How clinical outcomes reshape physician intent: a recursive extension of the Theory of Planned Behaviour using repertory grid technique

Background: The Theory of Planned Behaviour models behavioural intent as a function of attitude, social norms, and perceived behavioural control. Outcomes are treated as consequences of behaviour rather than inputs to subsequent decisions, leaving the theory unable to account for how prior outcomes reshape future intent. This limitation is particularly salient in physician work, where outcomes are both consequential and recurrent. This study examines whether clinical outcomes influence subsequent physician behavioural intent, identifies the perceived behavioural control facet through which this influence operates, and formalises the resulting recursive structure. Methods: This study combined qualitative evidence with formal modelling. Repertory Grid Technique interviews were conducted with 26 physicians across five clinical specialities at a Norwegian university hospital. Using triadic elicitation, 213 personal constructs were generated from evaluations of clinical cases and classified into 33 inductively derived categories, then mapped to TPB components using normalised variance analysis. Semi-structured interviews explored patterns of outcome attribution. The empirically supported feedback structure was formalised as a recursive extension of the Theory of Planned Behaviour, from which testable dynamic propositions were derived. Results: Constructs related to external perceived behavioural control (resource availability, time, continuity, cooperation, resource utilisation) collectively accounted for 26.3% of all constructs. Internal perceived behavioural control constructs (interpretation, experience, full responsibility, competency level, responsibility, confidence) accounted for 16.4%, yielding an external-to-internal ratio of 1.6:1. In semi-structured interviews, physicians predominantly attributed adverse outcomes to external factors and reported they would do “more of the same” if constraints were removed, indicating intact self-efficacy. From this empirically supported structure, the formal model predicts that asymmetric outcome updating produces a self-reinforcing degradation cycle under resource constraints, with degradation occurring faster than recovery. Conclusions: This study proposes a recursive extension of the Theory of Planned Behaviour, in which outcome feedback primarily operates through external perceived behavioural control under resource-constrained conditions. The framework generates testable predictions about intent dynamics and provides a basis for intervention design that targets perceived system capacity rather than individual confidence. Keywords: Theory of Planned Behaviour, perceived behavioural control, physician behaviour, outcome feedback, repertory grid technique, resource constraints, self-efficacy, learned helplessness