Table 1.

Definitions of context, mechanism and outcome

Context: Influences whether a mechanism is triggered or not. It may include macroeconomic conditions, cultural practices and interpersonal relations. Pawson [16] suggests that understanding context can include a focus on ‘(i) the individual players, (ii) their interrelationships, (iii) the institutional location and (iv) the surrounding infrastructure.’

Mechanism: Refers to what it is about social prescribing connector roles that cause outcomes. Mechanisms tend to be unobservable, ‘embodied in the subjects’ reasoning…’ [16]. Most usually in health services research, they can be conceptualised as a response (e.g. fear, reputation management, feeling valued, needing to appear competent) to resources provided by an intervention.

Outcome: From a realist perspective, ‘variations in programme performance are a crucial first step but outcome patterns considered alone are only surface “markers” or “traces”… the potential outward signals of inner workings of a programme in a particular manifestation’ [16]. Hence, we were not so much interested in the percentage or degree to which link workers/care navigators had worked (or not) but in how different outcomes were produced under different contexts.

Context-mechanism-outcome configuration (CMOC): This is the way in which causal explanations are presented in realist reviews. They are propositions that explain how an outcome is caused (O) ‘because of the action of some underlying mechanisms (M), which only comes into operation in particular contexts (C)’ [16]. CMOCs should be phrased in a manner that is testable.