Method

Introduction

Focus of the review

There is a large body of researc

Method

Study selection and data sources

Studies for this review were identified in the following way. First, searches were carried out of the five relevant academic databases: Web of Science, Social Care Online, Google Scholar, PubMed and EBSCO. The searches used a combination of key terms adjusted where necessary for individual databases (see Appendix 1for details). The relevant studies from this stage, together with other papers and reports expressing opinions about clustered housing, were then used as the basis for citation searches using Web of Science to identify other papers citing those discovered. Finally the references used in these papers were studied to locate any studies not already identified.

The criteria for inclusion in the review were as follows:

  • Papers were included if published in English
  • Papers could use either quantitative or qualitative methodology
  • Papers had to compare in some sense dispersed community housing and clustered housing (whether described as such or not, and so included studies of services which comprise a number of living units forming a separate community from the surrounding community, such as village or intentional communities, as well as those where several dwellings are grouped together)
  • Papers had to consider a population of residents who were disabled (including people with an enduring physical, sensory, mental health or intellectual impairment)
  • Papers had to have evaluated outcomes for residents and/or costs of service provision
  • Papers were included if published after 1990

19 papers met the above criteria and were included in the review (see Appendix 2).

Assessment and classification of studies

The methodological adequacy of studies included was assessed and reported using the standard classification developed by Khan et al (2001) at the Centre for Reviews and Dissemination at the University of York (see Box 3).

In addition, the evidence regarding quality and outcomes from the included studies was classified using the comprehensive range of quality of life outcomes identified by an international expert panel (Schalock et al., 2002) (see Box 1)

Evidence about costs was assessed against the three criteria set out in the European study Deinstitutionalisation and Community Living: Outcomes and Costs (Mansell et al., 2007b) - comprehensiveness, 'like-for-like' comparison and long term perspective (see Box 2)

Box 3 CRD Classification of methodology

Level

Description

1 Experimental studies

A study in which some conditions, particularly decisions concerning the allocation of participants to different intervention groups, are under the control of the investigator

2 Quasi-experimental studies

A study in which the allocation of participants to different intervention groups is controlled by the investigator but the method falls short of genuine randomisation and allocation concealment

3 Controlled observational studies

A study in which natural variation in interventions among study participants is investigated to explore the effect of the interventions on outcomes

3a Cohort studies

Comparison of outcomes between participants who have received an intervention and a group that has not

3b Case control studies

Comparison of interventions between participants with the outcome (cases) and those without the outcome (controls)

4 Observational studies without control groups

Cross-sectional studies - examination of the relationship between outcomes and other variables of interest as they exist in a defined population at one particular time; Before-and-after studies - comparison of findings in study participants before and after an intervention; Case series - description of a number of cases of an intervention and outcome (without comparison with a control group)

5 Expert opinion

Based on eg desk research or consensus

Data extraction

For the 19 papers included in the review data were extracted and recorded into a table according to authors, date of publication, publication title and title of study, country of research, affiliated institution, design, details of participants (including type of disability, number of participants in whole study and type of accommodation), measures used in the research, costing methodology (where applicable), reliability and validity measures (where used or reported), CRD evidence level, statistics used, discussion of statistical power, outcomes in each quality of life domain (emotional well-being, interpersonal relations, material well-being, personal development, physical wellbeing, self determination, social inclusion and rights), other findings, costs and finally methodological issues. Of the papers considered, six commented on costs (one of which considered costs and no outcome domains), though these six papers only related to three studies. Twelve papers looked at more than one of the above outcome domains.

Data presentation

Findings are presented at two levels. First of all, information about the papers and the studies they represent are presented at a general level, including a mapping of the quality of life domains and costs covered by each study. Secondly, each study included is presented in detail including a description of the participants, methods, the findings and the strengths, weakness and CRD quality rating for each study. For each study a summary table is provided to illustrate whether the findings highlight advantages for campus/clustered settings, dispersed housing or where no differences are found between the two settings.

In the presentation of individual studies, where inferential analysis has been conducted, differences between settings have been included where they are statistically significant. Tests of statistical significance estimate the likelihood that differences found between groups are real differences rather than being due to chance. In some cases authors present non-significant differences as representing a trend towards a particular result. These have not been included in the review on the basis that differences which do not reach statistical significance cannot be assumed to be other than due to chance. Where inferential analysis has not been conducted differences between settings as identified by the authors of the paper are reported.

Higher rates of psychotropic medication use have been presented in the summary tables as a disadvantage. Although medication may be appropriately given, it is arguable that better psychological and physical health would not require as much medication and there is evidence of overuse of psychotropic medication in residential services for people with intellectual disabilities (Ahmed et al., 2000, Brylewski and Duggan, 2004).