How do residents learn? The development of practice styles in a residency program

Purpose.-To determine if resident interactions that occur during rotations may influence resident learning and help explain practice styles in a residency program. Method.-We distributed clinical vignette surveys to all pediatric residents in a teaching hospital, eliciting practice styles for childhood fever without source and asthma, as well as resident demographics and learning styles. We measured residents' propensities to order tests for fever, empirically treat fever, diagnose asthma as serious, and aggressively treat asthma. For any 2 residents, we examined if similarity in propensities was associated with the number of months the residents had rotated together in the past year, controlling for demographics and learning styles. We also examined if the gender mix of each resident pair modified the association. Results.-Fifty-two of 62 residents (84%) responded. Practices varied across all vignettes and varied as much within as across training years. In adjusted regressions, 2 residents were 1.22 times (P = .03) and 1.30 times (P = .004) more likely to have similar propensities in fever testing and asthma diagnosis for each additional month they rotated together. By gender mix, associations were stronger for opposite-sex pairs than for same-sex pairs. Conclusions.-Practice variations emerged in a residency program within a single academic year. Practice styles were more similar among residents who worked together more frequently. However, the association was not uniform across gender combinations, suggesting the importance of interpersonal dynamics in learning. This finding may have major implications for practice variations, medical education, and quality improvement.