Assessing Social Determinants-Related Performance Bias of Machine Learning Models: A case of Hyperchloremia Prediction in ICU Population
This work addresses fairness and bias in medical ML models for ICU patients, highlighting the need for bias adjustment and subgroup reporting to mitigate health disparities.
The study evaluated four machine learning classifiers for predicting hyperchloremia in ICU patients and found that adding social determinants features improved overall model performance, but subgroup testing revealed significant performance disparities in 40 out of 44 model-subgroup comparisons.
Machine learning in medicine leverages the wealth of healthcare data to extract knowledge, facilitate clinical decision-making, and ultimately improve care delivery. However, ML models trained on datasets that lack demographic diversity could yield suboptimal performance when applied to the underrepresented populations (e.g. ethnic minorities, lower social-economic status), thus perpetuating health disparity. In this study, we evaluated four classifiers built to predict Hyperchloremia - a condition that often results from aggressive fluids administration in the ICU population - and compared their performance in racial, gender, and insurance subgroups. We observed that adding social determinants features in addition to the lab-based ones improved model performance on all patients. The subgroup testing yielded significantly different AUC scores in 40 out of the 44 model-subgroup, suggesting disparities when applying ML models to social determinants subgroups. We urge future researchers to design models that proactively adjust for potential biases and include subgroup reporting in their studies.