What contributes to good patient outcomes in the home treatment of the severely mentally ill: study protocol of a multi-centre analysis

BACKGROUND: While evidence is available that home treatment could be effective for treating severe mental illness, there is a lack of evidence on what exactly makes home treatment effective. The study presented here aims to develop recommendations for structures and processes in home treatment that are necessary for its effectiveness. METHODS/DESIGN: 14 provider networks of home treatment for severe mental illness will be analyzed and compared according to their structures, processes and patient-related outcomes. Data will be drawn from health care claims data, routine assessments of psychosocial functioning, and from questionnaires on structures and processes. The primary outcome will be psychosocial functioning; secondary outcomes, quality of life and days spent in hospital. The relation between structures and processes on one hand side and outcomes on the other side will be identified by multilevel analysis. In addition, focus groups with patients, relatives and network staff will be held to add further insight into relevant processes. All networks will receive individual quality reports, providing them with feedback on the results of this research and benchmarking them against the average. Based on this research, recommendations for processes and structures of home treatment will be developed. DISCUSSION: The research will use longitudinal data on outcomes routinely assessed since 2009 and claims data. Routine data is also used for the assessment of structures and processes. By way of additional questionnaires developed in discussion with providers, further relevant factors can be included. The approach of this study becomes more comprehensive by conducting focus groups with patients, relatives and providers and by having the chance to evaluate the results with the networks by providing feedback of results. Several factors such as outcomes related to regional availability of hospital beds or size of networks might limit this study.