Generalizability of predictive models for intensive care unit patients
This work addresses the generalizability of clinical predictive models for ICU patients, highlighting the importance of multi-center data sharing, though it is incremental in nature.
The study evaluated predictive models for ICU patients using the eICU database, finding that cross-validation across multiple centers accurately estimates performance in new hospitals and that a single multi-center model transfers well compared to hospital-specific retraining.
A large volume of research has considered the creation of predictive models for clinical data; however, much existing literature reports results using only a single source of data. In this work, we evaluate the performance of models trained on the publicly-available eICU Collaborative Research Database. We show that cross-validation using many distinct centers provides a reasonable estimate of model performance in new centers. We further show that a single model trained across centers transfers well to distinct hospitals, even compared to a model retrained using hospital-specific data. Our results motivate the use of multi-center datasets for model development and highlight the need for data sharing among hospitals to maximize model performance.