Supporting data for "Generalizable machine learning models for rapid antimicrobial resistance prediction in unseen healthcare settings"

The deployment of machine learning in clinical settings is often hindered by the limited generalizability of the models. Models that perform well during development tend to underperform in new environments, limiting their clinical utility. This issue affects models designed for the rapid identification of antimicrobial resistance, which is essential to guide treatment decisions. Traditional susceptibility tests can take up to three days, whereas integrating MALDI-TOF mass spectrometry with machine learning has the potential to reduce this to one day. However, model performance declines drastically in hospitals or time frames outside the training data. In this study, we use the publicly available DRIAMS dataset-comprising MALDI-TOF mass spectrometry profiles collected across multiple institutions in the Swiss healthcare system-to investigate strategies for improving the generalizability of machine-learning methods for rapid AMR prediction.