Herramientas para la evaluación de riesgos en la prevención de las úlceras de decúbito
Resumen
Antecedentes
Las herramientas de evaluación de riesgos de úlceras de decúbito son un componente del proceso de evaluación para identificar a los pacientes con riesgo de desarrollar una úlcera de decúbito. Muchas de las guías internacionales de prevención de las úlceras de decúbito recomiendan el uso de una herramienta de evaluación de riesgos, pero no se sabe si el uso de esta logra algún cambio en los resultados de los pacientes. Se realizó una revisión para proporcionar un resumen de las pruebas relevantes para la evaluación de riesgos de úlceras de decúbito en la práctica clínica.
Objetivos
Determinar si el uso de herramientas estructuradas y sistemáticas de evaluación de riesgos de úlceras de decúbito en cualquier ámbito de asistencia sanitaria reduce la incidencia de las úlceras de decúbito.
Métodos de búsqueda
En diciembre de 2013, para esta segunda actualización, se hicieron búsquedas en el registro especializado del Grupo Cochrane de Heridas (Cochrane Wounds Group); Registro Cochrane Central de Ensayos Controlados (Cochrane Central Register of Controlled Trials) (CENTRAL) (The Cochrane Library); Ovid MEDLINE; Ovid EMBASE; y EBSCO CINAHL.
Criterios de selección
Ensayos controlados aleatorios (ECA) que compararan el uso de herramientas estructuradas y sistemáticas de evaluación de riesgos de úlceras de decúbito, con herramientas de evaluación de riesgos de úlceras de decúbito no estructuradas o que utilizaron el criterio clínico sin ayuda alguna, o ECA que compararon el uso de diferentes herramientas estructuradas de evaluación de riesgos de úlceras de decúbito.
Obtención y análisis de los datos
Dos revisores de forma independiente evaluaron los títulos y los resúmenes de los estudios identificados mediante la estrategia de búsqueda para la elegibilidad, se obtuvieron las versiones completas de los estudios potencialmente relevantes y se les aplicaron los criterios de inclusión.
Resultados principales
En esta revisión se incluyeron dos estudios. Un estudio pequeño con asignación al azar grupal no encontró diferencias estadísticas en la incidencia de úlceras de decúbito entre pacientes evaluados por personal de enfermería mediante la herramienta de evaluación de riesgos Braden (n = 74) en comparación con pacientes evaluados por personal de enfermería que recibió formación y que luego utilizó una evaluación de riesgos no estructurada (n = 76) (CR 0,97; IC del 95%: 0,53 a 1,77) y pacientes evaluados por personal de enfermería mediante solo una evaluación de riesgos no estructurada (n = 106) (CR 1,43; IC del 95%: 0,77 a 2,68). El segundo fue un estudio controlado aleatorio simple ciego grande que comparó el efecto de la evaluación de riesgos sobre la incidencia de úlceras de decúbito mediante la herramienta de evaluación de riesgos Waterlow (n = 411), la herramienta de identificación de riesgos Ramstadius (n = 420) y ninguna evaluación de riesgos formal (n = 420). No hubo diferencia estadística en la incidencia de úlceras de decúbito entre los tres grupos (Waterlow 7,5% [n = 31]; Ramstadius 5,4% [n = 22]; criterio clínico 6,8% [n = 28] [CR 1,10; IC del 95%: 0,68 a 1,81; Waterlow versus evaluación de riesgos no formal], [CR 0,79; IC del 95%: 0,46 a 1,35; Ramstadius versus evaluación de riesgos no formal], [CR 1,44; IC del 95%: 0,85 a 2,44; Waterlow versus Ramstadius]).
Conclusiones de los autores
Se identificaron dos estudios que valoraron el efecto de la evaluación de riesgos sobre los resultados de los pacientes; en un estudio no hubo diferencias estadísticamente significativas en la incidencia de úlceras de decúbito en los pacientes evaluados mediante la herramienta de evaluación de riesgos Braden en comparación con los que recibieron una evaluación de riesgos no estructurada. Las limitaciones metodológicas de este estudio impiden que se establezcan conclusiones firmes. Sin embargo, un ECA adicional de alta calidad no identificó diferencias estadísticas en la incidencia de úlceras de decúbito cuando los pacientes se evaluaron mediante la herramienta de evaluación de riesgos Waterlow, la herramienta de evaluación de riesgos Ramstadius o mediante tan solo el criterio clínico. No existen pruebas confiables que indiquen que el uso de herramientas estructuradas y sistemáticas de evaluación de riesgos de úlceras de decúbito reduce la incidencia de dichas úlceras.
PICO
Resumen en términos sencillos
Herramientas de evaluación de riesgos para la prevención de las úlceras de decúbito
Las úlceras de decúbito (también conocidas como úlceras por presión o por compresión) son áreas de lesión localizada en la piel, el tejido subyacente o ambos, generalmente sobre una prominencia ósea, como resultado de la presión o la presión en combinación con la fricción (distorsión tisular resultante de apretar y rozar partes blandas entre estructuras óseas y la piel). Las úlceras de decúbito aparecen principalmente en pacientes con movilidad limitada, daño nervioso o ambos. La evaluación de riesgos de úlcera de decúbito forma parte del proceso utilizado para identificar a los pacientes con riesgo de desarrollar una úlcera de decúbito. Las evaluaciones de riesgos por lo general utilizan listas de verificación y su uso se recomienda en las guías de prevención de úlceras de decúbito. Esta revisión encontró dos estudios elegibles para inclusión. El primer estudio no encontró diferencias en el número de nuevas úlceras de decúbito en pacientes evaluados mediante la herramienta de evaluación de riesgos Braden en comparación con una herramienta de evaluación de riesgos no estructurada. Sin embargo, hubo limitaciones metodológicas en este estudio. El segundo estudio tampoco encontró diferencias en el número de nuevas úlceras de decúbito en pacientes evaluados con la herramienta de evaluación de riesgos Waterlow, la herramienta de evaluación de riesgos Ramstadius o mediante solo el criterio clínico. Este estudio no tuvo limitaciones metodológicas. Por lo tanto, hasta la fecha no hay estudios que indiquen que el uso de herramientas de evaluación de riesgos reduce el número de nuevas úlceras de decúbito.
Authors' conclusions
Background
Description of the condition
Pressure ulcers (also known as pressure injuries, bed sores, pressure sores and decubitus ulcers) are localised injury to the skin, underlying tissue or both, usually over a bony prominence, as a result of pressure or pressure in combination with shear (EPUAP 2009). They occur in people who do not have the ability to reposition themselves in order to relieve pressure on bony prominences (Moore 2011). This ability is often diminished in the very old, the malnourished and those with an acute illness (Moore 2012). Prevalence rates in long‐term care settings fluctuate from 8.8% to 53.2% and incidence rates vary from 7% to 71.6% (Moore 2011; Scott 2006). The most common anatomical sites for pressure ulcers to occur are the sacrum and the heels, and the majority are grade 1 or grade 2 in severity (Gallagher 2008; Moore 2011; Moore 2012). Furthermore, as age increases, so too does pressure ulcer prevalence and incidence (EPUAP 2009). Changing population demographics and the predicted rise in the number of older persons in the future (U.S. Census Bureau 2004) suggest that there will be a corresponding increase in the number of people with pressure ulcers unless effective preventative measures are put in place .
Pressure ulcers impact negatively on quality of life as it is known that individuals with pressure ulcers frequently experience pain, combined with fear, isolation and anxiety regarding wound healing (Fox 2002; Hopkins 2006; Spilsbury 2007). Importantly, it has also been shown that pressure ulcers are associated with an increased risk of death (Allman 1997). One study identified that the risk of dying for elderly patients with a pressure ulcer was three times greater than for those without a pressure ulcer (Berlowitz 1990) although it is probable that pressure ulcers are usually a consequence of poor health rather than a cause of death. Further prospective cohort studies examined the factors predictive of mortality in older individuals admitted to hospital. Both studies (Bo 2003; Alarcon 1999) conducted a detailed geriatric assessment of 659 and 352 patients on admission, and followed these patients for 10 months. Using regression analysis, the authors identified an odds ratio of 3.64 (P < 0.001) (Bo 2003) and 4.19 (P < 0.001) (Alarcon 1999), respectively, of death in acutely ill elderly people with pressure ulcers, indicating that in these individuals having a pressure ulcer increases the risk of dying.
Pressure ulcers are a significant financial burden to health care systems. Bennett and colleagues (Bennett 2004) estimated that the total annual cost for pressure ulcer management in the UK is £1.4 to £2.1 billion, which is equivalent to 4% of the total UK healthcare expenditure. Similar findings have been noted in the Netherlands, where pressure ulcers have been found to be the third most expensive health problem (Haalboom 2000). It has been suggested that the length of hospital stay is two to three times greater for those with a pressure ulcer, than for similar cases without (30.4 days compared to 12.8 days) (Allman 1997).
Globally, the economic impact of pressure ulcers has yet to be established. However, it is known that pressure ulcers are common (EPUAP 2002) and affect patients in both community (Margolis 2002) and hospital settings (Gallagher 2008). Therefore, it is reasonable to suggest that pressure ulcer prevention strategies that can reduce prevalence and incidence rates will have a positive impact on patients and the health service as a whole (Moore 2011).
Description of the intervention
Use of pressure ulcer risk assessment tools or scales is a component of the assessment process used to identify individuals at risk of developing a pressure ulcer (EPUAP 2009). Risk assessments generally use checklists that alert practitioners to the most common risk factors that predispose individuals to pressure ulcer development. These checklists are often developed into risk assessment tools, for example the Norton Scale (Norton 1975), the Waterlow risk assessment scale (Waterlow 1985) and the Braden scale (Braden 1987). It is argued that there is a lack of consensus regarding which variables are the most important indicators of risk (Gould 2002). Therefore, it is not surprising that there are currently almost 40 risk assessment scales in use, most of which are based on the seminal work of Norton 1975, or have been designed in response to a review of the literature (Defloor 2004). It is clear, however, that the risk factors that predispose an individual to developing a pressure ulcer will vary among patients in different clinical settings (Henoch 2003) and it may not be possible to design one risk assessment tool that will meet the needs of all patients in all clinical settings.
How the intervention might work
Use of a risk assessment tool is recommended by many international pressure ulcer prevention guidelines (EPUAP 2009; NICE 2001; Rycroft‐Malone 2000). The ideal risk assessment tool should be both reliable and valid, and sensitive and specific (NPUAP 1998). The tool must accurately identify those individuals who are at risk, as well as those not at risk ‐ and do this consistently (Defloor 2005). To date, there is little empirical evidence available concerning the reliability and validity of existing tools (Anthony 2008; Cullum 1995; Defloor 2004; Defloor 2005; Haalboom 1999; McGough 1999; Pancorbo‐Hidalgo 2006; Schoonhoven 2002). Assessing reliability and validity is a real challenge in clinical practice because risk assessment scales are used to identify those who would develop a pressure ulcer should no interventions be put in place. It is common to use different pressure ulcer prevention strategies once risk has been identified, which will therefore appear to alter the predictive ability of the scale (Defloor 2004; Halfens 2000). Different studies using the same risk assessment tools, but in diverse heath care settings with diverse patient populations and prevention strategies in use, report varying levels of sensitivity and specificity (Gould 2002). It is of relevance to note that the prevention strategies which were in use in these studies are often not stated (Halfens 2000). Lack of clear knowledge of the sensitivity and specificity of risk assessment tools has far‐reaching implications for practice, because clinical decisions ‐ such as the use, or not, of pressure ulcer preventative strategies ‐ are often made on the basis of the results of risk assessment, although it has been argued also that nurses often use their clinical judgement alone in deciding which preventative measures to use (Anthony 2008). Therefore, it is likely that some patients are receiving interventions that they do not require, and conversely others are not receiving interventions that they would benefit from (Defloor 2005). This inappropriate allocation of resources compounds the increasing burden of pressure ulcers, and adds to spiraling healthcare costs. It is important to note that the primary focus of interest for this systematic review is whether or not using a risk assessment tool makes any difference to pressure ulcer incidence, as such the review is not looking at the predictive validity of pressure ulcer risk tools.
Why it is important to do this review
Three systematic reviews that explored the effectiveness of pressure ulcer risk assessment tools in the prevention of pressure ulcers have been published previously. The Royal College of Nursing (UK) pressure ulcer prevention guidelines (NICE 2001) were based largely on the results of the review by McGough 1999. The first review searched from 1962 to 1995 (Cullum 1995), the second review from 1962 to 1999 (McGough 1999) and the third review from 1966 to 2003 (Pancorbo‐Hidalgo 2006). Two reviews (Cullum 1995; McGough 1999) restricted their inclusion criteria to studies published only in the English language; the third review (Pancorbo‐Hidalgo 2006) restricted the inclusion criteria to four languages: Spanish, English, French and Portuguese. The reviews concluded that they found no evidence that pressure ulcer risk assessment scales reduce the incidence of pressure ulcers. However, given the time since these reviews were written, and the language restrictions that were imposed, it is possible that other relevant literature was originally overlooked, or has been published in the meantime. Therefore, as the searches for these reviews are out of date, and the authors imposed language restrictions, it is timely to conduct a review with no language restrictions and recent searches in order to clarify the role of pressure ulcer risk assessment tools in clinical practice.
Objectives
The objective of this review was to answer the following question: does the use of structured, systematic pressure ulcer risk assessment tools, in any healthcare setting, reduce the incidence of pressure ulcers compared with no structured risk assessment or unaided clinical judgement?
Methods
Criteria for considering studies for this review
Types of studies
Randomised controlled trials (RCTs) comparing the use of structured, systematic, pressure ulcer risk assessment tools with no structured pressure ulcer risk assessment, or with unaided clinical judgement, or RCTs comparing the use of different structured pressure ulcer risk assessment tools were considered for this review. Studies that randomise individuals (RCTs) or cluster‐randomised trials (cluster‐RCTs) that randomise by groups, were eligible for inclusion.
Types of participants
Studies involving people without pressure ulcers, of any age, in any healthcare setting (primary, secondary and extended care) were eligible for inclusion.
Types of interventions
RCTs making the following comparisons were eligible for inclusion in this review.
-
Pressure ulcer risk assessment using a specific structured, systematic pressure ulcer risk assessment tool compared with no structured pressure ulcer risk assessment tool or unaided clinical judgement.
-
Comparisons between two different pressure ulcer risk assessment tools.
Types of outcome measures
Primary outcomes
The proportion of participants developing new pressure ulcers of any grade (for the purpose of this review a pressure ulcer was defined as a localised injury to the skin, underlying tissue or both, usually over a bony prominence, as a result of pressure, or pressure in combination with shear) (EPUAP 2009).
Secondary outcomes
-
The severity of new pressure ulcers.
-
Time to ulcer development.
-
Pressure ulcer prevalence.
Search methods for identification of studies
The search methods for the original version of this review can be found in Appendix 1
Electronic searches
In December 2013, for this second update, the following electronic databases were searched for reports of relevant studies:
-
The Cochrane Wounds Group Specialised Register (searched 19 December 2013);
-
The Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2013, Issue 11);
-
Ovid MEDLINE (1948 to November Week 3 2013);
-
Ovid MEDLINE (In‐Process & Other Non‐Indexed Citations, December 10 , 2013);
-
Ovid EMBASE (1980 to 2013 Week 50);
-
EBSCO CINAHL (1982 to 19 December 2013).
The following search strategy was used in The Cochrane Central Register of Controlled Trials (CENTRAL):
#1 MeSH descriptor: [Pressure Ulcer] this term only 524
#2 MeSH descriptor: [Skin Ulcer] this term only131
#3 decubitus or decubital 429
#4 skin near/3 breakdown* 82
#5 bedsore* or (bed next/1 sore*) 69
#6 decubitus next (ulcer* or sore*) 111
#7 pressure* next (wound* or sore* or ulcer* or injur* or damag*) 1237
#8 (#1 or #2 or #3 or #4 or #5 or #6 or #7) 1733
#9 MeSH descriptor: [Risk Assessment] this term only6724
#10 (anderson or braden or norton or knoll or waterlow or medley or maelor or arnold or gosnell) near (score* or scale* or tool* or assess*) 199
#11 risk near/2 assess* 14655
#12 (assess* or predict*) next (tool* or score* or scale*) 4087
#13 MeSH descriptor: [Nursing Assessment] explode all trees 497
#14 (knoll or norton or waterlow) next (modif*) 0
#15 birty* next para 1
#16 cubbin near jackson 2
#17 braden next dupa 1
#18 douglas next ward 1
#19 (wound* next assess*) near (tool* or score* or scale* or scoring or instrument* or equipment* or device*) 12
#20 (bed next sore*) near (tool* or score* or scale* or scoring or instrument* or equipment* or device*) 3
#21 decubit* near (tool* or score* or scale* or scoring or instrument* or equipment* or device*) 7
#22 (pressure next ulcer*) near (tool* or score* or scale* or scoring or instrument* or equipment* or device*) 117
#23 (pressure next sore*) near (tool* or score* or scale* or scoring or instrument* or equipment* or device*) 30
#24 bedsore* near (tool* or score* or scale* or scoring or instrument* or equipment* or device*) 1
#25 (pressure next injur*) near (tool* or score* or scale* or scoring or instrument* or equipment* or device*) 2
#26 (pressure next damag*) near (tool* or score* or scale* or scoring or instrument* or equipment* or device*) 1
#27 (pressure next wound*) near (tool* or score* or scale* or scoring or instrument* or equipment* or device*) 20
#28 #9 or #10 or #11 or #12 or #13 or #14 or #15 or #16 or #17 or #18 or #19 or #20 or #21 or #22 or #23 or #24 or #25 or #26 or #27 18066
#29 #8 and #28 424
The search strategies for Ovid MEDLINE, Ovid EMBASE and EBSCO CINAHL can be found in Appendix 2, Appendix 3 and Appendix 4 respectively. The MEDLINE search was combined with the Cochrane Highly Sensitive Search Strategy for identifying randomised trials in MEDLINE: sensitivity‐ and precision‐maximizing version (2008 revision); Ovid format (Lefebvre 2011). The EMBASE and CINAHL searches were combined with the trial filters developed by the Scottish Intercollegiate Guidelines Network (SIGN) (SIGN 2010). No date or language restrictions were applied.
Searching other resources
We searched citations in all retrieved and relevant studies identified by these strategies for further studies. We contacted experts in the wound care field, namely council members of the European Pressure Ulcer Advisory Panel, the European Wound Management Association, The National Pressure Ulcer Advisory Panel and the World Union of Wound Healing Societies to identify any studies not located through the primary search.
Data collection and analysis
Selection of studies
Two review authors independently assessed titles and, where available, abstracts of the studies identified by the search strategy for their eligibility (as identified in the selection criteria) for inclusion in the review. Two review authors obtained full versions of potentially relevant studies and screened these against the inclusion criteria independently. Any differences in opinion were resolved by discussion and, where necessary, reference to the Wounds Group editorial base.
Data extraction and management
For this update, one review author extracted and summarised trial data. Data entry was checked by the second review author independently. We extracted and summarised details of the eligible study using a data extraction sheet. Specifically, we extracted the following information:
-
author; title; source; date of study;
-
country; care setting;
-
inclusion and exclusion criteria;
-
participant baseline characteristics by group;
-
design details; study type; sample size;
-
allocation;
-
intervention details; concurrent interventions;
-
is risk assessment part of a wider assessment programme/package;
-
frequency of risk assessment; length of follow up;
-
patient length of stay;
-
which health professional administered the tool;
-
outcome measures;
-
verification of diagnosis;
-
analysis;
-
loss to follow up;
-
results; and conclusions.
Assessment of risk of bias in included studies
Two review authors independently assessed the included studies using the Cochrane Collaboration tool for assessing risk of bias (Higgins 2011). This tool addresses six specific domains: namely, sequence generation, allocation concealment, blinding, incomplete outcome data, selective outcome reporting and other issues (e.g. extreme baseline imbalance) (see Appendix 5 for details of criteria on which each judgement was based). We assessed blinding and completeness of outcome data for each outcome separately. We completed a 'Risk of bias' table for the eligible study. We have presented an assessment of risk of bias using a 'Risk of bias' summary figure (Figure 1). This display of internal validity indicates the weight the reader may give to the results of each study.
Methodological quality summary: review authors' judgements about each methodological quality item for each included study.
Assessment of heterogeneity
We planned to explore clinical heterogeneity by examining potentially influential factors, e.g. care setting or patient characteristics. Statistical heterogeneity was to be assessed using the I2 statistic (Higgins 2003). This examines the percentage of total variation across studies due to heterogeneity rather than to chance. Values of I2 over 75% indicate a high level of heterogeneity. We intended to carry out statistical pooling on groups of studies which were considered to be sufficiently similar. However, owing to the lack of homogeneity of the studies included, in terms of the interventions evaluated, statistical pooling was not relevant.
Data synthesis
We entered quantitative data into RevMan 5 (RevMan 2008) and analysed using the RevMan Analysis software. For dichotomous outcomes, we calculated risk ratio (RR) plus 95% confidence intervals (CI).
Results
Description of studies
Results of the search
The initial search identified 105 titles. Following independent review of the abstracts by the two review authors, we retrieved 10 citations in full. Two review authors independently assessed the papers and applied the inclusion and exclusion criteria. No papers were identified that met the inclusion criteria. Fifty‐two letters were written to wound care experts and 16 replies were received, yielding a response rate of 31%. We identified no further trials through this process. The search for the first update identified 98 citations. Following review of the abstracts, we retrieved one citation in full (Saleh 2009). For the second update 171 titles were identified. Following review of the abstracts 1 further study met the inclusion criteria (Webster 2011).
Included studies
Two studies met the inclusion criteria. The first study was published in 2009 (Saleh 2009). This cluster randomised study was conducted within a military hospital in Saudi Arabia and compared the effect of three different methods of pressure ulcer risk assessment on the incidence of pressure ulcers in hospitalised individuals with a Braden score of less than or equal to 18 (Braden 1987). The methods of risk assessment were: the Braden pressure ulcer risk assessment tool and training; unstructured risk assessment and training; and unstructured risk assessment alone (see Characteristics of included studies). The Braden pressure ulcer risk assessment scale comprises six sub‐scales: sensory perception, moisture, activity, mobility, nutrition and friction/shear. Each sub‐scale is ranked numerically from 1 to 4; a score of 4 indicates no problem with regard to the specific sub‐scale, whereas a score of 1 indicates a significant problem. The friction and shear sub‐scale is scored 1 to 3. The scores for each of the sub‐scales are totaled to give a final score ranging from 6 to 23; as scores become lower, predicted risk becomes higher (Braden 1987). Data were collected by the lead researcher, who was the Tissue Viability Nurse Specialist at the study site, and two staff nurses trained in the data collection procedure.
Saleh 2009 included nine wards within a military hospital and randomly allocated these wards into three groups. Group A nurses (the Braden scale group) received a mandatory wound care management study day, pressure ulcer prevention training programme and specific training on the application of the Braden scale. These nurses were required to implement the Braden scale on their patients in the post‐intervention stage. Group B nurses (the training group) were identical to group A but were not required to implement the Braden scale. Group C nurses (the clinical judgement group) received only a mandatory wound care management study day. Data were collected from all patients with Braden scores of ≤ 18 across the nine wards; follow up was for eight weeks. Patients were nursed on either standard foam mattresses, alternating pressure redistribution devices, gel overlay mattresses or air fluidised mattresses. Repositioning schedules were every two hours, three to four hours, or six hours. The procedure for allocation of mattresses and repositioning schedules are not reported by the study authors. Incidence was recorded as the development of a pressure ulcer during the study period. Pressure ulcers were identified according to the National Pressure Ulcer Advisory Panel (USA) pressure ulcer classification system (NPUAP 1998). Saleh 2009 did not identify the grade of pressure ulcer damage specifically for each participant but rather reported 'pressure ulcer present' ‐ yes or no.
The second study was published in 2011 (Webster 2011). This single blind RCT was conducted among 1231 patients admitted to internal medicine or oncology wards within a tertiary referral teaching hospital in Australia. Participants were allocated to either a Waterlow (n=410) or Ramstadius (n=411) screening tool group or to a clinical judgement group (n=410) where no formal risk screening instrument was used. (see Characteristics of included studies). The Waterlow risk assessment tool comprises 7 sub‐scales, Build/weight for height; skin type; nutrition; sex/age; continence; mobility; special risks. Each sub‐scale is scored individually according to an allocated score to each component of the sub‐scale, with the scores added to give an overall risk status. As scores become higher, the predicted risk become correspondingly higher (10+= low risk; 15+= high risk; 20+= very high risk) (Waterlow 1985). The main focus of the Ramstadius risk screening tool (Ramstadius 2000) is on mobility status, it is a non numerical tool and begins with the assessment of mobility as yes/no. If the patient can reposition themselves independently no further assessment is required and the patient is deemed not to be at risk. Conversely, if problems with mobility are identified, the patient is deemed to be at high risk and further assessment of risk factors, namely age, medication, skin integrity, temperature, decreased blood volume, dyspnoea and presence of an existing pressure ulcer is undertaken. No scores are given, rather an algorithm is provided to direct interventions which may be appropriate for the specific risk factor.
Following assignment to the screening method, a copy of the instrument was placed in the patient’s medical record for use by the ward nurse in their admission assessment. Research assistants who were trained in pressure ulcer staging, and who were blinded to the screening method, visually inspected patients for evidence of pressure ulcer formation daily (except weekends). Follow up was for four days. Pressure ulcers were staged according to the NPUAP pressure ulcer staging system (Black 2007). The pressure ulcers that developed were either stage 1 or stage 2, data collection ceased once a pressure ulcer was identified.The authors report that there were no differences in measured processes of care, including use of special mattresses, documentation of an explicit pressure care plan, referral to the specialist skin integrity nurse or referral to a dietician between the three groups.
Excluded studies
The Characteristics of excluded studies table summarises the 10 studies that were excluded from the review.
Risk of bias in included studies
See Figure 1, for the summary of the risk of bias of the included studies.
Allocation (selection bias)
Methods used for generating the allocation sequence and for concealing the group allocation were unclear in the study of Saleh 2009. However, Webster 2011 reports that a computer‐generated randomised list, with a phone randomisation method was used.
Blinding (performance and detection bias)
Saleh 2009 did not mention blinding in the study report, whereas Webster 2011 reports that the patient and the outcome assessor were blinded to group assignment.
Incomplete outcome data (attrition bias)
Saleh 2009 does not report if an intention‐to‐treat (ITT) analysis was undertaken. Webster 2011 reports that the number of participants allocated to each group were analysed for the primary outcome at the end of the study.
Selective reporting (reporting bias)
The two studies, Saleh 2009 and Webster 2011 report all outcomes mentioned in the methods of the papers, however we did not seek the trial protocols.
Other potential sources of bias
In the study of Saleh 2009 the groups were not comparable at baseline for medical diagnoses, pressure ulcer prevention practices, use of barrier creams and use of vitamin supplementary therapy. Furthermore, this was a cluster randomised study but the study authors did not report if they adjusted for the clustering in the sample size calculation and in the analysis. The study of Webster 2011 was not judged to be at risk from other potential sources of bias.
Based on this assessment the study of Saleh 2009 would be judged to be overall at high risk of bias, whereas the study of Webster 2011 would be judged to be overall at low risk of bias.
Effects of interventions
How the results are presented and what the terms mean
Results for dichotomous variables are presented as RR with 95% CI. Risk ratio is the rate of the event of interest (e.g. pressure ulcers developed) in the experimental group divided by the rate of this event in the control group and indicates the chances of pressure ulcer development for people in the experimental group compared with the control group. An RR of 1 means there is no difference in risk between the two study groups, an RR of < 1 means the event is less likely to occur in the experimental group than in the control group and an RR of > 1 means the event is more likely to occur in the experimental group than in the control group.
Comparison 01: Comparison between Braden pressure ulcer risk assessment and training compared with unstructured pressure ulcer risk assessment following the same training alone
This study randomly allocated nine wards into three groups; groups A, B and C. The randomisation resulted in unequal allocation across the groups and no explanation for this was given in the study report.
Following delivery of the training to the staff, Saleh 2009 enrolled 150 patients with a Braden score of ≤18, from six wards. Seventy‐four patients were in the Braden scale group (Group A), and 76 patients were in the training group (Group B). The ward, not the patient, was the unit of randomisation and therefore this would be a cluster RCT study design, however, it is unclear from the trial report if the analysis of data accounted for the clustering. The authors did not describe the characteristics of the participants in terms of age, gender or underlying health status specifically for each group. The patients were followed up for a period of eight weeks. Sixteen pressure ulcers developed in the Braden scale group and 17 pressure ulcers developed in the training group. There was no statistically significant difference in pressure ulcer risk between the groups (RR 0.97, 95% CI 0.53 to 1.77) (Analysis 1.1).
Comparison 02: Comparison between Braden pressure ulcer risk assessment and training compared with unstructured pressure ulcer risk assessment alone
Following delivery of the training to the staff, Saleh 2009 enrolled 106 patients from three wards with a Braden score of ≤ 18. These 106 patients were managed using unaided clinical judgement (Group C). The incidence of pressure ulcers in this clinical judgement group was compared with the 74 patients who were in the Braden scale group (Group A). The ward, not the patient, was the unit of randomisation, however, it is unclear from the trial report if the analysis of data accounted for the clustering. The authors did not describe the characteristics of the participants, in terms of age, gender or underlying health status, specifically for each group. The patients were followed up for a period of eight weeks. Sixteen pressure ulcers developed in the Braden scale group and 16 pressure ulcers also developed in the clinical judgement group. There was no statistically significant difference between the groups in terms of pressure ulcer risk (RR 1.43, 95% CI 0.77 to 2.68) (Analysis 2.1).
Comparison 03: Comparison between Waterlow pressure ulcer risk assessment and no formal risk assessment
Webster 2011 enrolled 411 participants in the Waterlow group and 410 participants into the group receiving no formal risk assessment. Following eligibility assessment, a research nurse allocated patients to study group using a phone, computer generated randomisation, randomisation was blocked and stratified by type of patient (oncology and medical) and by presence or absence of pressure ulcers on admission and mobility status (no pressure ulcer, pressure ulcer and able to move independently, no pressure ulcer but unable to move independently, and pressure ulcer and unable to move independently). The patient and the outcome assessor were blinded to group assignment.The incidence of hospital‐acquired pressure ulcers was similar between the groups (Waterlow 7.5% n=31; clinical judgement 6.8% (n=28). There was no statistically significant difference between the groups in terms of pressure ulcer risk (RR 1.10, 95% CI 0.68 to 1.81) (Analysis 3.1).
Comparison 04: Comparison between Ramstadius risk screening and no formal risk assessment
Webster 2011 enrolled 410 participants in the Ramstadius risk screening group and 410 participants into the no formal risk assessment group. Following eligibility assessment, a research nurse allocated patients to study group using a phone, computer generated randomisation, randomisation was blocked and stratified by type of patient (oncology and medical) and by presence or absence of pressure ulcers on admission and mobility status (no pressure ulcer, pressure ulcer and able to move independently, no pressure ulcer but unable to move independently, and pressure ulcer and unable to move independently). The patient and the outcome assessor were blinded to group assignment.The incidence of hospital‐acquired pressure ulcers was similar between the groups (Ramstadius 5.4% n=22; clinical judgement 6.8% (n=28). There was no statistically significant difference between the groups in terms of pressure ulcer risk (RR 0.79, 95% CI 0.46 to 1.35) (Analysis 4.1).
Comparison 05: Comparison between Waterlow pressure ulcer risk assessment and Ramstadius risk screening
Webster 2011 enrolled 411 participants in the Waterlow group and 410 participants in the Ramstadius risk screening group. Following eligibility assessment, a research nurse allocated patients to study group using a phone, computer generated randomisation, randomisation was blocked and stratified by type of patient (oncology and medical) and by presence or absence of pressure ulcers on admission and mobility status (no pressure ulcer, pressure ulcer and able to move independently, no pressure ulcer but unable to move independently, and pressure ulcer and unable to move independently). The patient and the outcome assessor were blinded to group assignment.The incidence of hospital‐acquired pressure ulcers was similar between the groups (Waterlow 7.5% n=31; Ramstadius 5.4% n=22). There was no statistically significant difference between the groups in terms of pressure ulcer risk (RR 1.44, 95% CI 0.85 to 2.44) (Analysis 5.1).
Discussion
Two eligible studies (Saleh 2009; Webster 2011) were included in this review. Saleh 2009 found no statistically significant differences in pressure ulcer incidence when patients were risk assessed using the Braden scale compared with a risk assessment following pressure ulcer prevention training, or when comparing risk assessment with using clinical judgement alone. Similiarly, Webster 2011 found no statistically significant differences in pressure ulcer incidence when patients were risk assessed using the Waterlow risk assessment tool, the Ramstadius risk screening tool, or using no formal risk assessment.
Some methodological issues require consideration and limit the conclusions that can be drawn from this review. In the study of Saleh 2009 randomisation was not at the individual level but rather at the unit level, where each ward served as the unit of randomisation and all patients within the ward were in the same group. This type of randomisation is called cluster‐randomisation (Medical Research Council 2002). Cluster‐randomised trials increase efficiency and study protocol compliance whilst avoiding contamination (Donner 2004). Contamination is said to occur when an intervention is given to an individual but may affect others within the trial (Puffer 2005) or when the intervention is given by accident to the control group.
The disadvantages of cluster‐randomisation is that all the individuals in the cluster cannot be assumed to be independent of one another and, furthermore, the analysis is not at the level of randomisation but is at the group level (Elley 2004). A way to overcome the disadvantages is to allow for the effects of clustering in the analysis of the data using, for example, regression models (Hahn 2005). Normally, with individual randomisation, one would expect there to be a variance in the responses within study groups. Clustering can exert an effect on this variance yielding a correlation of responses within the clusters. When cluster‐randomisation is used, this needs to be considered during both the sample size calculation and the data analysis. The study by Saleh 2009 was small and the authors did not report that they accounted for the use of cluster‐randomisation in either the sample size calculation, nor in the analysis (Saleh 2009). Conversely, the study of Webster 2011 randomised at the individual level, thereby enhancing the comparability between the study groups.
Concealment of group allocation was inadequately described in the study of Saleh 2009. Allocation concealment is a randomisation method that prevents the researcher influencing which group, experimental or control, a participant is allocated to (Higgins 2011), therein ensuring that the participant is assigned to a specific study group by chance (Higgins 2011). It has been suggested that lack of a clear description of allocation concealment leads to bias in assessing the outcome of studies (Moher 2001); the size of the effect could be overestimated and so give a false impression of the value of the intervention. The study of Webster 2011 used computer generated, phone randomisation, thereby minimising the risk of selection bias.
Blinding of the study is said to be complete if the investigators, the participants, the outcome assessor and the individual analysing the data have no idea which group the participant is allocated to (Higgins 2011). Saleh 2009 did not report blinding of the patient, the staff, the data collector or the data analyst. Whilst it would not have been feasible to blind care givers as they must know the allocation because they are conducting the risk assessment, it would have been possible to blind the outcome assessors and data analyst. Webster 2011 ensured that the patient and the outcome assessor were blinded to group assignment, thereby minimising the risk of performance bias and detection bias.
Intention‐to‐treat analysis (ITT) means that participants are analysed according to the group they were originally allocated to even if they did not adhere to the study protocol or complete the study. The rationale for using ITT analysis is two‐fold; it maintains treatment groups that are similar (apart from random variation) and therefore validates the use of randomisation (Hollis 1999), and allows for handling of protocol deviations, further protecting the randomisation process (Hollis 1999). In essence, omitting those who do not complete the study from the final analysis may bias the outcomes of the study because those who do not complete may do so because of adverse effects of the intervention (Montori 2001). Saleh 2009 did not report use of ITT; pressure ulcer incidence is reported for all patients in the post‐training groups but it is not clear whether any randomised patients withdrew. Conversely, Webster 2011 ensured that all the participants allocated to each group were analysed for the primary outcome, thereby minimising the risk of attrition bias.
Baseline data refers to the data collected from each participant before beginning the trial (Friedman 1996). This includes demographic information, medical condition, prognostic factors and, where appropriate, socioeconomic information. This allows the researcher to determine if participants in both arms of the study are comparable at the outset of the study (Friedman 1996) and allows those evaluating the study to determine if the characteristics of those participating in the study are similar to those normally encountered in clinical practice (Friedman 1996). Webster 2011 provides details of the baseline characteristics of the participants and does not identify statistically significant differences at baseline between the study groups. Saleh 2009 report that, overall, the groups were not comparable at baseline for medical diagnoses, pressure ulcer prevention practices, use of barrier creams and use of vitamin supplementary therapy. This was not an issue in the study of Webster 2011.
Use of pressure ulcer risk assessment tools or scales is a component of the assessment process used to identify individuals at risk of developing a pressure ulcer and is recommended by many international guidelines (AHCPR 1992; EPUAP 2009; NICE 2001). This review identified no RCT evidence to suggest that conducting a structured risk assessment makes any difference to pressure ulcer incidence. This finding is in keeping with previous reviews (Cullum 1995; McGough 1999; Pancorbo‐Hidalgo 2006) which also found a lack of published literature that reliably assesses whether the use of a risk assessment tool reduces the incidence of pressure ulcers.
Pressure ulcer risk assessment tools are widely used in clinical practice, although not necessarily in all healthcare settings (Anthony 2008; Defloor 2005) and as such it is impossible to unlearn what has been gained during the experience of using a risk assessment tool. This means that use of an individual's clinical judgement alone, without use of a risk assessment tool, will ultimately be influenced by prior knowledge of risk assessment tools. Thus, it is possible that within the clinical setting risk assessment follows a structured format similar to that of the current risk assessment tools even in the absence of a paper/electronic version of the tool (Anthony 2008). One therefore might not see a difference in pressure ulcer incidence because the tool does not add to the quality of the clinical judgement. Indeed, Defloor 2005 argues that if nurses act according to risk assessment scales, 80% of the patients would unnecessarily receive preventive measures. Furthermore, use of preventative measures impacts negatively on the predictive ability of the risk assessment tool. One may consider the presence of a pressure ulcer in an individual identified to be at risk to be a success of the risk assessment process; however, this actually indicates a failure of prevention methods (Defloor 2005). It would be interesting to determine what information is gathered using clinical judgement alone to assess whether this matches the data collected using structured risk assessment. If there were a relationship between the two methods of assessment then a reduction in pressure ulcer incidence, due to the introduction of structured risk assessment, would not be anticipated. Thus, in the studies of Webster 2011 and Saleh 2009, it is unclear what impact prior knowledge of pressure ulcer risk assessment had on the clinical judgement of the participants and as such this should be bone in mind in consideration of the generalisability of findings to other healthcare settings.
It has been argued that pressure ulcer risk assessment is in itself not an intervention but rather a precursor to the development of an appropriate plan of care to combat or reduce the impact of the risk factors identified (Lindgren 2002). Anthony 2008 suggests that if a risk assessment tool is working well, then a reduction in the incidence of pressure ulcers should follow. Presumably this means that the risk assessment is followed by appropriate risk intervention, and that these interventions are available and effective. It is evident from the literature, however, that this is not always the case (EPUAP 2002; Moore 2004). Indeed, EPUAP 2002 found that only 9.7% of patients in a pan‐European prevalence study were receiving adequate preventative measures in terms of repositioning and provision of pressure redistributing devices. Furthermore, in a survey of registered nurses Moore 2004 found that pressure ulcer prevention was not always a high priority, with some nurses admitting to being less interested in pressure ulcer prevention than other aspects of nursing care. Fundamentally, risk assessment alone will make no difference unless it is followed up by an intervention to combat risk, and these interventions need to be available. Interestingly, the method of risk assessment employed in the Webster 2011 study did not influence the interventions offered to patients, indeed, there were no differences in measured processes of care, including use of special mattresses, documentation of an explicit pressure care plan, referral to the specialist skin integrity nurse or referral to a dietician between the three groups. Thus, although risk assessment is suggested to be a precursor to planning and implementing care, it appears that this may not always be the case.
One cluster RCT at high risk of bias has explored the impact of pressure ulcer risk assessment on patient outcomes (Saleh 2009). However, methodological issues with the study make it difficult to draw firm conclusions. However, a further large RCT (Webster 2011) at low risk of bias, identified no statistical differences in pressure ulcer incidence when patients were assessed using either the Waterlow risk assessment tool, the Ramstadius risk assessment tool, or using clinical judgement alone. However, as the studies included here were within two specific clinical settings (military hospital and internal medicine or oncology) there is limited generalisability to other high risk groups for example elderly residents of care homes. Therefore, as yet, there is no RCT evidence to suggest that conducting pressure ulcer risk assessment makes any difference to the number of pressure ulcers that develop. If the use of risk assessment tools/scales continues to be used in clinical practice, in the absence of empirical knowledge regarding its effect on clinical outcomes, issues will arise concerning resource utilisation and this in turn will add to increasing healthcare costs.
Methodological quality summary: review authors' judgements about each methodological quality item for each included study.
Comparison 1 Comparison between Braden pressure ulcer risk assessment and training vs. unstructured pressure ulcer risk assessment following training alone, Outcome 1 Pressure ulcer incidence ‐ Braden risk assessment and training vs. unstructured risk assessment following training alone.
Comparison 2 Comparison between Braden pressure ulcer risk assessment and training vs. unstructured pressure ulcer risk assessment alone, Outcome 1 Pressure ulcer incidence ‐ Braden risk assessment and training vs. unstructured risk assessment alone.
Comparison 3 Comparison between Waterlow and no formal risk assessment, Outcome 1 Pressure Ulcer incidence: Waterlow versus no formal risk assessment.
Comparison 4 Comparison between Ramstadius and no formal risk assessment, Outcome 1 Pressure Ulcer incidence: Ramstadius versus no formal risk assessment.
Comparison 5 Comparison between Waterlow and Ramstadius, Outcome 1 Pressure Ulcer Incidence: Waterlow versus Ramstadius.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Pressure ulcer incidence ‐ Braden risk assessment and training vs. unstructured risk assessment following training alone Show forest plot | 1 | 150 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.97 [0.53, 1.77] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Pressure ulcer incidence ‐ Braden risk assessment and training vs. unstructured risk assessment alone Show forest plot | 1 | 180 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.43 [0.77, 2.68] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Pressure Ulcer incidence: Waterlow versus no formal risk assessment Show forest plot | 1 | 821 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.10 [0.68, 1.81] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Pressure Ulcer incidence: Ramstadius versus no formal risk assessment Show forest plot | 1 | 820 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.79 [0.46, 1.35] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Pressure Ulcer Incidence: Waterlow versus Ramstadius Show forest plot | 1 | 831 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.44 [0.85, 2.44] |
