From programme theory to logic models for multispecialty community providers: a realist evidence synthesis

Background: The NHS policy of constructing multispecialty community providers (MCPs) rests on a complex set of assumptions about how health systems can replace hospital use with enhanced primary care for people with complex, chronic or multiple health problems, while contributing savings to health-care budgets. Objectives: To use policy-makers’ assumptions to elicit an initial programme theory (IPT) of how MCPs can achieve their outcomes and to compare this with published secondary evidence and revise the programme theory accordingly. Design: Realist synthesis with a three-stage method: (1) for policy documents, elicit the IPT underlying the MCP policy, (2) review and synthesise secondary evidence relevant to those assumptions and (3) compare the programme theory with the secondary evidence and, when necessary, reformulate the programme theory in a more evidence-based way. Data sources: Systematic searches and data extraction using (1) the Health Management Information Consortium (HMIC) database for policy statements and (2) topically appropriate databases, including MEDLINE, MEDLINE In-Process & Other Non-Indexed Citations, PsycINFO, the Cumulative Index to Nursing and Allied Health Literature (CINAHL) and Applied Social Sciences Index and Abstracts (ASSIA). A total of 1319 titles and abstracts were reviewed in two rounds and 116 were selected for full-text data extraction. We extracted data using a formal data extraction tool and synthesised them using a framework reflecting the main policy assumptions. Results: The IPT of MCPs contained 28 interconnected context–mechanism–outcome relationships. Few policy statements specified what contexts the policy mechanisms required. We found strong evidence supporting the IPT assumptions concerning organisational culture, interorganisational network management, multidisciplinary teams (MDTs), the uses and effects of health information technology (HIT) in MCP-like settings, planned referral networks, care planning for individ