Exploring the effect of case management in homelessness per components: A systematic review of effectiveness and implementation, with meta‐analysis and thematic synthesis

BACKGROUND: Adequate housing is a basic human right. The many millions of people experiencing homelessness (PEH) have a lower life expectancy and more physical and mental health problems. Practical and effective interventions to provide appropriate housing are a public health priority. OBJECTIVES: To summarise the best available evidence relating to the components of case‐management interventions for PEH via a mixed methods review that explored both the effectiveness of interventions and factors that may influence its impact. SEARCH METHODS: We searched 10 bibliographic databases from 1990 to March 2021. We also included studies from Campbell Collaboration Evidence and Gap Maps and searched 28 web sites. Reference lists of included papers and systematic reviews were examined and experts contacted for additional studies. SELECTION CRITERIA: We included all randomised and non‐randomised study designs exploring case management interventions where a comparison group was used. The primary outcome of interest was homelessness. Secondary outcomes included health, wellbeing, employment and costs. We also included all studies where data were collected on views and experiences that may impact on implementation. DATA COLLECTION AND ANALYSIS: We assessed risk of bias using tools developed by the Campbell Collaboration. We conducted meta‐analyses of the intervention studies where possible and carried out a framework synthesis of a set of implementation studies identified by purposive sampling to represent the most ‘rich’ and ‘thick’ data. MAIN RESULTS: We included 64 intervention studies and 41 implementation studies. The evidence base was dominated by studies from the USA and Canada. Participants were largely (though not exclusively) people who were literally homeless, that is, living on the streets or in shelters, and who had additional support needs. Many studies were assessed as having a medium or high risk of bias. However, there was some consistency in outcomes across studies that improved confidence in the main findings. CASE MANAGEMENT AND HOUSING OUTCOMES: Case management of any description was superior to usual care for homelessness outcomes (standardised mean difference [SMD] = −0.51 [95% confidence interval [CI]: −0.71, −0.30]; p < 0.01). For studies included in the meta‐analyses, Housing First had the largest observed impact, followed by Assertive Community Treatment, Critical Time Intervention and Intensive Case Management. The only statistically significant difference was between Housing First and Intensive Case Management (SMD = −0.6 [–1.1, −0.1]; p = 0.03) at ≥12 months. There was not enough evidence to compare the above approaches with standard case management within the meta‐analyses. A narrative comparison across all studies was inconclusive, though suggestive of a trend in favour of more intensive approaches. CASE MANAGEMENT AND MENTAL HEALTH OUTCOMES: The overall evidence suggested that case management of any description was not more or less effective compared to usual care for an individual's mental health (SMD = 0.02 [−0.15, 0.18]; p = 0.817). CASE MANAGEMENT AND OTHER OUTCOMES: Based on meta‐analyses, case management was superior to usual care for capability and wellbeing outcomes up to 1 year (an improvement of around one‐third of an SMD; p < 0.01) but was not statistically significantly different for substance use outcomes, physical health, and employment. CASE MANAGEMENT COMPONENTS: For homelessness outcomes, there was a non‐significant trend for benefits to be greater in the medium term (≤3 years) compared to long term (>3 years) (SMD = −0.64 [−1.04, −0.24] vs. −0.27 [−0.53, 0]; p = 0.16) and for in‐person meetings in comparison to mixed (in‐person and remote) approaches (SMD = −0.73 [−1.25,−0.21]) versus −0.26 [−0.5,−0.02]; p = 0.13). There was no evidence from meta‐analyses to suggest that an individual case manager led to better outcomes then a team, and interventions with no dedicated case manager may have better outcomes than those with a named case manager (SMD = −0.36 [−0.55, −0.18] vs. −1.00 [−2.00, 0.00]; p = 0.02). There was not enough evidence from meta‐analysis to assess whether the case manager should have a professional qualification, or if frequency of contact, case manager availability or conditionality (barriers due to conditions attached to service provision) influenced outcomes. However, the main theme from implementation studies concerned barriers where conditions were attached to services. CHARACTERISTICS OF PERSONS EXPERIENCING HOMELESSNESS: No conclusions could be drawn from meta‐analysis other than a trend for greater reductions in homelessness for persons with high complexity of need (two or more support needs in addition to homelessness) as compared to those with medium complexity of need (one additional support need); effect sizes were SMD = −0.61 [−0.91, −0.31] versus −0.36 [−0.68, −0.05]; p = 0.3. THE BROADER CONTEXT OF DELIVERY OF CASE MANAGEMENT PROGRAMMES: Other major themes from the implementation studies included the importance of interagency partnership; provision for non‐housing support and training needs of PEH (such as independent living skills), intensive community support following the move to new housing; emotional support and training needs of case managers; and an emphasis on housing safety, security and choice. COST EFFECTIVENESS: The 12 studies with cost data provided contrasting results and no clear conclusions. Some case management costs may be largely off‐set by reductions in the use of other services. Cost estimates from three North American studies were $45–52 for each additional day housed. AUTHORS' CONCLUSIONS: Case management interventions improve housing outcomes for PEH with one or more additional support needs, with more intense interventions leading to greater benefits. Those with greater support needs may gain greater benefit. There is also evidence for improvements to capabilities and wellbeing. Current approaches do not appear to lead to mental health benefits. In terms of case management components, there is evidence in support of a team approach and in‐person meetings and, from the implementation evidence, that conditions associated with service provision should be minimised. The approach within Housing First could explain the finding that overall benefits may be greater than for other types of case management. Four of its principles were identified as key themes within the implementation studies: No conditionality, offer choice, provide an individualised approach and support community building. Recommendations for further research include an expansion of the research base outside North America and further exploration of case management components and intervention cost‐effectiveness.