Introduction Workplace-based assessment in competency-based medical education employs entrustment-supervision scales to suggest trainee competence. However, clinical supervision involves many factors and entrustment decision-making likely reflects more than trainee competence. We do not fully understand how a supervisor’s impression of trainee competence is reflected in their provision of clinical support. We must better understand this relationship to know whether documenting level of supervision truly reflects trainee competence.

Methods We undertook a collective case study of supervisor-trainee dyads consisting of attending internal medicine physicians and senior residents working on clinical teaching unit inpatient wards. We conducted field observations of typical daily activities and semi-structured interviews. Data was analysed within each dyad and compared across dyads to identify supervisory behaviours, what triggered the behaviours, and how they related to judgments of trainee competence.

Results Ten attending physician-senior resident dyads participated in the study. We identified eight distinct supervisory behaviours. The behaviours were enacted in response to trainee and non-trainee factors. Supervisory behaviours corresponded with varying assessments of trainee competence, even within a dyad. A change in the attending’s judgment of the resident’s competence did not always correspond with a change in subsequent observable supervisory behaviours.

Discussion There was no consistent relationship between a trigger for supervision, the judgment of trainee competence, and subsequent supervisory behaviour. This has direct implications for entrustment assessments tying competence to supervisory behaviours, because supervision is complex. Workplace-based assessments that capture narrative data including the rationale for supervisory behaviours may lead to deeper insights than numeric entrustment ratings.