Can artificial intelligence find better drug treatments?

To provide optimal patient care, doctors would have to comb through enough data to fill six million pages of text. Critical information can go undetected, leading to such dire consequences as adverse drug reactions. A Columbia startup, Droice, is using artificial intelligence to help doctors make sense of clinical data to prescribe better treatments.

Adverse drug reactions are the fourth-leading cause of death in the United States, costing close to $140 billion per year. Each drug performs differently depending on a patient’s age, sex, genetic profile, and co-conditions that include diseases, susceptibilities, and allergies. Doctors must also consider a multitude of relevant scientific literature, clinical trials, and FDA guidelines.

Droice predicts the performance of a treatment by combining medical research with an analysis of how that treatment has performed on millions of patients in the past. It then predicts how the same treatment will perform on a new patient. Droice analyzes possible treatment combinations for several common conditions, including diabetes, asthma, and heart disease, which together are responsible for 80 percent of all US prescriptions. In a single glance, doctors can see predictions for each drug treatment, backed by trusted scientific papers.