Reframing the evaluation of integrated care; examples from the NHS in England.

There is global interest in integrated care, often associated with how to improve system efficiency, strengthen clinical and cost-effectiveness, avoid gaps in patient care, and improve patient experiences and outcomes, through improved coordination across services. Despite considerable activity in both delivering and evaluating integrated care, evaluations have not greatly helped to understand how to 'do' it better. Evaluations of integrated care have often arrived at similar conclusions, frequently including the generic finding that results are patchy and context dependent. In this article, we explore and discuss these challenges to evaluation, how these challenges are understood in recent key publications, and suggest an alternative perspective. We explore technical inadequacies of evaluations (concerning definitions, metrics, and timing) as well as deeper problems (such as integrated care being dynamic and relational, and operating across multiple, larger systems). In re-framing how to evaluate integrated care, we propose an approach that involves a recursive evaluation architecture. This draws on systems thinking. This approach also recognises that we can better understand evaluations of integrated care as co-producing knowledge and applying this to learning and adaptation.