Tag
This paper introduces a framework to audit clinical prediction models by distinguishing between learner inefficiencies and measurement limitations, validated on multiple cohorts to guide whether to improve models or data collection.