Effects of 2- vs 4-Week Attending Physician Inpatient Rotations on Unplanned Patient Revisits, Evaluations by Trainees, and Attending Physician Burnout:  A Randomized Trial

IQR indicates interquartile range.^aFour rotations were added to attending physicians' prespecified rotation sequences after the study began and therefore were excluded.^bSeventeen discharges were carried out by an attending physician not assigned to the rotation.

Effect of rotation duration within attending physician, adjusting for ward team, crossover period, and the relatedness of grouped measures (details of mixed-effects models are shown in the eFigure). The P value for 30-day unplanned revisits is a test of inferiority vs noninferiority; P value for length of stay is a test of superiority. Outcomes with P <.003, a criterion for statistical significance that reflects a post hoc multiplicity adjustment for the 17 outcomes depicted here and in Figure 3 and Figure 4, are less likely to be chance findings. Error bars indicate 2-sided 95% CIs. The shaded area represents the noninferiority zone for the primary outcome, which is bounded only in the direction that favors 4-week rotations by a prespecified noninferiority margin (Δ, equal to 1.13).

Effect of rotation duration within attending physician, adjusting for ward team, crossover period, and the relatedness of grouped measures (details of mixed-effects models are shown in the eFigure). P values refer to superiority tests. Outcomes with P <.003, a criterion for statistical significance that reflects a post hoc multiplicity adjustment for the 17 outcomes depicted here and in Figure 2 and Figure 4, are less likely to be chance findings. Error bars indicate 2-sided 95% CIs. Percentages may not add to 100% because of rounding error. See eTable 1 and eTable 2 for the items that compose domains. Among 153 house staff and 253 medical students assigned to study rotations, 147 house staff (96%) and 229 medical students (91%) submitted a median of 6 (range, 1-11) and 2 (range, 1-5) evaluations; submitted 89% and 92% of all possible evaluations for 2-week study rotations and 88% and 89% of all possible evaluations for 4-week study rotations; and evaluated attending physicians a median of 12 (range, 4-30) times and 9 (range, 1-19) times, respectively. Two (3%) of 62 attending physicians were not evaluated by house staff. Medical students spent a median of 7 days (interquartile range [IQR], 4-14 days) on 2-week and 17 days (IQR, 7-21 days) on 4-week rotations. Missing domain scores (which occurred in 0.3%-0.7% and 0.2%-7.5% of domains from house staff and medical student evaluations, respectively) were not included in the denominators of score category percentages.

Effect of rotation duration within attending physician, adjusting for ward team, crossover period, and the relatedness of grouped measures (details of mixed-effects models are shown in the eFigure). P values refer to superiority tests. Outcomes with P <.003, a criterion for statistical significance that reflects a post hoc multiplicity adjustment for the 17 outcomes depicted here and in Figure 2 and Figure 3, are less likely to be chance findings. Error bars indicate 2-sided 95% CIs. Percentages may not add to 100% because of rounding error. See eTable 3 for the items that compose facets. A severity assessment from a 2-week rotation was missing from one attending physician who had completed 5 other assessments after 2-week rotations. Low, intermediate, and high overall raw mean scores were 5.8, 18.3, and 40.8 for emotional exhaustion; 1.0, 2.0, and 3.2 for the single-item summary of burnout; 14.6, 19.8, and 26.3 for inadequate workplace control; and 1.9, 5.6, and 9.0 for perceived stress, respectively.