Dr. Marina Sirota, Dr. Sergio Baranzini and colleagues have used machine learning to predict Alzheimer’s disease up to 7 years before symptoms appear. Finding that high cholesterol and, for women, osteoporosis, were both predictive of risk, the work demonstrates the promise of combining AI, knowledge graphs and clinical data to ID genetic markers and other determinants of risk.

Machine learning (ML) to predict treatment response often gives only a single estimate of average treatment effect, without quantifying uncertainty based on errors in the model or individual variation. Dr. Ahmed Alaa and colleagues here introduce a new method that can provide valid inference of treatment effects for individual patients, and a reliable uncertainty estimate, using any ML model.

Many factors related to medication switching are only captured in unstructured clinical notes and can be difficult to extract. Dr. Irene Chen and colleagues evaluated the abilities of GPT-4, with no previous training, to identify reasons women changed contraceptives, demonstrating that GPT-4 could accurately extract reasons for switching.

Dr. Ziad Obermeyer discovers that by training an algorithm to listen to patients—not doctors—researchers can diagnose and fix racial biases built into medical knowledge.

Dr. Ahmed Alaa et al develop a unified evaluation protocol, the echocardiographic task adaptation benchmark (ETAB). Echocardiography is one of the most commonly used non-invasive imaging techniques in Cardiology. ETAB is a benchmarking framework for developing deep learning models of echocardiograms using publicly available data sets.

Dr. Irene Chen, et al frame ethics of ML in healthcare through the lens of social justice. They describe ongoing efforts and outline challenges in a proposed pipeline of ethical ML in health, ranging from problem selection to post deployment considerations.

Dr. Adam Yala et al developed Tempo, a reinforcement-learning based framework for deriving personalized screening policies. Tempo policies were significantly more efficient than existing clinical guidelines across diverse test sets from four hospitals.