Study characteristics

Methods

Study design: Controlled before‐after study

Number of facilities: unclear (total of GreenHouse, control and excluded GreenHouse owned "legacy" homes: 190; 12 intervention and 178 control)

Participants

N = 74,449 (weighted sample)

Mean age: not reported.

% Female: not reported.

% Dementia: 33.9

Mean number of comorbidities: not reported
Country: USA

Inclusion criteria:

Nursing homes:

• that adopted the GreenHouse (GH) model over the period 2005 to 2010 (from list provided by The Green House Project)

• were in operation in the period prior to adoption ‐ non‐GH nursing home organisations matched at facility‐level by OSCAR (Online Survey, Certification, and Reporting)

• nursing home level data for each GH nursing home plus matched on state and year of adoption, using nearest neighbour matching

• matched on these covariates: nonprofit ownership, for‐profit ownership, government ownership, chain status, small size (75 beds or fewer), medium size (76–125 beds), large size (126 or more beds), rural location, above median Medicaid share, above median Medicare share, above median private‐pay share, and a nursing home‐level aggregate activities of daily living (ADL) score (0 if less than 4 on a scale of 0–5, 1 otherwise), with propensity score weighting

Exclusion criteria:

• Residents who were not entitled to Medicare Part A in the month of admission

• Residents who died during the month of admission

• Residents who were enrolled in the Medicare Advantage program during the month of admission

Interventions

Type of intervention: Home‐like model

Name of intervention: Green House model.

Design features: Small buildings (maximum 12 residents) and fit the style of surrounding neighbourhood

Other features that differed: Residents have more control over daily activities.

Control: Matched by multiple facility characteristics, state and year that did not adopt the Green House model, not a "small house" model

Outcomes

The following outcomes were reported:

Rehospitalisations, avoidable rehospitalisations, bedfast residents, catheter use, pressure ulcers (low risk and high risk) and use of physical restraints

Follow‐up: Up to five years

Notes

Study supported by the Robert Wood Johnson Foundation. Conflicts of interest: None. Ethical approval: not stated.

Data reported on "Legacy" units within the GreenHouse organisation (that were not GreenHouse model of care homes) were excluded.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Controlled before‐after study

Allocation concealment (selection bias)

High risk

Controlled before‐after study

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible

Blinding of outcome assessment (detection bias)
All outcomes

High risk

Blinding not feasible

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

Loss to follow‐up not stated

Selective reporting (reporting bias)

Unclear risk

No protocol; likely to be many fields in MDS not reported

Other bias

High risk

Method of selection of facilities unclear and potential residual confounding. Significant differences in baseline characteristics. Significant differences for many baseline outcome measures

Study characteristics

Methods

Study design: Controlled before‐after study

Number of facilities: 6 (3 intervention and 3 control)

Participants

N = 56 (intervention: N = 28; control: N = 28)

Mean age (SD): Intervention: 82.8 (5.0), control: 81.6 (5.0)

% Female: Intervention: 89.3, control: 82.1

% Dementia: not reported

Mean number of comorbidities: not reported

Country: Sweden

Inclusion criteria:

• Intervention residents: Alzheimer's, vascular dementia or mixed according to DSM‐III criteria (severity of dementia II to IV on Berger's scale)

• Control residents: matched by age, gender, diagnosis and level of dementia to the intervention residents

Exclusion criteria:

• Patients suffering from physical illness, clinically estimated to become terminal within 1 year

Interventions

Type of intervention: Home‐like model

Design features: 8‐10 residents, situated in ordinary blocks of flats in suburbs, specially adapted for people living with dementia. Flats had private areas with 1 or 2 rooms with toilet and shower, kitchen, living room, dining room and laundry room available to staff and residents. Staff/patient ratios were similar in the two types of care (0.69 in group living; 0.71 in traditional). However, the intervention staff were responsible not only for patient care but also for cooking, cleaning, washing, transportation and activating therapies.

Other features that differed: Staff familiar with resident biographies. A group living project was connected to a clinic, responsible for the geriatric care in a certain area corresponding to 20% of the persons 65 years old or older of the population in the community.

Control: Offered in three big long‐term care hospitals, mostly with wards of 50 patients and originally designed for acute medical care or rehabilitation. The care was usually organised in four separate groups in order to break the large‐scale design of the physical environment. Each of these groups included both somatic long‐term care patients with and without dementia. Staff training was seldom regularly scheduled and surveyed most aspects of geriatric care, favouring physical items.

Outcomes

The following outcomes (measurement scale) were reported:

Dyspraxia, hallucinations, lack of vitality, dysphasia, paranoia, aggressiveness, depression, clinical variations, restlessness, recent memory and identity (Organic Brain Syndrome Scale: OBS Scale)

Follow‐up: 6 months and 12 months

Notes

Sponsorship source: Swedish Medical Research Council, the Swedish Council for Social Research, Alzheimer Foundation and Medical Foundation at Lund University. Conflicts of interest: Not stated. Ethical approval: Not stated

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Controlled before‐after study

Allocation concealment (selection bias)

High risk

Controlled before‐after study

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible

Blinding of outcome assessment (detection bias)
All outcomes

High risk

Blinding not feasible

Incomplete outcome data (attrition bias)
All outcomes

High risk

21% loss to follow‐up

Selective reporting (reporting bias)

Low risk

All outcomes discussed were reported.

Other bias

High risk

Possible confounding, unadjusted results. Significant baseline differences in the time the participants had been institutionalised prior to relocation; participants in traditional facilities received more neuroleptic treatment. Differences in baseline outcomes e.g. dyspraxia and identity not controlled for.

Study characteristics

Methods

Study design: Controlled before‐after study

Number of facilities: 13 (7 intervention and 6 control)

Participants

N = 201 enrolled, 101 analysed (intervention: N = 50; control: N = 51)

Mean age (SD): Intervention: 83.8 (8.8), control: 83.5 (9.8)

% Female: Intervention: 62, control: 67

% Dementia: 59 (not reported by intervention and control groups)

Mean (SD) number of comorbidities: 4.3 (1.9) (intervention: 4.2 (1.9) and control: 4.4 (2.0))

Country: USA

Inclusion criteria:

Facilities

• Intervention: 2 to 3 pilot communities at each of 3 nursing home campuses for a total of 7 communities operated by one provider. Nursing and administrative staff chose communities with well‐functioning teams (2 communities each at 2 of the campuses, and 3 communities at the third campus) to pilot the culture change intervention to optimise the potential for a successful culture change transformation.

• Comparison: 2 communities at each campus were identified to serve as a comparison group (for a total of 6 comparison communities), selected by administrative and nursing leaderships’ clinical expertise to best match the culture change group by the level of care needed by elders, staff team functioning, number of elders in the community, and the environmental community structure.

Residents

• Living in community for at least 3 months, 60 years or older

Exclusion criteria:

• Not reported

Interventions

Name of intervention: Culture change model

Design features: Environmental changes were implemented in elder rooms and common areas, with a focus on person‐centred care. Elders and their family members were encouraged to individualise elder rooms with personal items, decoration, and pictures. Within the common areas, attention was given to creating a calm, peaceful environment. In the dining areas, new table cloths were purchased, centrepieces were placed on tables, art work decorated the walls, and water and juice were easily accessible to elders at all times. Hallways were decorated with painted murals and new wallpaper. The outsides of the elders’ rooms were individualised to facilitate easy room recognition for the elders. Additionally, homey nooks were created at the end of hallways with comfortable seating. Noise level was addressed by discontinuing the overhead paging system and turning off TVs and radios when they were not being actively used.

Other features that differed: Community co‐ordinators, education, organisational and community structure changes, meaningful activities and resident choice, family involvement, reduced floating and consistent staffing

Control: Continued to function along the nursing home’s pre‐culture change model, following the typical nursing home organisational structure and standard administrative and departmental hierarchy of care

Outcomes

The following outcomes (measurement scale) were reported:

Behaviour: forceful behaviours, physical agitation and verbal agitation (Cohen‐Mansfield Agitation Inventory: CMAI)

Follow‐up: 2 years

Notes

Sponsorship source not reported. Conflicts of interest: None. Ethical approval: Not stated

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Controlled before‐after study

Allocation concealment (selection bias)

High risk

Controlled before‐after study

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible

Blinding of outcome assessment (detection bias)
All outcomes

High risk

Blinding not feasible

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

Large loss to follow‐up; no differences between those lost and those with follow‐up, but this not reported by group allocation

Selective reporting (reporting bias)

High risk

Only reported behavioural outcomes, not changes in ADLs or cognition. CMAI overall score not reported and reported according to specific groupings

Other bias

High risk

Culture change communities selected based on "well‐functioning teams" to optimise potential for successful transformation. Significant baseline differences in ADLs and race reported. Significant differences in baseline outcome measurements

Study characteristics

Methods

Study design: Cluster‐randomised controlled trial.

Number of facilities: 38 (8 control, 10 PCE, 10 PCC, 10 PCE + PCC)

Participants

N = 601 (person‐centred environment (PCE): 154, person‐centred care (PCC): 155, PCE + PCC: 150, usual environment (UE, control): 142)

Mean age (SD): PCE: 84 (8), PCC: 84 (8), PCE + PCC: 84 (7), control: 86 (7)

% Female: PCE: 66, PCC: 67, PCE + PCC: 70, control: 77

% Dementia: 100

Mean (SD) number of comorbidities: not reported. % > 3 comorbidities: PCE: 35, PCC: 51, PCE + PCC: 55, control: 68

Country: Australia

Inclusion criteria:

Residential aged care home

• Government accreditation and building certification; high‐level care homes; accessible by sealed road, located within a 500 km radius of Sydney, Australia; with room for improvement in both PCE and PCC according to the Person‐Centred Environment and Care Assessment Tool (PCECAT), a validated 44‐item rating instrument with three domains designed for evaluation of residential aged care. The PCECAT 4‐point scale was rescored 0 (the best possible rating) and 1, 2, 3 (the worst possible ratings, ranked).

• A total “room for improvement score” (RFI) was calculated by summing across items (20 items in domain 2 (care services), and 19 in domain 3 (environment)). Homes that scored 1–3 for both care services and environment RFI were considered eligible.

Participants

• Self‐consent, proxy consent or Guardianship Tribunal consent

• Recorded dementia diagnosis

• Permanent stay

• Admission at least 3 months prior to baseline

• Assessed high care needs and presence of agitation

• Ability to participate over the life of the study (e.g. no florid mental illness or end‐stage dementia)

Exclusion criteria:

• Did not meet inclusion criteria

Interventions

Type of intervention: PCE or PCE + PCC

PCE: Two chief investigators with expertise in Person‐Centre Environment design and a Master of Design research student took responsibility for implementing the PCE interventions at each of the 10 PCE and 10 PCE + PCC sites. The Environment Audit Tool (EAT) was employed to evaluate the relationships between operations and space in terms of effectiveness and ideal resident care, and determining required environmental changes to meet PCE principles at the sites. Discussions of EAT findings were held with the home’s executive staff and managers to initially determine their understanding of the dysfunction generated for residents through the poor physical environment features identified. Planning then occurred with these senior staff to determine the best ways to undertake the most essential and inexpensive environmental changes required. Planned modifications to the environment were undertaken in some homes where feasible by a contracted building company. The environment interventions, agreed to by the managers and priced by the contractor, were as follows: (1) two facilities needed extensions of activity space made by covering balconies or areas that were previously open; (2) two facilities had changes made to internal walls that would allow better visual access to activity and bedroom spaces; (3) one facility was to be altered to provide access to a courtyard from a dining area needed for activity and group activities; (4) two facilities needed internal divisions with added partitions to reduce the overstimulation in larger group spaces; (5) two facilities had walls removed to make sitting areas visible to residents passing in the corridor; (6) one facility had fire doors relocated to improve access to the garden and (7) the remaining facilities all had some variation of external paving, new sitting areas in gardens or covered spaces in a landscaped exterior. All these changes were considered to provide maximum benefit in achieving improved support or staff undertaking PCC‐focused activities while engaging with residents.

PCC: Kitwood’s (1997) PCC principles, using experiential and adult learning approaches, were facilitated by two chief investigators with expertise in PCC approaches and one expert PCC trainer from Alzheimer’s Australia, employing train‐the‐trainer processes. Five staff (one Care Manager, one Registered Nurse, two Enrolled Nurses or Assistants in Nursing, one Diversion/Recreation Therapist) from each of the 10 PCC and 10 PCC + PCE homes were involved in the PCC training. The 32‐hour off‐site training occurred over 1 week, complemented by a further 32 hours of on‐site education and support to implement PCC in daily care practices and recreation activities. Prior experiences, case studies, role plays and simulations were utilised to develop awareness and insight of the relationship between care and the resident’s quality of life. The PCC trainer guided and supported PCC‐trained staff to employ PCC learning resources, mentoring and role modelling in educating all care and therapy staff in PCC. With the support of their managers and the PCC trainer, direct‐care staff members were assisted to develop person‐centred resident care and recreation activity plans, and to implement changes in care routines and procedures, with the focus on improving residents’ quality of life and reducing changed behaviours. Ongoing telephone support continued for PCC‐trained staff by the PCC trainer until post‐test.

Control: Regular monitoring of any unplanned changes to the environment

Outcomes

The following outcomes (measurement scale) were reported:

Quality of life (Dementia Quality of Life: DEMQOL), agitation (Cohen‐Mansfield Agitation Inventory: CMAI), quality of care (Quality of Interactions Schedule: QUIS)

Follow‐up: 8 months

Notes

Sponsorship source: National Health and Medical Research Council, Australia (funding source category 1), University of Technology Sydney, Australia (primary sponsor) and Australian Health Ministers‐States & Territories, Australia (secondary sponsor). Conflicts of interest: None. Ethical approval: Research ethics approval was granted by the University of Technology Sydney Human Research Ethics committee approval number: UTS‐HREC 2006‐269A in November 2007, and also by the participating residential care homes. Proxy consent was obtained for all participating residents and both written and verbal consent were obtained from a small number of residents who were able to understand and remember the study’s purpose and procedures prior to administering the measures that required their direct involvement.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Generated using a SAS program

Allocation concealment (selection bias)

Unclear risk

Unclear where the randomisation sequence was stored

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible

Blinding of outcome assessment (detection bias)
All outcomes

High risk

Blinding not possible on outcomes collected involving residents, family or staff

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

31% loss to follow‐up and reasons not described. Analysis compared completers and non‐completers.

Selective reporting (reporting bias)

Low risk

Published protocol and outcomes reported in protocol were the same as in the main paper.

Other bias

Low risk

No other instances for bias obvious from the study

Study characteristics

Methods

Study design: Controlled before‐after study

Number of facilities: 6 (3 intervention and number of control facilities unclear, but stated there was a control group "for each of the three residential groups")

Participants

N = 60 (Intervention: N = 27; control N = 33)

Mean age: Intervention: 82.9, control: 83.1

% Female: Intervention: 81.5, control: 78.5

% Dementia: 100

Mean (SD) number of comorbidities: not reported

Country: Germany

Inclusion criteria:

Intervention participants

• Residents from 3 dementia‐specific facilities

• Moderate or severe dementia (MMSE < 18)

• Barthel 25‐50

• Not excluded based on behaviour

Control participants

• Residents from home operated by same provider

• Matched by age

• Progression of dementia and mobility (Barthel) Cognition: MMSE; function: Barthel; length of stay in months

Exclusion criteria:

• Not reported

Interventions

Name of intervention: Residential group environment

Design features: This was the focus of residential groups for people with dementia. A residential living environment was created in small residential units that followed family structures. The size was between 6 and 12 and, exceptionally, up to 15 residents. A home‐like living environment, adapted to the residents, was created for all three segregated residential groups (13, 15, 15 residents). All three residential groups operated under the live‐in kitchen model, aimed at addressing the residents' multiple facets (activation, communication, emotion, chronological structuring).

Other features that differed: Nursing home typical organisational structures were replaced by small‐scale design, familiarity, communication, needs‐based activity and close human interaction (staff consistency and staff on duty). The daily routine was not dominated by the care activity, but relied on familiar day‐to‐day household tasks. Staff members interacted with the residents in a trusting respectful manner. The events of the day were aligned to the residents' mobility, cognitive abilities as well as their habits. The staff assigned to the participating residential groups and the control groups had comparable qualifications and rosters.

Control: Nursing home typical organisational structures.

Outcomes

The following outcomes (measurement scale) were reported:

Social behaviour (Nurses Observations Scale for Geriatric Patients: NOSGER), function (Barthel Index) and cognitive function (Mini Mental State Examination: MMSE)

Follow‐up: 12 months

Notes

Sponsorship source not reported. Conflicts of interest: Not stated
Ethical approval: Not stated

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

No randomisation

Allocation concealment (selection bias)

High risk

Controlled before‐after study

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible

Blinding of outcome assessment (detection bias)
All outcomes

High risk

Blinding not feasible

Incomplete outcome data (attrition bias)
All outcomes

High risk

Stated 158 residents participated and 111 available for all three surveys but results for 60 presented

Selective reporting (reporting bias)

High risk

Stated Cohen‐Mansfield Agitation Inventory measured but results not shown

Other bias

High risk

No adjustments made so potential for confounding. For baseline characteristics: no statistical tests completed, appeared to be some differences but unclear if statistically or clinically significant. Baseline differences in social behaviour

Study characteristics

Methods

Study design: Controlled before‐after study

Number of facilities: 8 (3 intervention and 5 control)

Participants

N = 119 enrolled (Intervention: N = 60, control: N = 59)

Mean age (SD): Intervention: 82.3 (5.7, mild cognitive impairment) 81.5 (7.4, severe cognitive impairment), control: 82.7 (8.0, mild cognitive impairment) 82.2 (8.0, severe cognitive impairment)

% Female: Intervention: 79.8 (mild cognitive impairment) 82.1 (severe cognitive impairment), control: 78.1 (mild cognitive impairment) 82.1 (severe cognitive impairment)

% Dementia: not reported

Mean (SD) number of comorbidities: not reported

Country: Spain

Inclusion criteria:

• Cognitive impairment

• Experimental group: resided in one of the 8 day or permanent cohabitation units, where the interventions relating to "Etxean Ondo" were incorporated

• Control group: The members of the control group were identified from five distinct centres, three of which coincided with the location of the cohabitation units.

Exclusion criteria:

• Absence of cognitive impairment, measured by Lobo's Cognitive Mini Examination (MEC > 29)

Interventions

Name of intervention: Etxean Ondo

Design features: Creation of domestic environments. "Comfortable, safe and accessible homelike environments, which expedite the daily life of the residents by integrating their important preferences, customs and activities" and "physical spaces were selected that were susceptible to be adapted to the features of domestic environments, favouring the incorporation of their own furniture and other decorative and important items both in public and private spaces".

Other features that differed: The development of important activities and organisational processes based on the daily life and the resources of residents, families and professionals. Support staff who volunteered to work in the units. Increased staff ratio "support staff ratio was increased, reducing the staff rotation between the different areas in the centres, and providing them with continuous professional development" and "periodical meetings of the technical staff (doctor, nurse, psychologist, etc.) with the support teams were set up, changing the decision‐making in relation to the care, with adaptations based on the information provided by the support staff, who act as "reference professionals" for the residents."

Control: Provision of public health services in accordance with the health needs of the residents, the formal registration of care tasks and activities, and the prioritisation of safety both in the design of the spaces and the organisation.

Outcomes

The following outcomes (measurement scale) were reported:

Quality of life (Quality of Life in Late‐Stage Dementia: QUALID for severe cognitive impairment or Fumat for mild cognitive impairment)

Follow‐up: 6 months

Notes

Sponsorship source not reported. Conflicts of interest: None
Ethical approval: Not stated

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

No randomisation

Allocation concealment (selection bias)

High risk

Controlled before‐after study

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible

Blinding of outcome assessment (detection bias)
All outcomes

High risk

Blinding not feasible

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

Did not report loss to follow‐up

Selective reporting (reporting bias)

Low risk

Outcomes reported as per methods

Other bias

High risk

No adjustments made so potential confounding. Stated statistically different quality of life measurements between groups at baseline. Possible contamination through professional staff "reduced staff rotation" mentioned which indicates there may have still been some rotation, plus technical staff meetings

Study characteristics

Methods

Study design: Controlled before‐after study

Number of facilities: 18 (14 corridor design and 4 non‐corridor design)

Participants

N = 105 (Corridor design: N = 66; non‐corridor design: N = 39)

Mean age (SD): Corridor design: 82.9 (5.3), non‐corridor design: not reported

% Female: Corridor design: 89, non‐corridor design: 87

% Dementia: 100

Mean (SD) number of comorbidities: not reported

Country: Sweden

Inclusion criteria:

• Living with dementia and admitted to group living units in Malmo, Sweden during study period

• Group living eligibility: dementia of Alzheimer's type or vascular dementia, care planning team judged home care situation as insufficient

Exclusion criteria:

• Not reported

Interventions

Type of intervention: building layout (comparison of group living units with a corridor design versus non‐corridor design (L‐shaped, square or H‐shaped))

Design features: Built for 6‐8 residents, with specially designed community area comprising living room, laundry, kitchen and dining room shared by the residents and staff. Each resident has a private area of approximately 25 m², furnished by the resident and included a toilet and shower. Located in ordinary blocks of flats outside institutions. Physical environment assessed by architect in standardised manner using Therapeutic Environment Screening Scale (TESS‐2)

Outcomes

The following outcomes (measurement scale) were reported:

Dyspraxia, hallucinations, lack of vitality, dysphasia, paranoia, aggressiveness, depression, clinical variations, restlessness, recent memory and identity (Organic Brain Syndrome Scale: OBS Scale)

Follow‐up: 12 months

Notes

Supported by the Swedish Council for Social Research. Conflicts of interest: Not stated. Ethical approval: Not stated

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Controlled before‐after study

Allocation concealment (selection bias)

High risk

Controlled before‐after study

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible

Blinding of outcome assessment (detection bias)
All outcomes

High risk

No blinding

Incomplete outcome data (attrition bias)
All outcomes

Low risk

> 90% follow‐up; unlikely to bias results

Selective reporting (reporting bias)

High risk

In the methods, there was mention of measuring ADLs and MMSE but results for these have not been reported. Also 6‐month data were not reported.

Other bias

Unclear risk

Potential residual confounding and baseline characteristics not shown by group. Significant differences in lack of vitality and restlessness at baseline. Adjusted analysis accounted for other symptoms.

Study characteristics

Methods

Study design: cluster‐randomised trial (participants served as own controls)

Number of facilities: 8

Participants

N = 52 

Mean age (SD): 85.1 (7.1)

% Female: 65.2%

% Dementia: 100

Number of comorbidities: not reported

Country: USA

Inclusion criteria: diagnosis of dementia according to Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition; a Mini Mental State Examination (MMSE) score between 4 and 24 points (indicating severe [≤ 10] to mild [< 25] dementia) or a Brief Interview for Mental Status (BIMS) score between 3 and 12 points (indicating severe [≤ 7] to moderate [8–12] cognitive impairment), depending on the particular facility’s evaluation procedures; and a score > 5 (indicating sleep disturbance) on the Pittsburgh Sleep Quality Index (PSQI) questionnaire

Exclusion criteria: Major organ failure, a major illness, a history of head injury, uncontrolled generalised disorders (e.g. diabetes), obstructing cataracts, macular degeneration, blindness, or used psychotropic medicine. Those with severe sleep apnoea or restless legs syndrome were also excluded

Interventions

Lighting designed to provide high circadian stimulus. Custom‐built floor luminaires, light boxes and light tables were used. Timers activated lights according to wake times and lights were placed in the person's bedroom or in the common area until 6 pm.

Outcomes

The following outcomes (measurement scale) were reported:

Quality of life (Minimum Data Set Activities of Daily Living Scale (MDS‐ADL)), behaviour (Cohen‐Mansfield Agitation Inventory (CMAI)) and depression (Cornell Scale for Depression in Dementia (CSDD))

Notes

Sponsorship source: This research was funded by the National Institute on Aging (grant #R01AG034157); the following manufacturers are acknowledged for their provision of in‐kind lighting products: GE Current, a Daintree company; OSRAM Sylvania; Ketra; and Sharp Corporation. Conflicts of interest: Neither the funding agency nor the in‐kind contributors had
any role in the design, methods, data analysis, or preparation of the manuscript.
Figueiro, Plitnick, Roohan, Sahin, and Rea received research grant support from the National Institutes of Health, Office of Naval Research, The United States General Services Administration, and industry (Acuity Brands; Axis Lighting; GE Current, a Daintree company; OSRAM Sylvania; Ketra; USAI Lighting; Armstrong Ceilings and Walls; Philips Lighting; Cree; View Glass; Marriott International). Kalsher received research grant support from the National Institutes of Health. Ethical approval: Approved by the Rensselaer Polytechnic Institute Institutional Review Board

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Unclear as to method of sequence generation

Allocation concealment (selection bias)

Unclear risk

Did not specify details

Blinding of participants and personnel (performance bias)
All outcomes

Unclear risk

Details of blinding not reported

Blinding of outcome assessment (detection bias)
All outcomes

Unclear risk

Details of blinding not reported

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

No flow of participants reported. Reported data not available for 4 participants due to nonadherence and some data were not usable

Selective reporting (reporting bias)

Low risk

Trial registered

Other bias

Low risk

Participants served as own controls.

Study characteristics

Methods

Study design: Controlled before‐after study

Number of facilities: 43 (25 SCU and 18 control)

Participants

N = 66 (Special care unit (SCU): N = 31; control: N = 35)

Mean age (SD): SCU: 81 (8), control: 81 (6)

% Female: SCU: 71, control: 80

% Dementia: 100

Mean (SD) number of comorbidities: SCU: 18 (4), control: 19 (3)

Country: Italy

Inclusion criteria:

Facilities

• NHs of East Lombardia region and Milano area that admitted residents with dementia with some degree of behavioural disturbance expected to stay at least 3 months

Patients

• Newly admitted

• Diagnosis of degenerative or vascular dementia according to accepted criteria, MMSE 21 or lower plus CDR 4 or lower plus at least one behavioural disturbance of mild to moderate severity

• MMSW 16 or lower, CDR score 2‐4, NPI total 24 or higher or score of 12 or more in at least one subscale

Exclusion criteria:

• MMSE 22 or higher, extended CDR scale 5 or higher (bed‐bound)

• 15 days or more between admission and communication

• Incomplete data provided by NH physician in first screening form

• History of mental insufficiency, psychosis or major depression

Interventions

Type of intervention: 10 two‐bed rooms, a large wandering area, a dining room, and a separate area for structured activity (physical and occupational therapy). Exit doors were secured by magnetic locks opening with a digital code. Noxious stimuli were minimised, and wall colours were made neutral. Way‐finding cues were used to help residents identify different areas.

Outcomes

The following outcomes (measurement scale) were reported:

Delusions, hallucinations, agitation, anxiety, euphoria/elation, disinhibition, irritability/lability, abnormal motor behaviour, sleep and global behaviour (NPI), agitation (CMAI), depression, cognitive function (MMSE and Clinical Dementia Rating), function (Bedford Alzheimer's Nursing Severity Scale), function (Barthel Index), falls in 3 months, physical restraints

Follow‐up: 3 months

Notes

Supported by European Commission (DGV). Conflicts of interest: Not stated. Ethical approval: Not stated

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Controlled before‐after study

Allocation concealment (selection bias)

High risk

Allocation not concealed

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible

Blinding of outcome assessment (detection bias)
All outcomes

High risk

Not blinded. Follow‐up assessment carried out by same interviewer who carried out baseline assessment whenever possible.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Appeared to be no loss to follow‐up over 3 months

Selective reporting (reporting bias)

Low risk

All relevant outcomes in the methods section were reported in the results section.

Other bias

High risk

Possibility of residual confounding. High risk of falls outcome only; low risk for other outcomes. Higher risk of falls (Tinetti balance and gait scale) in traditional arm. Differences in behavioural disturbances at baseline (although not statistically significant) and no adjustments made

Study characteristics

Methods

Study design: cluster‐randomised controlled trial

Number of facilities: 12 (6 intervention and 6 control)

Participants

N = 336 (173 intervention, 163 control)

Mean age (SD): 82.6 (10.1). Intervention: 82.7 (9.8), control: 82.5 (10.4)

% Female: 72.0%. Intervention: 72.3%, control: 71.8%

% Dementia: not reported

Mean comorbidities (SD): 2.9 (1.6). Intervention: 2.8 (1.5), control: 3.1 (1.7)

Country: USA

Inclusion criteria: 55 years of age or older, able to speak English, currently living in the nursing home, and scored ≤ 15 on the Mini‐Mental State Examination (MMSE)

Exclusion criteria: receiving hospice or sub‐acute rehabilitation

Interventions

Function and Behavior Focused Care for the Cognitively Impaired (FBFC). The FBFC Research Nurse worked with the facility Champion to assess the facility’s policies and the environment to identify opportunities for physical activity and engaging in functional tasks as well as barriers to these activities. For example, corridors were evaluated for wide, clear areas for walking and outdoor access was assessed. In addition, the FBFC Nurse and Champion collaborated with the activities director and rehabilitation staff (as available) to determine opportunities for exercise classes within the facility. Barriers to physical activity, such as policies that unnecessarily restrict movement for fear of falls, and environments that lack rest areas and age‐appropriate exercise materials also were assessed. Based on these assessments, modifications in policy and the environment were made, such as increasing availability of recreational activities that included physical activity (i.e. horseshoes, resistance bands, physical activity, BINGO), having adequate supply of chairs in dining rooms to allow for transfer out of wheelchairs for meals, having safe access to the outdoors, placing a bench in a hallway to have a resting place when walking, providing long‐handle sponges, no spill cups, adaptive utensils as appropriate, and changing the height of a toilet or bed to facilitate function.

Outcomes

The following outcomes (measurement scale) were reported:

Depression (Cornell Scale for Depression in Dementia (CSDD)), behaviour (Cohen‐Mansfield Agitation Inventory (CMAI)), resistiveness to care (Resistiveness to Care Scale) and function (Barthel Index)

Notes

Sponsorship source: National Institute on Aging grant R01 AG046217. Conflicts of interest: None. Ethical approval: Approved by a university‐based institutional review board

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Coin toss within matched pairs

Allocation concealment (selection bias)

Unclear risk

Not specified

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible

Blinding of outcome assessment (detection bias)
All outcomes

High risk

High for staff reported, functional ability, behaviour and mood

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

Large loss to follow‐up, 146 (30%) after randomisation due to MMSE but unclear which randomised group they belonged to. Other reasons: very roughly balanced follow‐up 61% intervention group, 73% control group (119/163)

Selective reporting (reporting bias)

Low risk

All reported as per methods

Other bias

Unclear risk

Differences in baseline outcome measures but no statistical significance reported

Study characteristics

Methods

Study design: Randomised trial (cross‐over)

Number of facilities: 8 (cross‐over trial)

Participants

N = 80 enrolled; N = 69 post‐intervention

Mean age (SD): 85.8 (7.5)

% Female: 86.3

% Dementia: not reported

Mean (SD) number of comorbidities: not reported

Country: England

Inclusion criteria:

• Resident of one of seven included care homes over 60 years of age

• Willing and able to give written informed consent or their family spend time each day in communal rooms where lights were installed

Exclusion criteria:

• Did not meet inclusion criteria

Interventions

Type of intervention: lighting (blue‐enriched lighting)

Design features: High colour temperature (17000 K) blue‐enriched white light in communal areas

Control: Low colour temperature (4000 K) white light

Outcomes

The following outcomes (measurement scale) were reported:

Behaviour and depression (Hospital Anxiety and Depression: HAD scale and Geriatric Depression Scale: GDS)

Follow‐up: 4 weeks

Notes

Sponsorship source: Cross‐Council New Dynamics of Ageing (NDA) Initiative

Conflicts of interest: There were no financial, personal, potential conflicts of interest in the conduct of the study or in the manuscript development. Although Philips Lighting supplied the light fitments, they had no part in the design of the protocol nor in the analysis of the data. Prof. Skene and Dr Middleton are co‐directors of Stockgrand Ltd and Prof. Skene has in the past received research grant support from Philips. Dr. Luc Schlangen is an employee of Philips Research.
Ethical approval: A favourable ethical opinion was obtained from the University of Surrey Ethics Committee and the care homes, whilst all research was carried out according to the Declaration of Helsinki 2008. Informed written consent was obtained from participants or their families where participants were unable to give consent themselves.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Described as randomised but did not specify randomisation

Allocation concealment (selection bias)

Unclear risk

Did not specify details

Blinding of participants and personnel (performance bias)
All outcomes

High risk

No blinding reported

Blinding of outcome assessment (detection bias)
All outcomes

High risk

No blinding reported

Incomplete outcome data (attrition bias)
All outcomes

High risk

69 completed study but 56 at most were analysed for included outcomes.

Selective reporting (reporting bias)

Unclear risk

Paper did not report details of published study protocol or trial registration.

Other bias

Unclear risk

Age and sex reported for each care home at baseline appeared to be different, but no statistical tests performed and no other characteristics reported. Baseline outcomes measures reported in little detail

Study characteristics

Methods

Study design: Controlled before‐after study

Number of facilities: 6 (3 intervention and 3 control)

Participants

N = 120 (Intervention: N = 57; control: N = 48)

Mean age (SD): Intervention: 86.1 (6.7), control: 87.7 (6.8)

% Female: Intervention: 67, control: 75

% Dementia: not reported

Mean (SD) number of comorbidities: not reported

Country: UK

Inclusion criteria:

• If informed assent provided (where the resident could not self‐consent), the interview only went ahead if the resident appeared happy and relaxed during the process.

Exclusion criteria:

• Living in the care home less than two months

Interventions

Type of intervention: dining

Design features: Food displayed for residents to see, fewer tables in dining room, tablecloths, flowers on table, white crockery with side plates for vegetables, drinks machine available at all times, biscuits, fruit, sandwiches and yoghurts on display available any time

Other features that differed: Increased choice at meals, increased number of hot meal options at breakfast and evening meal, choice of meal at mealtime with change of mind accommodated, use of buffet and Bain‐Marie to display options to residents, dining open for 90 minutes with several sittings of residents, visitors welcome, large variety of self‐service snacks available

Control: Limited choice at meals, only cold meal options available at breakfast and evening meal, residents make their meal selection in advance, meals at set times with single sitting, visitors rarely eat with residents, limited drinks and snacks available only on drinks trolley

Outcomes

The following outcomes (measurement scale) were reported:

Cognitive function (Mini Mental State Examination: MMSE), behaviour and depression (Hospital Anxiety and Depression: HAD scale) and adverse events (number of falls)

Follow‐up: 12 months

Notes

Funded by Norfolk City Council

Conflicts of interest: The research was funded by Norfolk County Council, and JB works for Norfolk County Council. These links did not affect the way that the data are presented. Ethical approval: Ethical approval for the study was obtained from the University of East Anglia, Faculty of Health, Ethics Committee. The trial was registered as ISRCTN86057119 (see http://www.controlled‐tri‐ als.com/ISRCTN86057119).

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Non‐randomised trial

Allocation concealment (selection bias)

High risk

Non‐randomised trial

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible

Blinding of outcome assessment (detection bias)
All outcomes

High risk

Primary outcome (falls) measured from the notes as reported by staff (who knew the allocation)

Incomplete outcome data (attrition bias)
All outcomes

High risk

Significant incomplete outcome data (which authors themselves noted)

Selective reporting (reporting bias)

Low risk

No evidence of selective reporting as all outcomes described in methods were in results

Other bias

High risk

Potential residual confounding. Significant differences in baseline characteristics between groups at baseline

Study characteristics

Methods

Study design: Repeated measures study with no control group

Number of facilities: 2 (repeated measures study)

Participants

N = 34

Mean age: 77.8

% Female: 33

% Dementia: 100

Mean (SD) number of comorbidities: not reported

Country: Canada

Inclusion criteria:

• A diagnosis of dementia

• Presence of one or more difficult to manage behaviours

• Living in a long‐term care setting

• Consent from their legal guardian to participate

• Moderate to severe dementia as indicated by Global Deterioration Scale/Functional Assessment Staging (GDS/FAST) (138, 139) scores of 5 to 7 and Mini Mental State Exam (MMSE) (139, 140) score of less than 20

• Residence on the unit for a minimum of four weeks

• A minimum of one difficult‐to‐manage behaviour such as delusions, hallucinations, agitation/aggression, depression/dysphoria, anxiety, elation/euphoria, apathy/indifference, disinhibition, irritability/lability, aberrant motor behaviour, sleep and night‐time behaviour disorders, and appetite and eating disorders

Exclusion criteria:

• Presence of intractable pain

Interventions

Type of intervention: Garden vignette

Design features: Designated area that contained clusters of gardening and nature‐related objects designed to both attract attention and encourage self‐determined interaction and exploration. Identical vignettes were established on each unit directly opposite each other but separated by a five‐foot high wall. Positioned in a highly visible, high traffic space. The vignette included all objects required to accomplish the activity of gardening: a sturdy garden centre table; soil, plastic pots, garden seeds, light plastic garden tools, and a plastic watering can; scented, colourful, edible plants; glossy gardening magazines with engaging pictures; and large artificial flowers to attract attention. When the garden vignette was in place, all residents had unobstructed exposure and access, 24 hours per day.

Outcomes

The following outcomes (measurement scale) were reported:

Behaviour (Neuropsychiatric Inventory for Nursing Homes: NPI‐NH and NPI‐NH‐Occupational Distress: NPI‐NH‐OD)

Follow‐up: Placed for 14 days then removed for 14 days and process repeated once

Notes

Sponsorship source: Canadian Nurses Foundation, Dr. Ann Beckingham Scholarship and the Alberta Registered Nurses Educational Trust Scholarship.

Conflicts of interest: Not stated.
Ethical approval: The Institutional Ethics Review Board of the University of Calgary granted ethics approval

Risk of bias

Bias

Authors' judgement

Support for judgement

Allocation concealment (selection bias)

High risk

Repeated measures study, no control group

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Lowest follow‐up for one outcome was 29/34 = 85%.

Selective reporting (reporting bias)

Low risk

All outcomes reported as stated in methods

Other bias

Low risk

Same study participants used as repeated measures study. Intervention removed during washout phases

Intervention independent of other changes (ITS)

Unclear risk

No reported compelling evidence that intervention was independent, nor that that intervention was not independent of other changes in time

Shape of the intervention effect pre‐specified (ITS)

High risk

Baseline measurement spread over 4 weeks, performed at time of admission, then intervention effect in second week of vignette rather than at the time of intervention

Intervention unlikely to affect data collection (ITS)

Low risk

Sources and methods of data collection were the same before and after the intervention.

Study characteristics

Methods

Study design: Cluster‐randomised controlled trial

Number of facilities: 4 wards in one nursing home randomised (2 Intervention wards and 2 control wards)

Participants

N enrolled: Intervention: 21, control: 17. N analysed: Intervention: 20, control: 14

Mean age (SD): Intervention: 82.6 (7.5), control: 78.2 (7)

% Female: Intervention: 66.6, control: 70

% Dementia: Not reported

Mean (SD) number of comorbidities: Intervention: 3 (1.2), control: 2.3 (1.3)

Country: The Netherlands.

Inclusion criteria:

• Resident in a nursing home from one of the four wards invited to join the study

• Older than 65 years

• Resident for more than 3 months at the start of the study

Exclusion criteria:

• Parenteral nutrition

• Terminal phase of a disease

• A specific exclusion criterion for the analyses of biochemical indicators of health was applied for the patients with severe anaemia.

Interventions

Type of intervention: Dining

Design features: Plant or flowers placed on every table and sufficient lighting. Background music chosen by the patients. Just before meals, tables were dressed up in the dining room with appropriate tablecloths and dinner plates, trays and covers removed from the table, carers out of patients’ sight. Other features that differed: Dishes served on dinner plate per course and per table, simultaneous start of the meal per table, and possibility of receiving help when necessary. Breakfast and supper served per table and at patients’ discretion: no ready‐to‐eat sandwiches. Continuous availability of coffee, tea, and soft drinks such as fruit juices outside meal periods. Rescheduling nursing staff timetable to have enough nurses at mealtime (one nurse for two patients). No walking around of the nursing staff in the dining room during meals. Medications handed out before the start of the meal to distinguish medical care and meals. No interference with patients’ meal: no questions or wishes for the next meal. Programme monitoring every trimester by both nursing staff and researchers. No cleaning activities in the dining room during meal consumption. Immediately after meals, tidying up the dining‐room for the social activities

Outcomes

The following outcomes (measurement scale) were reported:

Quality of life (Sickness Impact Profile: SIP and Dutch version of the Philadelphia Geriatric Center Moral Scale: PGCMS)

Follow‐up: 4 months, 8 months and 12 months

Notes

Sponsorship source: Not reported.

Conflicts of interest: Not stated. Ethical approval: The study protocol was approved by the ethical committee of the nursing home.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Randomisation method not stated

Allocation concealment (selection bias)

Unclear risk

No details reported

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible

Blinding of outcome assessment (detection bias)
All outcomes

Unclear risk

Not reported

Incomplete outcome data (attrition bias)
All outcomes

High risk

58% loss to follow‐up for quality of life outcomes

Selective reporting (reporting bias)

Unclear risk

Protocol or trial registration not mentioned

Other bias

High risk

Some clinical differences between groups in baseline characteristics. Statistical significance of differences not reported. Likely some contamination as wards randomised but within one nursing home

Study characteristics

Methods

Study design: Cluster‐randomised controlled trial

Number of facilities: 6 (3 intervention and 3 control)

Participants

N enrolled: Intervention: 133, control: 112. N analysed: Intervention: 95, control: 83

Mean age (SD): Intervention: 78 (11.1), control: 75 (9.9)

% Female: Intervention: 70, control: 55

% Dementia: 0

Mean (SD) number of comorbidities: Intervention: 3 (1.4), control: 3 (1.6)

Country: The Netherlands

Inclusion criteria:

• Residing in an eligible nursing home ward (medium‐sized, general population, two wards for chronic somatic disease, long‐term care, cover different parts of country, similar in organisational characteristics)

Exclusion criteria:

• Terminal phase of disease

• Needing total parenteral feeding

• Unable to give informed consent owing to a physical or mental condition

Interventions

Type of intervention: Dining

Design features: Tablecloths, normal plates and glasses, full cutlery, and flower arrangements

Other features that differed: Cooked meals served on table, greater choice at meals, residents decide when meal begins, staff sit down with residents

Control: No tablecloth, plastic cups, pre‐designed plate, divided into 3 sections, residents wear bibs. Cooked meals served individually on pre‐plated tray, residents choose meals two weeks before, staff do not sit down, no choice in meal times

Outcomes

The following outcomes (measurement scale) were reported:

Quality of life (Dutch Quality of Life of Somatic Nursing Home Residents questionnaire) and function (Nursing Home Physical Performance test)

Follow‐up: 6 months

Notes

Sponsorship source: Netherlands Organisation for Health Research and Development. Conflicts of interest: None. Ethical approval: This study was approved by the Ethical Committees of the nursing homes and the Medical Ethical Committee of Wageningen University.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Non‐random component in the sequence generation (e.g. using first letter of ward name)

Allocation concealment (selection bias)

Low risk

Randomisation at start of study

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible

Blinding of outcome assessment (detection bias)
All outcomes

Unclear risk

Allocation was blinded but did not describe details of whether the analyses were blinded

Incomplete outcome data (attrition bias)
All outcomes

High risk

Follow‐up < 80%

Selective reporting (reporting bias)

Low risk

All outcomes in methods reported in results

Other bias

High risk

Baseline differences between groups in age, gender and length of stay

Study characteristics

Methods

Study design: 2 x 2 factorial randomised trial

Number of facilities: 12 (6 intervention and 6 control)

Participants

N enrolled: intervention: 49, control: 45. N analysed: Intervention: 47, control: 40

Mean age (SD): Intervention: 85 (6), control: 85 (5)

% Female: Intervention: 91.8, control: 88.9

% Dementia: 100

Mean (SD) number of comorbidities: not reported

Country: The Netherlands

Inclusion criteria:

• Resident at one of 12 facilities who agreed to participate (assisted‐care facilities in which residents have their own apartment where they sleep and retreat, but spend most of the daytime in a common living room supervised by caregivers)

Exclusion criteria:

• Nil other

Interventions

Type of intervention: Lighting (bright light)

Design features: Light exposure was manipulated by installing a large number of ceiling‐mounted fixtures with Plexiglas diffusers containing an equal amount of Philips TLD 840 and 940 fluorescent tubes in the common living room. Lights were on daily between approximately 9 am and 6 pm. Light intensity was increased to ± 1000 lux between 10 am and 6 pm at the 6 light facilities (active condition)

Control: An equal number of fixtures were installed, but these contained only half of the tubes, accommodated concealed band‐stop filters, and were installed at a greater distance from the eyes.

Outcomes

The following outcomes (measurement scale) were reported:

Behaviour (Neuropsychiatric Inventory: NPI), depression (Cornell Scale for Depression in Dementia: CSDD), withdrawn behaviour (sub‐scale of the Multi Observational Scale for Elderly Subjects: MOSES), agitation (Cohen‐Mansfield Agitation Inventory: CMAI), positive and negative mood (Philadelphia Geriatric Centre Affect Rating Scale: PGCARS), cognitive function (Mini Mental State Examination: MMSE), and function (Nurse‐informant adaptation of the scale by Katz and colleagues)

Follow‐up: 6 weeks, 6 months, 12 months, 18 months and 24 months

Notes

Sponsorship source: Financial and material support were provided by the Netherlands Organisation for Health Research, The Hague, by grants 0028‐300‐30 and 907‐00‐012; the Netherlands Organisation for Scientific Research, The Hague, by grants 016.025.041 and 051.04.010; the Stichting De Drie Lichten, Leiden;Stichting RVVZ; Zeist by grant 01‐220; Japan Foundation for Aging and Health; Hersenstichting Nederland by grant 11F04‐2.47; Internationale Stichting Alzheimer Onderzoek by grant 05511. Philips Lighting BV, Braun, and Cambridge Neurotechnology supplied material for this study at reduced cost.

Conflicts of interest: Reported no financial disclosures
Ethical approval: The Medical Ethics Committees of Hospital De GelderseVallei, Ede, and the VU University Medical Center, Amsterdam, the Netherlands, approved the study.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Random number generator

Allocation concealment (selection bias)

Low risk

Managed by a research assistant external to the study

Blinding of participants and personnel (performance bias)
All outcomes

Low risk

Stated caregivers blinded and no significant difference when they asked caregivers to guess their facilities light status

Blinding of outcome assessment (detection bias)
All outcomes

Low risk

Assessors were blinded to allocation.

Incomplete outcome data (attrition bias)
All outcomes

High risk

High loss to follow‐up

Selective reporting (reporting bias)

Low risk

Registered on clinical trial registry ‐ all outcomes reported

Other bias

Low risk

No significant differences in baseline characteristics between groups. No differences in baseline outcome variables. Light exposure randomised by facility protecting against contamination

Study characteristics

Methods

Study design: Controlled before‐after study

Number of facilities: 26 (19 intervention and 7 control)

Participants

N enrolled: Intervention: 79, control: 132. N analysed: Intervention: 67, control: 97

Mean age (95% confidence interval): Intervention: 81.2 (79.7, 82.7), control: 83.6 (81.1, 86.1)

% Female: Intervention: 91.0, control: 73.2

% Dementia: 100

Mean (SD) number of comorbidities: not reported

Country: The Netherlands

Inclusion criteria:

• Psychogeriatric group living homes and psychogeriatric nursing homes or nursing homes with psychogeriatric units were selected.

Group living homes

• Maximum 6 residents

• Maximum 6 units

• Situated more than 200 m from the nursing home to which they belonged

• Prepared their own meals

• Built more than 2 years prior to the start of the study 

Traditional nursing homes

• Built according to the Dutch 1997 Building Regulation for Nursing Homes

• 20 residents per unit

Exclusion criteria:

• Residents not surviving to 6 months

Interventions

Type of intervention: home‐like model

Design features: Psychogeriatric group living homes. Criteria based on Concept Map that defined group living care (a) had a maximum of six residents; (b) had a maximum of six units; (c) were situated more than 200 meters from the nursing home to which they belonged; (d) prepared their own meals and (e) were built more than 2 years prior to the start of the study.

Control: Psychogeriatric nursing homes or nursing homes with psychogeriatric units. Built according to the Dutch 1997 Building Regulation for Nursing Homes, as these facilities offer, among other structural improvements, only single bedrooms. Large‐scale: more than 20 residents per unit were included in the study.

Outcomes

The following outcomes (measurement scale) were reported:

Quality of life (Dementia Quality of Life: DQoL), behaviour (Revised Memory and Behavior Problems Checklist: RMBPC and the Neuropsychiatric Inventory‐Questionnaire: NPI‐Q), cognitive function (Standardised Mini Mental State Examination: S‐MMSE), function (Interview for the Deterioration of Daily Living activities in Dementia: IDDD), social engagement (Revised Index of Social Engagement: RISE from the Resident Assessment Instrument: RAI) and physical restraints (nursing home physician or psychologist asked whether residents were prescribed one or more physical restraints)

Follow‐up: 6 months

Notes

Sponsorship source: Dutch ministry of Health Welfare and Sport, Foundation Het Zonnehuis, ActiZ organisation of care entrepreneurs

The authors had no conflicts of interests during any part of the study. Sponsors had no role in the design, collection, analysis and interpretation of the data, nor in writing the report and the decision to submit it for publication.
Ethical approval: The study was approved by the Medical Ethics Committee of the National Institute of Mental Health and Addiction.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Controlled before‐after study

Allocation concealment (selection bias)

High risk

Controlled before‐after study

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible

Blinding of outcome assessment (detection bias)
All outcomes

High risk

Blinding not feasible

Incomplete outcome data (attrition bias)
All outcomes

High risk

< 80% follow‐up

Selective reporting (reporting bias)

Low risk

All outcomes reported as stated in methods

Other bias

High risk

Potential residual confounding. Significant differences in cognition (MMSE) and depression scores at baseline

Study characteristics

Methods

Study design: Controlled before‐after study

Number of facilities: 112 (89 intervention and 23 control)

Participants

N enrolled: Intervention: 34, control: 22. N analysed: Intervention: 20, control: 13

Mean age (SD): Intervention: 83.4 (8.1), control: 81.2 (10.4)

% Female: Intervention: 76.9, control: 23.1

% Dementia: 100

Mean (SD) number of comorbidities: not reported

Country: Germany

Inclusion criteria:

• New residents of SHA and SCU with dementia in Berlin 1 July‐31 Dec 2008

• Established diagnosis dementia ‐ MMSE 24 or below

• Moving into SHA or SCU within 14 days

• Control (SCU): Admission criteria were eligibility to benefits under the long‐term care insurance scheme, a medical diagnosis of irreversible dementia and a score of less than 18 points according to the MMSE, severe behavioural problems according to the modified Cohen‐Mansfield Agitation Inventory and being able to participate in group activities and general group social life.

Exclusion criteria:

• Not reported

Interventions

Type of intervention: home‐like model

Design features: Small‐scale living shared housing arrangements were completely disconnected from traditional nursing homes. Often situated in large apartments in mostly urban settings, mostly 6‐8 people, which had typical structures of a flat with a kitchen, a living room and private bedrooms

Control: Special‐care unit for people with dementia. Admission criteria for special‐care unit for people with dementia in Berlin were eligibility to benefits under the long‐term care insurance scheme, a medical diagnosis of irreversible dementia and a score of less than 18 points according to the MMSE, severe behavioural problems according to the modified Cohen‐Mansfield Agitation Inventory and being able to participate in group activities and general group social life.

Outcomes

The following outcomes (measurement scale) were reported:

Quality of life (Dementia‐specific QUALIDEM), behaviour (Neuropsychiatric Inventory for Nursing Homes: NPI‐NH), cognitive function (Mini Mental State Examination: MMSE and Global Deterioration Scale: GDS to assess severity of dementia) and function (Barthel ADL Index)

Follow‐up: 6 months and 12 months

Notes

Sponsorship source: Grant of the German Federal Ministry of Health ‘Leuchtturmprojekt Demenz’

Conflicts of interest: Not stated
Ethical approval: The study was approved by the Medical Ethics Committee of Charite ́–Universitätsmedizin Berlin (application number EA1/109/08).

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Controlled before‐after study

Allocation concealment (selection bias)

High risk

Controlled before‐after study

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible

Blinding of outcome assessment (detection bias)
All outcomes

High risk

Blinding not feasible

Incomplete outcome data (attrition bias)
All outcomes

High risk

Large loss to follow‐up; not balanced between groups

Selective reporting (reporting bias)

Low risk

Outcomes reported as per methods

Other bias

High risk

Potential residual confounding. Significant differences in gender at baseline

Study characteristics

Methods

Study design: Controlled before‐after study

Number of facilities: 5 (3 intervention and 2 control)

Participants

N enrolled: Intervention: 41, control: 59. N analysed: Intervention: 25, control: 45

Mean age: Not reported

% Female: Not reported

% Dementia: Not reported

Mean (SD) number of comorbidities: not reported

Country: USA

Inclusion criteria:

Facilities

• Experimental: Texas nursing homes initiating the implementation process of the Eden Alternative™ model in their respective facilities

• Control: Texas nursing homes not initiating the Eden Alternative™ model agreed to participate.

Residents

• Facility social workers' assessment of the residents' cognitive ability to understand and complete questionnaires

• Agree to participate

Interventions

Name of intervention: Eden alternative

Design features: Human habitat model ‐ pets, plants and children. Imbued daily life with variety and spontaneity by creating an environment in which unexpected and unpredictable interactions and happenings can take place. Other features: Provided daily opportunities to give as well as receive care by promoting resident participation in the daily round of activities that are necessary to maintain the Human Habitat. De‐emphasised the role of prescription drugs in the residents' daily lives and committed those resources to the maintenance and growth of the Human Habitat. Leadership that placed the need to improve resident quality of life over and above the inevitable objections to change

Control: Traditional nursing home

Outcomes

The following outcomes (measurement scale) were reported:

Quality of life (Life Satisfaction Index: LSI)

Follow‐up: 6 months, 12 months and 18 months

Notes

Sponsorship source: Unclear

Conflicts of interest: Not stated
Ethical approval: Not stated

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Not randomised

Allocation concealment (selection bias)

High risk

Controlled before‐after study

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible

Blinding of outcome assessment (detection bias)
All outcomes

High risk

Blinding not feasible

Incomplete outcome data (attrition bias)
All outcomes

High risk

Loss to follow‐up 30% by second data collection point

Selective reporting (reporting bias)

Low risk

Outcomes reported as in methods

Other bias

High risk

Potential residual confounding. Some resident and staff baseline characteristics reported in Table 1, p. 15, no statistical comparisons performed. Differences in staff turnover between facilities. Difference in payer type and racial mix for Eden model. Significant difference in proportion of payers for Eden vs control (reviewer calc: Eden 301/410 vs 25/313 control; P < 0.0000001 Chi‐square Epionline). Some contamination; one control facility commenced implementing the Eden alternative and then abandoned.

Study characteristics

Methods

Study design: Controlled before‐after study

Number of facilities: 13 (9 intervention and 4 control)

Participants

N = 242 (Intervention: N = 93; control: N = 149)

Mean age (SD): Intervention: 87.2 (7.2), control: 85.8 (9.7)

% Female: Intervention: 73.1, control: 73.9

% Dementia: Intervention: 55.9, control: 50.0

Mean (SD) number of comorbidities: Intervention: 1.9 (1.2), control: 2.3 (1.4)

Country: USA

Inclusion criteria:

• Residing in included nursing home for at least six months

Exclusion criteria:

• Admitted for short‐term rehab or hospice at the start of their stay

Interventions

Type of intervention: home‐like model

Name of intervention: Green House model

Design features: Home‐like environment, Cluster of 2 or 3 homes with 10 residents each, private bedroom and bathroom, large common living and dining room, no nurses stations, medication carts, paging systems

Other features that differed: Organisational changes to support resident quality of life, care staff have diverse roles and greater autonomy and responsibility in daily care.

Control: Traditional large scale nursing homes, hospital‐like features such as nurses stations, medication charts, paging systems. Traditional hierarchical organisational structure and traditional care staff roles

Outcomes

The following outcomes (measurement scale) were reported:

Social engagement (Index of Social Engagement: ISE), depressive symptoms (Mood Scale Score: MSS), function (ADL long‐form scale) and cognitive function (Cognitive Performance Scale: CPS)

Follow‐up: Up to 18 months

Notes

Study partially supported by a grant from the Robert Wood Johnson Foundation (Grant 66360; PI: SDH) and the Clinical and Translational Science Award Program, through the NIH National Center for Advancing Translational Sciences (grant UL1TR000427; BJB)

Conflicts of interest: None

Ethical approval: Not stated

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Controlled before‐after study

Allocation concealment (selection bias)

High risk

Controlled before‐after study

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible

Blinding of outcome assessment (detection bias)
All outcomes

High risk

Blinding not feasible

Incomplete outcome data (attrition bias)
All outcomes

High risk

High loss to follow‐up (62%)

Selective reporting (reporting bias)

Unclear risk

Many outcomes on MDS so unclear how outcomes were decided

Other bias

High risk

Potential residual confounding. Significant difference in comorbidities at baseline