Our overall predictive ability was moderate, but was remarkably consistent across settings, datasets, and patient subsets (check out the supplemental analyses, there are a lot!)
Surprisingly, our ability to predict growth faltering was similar even in healthy controls (7/10)