Nonlinear Mixed Effect Modeling of Patient-Reported Outcomes to Improve Treatment Effect Estimation in an Oncology Clinical Trial

NLME modeling of EORTC QLQ-C30 GHS/QoL in the phase III coBRIM melanoma trial found significantly less longitudinal deterioration with cobimetinib plus vemurafenib (P = 0.036).

• Why it matters: PRO endpoints can show high variability that can obscure treatment effects, and conventional analyses may miss clinically meaningful differences over time.
• What they did: Using participant-level data from coBRIM (202 participants; 1,443 assessments), the study compared observed-score and MMRM approaches with pharmacology-informed NLME modeling that quantified between- and within-patient variability.
• The result: NLME also estimated significantly higher model-estimated GHS/QoL at Cycle 8 Day 1 (P = 0.049) and simulated a 10.4-point between-group difference (95% interval, -6.2 to 26.9), suggesting clinically important effects despite uncertainty.