Why Do Probabilistic Clinical Models Fail To Transport Between Sites?
This addresses a critical problem for healthcare AI practitioners by highlighting transportability issues, though it is incremental as it builds on existing concerns without presenting new empirical results.
The paper investigates why probabilistic clinical models often fail to generalize across different healthcare sites, identifying controllable and inherent sources of failure, and proposes a solution to isolate site-specific practices from disease patterns.
The rising popularity of artificial intelligence in healthcare is highlighting the problem that a computational model achieving super-human clinical performance at its training sites may perform substantially worse at new sites. In this perspective, we present common sources for this failure to transport, which we divide into sources under the control of the experimenter and sources inherent to the clinical data-generating process. Of the inherent sources we look a little deeper into site-specific clinical practices that can affect the data distribution, and propose a potential solution intended to isolate the imprint of those practices on the data from the patterns of disease cause and effect that are the usual target of probabilistic clinical models.