Pharmacometrics Golems: Exposure-Response Models in Oncology.

Pharmacometric models are prone to different types of bias, which can confound the analysis and challenge the credibility of causal inference. This holds particularly true for exposure-response analysis with time-to-event end points. With ever wider use of pharmacometric models and increased recognition of confounding factors, advanced methods addressing these biases are being developed and increasingly utilized. Herein, we provide a perspective highlighting the limitations introduced by the biases and considerations for future applications.