BackgroundThe Multispecialty Community Provider (MCP) model was introduced to the NHS as a primary care-led, community-based integrated care model to provide better quality, experience and value for local populations.ObjectivesThe three main objectives were to (1) articulate the underlying programme theories for the MCP model of care; (2) identify sources of theoretical, empirical and practice evidence to test the programme theories; and (3) explain how mechanisms used in different contexts contribute to outcomes and process variables.DesignThere were three main phases: (1) identification of programme theories from logic models of MCP vanguards, prioritising key theories for investigation; (2) appraisal, extraction and analysis of evidence against a best-fit framework; and (3) realist reviews of prioritised theory components and maps of remaining theory components.Main outcome measuresThe quadruple aim outcomes addressed population health, cost-effectiveness, patient experience and staff experience.Data sourcesSearches of electronic databases with forward- and backward-citation tracking, identifying research-based evidence and practice-derived evidence.Review methodsA realist synthesis was used to identify, test and refine the following programme theory components: (1) community-based, co-ordinated care is more accessible; (2) place-based contracting and payment systems incentivise shared accountability; and (3) fostering relational behaviours builds resilience within communities.ResultsDelivery of a MCP model requires professional and service user engagement, which is dependent on building trust and empowerment. These are generated if values and incentives for new ways of working are aligned and there are opportunities for training and development. Together, these can facilitate accountability at the individual, community and system levels. The evidence base relating to these theory components was, for the most part, limited by initiatives that are relatively new or not formally evaluated. Support for the programme theory components varies, with moderate support for enhanced primary care and community involvement in care, and relatively weak support for new contracting models.Strengths and limitationsThe project benefited from a close relationship with national and local MCP leads, reflecting the value of the proximity of the research team to decision-makers. Our use of logic models to identify theories of change could present a relatively static position for what is a dynamic programme of change.ConclusionsMultispecialty Community Providers can be described as complex adaptive systems (CASs) and, as such, connectivity, feedback loops, system learning and adaptation of CASs play a critical role in their design. Implementation can be further reinforced by paying attention to contextual factors that influence behaviour change, in order to support more integrated working.Future workA set of evidence-derived ‘key ingredients’ has been compiled to inform the design and delivery of future iterations of population health-based models of care. Suggested priorities for future research include the impact of enhanced primary care on the workforce, the effects of longer-term contracts on sustainability and capacity, the conditions needed for successful continuous improvement and learning, the role of carers in patient empowerment and how community participation might contribute to community resilience.Study registrationThis study is registered as PROSPERO CRD42016039552.FundingThe National Institute for Health Research Health Services and Delivery Research programme.