How feedback works in practice: Insights from a rapid realist review

Feedback is essential in health professions education, yet its diversity makes it difficult to study and theorize beyond specific instances. Building on their prior theoretical framing of feedback, the authors undertook a focused pattern informed realist exploration of how it might work, for whom, and under what circumstances, in order to develop realist programme theories.

A pattern informed rapid realist review of empirical studies on feedback was conducted using a recurring feedback pattern syntax as a guiding analytical lens. Through process mapping and context-mechanism-outcome configuration (CMOC) analysis, a convenience sample of articles describing feedback interventions across five modalities was reviewed. Causal explanations were refined into a series of realist programme theories.

The authors reviewed 35 feedback interventions described in the literature. Analysis generated 34 CMOCs, and from these, 10 programme theories and 2 provisional program theories were developed in terms of patterns from the earlier study (patterns capitalised for clarity). The SENSOR pattern was essential across all interventions but varied in form and influence. OUTCOMES were shaped less by SENSOR type and more by their interplay with JUDGEMENT, RECURSION and INTERPRETATION. Although FREQUENCY and RECURSION were often conflated, recursive cycles and longitudinal facilitation were more consistently associated with performance change than frequency alone. ASSESSMENT conditioned how feedback was experienced and acted upon, sometimes enabling learning but at other times constraining it.

This review developed and refined 10 programme theories along with 2 provisional programme theories. The findings suggest that feedback effectiveness arises from the interaction of elements, such as SENSORS, JUDGEMENT, RECURSION and FREQUENCY within specific contexts, rather than from any single feedback feature alone. The methods used in this study provide practical approaches for analysing heterogeneous feedback interventions, including the use of process mapping and PERFORMANCE MEASUREMENT-SENSOR-JUDGEMENT-ASSEMSSMENT (PmSJA) syntax, and offer some insight into where feedback cycles succeed, fail, and may be targeted for change.