RESUMO
INTRODUCTION: Assessment in simulation is gaining importance, as are scenario design methods increasing opportunity for assessment. We present our approach to improving measurement in complex scenarios using PARTS [Phase-Augmented Research and Training Scenarios], essentially separating cases into clearly delineated phases. METHODS: We created 7 PARTS with real-time rating instruments and tested these in 63 cases during 4 weeks of simulation. Reliability was tested by comparing real-time rating with postsimulation video-based rating using the same instrument. Validity was tested by comparing preintervention and postintervention total results, by examining the difference in improvement when focusing on the phase-specific results addressed by the intervention, and further explored by trying to demonstrate the discrete improvement expected from proficiency in the rare occurrence of leader inclusive behavior. RESULTS: Intraclass correlations [3,1] between real-time and postsimulation ratings were 0.951 (95% confidence interval [CI], 0.794-0.990), 1.00 (95% CI, --to--), 0.948 (95% CI, 0.783-0.989), and 0.995 (95% CI, 0.977-0.999) for 3 phase-specific scores and total scenario score, respectively. Paired t tests of prelecture-postlecture performance showed an improvement of 14.26% (bias-corrected and accelerated bootstrap [BCa] 95% CI, 4.71-23.82; P = 0.009) for total performance but of 28.57% (BCa 95% CI, 13.84-43.30; P = 0.002) for performance in the respective phase. The correlation of total scenario performance with leader inclusiveness was not significant (rs = 0.228; BCa 95% CI. -0.082 to 0.520; P = 0.119) but significant for specific phase performance (rs = 0.392; BCa 95% CI, 0.118-0.632; P = 0.006). CONCLUSIONS: The PARTS allowed for improved reliability and validity of measurements in complex scenarios.