Predictive medicine

Machine learning applied to electronic medical records can be used to create personalized lab-test reference ranges and to quantify disease risk, which will pave the way for precision medicine in clinical care.

A comprehensive overview of medical AI devices approved by the US Food and Drug Administration sheds new light on limitations of the evaluation process that can mask vulnerabilities of devices when they are deployed on patients.

The field of rheumatology is poised to categorize the phenotypes of systemic autoimmune diseases on the basis of measurable and quantifiable molecular signatures. Emerging efforts to identify similarities across diseases, predict clinical outcomes and predict response to therapy using quantitative, data-driven approaches could considerably change treatment paradigms.

Many widely used health algorithms have been shown to encode and reinforce racial health inequities, prioritizing the needs of white patients over those of patients of color. Because automated systems are becoming so crucial to access to health, researchers in the field of artificial intelligence must become actively anti-racist. Here we list some concrete steps to enable anti-racist practices in medical research and practice.

Researchers have developed an in silico (computer) platform that couples tissue adaptation with cellular and molecular interactions to simulate bone adaptation to mechanical loading and progress and treatment of metabolic bone diseases. What is the benefit of such in silico tools, and how can credibility of the simulation outcomes be established?

Testing drug safety in people who are pregnant remains a wicked problem, but in the transition toward big data and machine learning, target trials could afford a viable alternative to randomized, controlled trials.