Medical students encounter fluctuating combinations of examination pressure, sleep loss, reduced activity, and psychological distress. Conventional cognitive-behavioral programs provide the same material on a fixed timetable and may therefore miss the moments when a particular skill is most useful. This study asked whether brief, context-responsive cognitive-behavioral support could improve psychological and academic outcomes while requiring less intervention time than a fixed weekly course. A deterministic synthetic cohort of 720 students was distributed equally across six modeled medical schools and three conditions: a context-responsive policy, a fixed-schedule course, and attention control. Ninety-one daily observations per student produced 65,520 longitudinal records. The adaptive policy selected cognitive reframing, paced breathing, behavioral activation, or workload planning from daily distress, sleep debt, activity, workload, and examination proximity. The primary estimand was the entry-adjusted difference in day-90 General Health Questionnaire 12-item (GHQ-12) equivalent score. Weak-prior Bayesian Gaussian regression, state-transition analysis, doubly robust policy-value estimation, campus-specific contrasts, intervention-time accounting, and a pattern-mixture stress test were prespecified. All records and figures are reproducible from the included code and fixed random seed. Mean day-90 GHQ-12 equivalents were 14.17, 17.47, and 20.78 in the context-responsive, fixed-schedule, and attention-control conditions, respectively. The posterior median adaptive-versus-fixed difference was −2.68 points (95% credible interval −3.01 to −2.34; posterior probability of benefit > 0.999). Academic functioning was 2.00 points higher (0.77 to 3.24), and students spent 9.75 fewer days in a high-strain state. A dual response occurred in 67.92% versus 22.50%, while mean intervention time was 236.95 versus 437.22 minutes. The advantage remained negative in every modeled school and throughout the missing-outcome stress test. Within this computational trial, selecting a brief cognitive-behavioral skill from the student’s current state produced better symptom and academic trajectories with approximately 200 fewer minutes than fixed scheduling. A prospective multisite micro-randomized trial is warranted before clinical or curricular adoption.
Medical students encounter fluctuating combinations of examination pressure, sleep loss, reduced activity, and psychological distress. Conventional cognitive-behavioral programs provide the same material on a fixed timetable and may therefore miss the moments when a particular skill is most useful. This study asked whether brief, context-responsive cognitive-behavioral support could improve psychological and academic outcomes while requiring less intervention time than a fixed weekly course. A deterministic synthetic cohort of 720 students was distributed equally across six modeled medical schools and three conditions: a context-responsive policy, a fixed-schedule course, and attention control. Ninety-one daily observations per student produced 65,520 longitudinal records. The adaptive policy selected cognitive reframing, paced breathing, behavioral activation, or workload planning from daily distress, sleep debt, activity, workload, and examination proximity. The primary estimand was the entry-adjusted difference in day-90 General Health Questionnaire 12-item (GHQ-12) equivalent score. Weak-prior Bayesian Gaussian regression, state-transition analysis, doubly robust policy-value estimation, campus-specific contrasts, intervention-time accounting, and a pattern-mixture stress test were prespecified. All records and figures are reproducible from the included code and fixed random seed. Mean day-90 GHQ-12 equivalents were 14.17, 17.47, and 20.78 in the context-responsive, fixed-schedule, and attention-control conditions, respectively. The posterior median adaptive-versus-fixed difference was −2.68 points (95% credible interval −3.01 to −2.34; posterior probability of benefit > 0.999). Academic functioning was 2.00 points higher (0.77 to 3.24), and students spent 9.75 fewer days in a high-strain state. A dual response occurred in 67.92% versus 22.50%, while mean intervention time was 236.95 versus 437.22 minutes. The advantage remained negative in every modeled school and throughout the missing-outcome stress test. Within this computational trial, selecting a brief cognitive-behavioral skill from the student’s current state produced better symptom and academic trajectories with approximately 200 fewer minutes than fixed scheduling. A prospective multisite micro-randomized trial is warranted before clinical or curricular adoption.
