Learning and DiSentangling Patient Static Information from Time-series Electronic HEalth Record (STEER)
This work addresses privacy and fairness issues in healthcare machine learning by revealing and mitigating the unintended encoding of sensitive patient information, which is crucial for protecting patient data and ensuring equitable algorithmic outcomes.
The study systematically demonstrated that time-series electronic health record data and learned model representations can predict patient static information with high accuracy, achieving AUC scores up to 0.851 for biological sex, 0.869 for binarized age, and 0.810 for self-reported race, and proposed a variational autoencoder-based method to disentangle sensitive attributes to address privacy and fairness concerns.
Recent work in machine learning for healthcare has raised concerns about patient privacy and algorithmic fairness. For example, previous work has shown that patient self-reported race can be predicted from medical data that does not explicitly contain racial information. However, the extent of data identification is unknown, and we lack ways to develop models whose outcomes are minimally affected by such information. Here we systematically investigated the ability of time-series electronic health record data to predict patient static information. We found that not only the raw time-series data, but also learned representations from machine learning models, can be trained to predict a variety of static information with area under the receiver operating characteristic curve as high as 0.851 for biological sex, 0.869 for binarized age and 0.810 for self-reported race. Such high predictive performance can be extended to a wide range of comorbidity factors and exists even when the model was trained for different tasks, using different cohorts, using different model architectures and databases. Given the privacy and fairness concerns these findings pose, we develop a variational autoencoder-based approach that learns a structured latent space to disentangle patient-sensitive attributes from time-series data. Our work thoroughly investigates the ability of machine learning models to encode patient static information from time-series electronic health records and introduces a general approach to protect patient-sensitive attribute information for downstream tasks.