Michael Pencina

Observability and its impact on differential bias for clinical prediction models

2022 · Journal of the American Medical Informatics Association

OBJECTIVE: Electronic health records have incomplete capture of patient outcomes. We consider the case when observability is differential across a predictor. Including such a predictor (sensitive variable) can lead to algorithmic bias, potentially exacerbating health …

All models are local: time to replace external validation with recurrent local validation

2023 · arXiv (Cornell University)

External validation is often recommended to ensure the generalizability of ML models. However, it neither guarantees generalizability nor equates to a model's clinical usefulness (the ultimate goal of any clinical decision-support tool). External validation is …