniture and furnishings
Description of studies on furniture and furnishings:
One RCT has been included on hospital furniture (Table 22; Wilber 2005). This study was of 132 elderly out‐patients (66 assigned to a reclining chair, and 66 assigned to the standard gurney). Patients were 65 years or older (mean = 77.5 years old), including 55 males and 77 females. The study was conducted in the USA on patients admitted to an emergency department.
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| RCT; 2 parallel groups. | DESCRIPTION: 132 elderly out‐patients admitted to the emergency department, in Ohio, USA. | CHAIR GROUP: Remained on the gurney until initial physician and nursing evaluations complete. Patients were then assisted to a reclining chair. Patients were told "it is now time to move to the chair" and if they asked why, they were told they had the option of now moving to the chair. Patients were assisted to a position of comfort. Patients can sit on the reclining chairs with the hips and knees flexed. | PAIN: "The study nurse specifically instructed patients to rate pain associated with the gurney, rather than other sources." Measured on a NRS at three time points (time 0 = baseline; time 1 = one hour after randomisation; time 2 = two hours after randomisation). This outcome was dichotomised as favourable outcome (yes/no). A favourable outcome was considered as the patient having no pain at time 0 or time 1, or a decrease in pain from time 0 to time 1. An unfavourable outcome was defined as an increase in pain from time 0 to time 1, or no change in pain score if the patient complained of pain at time 0. | SDs for satisfaction were estimated from the 95% confidence interval (1.4, 2.8), assuming 66 patients per group. |
The chair group were assisted to a reclining chair where they could sit with their hips and knees flexed. The gurney group remained on a gurney (the Emergency Department bed), which has a thin foam mattress. Both groups were assisted to a position of comfort.
Outcomes assessed were pain and satisfaction.
A Cochrane systematic review has already been conducted to include studies on beds, mattresses, overlays, and cushions (Cullum 2008), and their influence on the incidence and prevention of pressure sores. This review on support surfaces (Cullum 2008) includes the secondary outcomes of patient comfort, as well as costs, durability, reliability, and acceptability of the devices assessed. Fifty‐two randomised controlled trials are included in Cullum 2008, and these will not be re‐assessed here. A further three studies evaluating pressure‐relieving mattresses were identified by the present review, which either have not been assessed (Beldon 2002), or do not include the primary outcomes of Cullum 2008 (Grindley 1996; Pring 1998). These studies have been omitted from the present review, since it is felt that including them without duplicating the work of Cullum 2008, will provide an unrepresentative view of the literature.
In total, we have tabulated 19 excluded studies (Table 23).
| Study ID | Reason for exclusion |
| Outcomes and policy changes | |
| Other systematic review in this area (Cullum 2008) | |
| Other systematic review in this area (Cullum 2008) | |
| Study design | |
| Not a research study | |
| Other systematic review in this area (Cullum 2008) | |
| Other systematic review in this area (Cullum 2008) | |
| Study design | |
| Test (not clinical) situation, policy change | |
| Study design | |
| Study design‐ inappropriate data | |
| Other systematic review in this area (Cullum 2008) | |
| Not health‐related outcomes | |
| Outcomes (validity and relevance) | |
| Other systematic review in this area (Cullum 2008) | |
| Policy change | |
| Study design | |
| Intervention | |
| Policy changes with cleaning of blankets |
Risk of bias in included studies on furniture and furnishings:
We included one parallel RCT (Wilber 2005), with randomisation conducted via a random numbers table, and sealed numbered packets (adequate allocation concealment). Patients were not aware of the true nature of the study but blinding of healthcare professionals was not done. Outcomes were patient‐reported, and we have judged these to be assessed blindly. Completeness of outcome data was achieved at 'time 1' (one hour after randomisation), but not 'time 2' (two hours after randomisation). A description of withdrawals and drop‐outs is reported (N = 28), and these were all due to being discharged before outcomes could be assessed. It is unclear whether the study is at risk of selective outcome reporting. Protection against contamination was not done.
Findings from studies on furniture and furnishings:
Pain
Wilber 2005 found that significantly more people in the chair group (64/66) had a favourable pain outcome compared with those in the gurney group (50/66), (RR 0.78, 95% CI 0.68 to 0.90, P value = 0.0007).
Satisfaction
Wilber 2005 (chair group = 66, gurney group = 66) found that participants in the chair group were significantly more satisfied than those in the gurney (MD −2.10 points on a 10 point numerical rating scale, 95% CI −2.80 to −1.40, P value < 0.00001).
Hospital noise
Description of studies on hospital noise:
There are no studies on interventions for hospital noise reduction included in our review. Our search revealed 23 reports that investigated hospital noise, and possible interventions to reduce this. None of these reports have been included in the review (Table 24).
| Study ID | Reason for exclusion |
| No intervention‐ observational study | |
| No intervention | |
| No intervention | |
| No intervention | |
| Review article | |
| Study design | |
| Review article | |
| Study design; outcomes | |
| No intervention | |
| No intervention in patients | |
| Study design | |
| Discussion article | |
| Study design | |
| Study design | |
| Study design | |
| Study design | |
| No intervention | |
| Review article | |
| No intervention | |
| Policy confound | |
| Policy change | |
| Study design (cross‐sectional, no intervention) | |
| Study design |
Lighting
Description of studies on lighting:
We included one CCT on hospital lighting conditions (Table 25; Walch 2005), resulting in a total sample of 89 patients (intervention group = 44, control group = 45). Mean age of the sample was 58.84 years old. There were 43 males and 46 females. The study was conducted in the USA on post‐operative patients.
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| CCT; 2 parallel groups. | DESCRIPTION: 89 post‐operative in‐patients recovering from elective cervical and lumbar spinal surgeries in single patient rooms, Pittsburgh, USA. | BRIGHT ROOM: On west side of the corridor. Received approximately 46% more natural sunlight than dim rooms (received on average 73,537 lux‐hours per day). | ANALGESIC CONSUMPTION: obtained via chart extraction and standardised to morphine equivalent mg/hr. | Analgesic consumption data extracted for entire LOS mg/hr. |
POMS: Profile of Mood States
The study investigated the effects of sunlight availability in the patients' rooms (as controlled by the aspect of the patients' rooms, e.g. east versus west facing). Rooms were classified as bright or dim and according to the report did not systematically differ in any other way.
Outcomes assessed were: analgesic consumption, pain, anxiety, stress, and depression.
We have tabulated 13 excluded studies on lighting (Table 26).
| Study ID | Reason for exclusion |
| Retrospective study | |
| Duplicate of Beauchemin 1998b | |
| Retrospective study | |
| Retrospective study | |
| Study design | |
| Study design | |
| Study design | |
| Setting | |
| Intervention | |
| Intervention | |
| Study design | |
| Setting; Study design | |
| Literature review |
Risk of bias in included studies on lighting:
One CCT (Walch 2005, where patients were allocated according to room availability) was included on lighting. Allocation concealment was not used. This study reports that healthcare professionals (assessing outcomes) were blinded to group allocation, and that patients were unaware of the study intervention, however, it is unclear how this was achieved, particularly when light readings were being taken in patient rooms. Completeness of outcome data was achieved on the day of surgery and the first and second post‐operative days, but not for the third to fifth post‐operative days (as patients were discharged). All patients included in the study were analysed on an intention‐to‐treat basis, and attrition was determined by length of stay in the ward. It is unclear if the study is at risk of selective outcome reporting. It is likely that patients were protected against contamination (as they stayed in single‐occupancy rooms) although no attempt was made to control patients' use of window blinds or overhead lighting, which may have affected their exposure.
Findings from studies on lighting:
Anxiety
Findings from non‐randomised studies:
Walch 2005 (N used in analysis: bright room = 29, dim room = 30) found no strong evidence for an effect for the outcome anxiety (MD 0.80 points on the Profile of Mood States (POMS) anxiety scale, 95% CI −0.56 to 2.16, P value = 0.25).
Pain
Findings from non‐randomised studies:
Walch 2005 (N used in analysis: bright room = 29, dim room = 30) found no strong evidence for an effect for the outcome pain (MD −1.30 points on the McGill Pain Questionnaire, 95% CI −2.58 to −0.02, reported P value = 0.058).
Pain medication requirement
Findings from non‐randomised studies:
Walch 2005 (bright room = 44, dim room = 45) found that patients in sunnier rooms consumed less analgesics (MD −0.90 mg/hr, 95% CI −1.80 to 0.00, reported P value = 0.047).
Other outcomes
Findings from non‐randomised studies:
Walch 2005 also reported on depression (MD 0.60 points on the Centre for Epidemiological Studies Depression Scale (CES‐D), 95% CI −1.95 to 3.15, P value = 0.64) showing no strong evidence for an effect, and perceived stress (MD −3.40 points on the Perceived Stress Scale (PSS), 95% CI −6.43 to −0.37, P value = 0.03), which favoured the patients residing in sunnier rooms.
Patient controls
Description of studies on patient controls:
There are no included studies on the provision of patient controls (no excluded studies have been tabulated). Some of the studies included for music interventions offered patients an element of choice over their 'music' environment, and we have explored this through subgroup analyses.
Technologies
Description of studies on technologies:
There are no studies on technologies included in the review. We have tabulated three excluded studies (Table 27) on alarm systems.
| Study ID | Reason for exclusion |
| Setting; study design | |
| Setting; outcomes | |
| Intervention |
Temperature
Description of studies on temperature:
One CCT (Table 28; Frank 1992) investigated the use of ambient room temperature in 97 patients (63 patients had an operation in a warm operating room and 34 patients had an operation in a cold operating room). Participants were 64.5 (range = 35 to 94) years old, and gender is not described. The study was conducted in the USA on surgical in‐patients.
| Study ID | Methods | Participants | Interventions | Outcomes |
| CCT; 2 parallel groups. | DESCRIPTION: 97 surgical in‐patients undergoing lower extremity vascular reconstruction in an ORa in Maryland, USA. | WARM OR: Patient underwent operation in one of a group of ORs with the ambient room temperature maintained at 24.5 (0.4) degrees Celsius. | ORAL TEMPERATURE: Measured with an electronic digital thermometer, pre‐operatively (before being transported to the OR), immediately post‐operatively, on arrival at the ICU, and every hour for 24 hours. |
aOR = Operating Room.
Intervention and control groups were: warm operating room of 24.5 ºc and cold operating room of 21.3 ºc.
Outcomes assessed were: re‐warming rate and oral temperature.
We have tabulated five excluded studies on temperature (Table 29).
| Study ID | Reason for exclusion |
| Study design | |
| Study design | |
| Study design | |
| Intervention‐ temperature of equipment not ambient room temperature | |
| Staff outcomes |
Risk of bias in included studies on temperature:
The included CCT (Frank 1992) used hospital scheduling rules and room availability to allocate patients to groups (allocation concealment not used), within these groups patients were also randomised to receive general or epidural anaesthesia. Oral temperature was measured with an electronic thermometer, but no blinding of group allocation was reported. Completeness of outcome data is reported for oral temperature (this is unclear for the outcome 'shivering'). However, the paper does not give a description of withdrawals and drop‐outs. It is unclear if the study is at risk of selective outcome reporting. Groups were protected against contamination.
Findings from studies on temperature:
Oral temperature
Findings from non‐randomised studies:
Frank 1992 found, that for patients undergoing operations with general anaesthesia, those assigned to warm operating rooms had less temperature loss than those assigned to colder rooms (MD −0.80 ºc, 95% CI −1.35 to −0.25, P value = 0.005). This difference was not apparent in patients undergoing operations with an epidural anaesthetic (MD 0.20 ºc, 95% CI −0.35 to 0.75, P value = 0.48).
Ward layout
Description of studies on ward layout:
There are no studies on ward layout included in the review; We have tabulated 10 excluded studies (Table 30).
| Study ID | Reason for exclusion |
| Not health‐related outcomes; policy changes | |
| Not health‐related outcomes; query validity | |
| Not a hospital setting | |
| Qualitative | |
| Design; outcomes | |
| Not health‐related outcome | |
| Outcomes | |
| Study design | |
| Confounding staffing differences | |
| Qualitative; participants |
Wayfinding
Description of studies on wayfinding:
There are no studies on wayfinding interventions included in the review; We have tabulated eight excluded studies (Table 31).
| Study ID | Reason for exclusion |
| Setting | |
| Study design | |
| Setting | |
| Study design‐ inappropriate data | |
| No intervention; comparison with healthy controls | |
| Not health‐related outcomes; conversation confounding | |
| Setting | |
| Participants not patients |
Windows
Description of studies on windows:
There are no included studies on the provision of windows included in the review. Three excluded studies have been tabulated (Table 32).
| Study ID | Reason for exclusion |
| Study design | |
| Study design, validity of outcomes | |
| Study design‐ retrospective |
Multifaceted interventions
Whole unit design
Description of studies on whole unit design:
Two CCTs with multifaceted interventions have been included in the review (Table 33; Kasmar 1968; Vaaler 2005). In these two studies there were 171 psychiatric participants. Kasmar 1968 does not make clear how many participants were assigned to each group (details of 115 participants are unknown), however in Vaaler 2005 there were 31 participants in a refurbished wing, and 25 in the traditional wing. Participants in Kasmar 1968 were aged 16 to 66 years old, it is unclear how many participants were under 18 years old (34% were aged 16 to 22 years), or if indeed this study should be excluded in the review for this reason. Participants in Vaaler 2005 were on average 37.07 years old. Overall, there were 84 males and 87 females included in the studies, with one conducted in the USA (Kasmar 1968) and the other in Norway (Vaaler 2005). One study was conducted in an out‐patient psychiatric treatment room (Kasmar 1968) and one in an in‐patient seclusion area (Vaaler 2005).
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| CCT; 8 parallel groups (2 x rooms, 2 x psychiatrists, 2 x time of data collection). | DESCRIPTION: 115 applicants for out‐patient psychiatric treatment at the Neuropsychiatric Institute, Los Angeles, USA. | 'BEAUTIFUL ROOM': Carpeted in burnt‐yellow carpeting and contained an abstract picture on one wall, a floor‐sized artificial plant, a wooden waste‐basket, and indirect lighting provided by a contemporary desk lamp. The room was neat and well kept.'UGLY ROOM': Carpetless, with beige asphalt floor tiling, overhead fluorescent lighting and was unkempt, with work papers strewn over the furniture and an overflowing grey metal wastebasket and ashtray. Both rooms were windowless offices of identical size (6 x 8 x 8 feet), wall covering, and colour. The furniture was the same (brown metal desks with beige formica tops, green leatherette desk chairs, and green and yellow leatherette side chairs. The offices varied only in decor. | ROOM RATINGS: not included in review. | Data not presented in enough detail for extraction. | |
| CCT; 2 parallel groups. | DESCRIPTION: 56 psychiatric in‐patients in the seclusion area of an acute psychiatric ward in Norway. | REFURBISHED WING: Redecorated and refurbished with the aim of looking, as much as security permitted, like an ordinary Norwegian home: Wainscoting walls, colourful wallpaper and paintings, lowered ceilings, multiple lighting spots, tasteful curtains, wardrobes, chairs, flowers, personal items, Italian ceramic‐tiled bathroom. | SYMPTOMS AND PSYCHOPATHOLOGY: Scored on the Positive And Negative Syndrome Scale (PANSS) for schizophrenia, with the time criterion of the last 24 hours. Scale has scores for total, positive, negative, and general symptoms. Assessed on admittance, day 3, and at discharge from seclusion. | Data extracted for change from baseline (beginning and end time‐points only).PANSS score for 'total' extracted for review. No significant difference was found on any of the subscales. |
Interventions were a 'beautiful room', neat and well‐kept, with burnt‐yellow carpeting, abstract picture, artificial plant, wooden waste‐basket, and indirect lighting provided by a contemporary desk lamp (Kasmar 1968), and a refurbished wing (Vaaler 2005), which was redecorated and refurbished to look like a Norwegian home, with wainscoting walls, colourful wallpaper and paintings, lowered ceilings, multiple lighting spots, "tasteful" curtains, wardrobes, chairs, flowers, personal items, and an Italian ceramic tiled bathroom. Control areas were an 'ugly room' (Kasmar 1968), which was carpetless, beige asphalt floor tiling, overhead fluorescent lighting, unkempt, with work papers strewn over the furniture and an overflowing grey metal wastebasket and ashtray, and a 'traditional wing' (Vaaler 2005), which at four years old was well kept with few signs of damage, had sparse furniture, grey walls, lacking pictures, no window curtains, single lamps in the 4 m‐high ceilings, bathrooms with grey, laminated paint, and patient rooms with a single bed and metal‐tubed chair.
Health‐related outcomes assessed by the studies were: mood ratings, psychiatric symptoms and psychopathology, function, violent behaviour, and length of stay.
We have tabulated 61 excluded studies (Table 34) on whole unit design.
| Study ID | Reason for exclusion |
| Descriptive article | |
| Descriptive article | |
| Study design | |
| Descriptive article | |
| No intervention | |
| No intervention | |
| CBA‐ Control sites and 'before' sites not clearly defined | |
| Review article | |
| Setting | |
| Policy confound | |
| Policy confound | |
| Study design | |
| Descriptive article | |
| Study design | |
| Review article | |
| Review article | |
| Study design | |
| Descriptive article | |
| Descriptive case study | |
| Study design | |
| Review article | |
| Study design | |
| Outcomes | |
| Outcomes | |
| Qualitative; setting | |
| Setting | |
| Literature review | |
| Setting | |
| Intervention | |
| Study design (same study as Lawson 2003) | |
| Study design (same study as Lawson 2003) | |
| Study design (before‐and‐after no contemporaneous control) | |
| Study design | |
| Study design (before‐and‐after no contemporaneous control) | |
| Study design | |
| Policy confound | |
| Qualitative study | |
| Study design | |
| Study design | |
| No intervention | |
| Confounding | |
| Literature review | |
| Study design | |
| Study design | |
| Study design | |
| Outcomes; unclear study design | |
| Policy confound | |
| Confounding staffing differences | |
| Literature review | |
| Intervention (care delivery systems) | |
| Study design | |
| Follow‐up study (Smylie 1971, Davidson 1971); study design | |
| Study design | |
| Study design (gender differences) | |
| Study design | |
| Unclear intervention/control | |
| Study design | |
| Qualitative interviews | |
| Overview article | |
| Study design | |
| Descriptive case studies |
Risk of bias in included studies on whole unit design:
Two CCTs are included on whole unit design. In Kasmar 1968, participants were assigned to one of eight conditions (two x room types, two x psychiatrists, two x time of data collection). For the purposes of this review, this study has been classified as a CCT (non‐randomised) as although there was an element of randomisation (method not described), it appears that participants were not randomly allocated to room type, as this was dependent on the previous allocation. The paper states that psychiatrist and room type were matched. In Kasmar 1968, allocation concealment is unclear. Blinding of healthcare personnel and assessment of outcomes was not done (and not feasible), and it is unclear if there was completeness of dataset and there is no description of withdrawals and drop‐outs. This study is at risk of selective outcome reporting. Protection against contamination seems likely.
In Vaaler 2005, patients were admitted to the wing with fewest patients, or (if there were even numbers of patients on each wing) to the wing which did not receive the last patient. Allocation concealment was not used. No blinding was possible in this study. It is unclear whether data was obtained for > 80% of participants. In Vaaler 2005, participants were not asked for consent due to their condition and all admitted patients were included apart from one with senile dementia. It is unclear if there were any withdrawals or drop‐outs. It is unclear if this study is at risk of selective outcome reporting. It is unclear if patients were retained to their wing, or if there was possible contamination.
Findings from studies on whole unit design:
Mood Ratings
Studies with insufficient data for extraction:
Kasmar 1968 found no strong evidence of an effect on mood between groups exposed to a 'beautiful room' or an 'ugly room'.
Psychiatric symptoms
Findings from non‐randomised controlled trials:
Vaaler 2005 found no strong evidence of an effect between a refurbished wing group and a traditional wing group in psychiatric symptoms (MD −7.60 points on the Positive And Negaitve Syndrome Scale (PANSS), 95% CI −16.81 to 1.61, P value = 0.11).
Symptoms and function
Findings from non‐randomised controlled trials:
Vaaler 2005 found no strong evidence of an effect between a refurbished wing group and a traditional wing group in function (MD −1.00 points, 95% CI −5.75 to 3.75, P value = 0.68) or symptoms (MD −2.00 points, 95% CI −8.69 to 4.69, P value = 0.56), as measured by the Global Assessment of Function Scale ‐split version (GAF‐S).
Violence
Findings from non‐randomised controlled trials:
Vaaler 2005 found no strong evidence of an effect between a refurbished wing group and a traditional wing group in violence (MD −0.04 points on the Brøset Violence Checklist, 95% CI −0.69 to 0.61, P value = 0.90).
Length of stay
Findings from non‐randomised controlled trials:
Vaaler 2005 found no strong evidence of an effect for length of stay between a refurbished wing group and a traditional wing group (MD −1.80 days, 95% CI −6.18 to 2.58, P value = 0.42).
Discussion
Summary of main results
Providing positive distracters
We found studies for a number of interventions which could be used as positive distractions in hospital environments. These interventions included: music, audiovisual distractions, visual distractions, decoration, and aromas. Studies reported on a variety of health‐related outcomes, and here we will focus on patient‐reported anxiety and pain, anxiolytic and pain medication requirements, blood pressure, heart rate, and respiration rate.
In general, the findings support the use of music for reduction of patient‐reported anxiety, particularly in the pre‐procedure period which was favourable compared with both standard care and blank headphones. For use during medical procedures, music was preferable to standard care but there was no strong evidence of an effect when music was compared with a blank tape with headphones (note, we have not statistically assessed the difference between the comparison of 'music versus standard care' and 'music versus blank tape with headphones'). This may imply that it is the reduction of unpleasant noise, rather than the addition of music, which accounts for significant findings when music is compared with standard care, or that the studies using standard care as a control group were at higher risk of bias (as healthcare personnel were not blinded to study groups). Alternatively, the attention associated with providing patients with a blank tape and headphones may explain why this is just as effective as providing music. Some studies reported on anxiolytic medication requirements, but these findings did not always parallel the findings for patient‐reported anxiety. For physiological measures (heart rate, blood pressure, and respiration rate), findings are less positive. On these outcomes mixed results are obtained when comparing music with standard care. However, as with some of the findings for anxiety, when comparing music to blank tape (or headphones), studies tend to show no strong evidence of an effect. Only one study (comparing music with pre‐recorded operating room noise) found music to increase systolic blood pressure, and all other outcomes reported by all studies, if not positively in favour of music, demonstrated that the addition of music did not do any harm. Evidence on music for pain relief is reported in Cepeda 2006.
It is unclear whether patients should be provided with a choice of music or provided with set pieces. For some outcomes and comparisons, set pieces seem preferable, whereas for others a choice of music appears preferable. This may reflect that it is not 'having a choice' which is most important but rather the 'content' of the music that makes the difference. The geographical location of studies does not appear to influence the heterogeneity of the findings so it appears that patients may react similarly to music interventions regardless of location. In this review, we have not explored the relationship between musical content and country of study, or how the content of music relates to the study findings.
There were fewer studies that investigated the use of audiovisual distractions. Three audiovisual studies had mixed results (showing positive findings and no difference) for anxiety compared with standard care. A number of explanations may explain these differences, including the type of audiovisual distraction (the dynamic distractions were positive and the static distraction showed no strong evidence of an effect). The studies on audiovisual distraction for patient‐reported pain were all positively in favour of audiovisual distraction (compared with standard care) but to different degrees. One study reporting on sedation medication requirements had results in the same positive direction as the patient‐reported pain findings. Two studies had heterogeneous findings for audiovisual distraction versus audio distraction for pain. Audiovisual distraction was no better than audio alone for anxiety, heart rate, and blood pressure.
One study compared audiovisual distraction with visual distraction alone on the outcomes pain and pain medications, and found audiovisual distraction to be preferable. One study found that audiovisual distraction was no better than scheduled rest for anxiety, pain, blood pressure and heart rate. The one study included on decoration had small study groups and found no difference between groups on anxiety, pain medication and length of stay. One study on aromas reported a positive effect for anxiety, whilst another found improvements in agitation levels. One study found no strong evidence of aroma effects for depression, fatigue, and general health.
On the whole, it appears that some positive distractions in hospital may prove worthwhile for improving patient‐reported outcomes such as pain and anxiety, however, the benefit of these interventions for physiological outcomes (heart rate, blood pressure, and respiration rate) has less support. It is also unclear if the benefits that positive distractions may have for patient‐reported outcomes translate directly in to reduced medication usage.
Reducing environmental stressors by implementing physical changes
Included in the review are studies that looked at lighting, air quality, temperature, bedroom types, flooring, and furniture; each reporting on a range of relevant outcomes.
Patient rooms facing directions which allow more sunlight in were found to be no different than darker rooms for reducing anxiety and pain (one study reporting on each). However, pain medication requirement was found to be less in sunnier rooms (findings from one study), as was perceived stress.
There was no strong evidence of an effect of air quality, as controlled by various air conditioning systems, on various measures of infection (number of infections, cases of septicaemia, incidence of invasive aspergillosis, number of septic episodes, presence of resistant bacteria, and acquisitions of Staphylococcus aureus), mortality, or remission. Additionally one study reported that there were cases of poor psychologic tolerance of confinement in a laminar air flow room. One study on temperature of operating theatres found that warmer rooms were beneficial for patients under general anaesthesia but made no difference to those with epidural anaesthesia in terms of temperature loss.
One study included on single‐ and multi‐bed rooms found no difference between room types on infection rates. Two studies were included on flooring and each investigated different outcomes. One study found no difference between carpet and linoleum for number of falls and fallers, independence, and length of stay. However, elderly people had a faster gait and longer stride on carpeted floor compared to vinyl. One study included on furniture, found reclining chairs to produce less patient‐reported pain and more satisfaction than hospital gurneys.
These studies represent a diverse range of physical modifications that could be made to hospital environments. There are few studies included for each intervention, and many of the studies included are at unclear or high risk of bias. Some environmental interventions appear to have positive effects, whilst others provide no strong evidence of an effect and may even indicate harmful effects.
Multifaced interventions
Two studies conducted on psychiatric populations found no differences between environments. This finding is important as it suggests that providing sparsely decorated and furnished rooms for psychiatric populations is not necessary for the control of psychiatric symptoms, and providing a pleasant environment at least does not do any harm.
Overall completeness and applicability of evidence
The majority of studies included in this review explored the use of music. Few studies were found to meet the inclusion criteria for other forms of environmental intervention. The inclusion criteria were broad, making the review applicable to all types of adult patients attending hospital for any reason. The downside to this inclusiveness means that much heterogeneity exists between the studies, so interpretation and application of the findings requires that attention is paid to the specific scenarios in which the included studies were conducted. There were a number of environmental interventions which were searched for but no studies meeting the methodological criteria were found. This means there are gaps in the evidence for interventions on: art, nature, ceilings, windows, and layout. It is sometimes argued that it is too logistically complex to conduct studies with good methodological designs on environmental interventions, however, this review has demonstrated that it is possible (as some studies have been included), however, for some interventions these types of studies have yet to take place.
To assist with the manageability of the review process, the inclusion criteria for this review was limited to the hospital setting; other evidence exists pertaining to, for example, dentists, nursing and care home environments, which has not been systematically assessed by this review. Furthermore, it is possible that the hospital environment may also impact on staff, having more indirect effects on patient care and outcomes; this evidence is currently lacking however, as summarised in another Cochrane systematic review focusing on the effects of the healthcare environment on staff outcomes (Tanja‐Dijkstra 2011).
Studies included in this review have a wide geographical spread, covering Australasia, North America, Asia, and Europe. There are no studies included from Africa or South America, and the transferability of the findings of this review to low‐to‐middle‐income countries is unknown. Heterogeneity between included studies was not easily explained by geographical location implying that this factor may not be so important (i.e. the findings are transferable across the locations assessed).
We have not undertaken to formally include economic evaluations as part of this review. For the majority of interventions pertaining to the sensory hospital environment, the evidence‐base for effectiveness is limited due to lack of studies and poorly rated studies, this should not be confused however with evidence of no effect; it may be cost‐effective to implement changes to the sensory environment, however this is not known and more robust evidence of effectiveness is required.
Quality of the evidence
The quality of the evidence is quite varied. Much of the evidence was rated poorly due to a lack of clarity in the reporting as opposed to definitively being at high risk of bias (although this could have also been the case). A number of non‐randomised controlled trials were found, and we did not summarise these for music interventions since there were many randomised trials conducted in this area. However, we have reported on non‐randomised controlled trials for other interventions where the body of evidence is much smaller. Whereas sometimes the approach of not randomly allocating patients to groups is understandable for logistical reasons (e.g. when patient allocation is dictated by hospital room availability); in other cases there is no real valid reason why randomisation could not have been used (instead of for example, alternate days).
Potential biases in the review process
Searching for studies in this area has proved particularly problematic, which is partially due to fact that many words associated with environmental interventions are used in other contexts with different meanings (for example, "art", "floor", "ventilation"), resulting in searches with very low specificity. It is hoped that, most importantly, the sensitivity of the search has been maximised through the wide range of sources searched, however, there is no way to guarantee the capture of all relevant studies. The review does now require updating and 69 studies are listed in Studies awaiting classification from an updated search. It is possible that once further relevant studies have been incorporated into the review findings, that the conclusions may change. All stages of selecting, appraising, and collecting data from studies in this review have been conducted independently by at least two people, in order to minimise bias and improve the robustness of decisions. The protocol for this review was written prior to the release of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2008), and we have tried to align the review with the updated guidance, which has involved deviating somewhat from the original protocol (See: Differences between protocol and review).
This review set out with a very broad question; in particular, including all "health‐related outcomes" may be seen as problematic as with multiple analyses there is higher potential for an analysis to be significant by chance alone. This issue, which is a problem related to random error as opposed to bias, is further manifested amongst individual studies, some of which contain multiple tests and many outcomes. We have handled this issue by reporting up to five relevant outcomes for each comparison and grouping the remaining reported outcomes for that intervention under a heading "other outcomes"; where only a minority of studies have reported an outcome, this is also highlighted in the text, so users of this review should be particularly wary of chance findings in these circumstances.
Agreements and disagreements with other studies or reviews
There have been a series of other reviews conducted in this area, although the foci of these reviews are not directly comparable to the inclusion and exclusion criteria of this current review. A number of reviews have been published (including in the grey literature) covering a broad range of environmental interventions (Dijkstra 2010; Ulrich 2004a and Ulrich 2008; Van den Berg 2005).
Ulrich 2004a conducted a review on "the role of the physical environment in the hospital" on patient (of all ages) and staff outcomes, which was later revised and updated (Ulrich 2008). The updated narrative review places more emphasis on the quality of studies than the original, however, the criteria against which studies were assessed is not clear ("each study was evaluated in terms of its research design and methods and whether the journal was peer‐reviewed"). The findings additionally draw on a range of evidence (excluding non‐English language studies), including qualitative studies, surveys, other literature reviews, and observational studies, and including indirect evidence (e.g. surface and air contamination levels as a proxy for hospital‐acquired infections which in turn is linked to reduced length of stay, and studies conducted in non‐hospital settings), and multifaceted interventions which have an element of environmental modification.
A series of interventions are advocated in Ulrich 2008, which are not supported by the present review either through lack of evidence or interpretation of the findings. These include the following.
-
Access to nature (e.g. large windows, art, and technological audio‐visual distractions) for pain, stress (but with limited evidence on nature art), length of stay, and satisfaction. Some of the studies reported to favour nature in Ulrich 2008 were included under audiovisual distractions in the present review (as the studies assessed pictures and sounds rather than real‐life nature and the findings do not support the use of nature per se due to the choice of comparison groups).
-
Daylight for pain (with limited evidence) and depression (drawing on a number of studies which utilised bright light treatment for depression, which were not included in the present review), and length of stay. The non‐significant findings (Walch 2005: pain and anxiety) highlighted in the present review are over‐interpreted by Ulrich 2008 as favouring sunnier rooms.
-
Single‐occupancy rooms for infection rates, reduction of stress, and improvement of sleep.
-
Rooms with better air quality, easy‐to‐clean surfaces, alcohol‐based hand‐rub dispensers in accessible locations, and carefully maintained water systems for reduction of infection rates.
-
Limited indirect evidence around the influence of noise, lighting, and acuity‐adaptable single patient rooms on medical errors.
-
Sound‐absorbing ceiling tiles and eliminating or reducing noise sources (e.g. adopting a noiseless paging system) for sleep.
In line with the current review, Ulrich 2008 found no conclusive evidence linking environmental interventions with reduced falls, and states there is little research assessing wayfinding systems on healthcare outcomes. Ulrich 2008 also assessed outcomes related to speech privacy and confidentiality, communication and social support for patients and family, and staff‐related outcomes, which were not covered in the present review. On the whole, compared to the present review, Ulrich 2008 is a more inclusive review, with the resulting conclusions reporting more confidently on the positive impact that hospital environments may have on patient outcomes. The present review has more stringent inclusion criteria and assessment of risk of bias and as such does not draw the same degree of positive conclusions.
Van den Berg 2005 reviewed studies on nature, daylight, fresh air, and quiet in healthcare settings. Van den Berg 2005 included a range of healthcare settings (including non‐clinical settings), age groups, and staff and patient health outcomes. Van den Berg 2005 took a more systematic approach than Ulrich 2004a by setting out the inclusion and exclusion criteria, and relating the findings to a methodological critique and relevance. Like the present review, Van den Berg 2005 concludes that there is insufficient evidence to support the use of hospital gardens for improved clinical outcomes. Van den Berg 2005 concludes that there is sufficient evidence that viewing nature can reduce stress and pain and that there is solid evidence that ventilation of fresh air is associated with improved health (these conclusions are not fully supported by the present review). Van den Berg 2005 reports that there is weak evidence to support the use of indoor plants (the present review included no studies of this nature) and that there is weak and inconclusive evidence for the health benefits of daylight (concurring with the present review findings). Van den Berg 2005 concludes that there is some evidence to support the promotion of quiet in healthcare settings, and this is supported in part by the present review, the findings of which suggest that the provision of music may reduce the impact of stressful noise.
Dijkstra 2006 also reviewed the evidence in this area. The inclusion/exclusion criteria of Dijkstra 2006 differ from the current review although there is some overlap in the studies included in each. Dijkstra 2006 only includes two studies on music and one on ocean sounds, but also includes studies that did not meet the methodological inclusion criteria of the present review. Dijkstra 2006 included more studies on multifaceted interventions, which did not meet the criteria of the present review; from these studies Dijkstra 2006 concludes that there is support for the notion that the environment impacts upon patient well‐being. With regard to single environmental interventions, the need for further research is suggested. Dijkstra 2006 did not incorporate any meta‐analyses, and included evidence on televisions, seating arrangements, and layout which did not meet the criteria for the present review. Despite these differences, in very broad terms, the overall conclusions (that there is a general lack of high quality evidence for many environmental interventions) are similar.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Funnel plot of comparison: 5 Music versus standard care, outcome: 5.1 Anxiety.
Comparison 1 Audiovisual distraction versus music, Outcome 1 Anxiety.
Comparison 1 Audiovisual distraction versus music, Outcome 2 Pain.
Comparison 2 Audiovisual distraction versus standard care, Outcome 1 Anxiety.
Comparison 2 Audiovisual distraction versus standard care, Outcome 2 Pain.
Comparison 3 Music versus blank tape / headphones only, Outcome 1 Anxiety.
Comparison 3 Music versus blank tape / headphones only, Outcome 2 Heart rate.
Comparison 3 Music versus blank tape / headphones only, Outcome 3 Blood pressure.
Comparison 3 Music versus blank tape / headphones only, Outcome 4 Respiration rate: with choice of outcomes.
Comparison 3 Music versus blank tape / headphones only, Outcome 5 Skin conductance.
Comparison 3 Music versus blank tape / headphones only, Outcome 6 Stress hormones.
Comparison 4 Music versus scheduled rest, Outcome 1 Anxiety.
Comparison 4 Music versus scheduled rest, Outcome 2 Heart rate.
Comparison 4 Music versus scheduled rest, Outcome 3 Blood pressure.
Comparison 4 Music versus scheduled rest, Outcome 4 Respiration rate.
Comparison 5 Music versus standard care, Outcome 1 Anxiety.
Comparison 5 Music versus standard care, Outcome 2 Heart rate.
Comparison 5 Music versus standard care, Outcome 3 Blood pressure.
Comparison 5 Music versus standard care, Outcome 4 Respiration rate.
Comparison 5 Music versus standard care, Outcome 5 Sedation requirements (anxiolytics).
Comparison 5 Music versus standard care, Outcome 6 Peripheral skin temperature.
Comparison 5 Music versus standard care, Outcome 7 Oxygen saturation.
Comparison 5 Music versus standard care, Outcome 8 Mood state.
Comparison 5 Music versus standard care, Outcome 9 Satisfaction.
Comparison 5 Music versus standard care, Outcome 10 Length of stay / Examination time.
Comparison 6 Music versus white noise, Outcome 1 Heart rate.
Comparison 6 Music versus white noise, Outcome 2 Blood pressure.
| Study ID | Methods | Participants | Interventions | Outcomes |
| RCT; 3 parallel groups. | DESCRIPTION: 313 patients undergoing radiotherapy, in Australia. | FRAGRANT PLACEBO: Patients were administered the carrier oil with low‐grade essential oils. The carrier oil was sweet almond cold‐pressed pure vegetable oil. The low‐grade fractionated oils (lavender, bergamot, and cedarwood) were of unknown purity (supplied by Naturistics, Hornsby, Australia). These fractionated oils were diluted with the carrier oil in a ratio of 1:2. | ANXIETY, DEPRESSION, and FATIGUE: Measured via the Hospital Anxiety and Depression Scale (HADS), and the Somatic and Psychological Health Report (SPHERE), which is composed of the General Health Questionnaire (GHQ) and Symptoms of Fatigue and Anergia (SOFA) scales. | |
| CCT; Cross‐over trial, 2 conditions. | DESCRIPTION: 15 psychiatric inpatients in the communal area of a long‐stay hospital psychogeriatric ward for patients with behavioural problems, in England. | LAVENDER: The communal area of the unit was diffused with a standard concentration of lavender oil (2%), using three aroma‐streams for a period of two hours between the period of 4pm and 6pm. | AGITATION: Measured on the 16‐point Pittsburgh Agitation Scale by a blinded observer for the final hour of each two hour study period. Outcomes are presented as median scores for each patient in each condition. | |
| SD: standard deviation | ||||
| Study ID | Reason for exclusion |
| Intervention | |
| Study design | |
| Study design | |
| Study design | |
| Setting and population | |
| Editorial | |
| Study design | |
| Outcomes | |
| Study design | |
| Data unsuitable for cross‐over study | |
| Commentary | |
| Setting | |
| Setting | |
| Aromas administered via nasal cannula, judged to be too invasive to constitute an 'environmental' intervention. | |
| Intervention; outcome not validated |
| Study ID | Reason for exclusion |
| Qualitative report | |
| Descriptive article | |
| Participants | |
| Intervention interactive | |
| Qualitative report | |
| News article | |
| Qualitative | |
| Descriptive article | |
| Narrative article | |
| Commentary | |
| No group comparisons presented, unable to obtain further details from authors | |
| Study design | |
| Study design (part of same study as Staricoff 2001) | |
| Conference abstract, unable to obtain further details from author | |
| Setting | |
| Setting |
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| RCT; 3 parallel groups. | DESCRIPTION: 96 in‐patients in the cardiovascular ICU and progressive care units having undergone elective coronary artery bypass grafting, in USA. | MUSIC GROUP: Choice of 5 tapes: 'Country Western Instrumental' or 'Fresh Aire' by Mannheim Steamroller, 'Winter into Spring' by George Winston, or 'Prelude' or 'Comfort Zone' by Steven Halpern. Played via headphones for 30 minutes. | Barnason 1995 reports: | Data extracted for state anxiety (STAI measure). Pain data are extracted for the end values on the VRS. | |
| RCT; 2 parallel groups. | DESCRIPTION: 80 in‐patients and out‐patients undergoing FB in Mariland, USA. | BEDSCAPES GROUP: Countryside river scene with associated sounds of nature played through headphones. Intervention available before, during, and after FB procedure. The scene was mounted by the bedside in the recovery area and on the ceiling in the procedure room. | ANXIETY: State anxiety via STAI; | SDs for anxiety have been estimated from P value of a t‐test. | |
| RCT; 3 parallel groups. | DESCRIPTION: 157 day‐patients undergoing colonoscopy in an Endoscopy Suite in Hong Kong, China. | VISUAL DISTRACTION: Eyetreck system (olympus) with preset home made movie (mainly scenic views), patient wears earphones but with no sound. | PAIN: scored using 10 cm VAS; | ||
| RCT; 3 parallel groups. | DESCRIPTION: 37 patients undergoing flexible sigmoidoscopy in Calfornia, USA. | AUDIOVISUAL: Virtual‐i glasses, personal display system showing an ocean shoreline with corresponding sounds (via headphones) | DISCOMFORT: Measured via VAS which asked patients to rate their level of abdominal discomfort from faint to severly intense. | Data for discomfort entered as pain scores. | |
| RCT; 2 parallel groups. | DESCRIPTION: 17 in‐patients undergoing burns care treatment and dressings change in a Burn Special Care Unit, Cincinnati, USA. | AUDIOVISUAL GROUP: "Muralvision" (Muralvision Studios, Inc., Eugene, Ore.)‐ on a bedside television, video programmes composed of scenic beauty (ocean, desert, forest, flowers, waterfalls, and wildlife) with accompanying music. | PAIN: Measured via the McGill Pain Questionnaire, with the Pain Rating Index and Present Pain Intensity scales. | For data extraction in the review, change scores from baseline were calculated and associated estimated standard deviations, using the F statistics provided. | |
| SD: standard deviation; STAI: State Trait Anxiety Inventory; VAS: visual analogue scale | |||||
| Study ID | Reason for exclusion |
| Intervention | |
| Study design | |
| Study design | |
| Intervention‐ interactive virtual reality | |
| Intervention‐ interactive virtual reality | |
| Intervention‐ patient education video | |
| Inadequate information provided (Ulrich study) | |
| Intervention interactive | |
| Unclear validity of outcomes | |
| Intervention interactive | |
| Intervention interactive | |
| Intervention‐ snoezelen | |
| Population, < 90% over 18 years old. | |
| Setting | |
| Setting | |
| Participants not adults |
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| RCT; 4 parallel groups. | DESCRIPTION: 39 in‐patients (10 post‐operative for cardiac surgery, 29 undergoing cardiac observations) in a cardiac care unit in Florida, USA. | BEIGE: Walls remained original colour of beige (similar to Sherwin Williams colour SW6658) in four rooms. | ANXIETY: Measured via STAI on day of discharge (after 2 to 5 days in hospital). Presented as mean (SD). No significant differences reported. | Partients in this study were not approached for informed consent until Day 3 of the study. | |
| STAI: State Trait Anxiety Inventory | |||||
| Study ID | Reason for exclusion |
| Outcomes not validated | |
| Qualitative report | |
| Setting | |
| Study design | |
| Study design | |
| Participants | |
| Descriptive article | |
| Study design | |
| Descriptive article | |
| Outcomes | |
| Counfounding | |
| Staffing confound |
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| RCT; 2 parallel groups. | DESCRIPTION: 40 day‐patients undergoing ophthalmic surgery in New York, USA. | MUSIC GROUP: Patient choice of 22 types of music (e.g. soft hits, classical guitar, chamber music, folk music, popular singers from 1940's and 1950's), played via headphones throughout pre‐operative, surgical, and post‐operative periods. | PHYSIOLOGICAL: HR and BP measured via Propaq Monitor (Protocol Systems, Inc., Beaverton, OR) every 5 minutes during pre‐operative, surgical, and post‐operative period. Averages for the last three recorded measures within each time period were used for analysis in the paper. For purposes of review, data is extracted for the mean post‐operative scores only. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 118 out‐patients undergoing colonoscopy in a Digestive Endoscopy Unit in Spain. | MUSIC GROUP: Series of classical tracks (e.g. Bach, Grieg, Mozart, Delibe, Faure, and Mendelssohn) played via headphones during procedure. | ANXIETY: State anxiety measure pre and post procedure using the STAI; Reported as post ‐ pre difference with 95% CI. | There were no significant differences between groups regarding abnormal events. This data has not been extracted for the review. | |
| RCT; 3 parallel groups. | DESCRIPTION: 83 in‐patients undergoing cardiac catheterization in Germany. | MUSIC GROUP: Music was played via headphones, which were worn half on so patients could still hear the medical personnel, during the cardiac catheterization. A music therapist was present only to control the volume. Music played was "Entspannung" [relaxation] by Markus Rummel, composed specially for relaxation. | ANXIETY: Measured via the STAI before and after cardiac catheterization. Unclear whether post measurements were taken on the following day after cardiac catheterization. | ||
| RCT; 3 parallel groups. | DESCRIPTION: 90 patients undergoing urological surgery with spinal anaesthesia and PCS in Connecticut (USA) and Beirut (Lebonon). | MUSIC GROUP: Patients brought own music from home. | PROPOFOL REQUIREMENTS: Recorded as mg/kg/min and % of patients not using any propofol. Unclear if data presented are the SDs, and if the data presented as mg/kg/min is based on the total N or % of patients who used propofol. | Data not extracted for meta‐analysis. | |
| RCT; 2 parallel groups. | DESCRIPTION: 107 patients undergoing diagnostic coronary angiography or a percutaneous intervention procedure, in Ontario, Canada. | MUSIC GROUP: patient selected music (classical, soft rock, relaxation, country, own/other) played via headphones before, during, and after procedure, continued as the patient desired. | ANXIETY: State Anxiety via STAI pre and post procedure; | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 3 parallel groups. | DESCRIPTION: 96 in‐patients in the cardiovascular ICU and progressive care units having undergone elective coronary artery bypass grafting, in USA. | MUSIC GROUP: Choice of 5 tapes: 'Country Western Instrumental' or 'Fresh Aire' by Mannheim Steamroller, 'Winter into Spring' by George Winston, or 'Prelude' or 'Comfort Zone' by Steven Halpern. Played via headphones for 30 minutes. | Barnason 1995 reports: | Data extracted for state anxiety (STAI measure). See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 20 patients undergoing regional anaesthesia in North Carolina, USA. | NATURE TAPES: choice of sounds of birds, the ocean, a lagoon, or deeply resonant chimes, played for the duration of the surgery. | STATE ANXIETY: STAI administered pre‐operatively and one hour post‐operatively to calculate the change score. | Scores given for state anxiety are outside of the normal range for this questionnaire (20‐80). Method of calculating scores is not described. | |
| RCT; 3 parallel groups. | DESCRIPTION: 95 in‐patients undergoing coronary artery bypass (n = 92) or valvular heart surgery (n = 3) in Ohio, USA. | TAPED THERAPEUTIC SUGGESTIONS: excluded from review. | POST‐OPERATIVE STAY (days); Data extracted for analyses. | See Cepeda 2006 for details on music for pain relief. | |
| RCT: 3 parallel groups. | DESCRIPTION: 156 in‐patients undergoing chest tube removal after open heart surgery in Virginia, USA. | MUSIC GROUP: Patients preselected music they would prefer to hear from a library of 10 pre‐recorded music cassettes with no lyrics. Cassettes were produced by students in a music therapy programme under the supervision of a music therapist. Patients listened to the music via headphones for 10 minutes prior to and then during chest tube removal. | PAIN INTENSITY: Measured using a 10 cm NRS at 3 time points: (1) 10 minutes prior to chest tube removal, (2) immediately after chest tube removal, (3) 15 minutes after chest tube removal. | Report states that 18 participants had missing physiological data. Unclear from which groups these belonged, so for purposes of data extraction, it has been assumed that 6 participants were missing from each group. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 170 pre‐operative patients to undergo vascular angiography in California, USA. | MUSIC GROUP: Selection of 5 categories (classical, jazz, rock, country western, easy listening), played via headphones for 15 minutes prior to angiography. Patients could continue to listen to music during the angiography after collection of outcomes. | ANXIETY: Measured via STAI before and after 15 minute study period. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 140 in‐patients in the cardiac units with intravascular sheaths or an intra‐aortic balloon pump (IABP) in place, in USA. | MUSIC GROUP: 30 minutes of music through headphones. Mixture of symphonic music and nature sounds selected by the researchers. Same music played to all those in the music group. | PAIN PERCEPTION: measured via 10 mm VAS pre‐ and post‐intervention. | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 193 day‐patients undergoing lithotripsy for renal stones (America). | MUSIC GROUP: Music of type preferred by patient played via headphones, starting 10 minutes prior to procedure and continuing until 10 minutes after lithotripsy is complete. Patients wore additional ear protectors to protect patients from the noise of the lithotriptor. | PAIN INTENSITY: rated verbally every 5 minutes throughout the lithotripsy on a NRS (0‐10). | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 220 female out‐patients undergoing colposcopy in China. | MUSIC GROUP: CD compilation of slow‐rhythm music (instrumental ballad). Patients could choose to listen to any song(s) within the compilation. Music played through speakers during the examination. Prior to onset of study women attending the clinic were surveyed on musical preferences to inform compilation disc. | ANXIETY: State anxiety via Chinese version of STAI measured pre and post colposcopy; | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 43 in‐patients undergoing application of a C‐clamp (which applies pressure to stop bleeding when sheaths are removed after percutaneous coronary interventions), in an ICU, Hong Kong, China. | MUSIC GROUP: 3 choices of soft, slow music without lyrics (slow rhythmic songs, Chinese slow rhythmic music, Western slow rhythmic music). Music played via headphones during the application of the C‐clamp (approximately 45 minutes). | VITAL SIGNS: BP, HR, RR, and oxygen saturation, recorded at baseline, 15, 30, and 45 minutes via a bedside monitor. | Outcomes extracted for end time‐point only. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 64 patients undergoing cesarean section in Taiwan. | MUSIC GROUP: Choice of western classical, new age, or Chinese religious music, played via headphones. Participants listened to music for at least 30 minutes from start of anaesthesia to end of surgery. Volume low enough to allow mutual conversation with the researcher. | ANXIETY: measured via a 10 cm VAS, researcher filled it in after asking participant to indicate how they were feeling. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 20 in‐patients receiving mechanical ventilation in private patient room areas in critical care units in Midwest USA. | MUSIC GROUP: Selection of classical music played for 30 minutes via headphones. Patients instructed to close eyes and concentrate on the music. | MOOD: Short‐form POMS. | No SDs are presented for the post‐intervention physiological outcomes. Significant differences were found for HR and RR in favour of music. Other variables did not differ significantly. | |
| RCT; 2 parallel groups. | DESCRIPTION: 54 in‐patients receiving mechanical ventilation in one of 4 Intensive Care Units, USA. | MUSIC GROUP: choice of non‐lyric tapes 60‐80 bpm, classical, new age, country western, religious, and easy listening (30 mins). | ANXIETY: State anxiety via short form STAI (6 items); | Due to the lack of clarity over withdrawals and drop‐outs, it is unclear how many people were analysed in each group for the outcome anxiety. The degrees of freedom stated (49) suggests 51 observations were made, and we have assumed that N = 25 in the music group and N = 26 in the control group, based on the descriptions given for 3 of the withdrawals in the paper. | |
| RCT; 2 parallel groups. | DESCRIPTION: 64 out‐patients undergoing flexible sigmoidoscopy (FS) in Midwest USA. | MUSIC GROUP: choice of music (classical, country‐western, new‐age, easy listening, pop, rock, religious, jazz, era‐specific, motion picture soundtracks), played via headphones during FS procedure. Patients instructed to concentrate on music and that the investigator would meet with them afterwards to discuss their experiences. | ANXIETY: State anxiety measured via STAI | ||
| RCT; 2 parallel groups. | DESCRIPTION: 68 pre‐operative patients undergoing extracorporeal shock wave lithotripsy (ESWL) in Taiwan. | MUSIC GROUP: listened to natural music via headphones for 5 minutes prior to ESWL. | BLOOD PRESSURE: Method of data collection unclear. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 60 in‐patients and out‐patients undergoing flexible fibreoptic bronchoscopy (FB) in California, USA. | MUSIC GROUP: "Relax" (Expansion Records, Manchester, UK) consisting of piano improvisations (60bpm), played via headphones during FB procedure. | ANXIETY: State and Trait anxiety measure via STAI. | ||
| RCT; 3 parallel groups. | DESCRIPTION: 180 pre‐operative day patients scheduled for day surgery in Australia. | MUSIC GROUP: Choice of classical, jazz, country & western, new age, easy‐listening, and "other" mostly by contemporary artists. Music played for 30 minutes via headphones. | ANXIETY: State anxiety measured via STAI. | 95% Confidence Intervals are presented in the paper based on logarithmically transformed scores. Standard deviations were calculated by first back‐translating data to a log scale before utilising the CIs to estimate the SDs. Data entered into the review analysis are expressed as the natural log. | |
| RCT; 4 parallel groups. | DESCRIPTION: 121 patients undergoing elective cataract extraction in Canada. | MUSIC GROUP: Classical music accompanied by soothing sounds of nature, played via headphones intra‐operatively. | ANXIETY: Measured via STAI before and after surgery. | Data reporting unclear. Cannot extract SDs. | |
| RCT; 3 parallel groups. | DESCIPTION: 90 pre‐operative in‐patients waiting for dental restoration surgery under general anaesthesia, or orthopedic surgery, in Germany. | MUSIC GROUP: listened to 45 minutes of music pre‐operatively. Choice of music arranged the evening before surgery. 11 patients chose classical, 19 chose pop music. | ANXIETY: Measure via STAI and a tick‐box anxiety scale tailored for the clinic. There was a significant decrease in state anxiety from pre‐treatment to post‐treatment in the music group (change score = 2.2). There was no significant change in anxiety for the 'no medication' control group (change score = 0.633). | Not enough information for data extraction. | |
| RCT; Cross‐over study. | DESCRIPTION: 24 in‐patients with MI (N = 12) and other cardiac conditions (N=12) in a Critical Care Unit, Ontario, Canada. | MUSIC: 3 pieces each lasting approximately 12 minutes played in randomised Latin square design via headphones. a) Symphony no.6 by Beethoven; b) Eine Kleine Nachtmusik by Mozart; and c) The Moldau by Smetana. | PSYCHOLOGIC: Questionnaire not fully validated (content validity only). Not extracted for review. | Significant order effects for HR variability (P = 0.03) and heart rhythm (ectopy). | |
| RCT; 3 parallel groups. | DESCRIPTION: 143 out‐patients (93 included in present review) undergoing screening mammography in a clinic of a tertiary care hospital in Massachusetts, USA. | RELAXATION GROUP: excluded from review, as taped instructions constitute a psychological therapy. | ANXIETY: recorded before and after study period using the STAI, and at the end of the study with a Likert scale (1‐10) asking to rate level of anxiety felt during the procedure. | Investigator identified a possible floor effect. See Cepeda 2006 for details on music for pain relief (this study is not yet included in Cepeda 2006). | |
| RCT; 3 parallel groups. | DESCRIPTION: 56 in‐patients with unstable angina pectoris or acute MI at coronary care unit in Australia. | MUSIC GROUP: Received two or three 30‐minute sessions of light classical music (Bonny,Music Rx) played via headphones. | ANXIETY: Measured at pre and post test with three psychologic scales, 1) STAI, 2) Hospital Anxiety and Depression Scale, and 3) Linear Analogue Anxiety Scale (VAS). STAI scores extracted for review. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 33 in‐patients undergoing bone marrow transplant chemotherapy, in Columbus, USA. | MUSIC GROUP: Listened to 45 minute recording of self‐selected music via headphones at 6, 9, and 12 hours after the start of each infusion as an adjunct to antiemetic therapy. A variety of music selections was available and patients were encouraged to bring their favourite music. | NAUSEA: Measured on a VAS in the pictorial form of a thermometer (questionable validity). | Data not in sufficient detail for extraction. | |
| RCT; 2 parallel groups. | DESCRIPTION: 11 patients undergoing range‐of‐motion exercises as part of acute care rehabilitation for burns, Virginia, USA. | MUSIC GROUP: Choice of 6 cassette tapes (Lifescapes series) that met the guidelines for music in medical settings. Music was played during the range‐of‐motion exercises. | PAIN: Measured via VAS before and after rehabilitation exercises. There was a statistically significant increase in pain from pre‐treatment to post‐treatment in both groups (P = 0.04). There was no significant difference between groups (P = 0.38). | Data not in sufficient detail for extraction. See Cepeda 2006 for details on music for pain relief. | |
| RCT (post‐test only); 3 parallel groups. | DESCRIPTION: 15 pre‐operative patients in the waiting room for elective same‐day surgery in Pittsburgh, USA. | HUMOUR GROUP: excluded from review. | ANXIETY: Measured via a VAS after the intervention period. | ||
| RCT (post‐test only); 3 parallel groups. | DESCRIPTION: 46 pre‐operative patients scheduled for same‐day surgery in Pittsburgh, USA. | HUMOUR GROUP: excluded from review. | ANXIETY: Measure post‐intervention via VAS. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 50 pre‐operative patients to undergo septorhinoplastic surgery in Turkey. | MUSIC GROUP: Patients brought own CD or tape from home, those who forgot were provided with a 'suitable replacement', played via headphones during pre‐operative period, as patients were being sedated. | "modified" OAA/S: not included in review. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 93 in‐patients scheduled for laparoscope surgery, in Beijing, China. | MUSIC GROUP: Choice of 6 types of music (Western classical; light; pop; folk; folksong; opera) with 30 minutes listening time. Patients listened to music via headphones 1‐2 hours before their operation, whilst lying in bed. | ANXIETY: Measured via STAI before and after intervention period. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 78 patients undergoing colonoscopy in Kerala, India. | MUSIC GROUP: Choice of 6 tapes played via headphones during colonoscopy. Selection of: popular film songs (based on carnatic classical ragas), classical music, devotional songs, folk songs, soft instrumental music, bioacoustics (soft instrumental music with nature sounds). | SEDATION: Dose of midazolam (2 mg given on demand). | Data reported as median and range. See Cepeda 2006 for details on music for pain relief (this study is not yet included in Cepeda 2006). | |
| RCT; 2 parallel groups. | DESCRIPTION: 198 out‐patients awaiting gastrointestinal procedures (colonoscopy or esophagogastroduodenoscopy [EGD]) in California, USA. | MUSIC GROUP: Patient selected music (classical, rock, jazz, country western, easy listening) for 15 minutes prior to medical procedure. Were allowed to continue listening to music after outcome measures were taken. | ANXIETY: state anxiety measured via STAI. | This study has questionable clinical relevance: "To avoid introducing more anxiety with the explanation of the [gastrointestinal] procedure, the consent process for the [gastrointestinal] procedure was delayed until after the patients completed the music study". | |
| RCT; 3 parallel groups. | DESCRIPTION: 60 in‐patients who have undergone surgery with general anaesthesia, in a PACU in Iowa, USA. | MUSIC GROUP: Choice of three instrumental tapes: Calm classical (e.g. Bach, Debussy, Pachelbel), Stimulative classical (e.g. Strauss, Tschaikovsky), and calm popular music (e.g. piano solos by George Winston, guitar solos by William Ackerman and Steve Halpern). Patients decided which music they would like to listen to in the PACU in the pre‐operative visit. Played via headphones until discharge from PACU. | PAIN: Measured via 10cm VAS every 15 minutes while in the PACU. | Data insufficient for extraction. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 55 day‐patients undergoing laparoscopic sterilization or laparoscopic tubal dyeing as part of a fertility programme in Sweden. | MUSIC GROUP: Panpipe music played via headphones for 30 minutes pre‐operatively and 30 minutes post operatively. | PHYSIOLOGICAL MEASURES: RR, BP, and HR, were measured by an attending nurse blinded to group allocation, before and after each 30 minute session (pre‐ and post‐surgery). | Sensitivity analyses conducted using data extracted for pre‐operative and post‐operative scores for anxiety, RR, BP, and HR. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 3 parallel groups. | DESCRIPTION: 110 in‐patients and out‐patients undergoing IV catheter insertion in Southwestern public and private hospitals, USA. | SALINE GROUP: excluded from review as confounding non‐environmental intervention. | PAIN INTENSITY: measured via 100mm VAS. | See Cepeda 2006 for details on music for pain relief. Insertion difficulty is not included in the present review as a health‐related outcome. | |
| RCT; 3 parallel groups. | DESCRIPTION: 76 in‐patients undergoing surgery with general anaesthetic, in Sidcup, UK. | MUSIC GROUP: Classical music ('Adagio' Karajan, Deutsche Grammophon, 445 282‐4). Played via headphones during surgery. | FENTANYL REQUIIREMENTS: This served as an indication of the adequacy of nociception control provided during the operation. Fentanyl (ųg) was given intravenously if intra‐operative BP or HR increased by 20% or more above baseline values for more than 5 minutes. | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 34 out‐patients undergoing urologic procedures using spinal anaesthesia, USA. | MUSIC GROUP: All patients requested to bring their favourite CD to hospital on morning of surgery. A suitable substitution was provided to those who did not have access to their favourite CD. Music played via occlusive headphones intraoperatively. | BP (mm Hg), HR (bpm), PROPOFOL REQUIREMENTS (mg/min): recorded every 10 minutes for duration of surgery; | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 43 patients undergoing lithotripsy, USA. | MUSIC GROUP: Patients brought own music from home to listen to via occlusive headphones during surgery. | BP (mm Hg), HR (bpm), ALFENTANIL REQUIREMENTS: recorded every 15 mins. | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 3 parallel groups. | DESCRIPTION: 163 in‐patients undergoing hysterectomy in Vienna, Austria. | MUSIC GROUP: Patient choice of classical (Bach), Entertainment ("Musik zum Träumen" [Music for Dreaming]), or Relaxation (Oliver Shanti "Rainbow Way"). Played via headphones for 45 minutes beginning at the start of the operation (at the abdominal incision). | PAIN: Pain intensity measured via 10 cm VAS once a day for 5 days. There were no differences between groups. | Where means and SDs are presented (for length of stay), there are missing data from 23 participants so it is unclear how many people are in each group. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 104 patients undergoing gastrointestinal endoscopy in India. | MUSIC GROUP: Classical Indian instrumental music played for 10 minutes prior to procedure and then throughout the procedure. | PHYSIOLOGICAL OUTCOMES: BP, HR, and RR measured at beginning of the consultation and at the end of the procedure. Methods unclear. Data presented as change scores. | ||
| RCT; 3 parallel groups. | DESCRIPTION: 58 patients undergoing noxious medical procedures (e.g. tissue biopsy, vascular port placement) at an oncology clinic in Midwestern USA. | MUSIC GROUP: Patient choice from selection of CDs (pop, rock, easy listening, classical, religious hymns, jazz or blues, country), listened via headphones prior to and during procedure. | PAIN: Pain intensity measured via a NRS prior to treatment, during treatment (retrospectively), and post treatment. | Values provided are adjusted scores (from analysis of covariance), adjusted for baseline scores, medications, and gender. Numbers reported as SDs in the text although too small so assumed to be the standard errors, which reflect the non‐significant findings described. See Cepeda 2006 for details on music for pain relief | |
| RCT; 3 parallel groups. | DESCRIPTION: 165 out‐patients undergoing elective colonoscopy in Hong Kong, China. | MUSIC ALONE: excluded from review as no appropriate control (i.e. music with no PCS). | PAIN SCORE: 10 mm VAS; | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 64 in‐patients on mechanical ventilation in an Intensive Care Unit in Hong Kong, China. | MUSIC GROUP: Choice of Chinese classical music, religious music (Buddhist and Christian), Western classical, natural sounds with slow beats. Played via headphones. | ANXIETY: State anxiety using Chinese STAI ‐short version (6 items). Participants responded to questions by holding up corresponding number of fingers. | ||
| RCT; 3 parallel groups. | DESCRIPTION: 37 patients undergoing flexible sigmoidoscopy in Calfornia, USA. | AUDIOVISUAL: Virtual‐i glasses, personal display system showing an ocean shoreline with corresponding sounds (via headphones) | DISCOMFORT: Measured via VAS which asked patients to rate their level of abdominal discomfort from faint to severely intense. | Data extracted for review on anxiety and anger. Arousal and attention not considered health‐related outcomes. | |
| RCT; 2 parallel groups. | DESCRIPTION: 50 in‐patients and out‐patients undergoing non‐oncologic surgery under spinal anaesthesia in Canada. | MUSIC GROUP: Choice of pop, jazz, classical, and new age played via non‐occlusive headset. Patients received the music prior to, during and after surgery. | PCS: Amount of midazolam consumed during the perioperative period was recorded. Data extracted for total midazolam consumed (mg). | Estimated anxiety data extracted. | |
| RCT; 2 parallel groups. | DESCRIPTION: 35 in‐patients in one of 5 Intensive Care Units having had an acute myocardial infarction, Midwestern USA. | MUSIC GROUP: 3 sessions of music listening on 1st, 2nd, and 3rd day (or 2nd‐4th day) of hospitalisation. Music session consisted of listening to 3 pieces of music each session: Bach's 'Largo', Beethoven's 'Largo', and Dubussy's 'Prelude to the Afternoon Faun'. Each session lasted approximately 22 minutes. | STATE ANXIETY: measured with the STAI at two time points: (1) during the first 48 hours of admission, and (2) on the 3rd or 4th day of hospitalisation. For the music group, these measurements were taken prior to the first music session and at the end of the 3rd music session. | ||
| RCT; 3 parallel groups. | DESCRIPTION: 45 patients undergoing femoral angiography in Boston, USA. | RELAXATION TAPE: Progressive muscle relaxation. Excluded from review. | STATE ANXIETY: Measured via STAI immediately pre‐ and post‐procedure. | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 44 post‐operative in‐patients in an Orthopaedic Department, Japan. | MUSIC GROUP: Choice of: Western classical music, Gagaku (Japanese traditional court music), Noh songs, or Enka. All participants choose Enka (a melodramatic and representative genre of Japanese popular songs, usually about sad aspects of life, irrecoverable destiny, and desertion by a lover, sung with a slow tempo), which is popular among elderly Japanese people. There were 10 Enka CDs to choose from. Music played via headphones for 20 minutes whilst lying in bed. | PAIN: Measured via 10 cm VAS and the Wong/Baker Faces Pain Rating Scale (includes 6 categories of facial expressions ranging from '0, a happy smiling face' and '5, a tearful face'). | See Cepeda 2006 for details on music for pain relief (study not yet included in Cepeda 2006). | |
| RCT; 4 parallel groups. | DESCRIPTION: 52 in‐patients undergoing various surgical procedures in USA. | MASSAGE THERAPY: excluded from review. | STATE ANXIETY: measured via the short‐form STAI‐6 in the PACU (after surgical procedure). | See Cepeda 2006 for details on music for pain relief (study not included in Cepeda 2006). | |
| RCT; 2 parallel groups. | DESCRIPTION: 38 emergency department admissions undergoing laceration repair in Pittsburgh, USA. | MUSIC GROUP: Listened to music via headset. Choice of 50 styles and artists to choose from. | ANXIETY: State anxiety measured via STAI before and after laceration repair. SDs presented in graphical format. Estimated figures extracted for purposes of review. | SDs for pain and anxiety are estimates from graphs. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 30 patients undergoing surgery with general anaesthesia, Canada. | MUSIC GROUP: Selected a CD preoperatively from a choice of 4: classical, jazz, new‐age, popular piano music. Selected listening volume. Anesthesiologist started CD once patients were anaesthetised. Music played via headphones. | Repeated measures observations (4 time points)‐ T1: immediately after arterial line insertion, T2: 5 min after peritoneal incision, T3: at skin closer, T4: 30 min after arrival in the recovery area. | Data extracted for end time points only. See Cepeda 2006 for details on music for pain relief | |
| RCT; 2 parallel groups. | DESCRIPTION: 28 post‐operative in‐patients who have undergone elective abdominal hysterectomy, USA. | MUSIC GROUP: Four easy listening selections (out of an original 10) were selected by nursing graduate students to be harmonious, pleasant, and calming. Patients listened to music for 10 minutes via headphones and were requested to close eyes on the first and second post‐operative day. | PAIN: measured via VAS on first and second post‐operative day (pre and post 10 minute intervention period). | Outcomes extracted for 2nd post‐operative day as only 6 and 5 outcome measures were obtained for the intervention and control group on the 1st post‐operative day (all 28 participants completed the 2nd day of testing). See Cepeda 2006 for details on music for pain relief. | |
| RCT; 3 parallel groups. | DESCRIPTION: 90 (58 included in present review) in‐patients undergoing elective hysterectomy in Sweden. | MUSIC GROUP: relaxing music accompanied by sea waves (ref: Uneståhl L‐E. Avslappningmusik. Träningsprogram för kropp och själ. [Relaxation music. Training programme for body and soul]. Örebro, Sweden Veje International AB: 1970). Played through headphones from time of skin incision to time of wound closure. | PAIN INTENSITY: 10‐point VAS used every hour for the first 24 post‐operative hours, and every 3 hours after that until the patient felt no pain. Data presented as group mean score of patients' median score. | Data extracted for responses given on the day of surgery. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 3 parallel groups. | DESCRIPTION: 151 day patients undergoing surgery of varicose veins or inguinal hernia repair under general anaesthesia, in Sweden. | INTRA‐OPERATIVE MUSIC GROUP: Music played via headphones from end of induction of anaesthesia until wound dressing via headphones. Music was soft instrumental comprising 7 melodies of a new‐age synthesizer. Post‐operatively, in the PACU, patients were exposed for 1 hour to a blank (silent) disc via headphones. | PAIN INTENSITY: Post‐operatively on a NRS (0 to10), every 30 minutes for 2 hours in the PACU. | Anxiety and fatigue not presented in enough detail. Findings on night sleep not reported at all. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 3 parallel groups. | DESCRIPTION: 182 (125 included in present review) day patients undergoing surgery for varicose veins or inguinal hernia repair under general anaesthesia, from two hospitals in Sweden. | MUSIC GROUP: Soft classical music via headphones which allow conversation to take place. Played on auto‐reverse from the time of arrival at the PACU until the patient wanted to stop listening. | PAIN INTENSITY: taken every 30 minutes until patient reports a pain level of ≤ 3 on a VAS (0 to 10). Data presented as the mean of median scores in 120 minutes. | There was a significant difference in the length of time patients listened to their allocated tape. See Cepeda 2006 for details on music for pain relief. | |
| RCT; 3 parallel groups. | DESCRIPTION: 75 day patients undergoing surgery of open Lichtenstein inguinal hernia repair, in Sweden. | INTRA‐OPERATIVE MUSIC GROUP: exposed to new‐age synthesizer music during operation, and a sham (silent) CD for 1 hour post‐operatively in the PACU. | PAIN: assessed by NRS (0 to 10), 30 minutes before anaesthesia and 1 hour after admission to PACU. | Data extracted for post‐operative music group versus control group. See Cepeda 2006 for details on music for pain relief (study not yet included in Cepeda 2006). | |
| RCT; 2 parallel groups. | DESCRIPTION: 36 in‐patients with bronchial asthma, in Poland. | MUSIC GROUP: Underwent 10 day rehabilitation programme, including 45 minutes each day of: exercise of breath control, correction of respiratory pattern, training of diaphragm and additional respiratory muscles, plus an additional 15 minutes of music listening (C.M. Weber "Adagio"; J.S. Bach "Air on a G‐string"; V.A. Mozart "Andante z Divertimenta D‐dur"). | ANXIETY: Measured via the STAI at baseline and after the 10th session. | Data not provided in sufficient detail for extraction. | |
| RCT; 3 parallel groups. | DESCRIPTION: 104 day‐patients scheduled for elective surgical procedures (gynaecology, general surgery, urology) in a Day Surgery Unit in England. | MUSIC + BINAURAL BEAT: taken to a quiet environment pre‐operatively and asked to listen to music with their eyes closed; 30 minute soundtrack (Holosync Solution, 'Awakening Prologue', Centerpointe Research Institute, Beaverton, OR) which produces binaural beats (through two similar pure tones being presented separately to each other). | STATE ANXIETY: measured via the STAI before and after intervention period. Results converted to percentages instead of presenting STAI score. | It is unclear if > 90% of participants are 18 years old or over. | |
| RCT; 2 parallel groups. | DESCRIPTION: 50 out‐patients undergoing flexible sigmoidoscopy in Maryland, USA. | MUSIC GROUP: Choice of 20 tapes (classical/country‐western/popular/rhythm and blues/gospel) played via headphones throughout the procedure. | ANXIETY: Measured via STAI. An analysis of variance identified a significant difference between groups (P < 0.002): | No SDs reported. Insufficient data for extraction. | |
| RCT; 2 parallel groups. | DESCRIPTION: 110 in‐patients giving birth for the first time (to a singleton fetus), in Southern Thailand. | MUSIC: patient choice of western music without lyrics, including synthesizer, harp, piano, orchestra, and jazz (60‐80 bpm). Patients listened to their choice via headphones for the first 3 hours of the active phase of labour, starting when cervical dilation was 3 or 4 cm with uterine contraction of 30 to 60 seconds. Women could stop listening to music for 10 minutes if they wished. | PAIN SENSATION: measured using 100 mm VAS at four time points‐ once at the start of the study before the treatment period, and then every hour for 3 hours. | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 119 patients undergoing diagnostic endoscopy in Germany. | MUSIC GROUP: Played Radio Arabella (105.2 MHz) throughout the procedure. This is a well‐known regional broadcasting company in Munich, and plays various trends of pop, rock, soul, and 'Deutsche Schlager' music, specialising in 'oldies'. | SEDATION & ANALGESIA: Number of patients requiring midazolam and pethidine, and amount administered. It is unclear from the report if sedation is an outcome measure or baseline characteristic. | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 2 parallel groups. | DESCRIPTION: 30 patients undergoing cerebral angiography in Hannover, Germany. | MUSIC GROUP: Choice of nine tapes (international pop, German pop, oldies, meditation, rock, techno, instrumental, classic, and traditional). Played via stereo speakers throughout the angiogram; patients could adjust the volume. | ANXIETY: Measured via STAI the evening before the angiography and during the angiography. | Data not sufficient for extraction. | |
| RCT; 2 parallel groups. | DESCRIPTION: 86 post‐operative in‐patients who have undergone cardiac surgery, in cardiovascular units, Midwest USA. | MUSIC GROUP: Participants advised to clear minds and allow muscles to relax. They were given a choice of easy listening, classical, or jazz music, played via headphones, for 20 minutes whilst participant remained in bed. Music qualities were: no dramatic changes, consonance, instrumental, 60 to 70 bpm. The environment was made conducive to rest. | HR: Recorded on bedside monitor. No differences between groups. | Paper only reports data for the first 3 sessions due to missing data. SDs obtained from authors (unpublished). Data extracted for morning session of the first post‐operative day. See Cepeda 2006 for details on music for pain relief (study not yet included in Cepeda 2006). | |
| RCT; 2 parallel groups. | DESCRIPTION: 42 male out‐patients undergoing radiation therapy in a Veterans Affairs Hospital in Southeastern USA. | MUSIC GROUP: Patient choice of rock and roll, big band, country and western, classical, easy listening, Spanish, religious. Four to six tapes were available in each category. Each patient could choose one category for the duration of the study. Patients listened via headphones to the music before and during the radiation simulation appointment, and during daily radiation treatments for the duration of course of therapy. | STATE ANXIETY: measured using the STAI at 5 time points: (1) time of evaluation; (2) post simulation appointment; (3) end of first week of treatment; (4) end of third week of treatment; and (5) end of the fifth week or end of radiation therapy. | Data extracted for end time point. | |
| RCT; 2 parallel groups. | DESCRIPTION: 32 out‐patients undergoing colonoscopy, USA. | MUSIC GROUP: Patient choice of classical, jazz, pop rock, easy listening, played via headphones throughout pre‐sedation and procedure. | SEDATION REQUIREMENTS (mg). Amount of Versed and Demerol administered was recorded (not a primary outcome). | State anxiety data extracted from graph in article, bars taken as standard errors and converted into SD. | |
| RCT; 2 parallel groups. | DESCRIPTION: 17 in‐patients with HIV/AIDS staying in the HIV/AIDS ward, London, UK. | MUSIC GROUP: Attended a live music concert played in the public area on the hospital ground floor. Unclear what type of music was played. | LEVELS OF CD4 AND CD8 LYMPHOCYTES (cells/mm3): measured within one hour before and after the study period. | Data insufficient detail. | |
| RCT; 2 parallel groups. | DESCRIPTION: 120 in‐patients undergoing surgery with epidural anaesthesia in Taiwan. | MUSIC GROUP: Walkman music via headphones. Choice of 5 types (Manderin pop song, local Taiwanese folksong, Western, Classical, Buddhist hymn). Music played throughout operation. | Subjective feelings of anxiety and sedation: Method of assessment unclear (data EXCLUDED from review). | ||
| RCT; 3 parallel groups. | DESCRIPTION: 61 in‐patients who have had an elective abdominal hysterectomy using general anaesthesia in Arizona, USA. | MUSIC GROUP: patient choice brought from home or from selection provided (classical; jazz; light rock; country; rock and roll; easy listening; gospel), played via headphones. | PAIN INTENSITY: Two measures used: a 9‐inch VAS, and a 10‐point NRS. Measures were taken every 15 minutes for the duration of PACU stay (unspecified). Results of VAS are reported as a mean value for each group; results of NRS are reported as mean rating at 1 hour and at discharge from PACU. | See Cepeda 2006 for details on music for pain relief. | |
| RCT; 3 parallel groups. | DESCRIPTION: 30 pre‐operative in‐patients scheduled to undergo cardiac catheterization (CC) in Hong Kong, China. | SENSORY INFORMATION: excluded from review. | STATE ANXIETY: measured via STAI before and after intervention period; | Outcomes extracted for T1 ‐ after the study intervention but before cardiac catheterization, as the study has been categorised in the "pre‐procedure" subgroup. | |
| RCT; 2 parallel groups. | DESCRIPTION: 200 patients undergoing flexible bronchoscopy in Slovenia. | MUSIC GROUP: Easy listening and relaxation ambient music selected by the investigator, started immediately after the beginning of the procedure and stopped when the procedure is over. | PHYSIOLOGICAL: HR and BP recorded before and after the procedure, method of data collection not described. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 59 in‐patients undergoing elective laparoscopic cholecystectomy in an operating theatre in Japan. | NATURE SOUNDS: Patients were played their choice of nature sounds via headphones throughout operation, after induction of anaesthesia until the last suture of surgery. | PHYSIOLOGICAL MEASURES: BP and HR was recorded through 'non‐invasive' methods using the Philips patient monitoring system for anaesthesia [unpublished]. Outcomes were measured pre‐anaesthesia, at the start of surgery, at gallbladder removal, at the end of surgery, at extubation, at the end of anaesthesia, and in the PACU. | Data insufficient detail for extraction. Authors unable to provide means and SDs for physiological measures. | |
| RCT; 2 parallel groups. | DESCRIPTION: 60 in‐patients undergoing coronary artery bypass graft or vascular surgery, in the OR and ICU, Florida, USA. | MUSIC GROUP: Choice of 6 discs played during surgery and post‐operatively. Family encouraged to bring in additional music choices post‐operatively. Music selection from Prescriptive Music Inc., ('Clarity'‐ melodies from classical motion pictures; 'Timeless'‐ heartfelt originals, 'Towards'‐ piano improvisation, 'Interlude'‐ piano music by Mozart, 'Universe'‐ synthesized compositions, 'Essence'‐ cello and piano). | ANXIETY: Measured via the STAI on the 3rd post‐operative day. | ||
| RCT; 3 parallel groups. | DESCRIPTION: 61 post‐operative in‐patients following open heart surgery undergoing 30 minutes chair rest, in the surgical intensive care unit of a rural Midwestern hospital, USA. | MUSIC GROUP: Instructed to listen and follow music and allow it to distract and relax. Played through headphones. 6 choices: synthesizer, harp, piano, orchestra ,slow jazz, flute. 30 second excerpts provided for choice. Phone unplugged, blinds closed, lights dimmed and door closed. Do not disturb sign placed on door. | ANXIETY: Anxiety about chair rest measured via 100 mm VAS. | Change scores extracted for anxiety. | |
| RCT; 2 parallel groups. | DESCRIPTION: 93 pre‐surgical out‐patients, to undergo elective surgery (ear‐nose‐throat; orthopedics; plastics; or other general minor surgery), in USA. | MUSIC GROUP: Music played for 30 minutes via headphones prior to surgery in a hospital isolation room. Participants brought their own choice of music from home. | ANXIETY: Measured via STAI before and after experiment. | Outcomes are expressed as a percentage of the baseline score (mean and SD). | |
| RCT; 2 parallel groups. | DESCRIPTION: 40 in‐patients in acute care unit for acute myocardial infarction in Midwest USA. | MUSIC GROUP: listened to 4 adagios selected by the investigator for 25 minutes. Primarily string composition, low‐pitched, simple and direct musical rhythm, tempo approximately 60 bpm. Played via headphones. | ANXIETY: state anxiety measured via STAI. | ||
| RCT; 3 parallel groups. | DESCRIPTION: 45 in‐patients who had an acute MI, in private rooms in an ICU in Midwest USA. | MUSIC GROUP: Experimenter selected classical music, played via headphones. Asked to assume a comfortable position in bed, lights lowered, telephones unplugged, curtains drawn, doors closed, advised to clear mind and let muscles relax. 20 minutes. | PHYSIOLOGICAL: HR (chart extraction), RR (auscultation with a stethoscope for 30 seconds), BP (non‐invasive automatic oscillometric BP cuff), measured pre, immediately post and at 1 and 2 hours post the intervention period. Heart rate variability determined using power spectral analysis and fast Fourier transform from 3 hours of continuous electrocardiographic data (30 minutes pre to 2 hours post intervention period). | Data presented as mean and SE. SEs converted into SDs for purposes of review. | |
| RCT; 2 parallel groups. | DESCRIPTION: 50 day patients scheduled for elective same‐day surgery of gynaecological procedures (e.g. exploratory laparoscopies, laparoscopic tubal ligation, ovarian cysts excision, and intrauterine device removal), in New Jersey, USA. | MUSIC GROUP: Choice of classical, country, jazz, popular, or show music, played via headphones. | ANXIETY: State and trait anxiety via STAI on entry and exit of holding area. Results reported as Mean +/‐ SEM. | Anxiety data presented as mean and SEs. SEs converted into SDs for purposes of review. | |
| RCT; cross‐over study (2 allocation groups). | DESCRIPTION: 20 ventilator‐dependent in‐patients in an ICU in Hong Kong, China. | MUSIC GROUP: Choice of 7 cassettes (Chinese folk song; Chinese instrumental music; Chinese music with western instruments; Buddhist music; Western classic; Western movie music; piano music). Participants instructed to close eyes and focus on the flow of the music. | STATE ANXIETY: short‐form STAI, patients held up appropriate number of fingers in response to questions. Taken before and after each intervention. | This cross‐over study has not reported patient‐specific differences between the two intervention measurements. Therefore, the data has been extracted as if it is independent groups (as there was not enough data to calculate a correlation coefficient). | |
| RCT; 2 parallel groups. | DESCRIPTION: 39 patients undergoing eye operations with anaesthetics and analgesics in China. | MUSIC GROUP: Given a choice of 3 musical styles (pop, light, classical) with each type including 4 songs. Patients were asked to listen to the music 2 to 3 times the night before the operation and relax (not clear if patients were in‐patients). Patients then listened to the music throughout their operation. | ANXIETY: STAI presented as change scores. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 66 pre‐operative surgical patients in the OR holding area in Hong Kong, China. | MUSIC GROUP: Slow music played via headphones for 20 minutes pre‐operatively. Participants had a choice of 3 tapes‐ Chinese instrumental, Western instrumental, or Western and Chinese slow songs. | ANXIETY: State anxiety measured via C‐STAI pre‐ and post‐intervention period. | ||
| RCT; 2 parallel groups. | DESCRIPTION: 110 in‐patients undergoing abdominal hysterectomy in China. | MUSIC GROUP: Patient choice of music they felt to be calming and comforting, played via headphones intraoperatively, from 6 minutes before skin incision and 1 minute before the loading dose of anaesthesia to wound closure | PHYSIOLOGICAL: BP, HR, Bispectral Index recorded at 10‐minute intervals during surgery. Results reported as baseline and mean score during surgery (during target sedation period). | See Cepeda 2006 for details on music for pain relief (study not yet included in Cepeda 2006). | |
| RCT; 3 parallel groups. | DESCRIPTION: 75 in‐patients with suspected myocardial infarction in a coronary care unit in Midwest USA. | MUSIC GROUP: 30 minutes of self‐selected music choice of instrumental tapes: Halpern relaxation tape, classical music, country and western. Played via headphones. | ANXIETY: state anxiety measured via STAI | ||
| RCT; 2 parallel groups. | DESCRIPTION: 40 in‐patients with chronic cancer pain in acute care, Midwestern USA. | MUSIC GROUP: Choice of 10 relaxing instrumental tapes. Participants without a preference were given a Halpern antifrantic tape. The researcher suggested to the participant that the music would help them relax and reduce their pain. Participants listened to the music for 30 minutes via headphones. Participants lay on their beds and the lights were dimmed. | PAIN: Measured before and after 30 minute test period via McGill Pain Questionnaire (Pain Rating Index, Number of Words Chosen, and Present Pain Index). Additionally, pain intensity was measured via a 10 mm [sic] VAS. | See Cepeda 2006 for details on music for pain relief. | |
| BP: blood pressure; HR: heart rate; PACU: post‐operative care unit; POMS: Profile of Mood States; RR: respiration rate; SD: standard deviation; STAI: State Trait Anxiety Inventory; VAS: visual analogue scale; VRS: verbal rating scale; | |||||
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| CCT; 2 parallel groups; | DESCRIPTION: 42 pre‐operative patients scheduled for ambulatory surgery in a Midwestern city hospital, USA. | MUSIC GROUP: Choice of 20 tapes (classical, environmental, new age, country‐western, general easy listening), played via headphones whilst resting in recliner chairs. Intervention lasted 15‐30 minutes depending on how long patient had left before surgery. | PHYSIOLOGICAL MEASURES: measured using "standard noninvasive technology" before and after the intervention period. | ||
| CCT; 2 parallel groups; | DESCRIPTION: 301 outpatients and inpatients undergoing colonoscopy or esophagogastroduodenoscopy (EGD) in Switzerland. | MUSIC GROUP: Patients could choose between "light" and "classical" music which was played in the background during the examination. | GENERAL EVALUATION: Not extracted for review. | Paper reports outcomes as medians. Authors provided means and standard deviations on request for purposes of review. | |
| CCT; 2 parallel groups; | DESCRIPTION: 44 pre‐operative in‐patients in the holding area before total joint replacement, in Florida, USA. | MUSIC GROUP: Listened to music via headphones for a minimum of 30 minutes in the pre‐operative holding area. Music consisted of music‐enhanced nature sounds (sea, thunder, rainstorms, wind, and waterfalls). | The paper provides descriptive statistics only. | ||
| CCT; 2 parallel groups; | DESCRIPTION: 49 out‐patients undergoing bronchoscopy, USA. | MUSIC GROUP: Played new wave music 'Reflections of Passion' by Yanni, via headphones for the duration of bronchoscopy procedure. | PHYSIOLOGICAL: Methods of obtaining data unclear. There were no significant differences between groups on any of the physiological parameters. | ||
| CCT; 2 parallel groups; | DESCRIPTION: 158 inpatients with cardiovascular and respiratory illnesses participating in a therapeutic exercise programme, in the Republic of Uzbekistan. | MUSIC GROUP: Were played classical and pop music during the two week exercise programme. | PULSE: Assessed increase in pulse rate during exercise in the first and second weeks of the exercise programme. Method of data collection unclear. In week one, the music group worked harder, as demonstrated by increased pulse. | Authors conclude that cardiovascular patients should use music from the 1st week of exercise, and respiratory patients should use music from the 2nd week of exercise for better adaptation. | |
| CCT; 2 parallel groups; | DESCRIPTION: 24 pre‐operative day surgery patients waiting for surgery under general anaesthesia, in Texas, USA. | MUSIC GROUP: Choice of "easy listening" music selections recorded by the medical staff played via headphones for 20 minutes pre‐operatively. | ANXIETY: Measured via the STAI and a VAS. There was no significant difference between groups on either measure. STAI: | ||
| CCT; 2 parallel groups; | DESCRIPTION: 65 in‐patients undergoing rehabilitation (physiotherapy, balneotherapy, re‐education, and physical exercise) for lower back pain, France. | MUSIC GROUP: Music provided for first 4 days of 12 day hospitalisation. Music played for 20 minutes in the afternoon after physical therapy via headphones in a silent room. Patients were given a choice of music, with each choice arranged to have progressive relaxation with re‐awakening period at the end. | PAIN: Measured via VAS at baseline, day 5, and day 12, plus immediately pre and post therapy sessions. Day 5 outcomes did not significantly differ between groups: | ||
| CCT; 2 parallel groups; | DESCRIPTION: 101 in‐patients and out‐patients waiting for cardiac catheterizations in a cardiac telemetry unit, USA. | MUSIC GROUP: Listened to 20 minutes of 'Trance‐Zendance' by Halpern. Played via headphones prior to cardiac catheterization. | ANXIETY: Measured via STAI pre‐ and post‐ 20‐minute intervention period. There was a significant difference in anxiety scores in favour of the music group: | ||
| CCT; 2 parallel groups; | DESCRIPTION: 20 pre‐operative patients scheduled for breast biopsy in a holding area in Kentucky, USA. | MUSIC GROUP: Choice from selection of "new age" music listened to via headphones for 20 minutes pre‐operatively. No other music types were offered and patients' preferences were not solicited. | ANXIETY: Measured via the STAI pre‐ and post‐ 20‐minute study period. | ||
| CCT; 2 parallel groups; | DESCRIPTION: 34 in‐patients undergoing elective lumbar microdiscectomy procedures consented to participate (although only 10 analysed), Kentucky, USA. | MUSIC GROUP: Participants had a choice of 3 cassettes (country, instrumental, classical), with music 60‐80 bpm, played via headphones at a volume pre‐selected by participants. Music began 30 minutes before the end of the surgery and continued without interruption for one hour in PACU. | No data reported in paper. The paper reports there were no differences between the two groups on any of the outcome measures. | ||
| CCT; 2 parallel groups; | DESCRIPTION: 33 out‐patients awaiting arthroscopic procedures in Philadelphia, USA. | MUSIC GROUP: 20 minutes of classical music (tape 3 of Music Rx, developed by Bonny), played via audiocassette played placed 1 foot away during the waiting period. | BP: Measured via a Dinamap monitor before and after the 20 minute study period. There were no significant differences between the groups. Standard deviations are not provided. | ||
| CCT; 2 parallel groups; | DESCRIPTION: 113 pre‐operative day patients undergoing cytoscopy, cauterisation, or endoscopy, in China (Hong Kong).NUMBERS: Music group = 58, Control group = 55.AGE, mean (SD): Music group = 50.0 (15.5), Control group = 51.9 (14.4) years old.GENDER (male/female): Music group = 31/27, Control group = 27/28.ETHNICITY: not described.INCLUSION CRITERIA: 18 years or older; undertaking noninvasive day procedures with regional or local anaesthetic.EXCLUSION CRITERIA: Cognitive disability; hearing impairment; received preoperative sedatives; received a colon preparation; pre‐existing co‐morbid illness; did not have sufficient time to participate. | MUSIC GROUP: Choice of eastern and western style easy listening music and Chinese pop music (10 CDs and 10 mini‐discs) played via headphones in reclining chairs for 20‐40 minutes pre‐operatively.CONTROL GROUP: Undertook usual pre‐procedural relaxing activities (e.g. reading, watching TV) in the waiting room. | STATE ANXIETY: Measured via STAI pre‐ and post‐intervention period. Only the music group had a significant drop in anxiety. Post‐intervention scores:Music group = 42.5 (5.7) Control group = 46.4 (6.5) PHYSIOLOGICAL OUTCOMES: Measured via "standard non‐invasive instruments" pre‐ and post‐intervention period. There were no significant differences between groups.BP scores: Music group: Systolic BP = 124.2 (21.1), Diastolic BP = 70.0 (10.8) Control group: Systolic BP = 129.4 (25.6), Diastolic BP = 72.0 (11.1) PULSE scores: Music group = 71.1 (10.4) Control group = 70.1 (8.6) RR:Music group = 16.6 (1.0) Control group = 16.7 (1.0) | ||
| CCT; 2 groups (non‐parallel), intervention group was cross‐over (2 types of music played in same order for all participants), control group may have been recruited post‐hoc (UNCLEAR). | DESCRIPTION: 45 patients with disease of the lungs or chest being treated at the Department of Chest Surgery in Zakopane, Poland. | MUSIC GROUP: Cross‐over study. Music was played via headphones. Patients were played Debussy's 3rd part of Bergamasque Suite (Clair de Lune) on the piano (soothing music), and then Bartok's Wonderful Mandarin (nerve racking/exciting music). First, there was a 3‐minute rest in the recumbent position then, before, in between, and after each musical piece was a 3‐minute pause. | Parameters were measured at 5 time points in the music group: 1) Following 3‐min rest; 2) Following soothing music; 3) Following 3‐min pause; 4) Following exciting music; 5) Following 3‐min pause. Parameters were measured at 4‐time points in the control group (to follow same time scale as the music group)‐ a final 5th measurement was not taken for the control group. | Data not sufficient for extraction. No SDs reported. Paper reports significance as P < 0.1. | |
| CCT; 2 groups (non‐parallel), intervention group was cross‐over (2 types of music played in same order for all participants), control group may have been recruited post‐hoc (UNCLEAR). | DESCRIPTION: 45 patients with disease of the lungs or chest being treated at the Department of Chest Surgery in Zakopane, Poland. | MUSIC GROUP: Cross‐over study. Music was played via headphones. Patients were played Debussy's 3rd part of Bergamasque Suite (Clair de Lune) on the piano (soothing music), and then Bartok's finale of the Wonderful Mandarin (nerve racking/exciting music). First there was a 3‐minute rest in the recumbent position, then a 3‐minute pause before the first musical piece. After the first piece of music, there was a 4‐minute pause before the second musical piece was played. | Parameters of airway resistance were measured at three time points in both groups: 1) Following 3‐min rest; 2) Following relaxing music; 3) Following exciting music. | Data not sufficient for extraction. | |
| CCT; 2 parallel groups; | DESCRIPTION: 80 patients undergoing minor surgery in a day‐surgery ward in China. | MUSIC GROUP: Choice of 3 types of music with slow rhythms: classical music (concertos and sonatas), contemporary popular music (e.g. "The heart will go on"), Chinese popular music (e.g. "Night plane"). Music had 45 minutes running time and played via headphones for duration of surgery. | ANXIETY: measured via C‐STAI post‐surgery. Patients were asked to fill out the questionnaire by thinking retrospectively over procedure. | ||
| CCT; 2 parallel groups; | DESCRIPTION: 17 day‐patients undergoing arthroscopic surgery, USA. | MUSIC GROUP: Choice of 4 musical tapes (classical tapes from 'Music Rx', Bonny; popular tapes from 'Music Rx'; New Age tape by Steven Halpern called "Dawn"; easy listening selections assembled by investigator). Music played via headset and auto‐reverse cassette player. From administration of pre‐operative medication to PACU. Participants were told to restart the music if they desired after their return to the ambulatory surgery unit. | ANXIETY: state anxiety measured via STAI approximately 2 hours post‐operatively. Non‐significant. | SDs are estimated from the t‐value. | |
| CCT; 2 parallel groups; | DESCRIPTION: 63 patients undergoing dialysis, Florida, USA. | MUSIC GROUP: Choice of classical, pop, rock, jazz, country/western, gospel, easy listening, swing, and bluegrass played via headphones during dialysis treatment. Music was played for 1 hour beginning 30 minutes after the onset of dialysis treatment, there then was an hour of no music, followed by another hour of music. | BP: Measured after each hour of dialysis treatment. BP was recorded daily from each patient's chart for a 2‐week baseline period and a 3‐week treatment period. The two groups did not significantly differ on systolic and diastolic BP readings for onset through final readings during the treatment period. | Data not sufficient for extraction. | |
| CCT; 2 parallel groups; | DESCRIPTION: 88 pregnant women attending a high‐risk antenatal clinic, London, UK. | MUSIC GROUP: One or two musicians playing in one corner of the waiting room with the chairs arranged in a semi‐circle around them. Harp, clarinet, or guitar were preferred, violin and cello not welcomed. Does not explain in depth what was played. | BP: Obtained by the clinician (method unclear) at the beginning of the consultation (after the waiting room experience). There was no significant difference between the groups. | This is one of a series of studies. Two more studies were conducted in the antenatal clinic that do not need inclusion criteria for the review due to the study design (before and after). | |
| CCT; 2 parallel groups; | DESCRIPTION: 9 in‐patients waiting for elective surgery in a theatre holding area, China. | MUSIC GROUP: Choice of: slow rhythmical songs; Chinese slow rhythmical music; Western slow rhythmical music. Played via headphones for 20 minutes pre‐operatively. | BP: Measured via a calibrated Dinamap BP monitor before and after 20 minute study period. No significant differences between groups. | ||
| CCT; 3 parallel groups; | DESCRIPTION: 76 patients presenting to an emergency department with minor musculoskeletal trauma, in Midwest USA. | MUSIC GROUP: Provided with a Walkman tape player and choice of music (classical, country, rock, pop, and jazz), or were allowed to listen to the radio if they preferred. Patients also received standard care. | PAIN: Pain intensity measured at 0, 30, and 60 minutes via 10‐point NRS. There was no statistical differences between groups at any time interval. All groups showed significant improvement from baseline. | No SDs provided. | |
| CCT; 2 parallel groups; | DESCRIPTION: 57 post‐operative in‐patients who have undergone elective nasal surgery. | MUSIC GROUP: Choice of Chinese and Western various music types and patients encouraged to bring music of their own choice. Listened for 30 minutes on four occasions: post‐operatively (T1), again 4 hours later (T2), on the first post‐operative day at 8am (T3) and again at noon (T4). | PAIN: Measured via VRS at baseline and after each intervention session. Music group gave significantly lower pain ratings at all four time points. T4 data: | ||
| CCT; 2 parallel groups; | DESCRIPTION: 60 post‐operative in‐patients after coronary artery bypass graft in a progressive care area, USA. | SOUNDS GROUP: Marsona Sound Conditioner (providing white noise in the form of rain, ocean waves, or a waterfall). 56/60 patients chose ocean sounds. Sounds were played at the bedside throughout the night (switched on between 20.30 and 21.00 hours). Sounds played for 3 nights. | SLEEP: Pre‐test evaluated for patients on their usual sleep at home, data collected on admission, prior to surgery. Assessed on the fourth day post‐transfer for the quality of sleep the previous night. Measured via the Richards‐Campbell Sleep Questionnaire, which includes VAS for sleep depth, latency to sleep onset, awakening, return to sleep, quality of sleep, and a total sleep score. No SDs reported. | ||
| CCT; 2 parallel groups; | DESCRIPTION: 50 patients undergoing surgery with local anaesthetic, Poland. | MUSIC GROUP: Played music before during and after surgery. Patients chose music from a selection of instrumental tapes (music was soft, bright, avoiding high‐pitched sounds, classical). | ANXIETY: Measured via STAI the day before and directly after surgery in the recovery room. | Data not in sufficient detail for extraction. | |
| CCT; 2 parallel groups; | DESCRIPTION: 57 in‐patients undergoing surgery with local anaesthetic, Japan. | MUSIC GROUP: Were played a set piece (by Elgar) from when they entered the theatre to the end of surgery. | ANXIETY: Measured via STAI before and after surgery. Participants were requested to think back to how they felt during surgery in the post‐treatment questionnaire. | ||
| CCT; 3 parallel groups; | DESCRIPTION: 30 pre‐operative patients waiting for transurethral resection of the prostate, in a theatre holding area, Hong Kong, China. | MUSIC GROUP: 20 minutes of slow rhythm soft music played pre‐operatively via headphones. Choice of three tapes that had been judged by a panel of 3 musicians (slow rhythm songs, Chinese slow rhythm music, Western slow rhythm music). No nurse presence. | PHYSIOLOGICAL: BP and HR recorded on an automated BP monitor (Dinamap 1846‐SX) before and after 20 minute study period. Paper reports no significant differences between groups (Kruskal‐Wallis test) | ||
| BP: blood pressure; HR: heart rate; RR: respiration rate; STAI: State Trait Anxiety Inventory; VAS: visual analogue scale | |||||
| Study ID | Reason for exclusion |
| No data presented‐ summary paper | |
| Validity of outcomes | |
| Setting | |
| Validity of outcomes | |
| Outcomes not reported for relevant groups; data collection methods unclear. | |
| Study design | |
| Intervention provided outside of hospital | |
| Study design | |
| Setting and population | |
| Intervention | |
| Intervention | |
| Intervention‐ relaxation technique | |
| Intervention | |
| Questionable validity of outcome, relevant data not presented | |
| Setting | |
| Setting | |
| Qualitative report | |
| Outcome measure | |
| Outcomes | |
| Setting | |
| Intervention | |
| Setting | |
| Intervention not well defined | |
| Intervention provided during education programme | |
| Study design | |
| Music therapist confound; group differences in timing of data collection | |
| Intervention included coaching of participants | |
| Setting not hospital | |
| Study design | |
| Study design | |
| Intervention interactive music therapy | |
| Interventions not suitable for inclusion | |
| Invention group provided 20mins coaching. | |
| Intervention group provided reinforcement and training‐ bias | |
| Intervention group provided coaching on relaxing | |
| Secondary analysis of previous study | |
| Secondary analysis of previous study | |
| Secondary analysis of previous study | |
| Setting; qualitative | |
| Qualitative | |
| Relaxation (psychological) technique use with intervention | |
| Outcomes not health‐related | |
| Intervention‐ taught techniques | |
| Outcomes | |
| Case study | |
| Intervention not well defined | |
| Intervention not well defined | |
| Policy confound (restraints use) | |
| Policy confound (restraints use) | |
| Policy confound (restraints use) | |
| Outcome measure | |
| Outcome measure | |
| Study design | |
| Conference abstract‐ not enough detail | |
| Study design | |
| Intervention not well defined | |
| Data unsuitable for cross‐over study | |
| Setting | |
| Setting | |
| Intervention not well defined | |
| Validity of outcomes | |
| Study design‐ selection of participants by matching, different wards assigned to different conditions. | |
| Unable to clarify discrepancies in data with author | |
| Setting | |
| Setting; Duplicate | |
| Intervention began before admission | |
| Unclear methods and data | |
| Study design | |
| Intervention not well described (CCT) | |
| Intervention not well described (CCT) | |
| Outcomes not validated/reliable | |
| Study design | |
| Lack of information | |
| Confounding | |
| Intervention not well described | |
| Intervention not well described | |
| Intervention not well described, no data presented (CCT) | |
| Setting |
| Study ID | Reason for exclusion |
| Study design | |
| Study design | |
| Setting | |
| Inappropriate methods and analysis | |
| Outcomes not patient‐related | |
| Intervention | |
| Patients allocated retrospectively to music or control, depending on whether or not they had noticed or heard any music playing (unpublished information). | |
| Intervention not well described | |
| Data collection tool not validated | |
| Outcomes not validated | |
| Policy change‐ staff asked to remain quiet on intervention days | |
| Therapy as treatment | |
| Setting | |
| Overview‐ insufficient detail | |
| Music not well described; cross‐over trial with and with‐out vibration | |
| Setting | |
| Study design | |
| Study design | |
| Study design | |
| Study design | |
| Intervention not well defined | |
| Outcomes not validated | |
| Study design | |
| Setting | |
| Intervention | |
| Outcomes not validated | |
| Study design | |
| Study design; outcomes | |
| Insufficient information (intervention and data) | |
| Study design | |
| Study design | |
| Test (not clinical) situation, healthy controls | |
| Setting | |
| Intervention not well defined; validity of outcomes unclear | |
| Inappropriate control‐ drugs | |
| Duplicate study (Lee 2005) |
| Outcome | Detailed RCTs (N) | Participants (N) | Heterogeneity (%) | Results | Other RCTs | Participants (N) | Findings |
| Oxygen saturation | N = 1 | Total = 50 | N/A | MD 1.60, 95% CI 0.05 to 3.15, P value = 0.04 | N = 1 | Total = 20 | No significant difference between groups. |
| Airway pressure | None | N/A | N/A | N/A | N = 1 | Total = 20 | No significant differences between groups. |
| Skin conductance | Total = 186 | I2 = 81.7% (one study significant, one study non‐significant) | SMD 0.13, 95% CI −0.55 to 0.80, P value = 0.71 (Analysis 3.5). | None | N/A | N/A | |
| Heart rate variability (RR intervals, low and high frequency bands, total power, low/high frequency ratio) | None | N/A | N/A | N/A | N = 1 | Total = 68 | Heart rate variability data (high frequency power, logarithm of low frequency, and high/low frequency ratio) showed significant positive changes in the music group but not control group. Other variables (RR intervals, low and high frequency bands, low frequency power, and logarithm of high frequency) were non‐significant. |
| Bispectral index (mean, time to reach BIS 60) | N = 2 | Total = 160 | N/A (studies reported different outcomes) | Bispectral index: MD 1.00, 95% CI −1.27 to 3.27, P value = 0.39; | None | N/A | N/A |
| Stress hormones (Cortisol, epinephrine, norepinephrine, adrenocorticotropic hormone‐ ACTH) | N = 3 | Total = 216 | Cortisol I2 = 51.8% (3 studies) | Cortisol: SMD −0.32, 95% CI 0.73 to 0.09, P value = 0.13; | N = 1 | Total = 75 | No differences in cortisol levels at any time between groups. Change scores at 2 hours post‐operatively were significantly greater in the post‐operative music group than control. Blood glucose levels did not differ between groups at any time. |
| Mood | N = 1 | Total = 20 | N/A | MD −8.50, 95% CI −18.55 to 1.55, P value = 0.10 | None | N/A | N/A |
| Abnormal events (hypoxaemia, hypotension, hypertension, bradycardia, tachycardia, respiratory depression, pruritis) | N = 1 | Total = 118 | N/A | Cardio‐respiratory incidents = 0; | N = 1 | Total = 193 | Vomiting (Intra‐operatively/PACU): Music = 0/0%, Control = 2.2/0% Unclear missing data as presented in % values. |
| Headache | N = 1 | Total = 115 | N/A | MD 0.00, 95% CI −0.15 to 0.15, P value = 1.00 | None | N/A | N/A |
| Fatigue | N = 1 | Total = 115 | N/A | MD −0.30, 95% CI −0.78 to 0.18, P value 0.22 | N = 1 | Total = 151 | There were no significant differences between groups. |
| Urinary problems | N = 1 | Total = 115 | N/A | MD −0.10, 95% CI −0.40 to 0.20, P value = 0.51 | None | N/A | N/A |
| Well‐being | N = 1 | Total = 115 | N/A | MD 0.30, 95% CI 0.02 to 0.58, P value = 0.03 in favour of music group. | None | N/A | N/A |
| Nausea | N = 2 | Total = 308 | N/A (different methods of measurement) | MD −0.20, 95% CI −0.50 to 0.10, P value = 0.19 | N = 1 | Total = 151 | There were no significant differences between groups. |
| Satisfaction | N = 1 | Total = 110 | N/A | MD 1.60, 95% CI 1.29 to 1.91, P value < 0.00001 | N = 2 | Total = 344 | There were no significant differences between groups. |
| Length of stay | N = 1 | Total = 61 | N/A | MD 0.00, 95% CI −0.99 to 0.99, P value = 1.00 | N = 2 | Total = 118 | Harikumar 2006 reports that recovery time was significantly longer (difference in medians = 10 minutes) in the control group. Heitz 1992 reports no significant differences between groups. |
| Activities of daily living | N = 1 | Total = 61 | N/A | MD −0.30, 95% CI −2.63 to 2.03, P value 0.80 | None | N/A | N/A |
| Serum interleukins (IL‐6) | N = 1 | Total = 110 | N/A | MD −7.40, 95% CI −22.61 to 7.81, P value = 0.34 | None | N/A | N/A |
| Induction time of sedation (minutes) | N = 1 | Total 110 | N/A | MD −6.00, 95% CI −10.49 to −1.51, P value = 0.009 in favour of music. | None | N/A | N/A |
| CI: confidence interval; MD: mean difference; SMD: standardised mean difference; PACU: post‐operative care unit; RR: respiration rate | |||||||
| Outcome | Detailed RCTs (N) | Participants (N) | Heterogeneity (%) | Results | Other RCTs | Participants (N) | Findings | Comments |
| Skin temperature | N = 4 | Total = 298 | I2 = 0% | SMD 0.15, 95% CI −0.08 to 0.37, P value = 0.21; | None | N/A | N/A | |
| Oxygen saturation | N = 3 | Total = 150 | I2 = 79%; | MD −0.71% | None | N/A | N/A | |
| Requirement for oxygen supplementation | N = 1 | Total = 119 | N/A | OR 0.49, | None | N/A | N/A | |
| Blood flow characteristics | N = 1 | Total = 44 | N/A | Blood flow: MD −2.40 ml/min/100g, 95% CI −7.45 to 2.65, P value = 0.35; | None | N/A | N/A | |
| Bispectral index | N = 1 | Total = 39 | N/A | MD 0.22 BIS value, | None | N/A | N/A | It is unclear if this study meets the review inclusion criteria. Music group may have received intervention prior to coming to hospital. |
| Lung Function: (dyspnoea, tidal volume, minute ventilation, oxygen consumption, airway resistance). | None | N/A | N/A | N/A | N = 1 | Total = 36 | No significant differences. | |
| Stress hormones (Cortisol, Prolactin). | N = 1 | Total = 26 | N/A | Cortisol: MD 7.29, 95% CI −7.37 to 21.95, P value = 0.33; | N = 1 | Total = 30 | Cortisol significantly increased in the control group and remained unchanged in the music group. Catecholamines were non‐significant. | |
| Mood | N = 2 | Total = 170 | I2 = 0% | MD −1.18, | None | N/A | N/A | |
| Anger | N = 1 | Total = 24 | N/A | MD −1.80, | None | N/A | N/A | |
| Depression | N = 1 | Total = 39 | N/A | MD −3.29, | None | N/A | N/A | It is unclear if this study meets the review inclusion criteria. Music group may have received intervention prior to coming to hospital. |
| Fatigue | None | N/A | N/A | N/A | N= 1 | Total = 24 | No significant difference between groups. | |
| Uncertainty | N = 1 | Total = 30 | N/A | MD −3.53, 95% CI −12.15 to 5.09, P value = | None | N/A | N/A | |
| Satisfaction | N = 2 | Total = 174 | I2 = 0% | MD 0.46, 95% CI 0.16 to 0.76, | None | N/A | N/A | |
| Nausea | None | N/A | N/A | N/A | N = 1 | Total = 33 | The paper reports that the music group had significantly less nausea and vomiting than the control group (Mann‐Whitney U test, P < 0.017). | |
| Length of stay | N = 2 | Total = 153 | I2 = 0% | MD −6.00 minutes, 95% CI −10.72 to −1.28, P value = 0.01 in favour of music group (Analysis 5.10). | N = 1 | Total = 60 | Findings were non‐significant. | |
| Intubation time | N = 1 | Total = 60 | N/A | MD −200.20 minutes, 95% CI −391.03 to −9.37, P value = 0.04 in favour of music group. | None | N/A | N/A | |
| CI: confidence interval; MD: mean difference; SMD: standardised mean difference | ||||||||
| Study ID | Reason for exclusion |
| Setting | |
| Commentary | |
| Outcomes | |
| Study design | |
| Study design‐ retrospective matched pairs |
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| CCT; 2 parallel groups. | DESCRIPTION: In‐patients, predominantly with leukaemia, multiple myeloma, and malignant non‐Hodgkins lymphoma, admitted to the haematology‐oncology unit in Bonn, Germany. | AIR FILTRATION: 3 rooms (single or double) were fitted with portable air filtration units (NSA model 7100A/B Environment Air System, National Safety Association Ltd., Memphis, USA). These units have a 95% filtering capacity for particles > 0.3 micrometers. They have two settings (high/low). Flow rate: High = 168, Low = 112 cubic m/hr. | INVASIVE ASPERGILLOSIS: Counts of cases (confirmed and suspected) were achieved via ward liaison, targeted chart reviews, and consultation with medical staff. | Compliance to the air filtration units was low; they were generally run on low due to the noise generated, and sometimes stopped during the night. | |
| RCT; 2 parallel groups. | DESCRIPTION: 45 inpatients undergoing treatment for acute leukaemia or bone marrow aplasia, in Belgium. | ISOLATION: Patients were isolated in a laminar air flow room. The isolation unit consisted of a normal bed enclosed in a plastic tent. The area available for the patient was about 5 sq metres. Air was pumped vertically through high‐efficiency filters positioned on the ceiling. All procedures on the patient were done through plastic gloves on the sides of the tent. All items entering the unit were either gas or steam sterilized, passing through locks irradiated by ultraviolet light. The tent was only opened when absolutely necessary. Any person approaching the isolated patients wore sterile boots, gloves, gown, and mask. Patients remained in isolation until either the bone marrow showed haematologic remission or myeloid proliferation and maturation. The study period ended when the patient was no longer in isolation. | BACTERIAL INFECTIONS: Fungal cultures of stools were performed at least once a week. Data is reported as a count of the number of patients in which suppression of bacterial growth from stool cultures was achieved. Cultures were also obtained from the nose, gingiva, throat, ear, and axilla. Results report the number of days with infection (per 1000 days with severe neutropenia). | ||
| CCT; 2 parallel groups. | DESCRIPTION: 1726 in‐patients residing in the wards of a surgical unit (urological surgery/general surgery), in Scotland, UK. | OPEN WARDS: Two wards (one male, one female) with natural ventilation (and no mechanical ventilation). The male ward had partitioned 4‐6 bedded bays, housing a total of 28 beds, and female ward was completely open plan with 29 beds. | STAPHYLOCOCCUS AUREUS (Staph. aureus): Nasal acquisition of Staph. aureus was monitored via swabs taken on admission, and then every Monday and Thursday thereafter. Data reported as the number of acquisitions, and the rate per 100 patient‐week. The number of tetracycline‐resistant strains of Staph. aureus and delayed acquisitions was also determined. |
| Study ID | Reason for exclusion |
| Study design | |
| Setting | |
| Outcomes | |
| Study design (no control) | |
| Study design (matched pairs) | |
| Policy confound (sterile versus clean attire) | |
| Study design (computer modelling) | |
| Confounding | |
| Patients‐ mix of adults and children, data inseparable | |
| Preliminary report of Rodriguez 1978; policy confound | |
| Sham operations | |
| One patient health‐related outcome reported (surgical site infections), but not enough detail on how it was assessed or if it was noted on an 'ad hoc' basis. | |
| Not patient outcomes | |
| Study design | |
| Not patient outcomes; test scenario | |
| Not patient outcomes; test scenario | |
| Study design | |
| Some relevant subgroups within study however cannot separate data (number of re‐operations in each subgroup). Additionally, there are discrepancies in reporting with Lidwell 1984 with 185 patients switching subgroups. | |
| Same study as Lidwell 1982. Non‐transparant reporting of findings‐ data not usable. Discrepancies between Lidwell 1982 and Lidwell 1984. | |
| Population‐ age | |
| Preliminary report of Lidwell 1982 and Lidwell 1984 | |
| Study design | |
| Retrospective study; population (age) | |
| Policy (decontamination) changes | |
| Mock setting | |
| Policy confounding | |
| Policy confounding | |
| Study design | |
| Study design | |
| Not patient outcomes | |
| Policy confounding |
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| CCT; 2 parallel groups. | DESCRIPTION: 3327 patient‐weeks were analysed. Participants were in‐patients on one of two medical wards in the UK. | OPEN WARD: Two identical wards each containing 6 four‐bed rooms, which opened up on to the corridor. The four‐bed rooms each had three proper walls with the fourth side being open to the corridor except for low dividing walls (3 ft.) on each side of the entrance. | PATIENT NASAL AQUISITION OF STAPHYLOCOCCUS AUREUS: A nasal swab was taken from patients as soon as possible after arriving on the ward (within 3 days). Subsequently a swab was taken from each patient on a set day of the week. | Ratio of nasal carriage rates in single rooms to 4‐bed bays: |
| Study ID | Reason for exclusion |
| Review article | |
| Participants (age) | |
| Cross‐sectional survey and qualitative interviews | |
| Confounding | |
| Retrospective study | |
| Policy changes | |
| Preliminary report of Rodriguez 1978; policy confound | |
| Healthy participants received intervention | |
| Study design | |
| Study design | |
| Study design | |
| Study design, outcomes | |
| Study design | |
| Outcome measure | |
| Participants | |
| Study design | |
| Study design; confounding | |
| Intervention | |
| Study design | |
| Confounding | |
| Study 1: inappropriate control; Study 2: intervention | |
| Study design | |
| Study design | |
| Study design; unclear intervention | |
| Participants; Study design | |
| Setting | |
| Study design | |
| Policy changes | |
| Patients‐ 38% <10 years old | |
| Study design | |
| Intervention (non‐environmental changes) | |
| Policy confounding | |
| Policy confounding | |
| Policy change | |
| Study design | |
| Study design | |
| Policy changes | |
| Study design | |
| Study design | |
| Study design‐ no control | |
| Validity of outcomes | |
| Policy confounding |
| Study ID | Reason for exclusion |
| Participants (students) | |
| Study design |
| Study ID | Methods | Participants | Interventions | Outcomes |
| RCT; 2 x 2 design (2 x flooring, 2 x therapy). | DESCRIPTION: 54 in‐patients in an elderly care rehabilitation ward, in England. | LINOLEUM FLOOR: Patient admitted to one of two four‐bed bays with latex vinyl square tile flooring. | NUMBER OF FALLERS: obtained from the accident report forms. | |
| RCT, cross‐over design, 2 conditions. | DESCRIPTION: 58 elderly hospital patients in England. | CARPETED CORRIDOR: Each patient walked along the corridor towards a staff member standing 15 m ahead. Chalk marks were drawn at an interval of 10 m. | GAIT SPEED and STEP LENGTH: The number of steps taken was counted from the time the patient crossed the first chalk line until the distal line was crossed by the leading forefoot, and elapsed time was measured by a stopwatch. |
| Study ID | Reason for exclusion |
| Paediatric hospital | |
| Study design; outcomes | |
| Qualitative | |
| Study design | |
| Study design | |
| Study design | |
| Qualitative evaluation |
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| RCT; 2 parallel groups. | DESCRIPTION: 132 elderly out‐patients admitted to the emergency department, in Ohio, USA. | CHAIR GROUP: Remained on the gurney until initial physician and nursing evaluations complete. Patients were then assisted to a reclining chair. Patients were told "it is now time to move to the chair" and if they asked why, they were told they had the option of now moving to the chair. Patients were assisted to a position of comfort. Patients can sit on the reclining chairs with the hips and knees flexed. | PAIN: "The study nurse specifically instructed patients to rate pain associated with the gurney, rather than other sources." Measured on a NRS at three time points (time 0 = baseline; time 1 = one hour after randomisation; time 2 = two hours after randomisation). This outcome was dichotomised as favourable outcome (yes/no). A favourable outcome was considered as the patient having no pain at time 0 or time 1, or a decrease in pain from time 0 to time 1. An unfavourable outcome was defined as an increase in pain from time 0 to time 1, or no change in pain score if the patient complained of pain at time 0. | SDs for satisfaction were estimated from the 95% confidence interval (1.4, 2.8), assuming 66 patients per group. |
| Study ID | Reason for exclusion |
| Outcomes and policy changes | |
| Other systematic review in this area (Cullum 2008) | |
| Other systematic review in this area (Cullum 2008) | |
| Study design | |
| Not a research study | |
| Other systematic review in this area (Cullum 2008) | |
| Other systematic review in this area (Cullum 2008) | |
| Study design | |
| Test (not clinical) situation, policy change | |
| Study design | |
| Study design‐ inappropriate data | |
| Other systematic review in this area (Cullum 2008) | |
| Not health‐related outcomes | |
| Outcomes (validity and relevance) | |
| Other systematic review in this area (Cullum 2008) | |
| Policy change | |
| Study design | |
| Intervention | |
| Policy changes with cleaning of blankets |
| Study ID | Reason for exclusion |
| No intervention‐ observational study | |
| No intervention | |
| No intervention | |
| No intervention | |
| Review article | |
| Study design | |
| Review article | |
| Study design; outcomes | |
| No intervention | |
| No intervention in patients | |
| Study design | |
| Discussion article | |
| Study design | |
| Study design | |
| Study design | |
| Study design | |
| No intervention | |
| Review article | |
| No intervention | |
| Policy confound | |
| Policy change | |
| Study design (cross‐sectional, no intervention) | |
| Study design |
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| CCT; 2 parallel groups. | DESCRIPTION: 89 post‐operative in‐patients recovering from elective cervical and lumbar spinal surgeries in single patient rooms, Pittsburgh, USA. | BRIGHT ROOM: On west side of the corridor. Received approximately 46% more natural sunlight than dim rooms (received on average 73,537 lux‐hours per day). | ANALGESIC CONSUMPTION: obtained via chart extraction and standardised to morphine equivalent mg/hr. | Analgesic consumption data extracted for entire LOS mg/hr. | |
| POMS: Profile of Mood States | |||||
| Study ID | Reason for exclusion |
| Retrospective study | |
| Duplicate of Beauchemin 1998b | |
| Retrospective study | |
| Retrospective study | |
| Study design | |
| Study design | |
| Study design | |
| Setting | |
| Intervention | |
| Intervention | |
| Study design | |
| Setting; Study design | |
| Literature review |
| Study ID | Reason for exclusion |
| Setting; study design | |
| Setting; outcomes | |
| Intervention |
| Study ID | Methods | Participants | Interventions | Outcomes |
| CCT; 2 parallel groups. | DESCRIPTION: 97 surgical in‐patients undergoing lower extremity vascular reconstruction in an ORa in Maryland, USA. | WARM OR: Patient underwent operation in one of a group of ORs with the ambient room temperature maintained at 24.5 (0.4) degrees Celsius. | ORAL TEMPERATURE: Measured with an electronic digital thermometer, pre‐operatively (before being transported to the OR), immediately post‐operatively, on arrival at the ICU, and every hour for 24 hours. | |
| aOR = Operating Room. | ||||
| Study ID | Reason for exclusion |
| Study design | |
| Study design | |
| Study design | |
| Intervention‐ temperature of equipment not ambient room temperature | |
| Staff outcomes |
| Study ID | Reason for exclusion |
| Not health‐related outcomes; policy changes | |
| Not health‐related outcomes; query validity | |
| Not a hospital setting | |
| Qualitative | |
| Design; outcomes | |
| Not health‐related outcome | |
| Outcomes | |
| Study design | |
| Confounding staffing differences | |
| Qualitative; participants |
| Study ID | Reason for exclusion |
| Setting | |
| Study design | |
| Setting | |
| Study design‐ inappropriate data | |
| No intervention; comparison with healthy controls | |
| Not health‐related outcomes; conversation confounding | |
| Setting | |
| Participants not patients |
| Study ID | Reason for exclusion |
| Study design | |
| Study design, validity of outcomes | |
| Study design‐ retrospective |
| Study ID | Methods | Participants | Interventions | Outcomes | Notes |
| CCT; 8 parallel groups (2 x rooms, 2 x psychiatrists, 2 x time of data collection). | DESCRIPTION: 115 applicants for out‐patient psychiatric treatment at the Neuropsychiatric Institute, Los Angeles, USA. | 'BEAUTIFUL ROOM': Carpeted in burnt‐yellow carpeting and contained an abstract picture on one wall, a floor‐sized artificial plant, a wooden waste‐basket, and indirect lighting provided by a contemporary desk lamp. The room was neat and well kept.'UGLY ROOM': Carpetless, with beige asphalt floor tiling, overhead fluorescent lighting and was unkempt, with work papers strewn over the furniture and an overflowing grey metal wastebasket and ashtray. Both rooms were windowless offices of identical size (6 x 8 x 8 feet), wall covering, and colour. The furniture was the same (brown metal desks with beige formica tops, green leatherette desk chairs, and green and yellow leatherette side chairs. The offices varied only in decor. | ROOM RATINGS: not included in review. | Data not presented in enough detail for extraction. | |
| CCT; 2 parallel groups. | DESCRIPTION: 56 psychiatric in‐patients in the seclusion area of an acute psychiatric ward in Norway. | REFURBISHED WING: Redecorated and refurbished with the aim of looking, as much as security permitted, like an ordinary Norwegian home: Wainscoting walls, colourful wallpaper and paintings, lowered ceilings, multiple lighting spots, tasteful curtains, wardrobes, chairs, flowers, personal items, Italian ceramic‐tiled bathroom. | SYMPTOMS AND PSYCHOPATHOLOGY: Scored on the Positive And Negative Syndrome Scale (PANSS) for schizophrenia, with the time criterion of the last 24 hours. Scale has scores for total, positive, negative, and general symptoms. Assessed on admittance, day 3, and at discharge from seclusion. | Data extracted for change from baseline (beginning and end time‐points only).PANSS score for 'total' extracted for review. No significant difference was found on any of the subscales. |
| Study ID | Reason for exclusion |
| Descriptive article | |
| Descriptive article | |
| Study design | |
| Descriptive article | |
| No intervention | |
| No intervention | |
| CBA‐ Control sites and 'before' sites not clearly defined | |
| Review article | |
| Setting | |
| Policy confound | |
| Policy confound | |
| Study design | |
| Descriptive article | |
| Study design | |
| Review article | |
| Review article | |
| Study design | |
| Descriptive article | |
| Descriptive case study | |
| Study design | |
| Review article | |
| Study design | |
| Outcomes | |
| Outcomes | |
| Qualitative; setting | |
| Setting | |
| Literature review | |
| Setting | |
| Intervention | |
| Study design (same study as Lawson 2003) | |
| Study design (same study as Lawson 2003) | |
| Study design (before‐and‐after no contemporaneous control) | |
| Study design | |
| Study design (before‐and‐after no contemporaneous control) | |
| Study design | |
| Policy confound | |
| Qualitative study | |
| Study design | |
| Study design | |
| No intervention | |
| Confounding | |
| Literature review | |
| Study design | |
| Study design | |
| Study design | |
| Outcomes; unclear study design | |
| Policy confound | |
| Confounding staffing differences | |
| Literature review | |
| Intervention (care delivery systems) | |
| Study design | |
| Follow‐up study (Smylie 1971, Davidson 1971); study design | |
| Study design | |
| Study design (gender differences) | |
| Study design | |
| Unclear intervention/control | |
| Study design | |
| Qualitative interviews | |
| Overview article | |
| Study design | |
| Descriptive case studies |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Anxiety Show forest plot | 2 | 87 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.24 [‐1.05, 0.56] |
| 2 Pain Show forest plot | 2 | Std. Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 2.1 Endoscopic procedural pain | 1 | Std. Mean Difference (IV, Random, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.2 post‐operative pain | 1 | Std. Mean Difference (IV, Random, 95% CI) | 0.0 [0.0, 0.0] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Anxiety Show forest plot | 3 | Std. Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 2 Pain Show forest plot | 3 | Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Anxiety Show forest plot | 10 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 1.1 Pre‐procedural anxiety | 4 | 361 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.82 [‐1.03, ‐0.60] |
| 1.2 Procedural anxiety | 3 | 183 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.12 [‐0.47, 0.23] |
| 1.3 Post‐operative anxiety | 3 | 220 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.04 [‐0.52, 0.44] |
| 1.4 ICU anxiety | 1 | 64 | Std. Mean Difference (IV, Random, 95% CI) | 0.17 [‐0.32, 0.66] |
| 2 Heart rate Show forest plot | 8 | 543 | Mean Difference (IV, Random, 95% CI) | 0.40 [‐1.02, 1.82] |
| 3 Blood pressure Show forest plot | 8 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 3.1 Systolic | 7 | 533 | Mean Difference (IV, Random, 95% CI) | ‐0.40 [‐2.48, 1.67] |
| 3.2 Diastolic | 6 | 478 | Mean Difference (IV, Random, 95% CI) | ‐0.35 [‐2.08, 1.39] |
| 3.3 Arterial | 1 | 30 | Mean Difference (IV, Random, 95% CI) | 4.0 [‐5.33, 13.33] |
| 4 Respiration rate: with choice of outcomes Show forest plot | 2 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 4.1 Post‐operative period for Ikonomidou 2004 and final scores (both studies) | 2 | 119 | Mean Difference (IV, Random, 95% CI) | ‐1.72 [‐1.00, ‐0.44] |
| 4.2 Post‐operative period for Ikonomidou 2004 and change score for Lee 2005 | 2 | 119 | Mean Difference (IV, Random, 95% CI) | ‐2.48 [‐3.85, ‐1.11] |
| 4.3 Pre‐operative period for Ikonomidou 2004 and final scores (both studies) | 2 | 119 | Mean Difference (IV, Random, 95% CI) | ‐0.92 [‐1.92, 0.09] |
| 4.4 Pre‐operative period for Ikonomidou 2004 and change scores for Lee 2005 | 2 | 119 | Mean Difference (IV, Random, 95% CI) | ‐2.09 [‐4.51, 0.34] |
| 5 Skin conductance Show forest plot | 2 | Std. Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 6 Stress hormones Show forest plot | 3 | Std. Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 6.1 Cortisol | 3 | 216 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.32 [‐0.73, 0.09] |
| 6.2 Epinephrine | 2 | 123 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.02 [‐0.38, 0.33] |
| 6.3 Norepinephrine | 2 | 123 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.08 [‐0.44, 0.27] |
| 6.4 Adrenocorticotropic hormone (ACTH) | 1 | 30 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.44 [‐1.17, 0.28] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Anxiety Show forest plot | 8 | Std. Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 1.1 Procedural anxiety | 1 | Std. Mean Difference (IV, Random, 95% CI) | 0.0 [0.0, 0.0] | |
| 1.2 ICU/CCU anxiety | 7 | Std. Mean Difference (IV, Random, 95% CI) | 0.0 [0.0, 0.0] | |
| 2 Heart rate Show forest plot | 4 | 180 | Mean Difference (IV, Random, 95% CI) | ‐2.76 [‐6.65, 1.13] |
| 3 Blood pressure Show forest plot | 4 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 3.1 Systolic | 3 | 140 | Mean Difference (IV, Random, 95% CI) | ‐1.51 [‐6.65, 3.63] |
| 3.2 Diastolic | 2 | 110 | Mean Difference (IV, Random, 95% CI) | ‐5.29 [‐8.78, ‐1.79] |
| 3.3 Arterial | 1 | 40 | Mean Difference (IV, Random, 95% CI) | ‐4.75 [‐13.98, 4.48] |
| 4 Respiration rate Show forest plot | 3 | 110 | Mean Difference (IV, Random, 95% CI) | ‐2.04 [‐3.43, ‐0.66] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Anxiety Show forest plot | 29 | 1812 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.55 [‐0.74, ‐0.36] |
| 1.1 Pre‐procedural anxiety | 9 | 744 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.37 [‐0.62, ‐0.12] |
| 1.2 Procedural anxiety | 15 | 865 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.69 [‐1.02, ‐0.36] |
| 1.3 ICU/post‐operative anxiety | 5 | 203 | Std. Mean Difference (IV, Random, 95% CI) | ‐0.58 [‐1.01, ‐0.15] |
| 2 Heart rate Show forest plot | 21 | 1653 | Mean Difference (IV, Random, 95% CI) | ‐2.72 [‐4.70, ‐0.74] |
| 3 Blood pressure Show forest plot | 20 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 3.1 Systolic | 18 | 1437 | Mean Difference (IV, Random, 95% CI) | ‐1.76 [‐5.09, 1.56] |
| 3.2 Diastolic | 17 | 1407 | Mean Difference (IV, Random, 95% CI) | ‐0.97 [‐2.58, 0.63] |
| 3.3 Arterial | 2 | 186 | Mean Difference (IV, Random, 95% CI) | ‐9.86 [‐12.06, ‐7.65] |
| 4 Respiration rate Show forest plot | 9 | Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 5 Sedation requirements (anxiolytics) Show forest plot | 3 | Mean Difference (IV, Random, 95% CI) | Totals not selected | |
| 5.1 Midazolam (aka: Versed, Hypnovel, Dormicum, Dormonid), mg | 3 | Mean Difference (IV, Random, 95% CI) | 0.0 [0.0, 0.0] | |
| 6 Peripheral skin temperature Show forest plot | 4 | 298 | Std. Mean Difference (IV, Random, 95% CI) | 0.15 [‐0.08, 0.37] |
| 7 Oxygen saturation Show forest plot | 3 | 150 | Mean Difference (IV, Random, 95% CI) | ‐0.71 [‐1.75, 0.32] |
| 8 Mood state Show forest plot | 2 | 170 | Mean Difference (IV, Random, 95% CI) | ‐1.18 [‐2.17, ‐0.19] |
| 9 Satisfaction Show forest plot | 2 | 174 | Std. Mean Difference (IV, Random, 95% CI) | 0.46 [0.16, 0.76] |
| 10 Length of stay / Examination time Show forest plot | 2 | 153 | Mean Difference (IV, Random, 95% CI) | ‐6.00 [‐10.72, ‐1.28] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Heart rate Show forest plot | 2 | 144 | Mean Difference (IV, Random, 95% CI) | 4.67 [‐0.76, 10.10] |
| 2 Blood pressure Show forest plot | 2 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 2.1 Systolic | 2 | 144 | Mean Difference (IV, Random, 95% CI) | ‐1.80 [‐8.59, 5.00] |
| 2.2 Diastolic | 2 | 144 | Mean Difference (IV, Random, 95% CI) | 0.99 [‐7.60, 9.58] |
