Complex interventions have components which can vary in different contexts. Using the Realistic Evaluation framework, this study investigates how a complex health services intervention led to developments in shared care for people with long-term mental illness.
A retrospective qualitative interview study was carried out alongside a randomised controlled trial. The multi-faceted intervention supported by facilitators aimed to develop systems for shared care. The study was set in London. Participants included 46 practitioners and managers from 12 participating primary health care teams and their associated community mental health teams. Interviews focussed on how and why out comes were achieved, and were analysed using a framework incorporating context and intervening mechanisms.
Thirty-one interviews were completed to create 12 case studies. The enquiry highlighted the importance of the catalysing, doing and reviewing functions of the facilitation process. Other facets of the intervention were less dominant.
The intervention catalysed the allocation of link workers and liaison arrangements in nearly all practices. Case discussions between link workers and GPs improved individual care as well as helping link workers become part of the primary care team; but sustained integration into the team depended both on flexibility and experience of the link worker, and upon selection of relevant patients for the case discussions. The doing function of facilitators included advice and, at times, manpower, to help introduce successful systems for reviewing care, however time spent developing IT systems was rarely productive. The reviewing function of the intervention was weak and sometimes failed to solve problems in the development of liaison or recall.
Case discussions and improved liaison at times of crisis, rather than for proactive recall, were the key functions of shared care contributing to the success of Mental Health Link. This multifaceted intervention had most impact through catalysing and doing, whereas the reviewing function of the facilitation was weak, and other components were seen as less important.
Realistic Evaluation provided a useful theoretical framework for this process evaluation, by allowing a specific focus on context. Although complex interventions might appear 'out of control', due to their varied manifestation in different situations, context sensitive process evaluations can help identify the intervention's key functions.