Machine Learning, Real-World Data Find New Uses for Existing Drugs

Machine learning and real-world data could help researchers identify new uses for existing drugs, leading to accelerated drug repurposing, according to a study published in Nature Machine Intelligence.

Drug repurposing is an effective strategy to find new purposes for existing drugs, offering the quickest transition from research to clinical care. Developers have used this method to find new purposes for Botox injections – a medication first used to treat crossed eyes but that now treats migraines and can reduce the appearance of wrinkles.

Drug repurposing could lower the risk associated with safety testing of new medications and dramatically reduce the time it takes to get a drug to market. However, discovering new uses for existing medications requires researchers to conduct time-consuming and expensive randomized controlled trials to make sure that a drug that’s effective for one disorder will be useful to treat other conditions.

Researchers set out to develop a machine learning framework that combines large patient care-related datasets with high-powered computation to arrive at repurposed drug candidates and the estimated effects of those existing medications on a defined set of outcomes.

Machine learning methods can easily account for hundreds, or thousands, of human differences within a large population that could influence how medicine works in the body, the team noted. These factors, ranging from sex, age, and race, disease severity, and the presence of other illnesses, function as parameters in the machine learning algorithm on which the framework is based.  

“Real-world data has so many factors. This is the reason we have to introduce the deep learning algorithm, which can handle multiple parameters," said senior author Ping Zhang, assistant professor of computer science and engineering and biomedical informatics at Ohio State.

"If we have hundreds or thousands of confounders, no human being can work with that. So we have to use artificial intelligence to solve the problem. We are the first team to introduce use of the deep learning algorithm to handle the real-world data, control for multiple confounders, and emulate clinical trials.”

The group used insurance claims data on nearly 1.2 million heart disease patients, which provided information on their assigned treatment, disease outcomes, and various values for potential confounders.

The machine learning algorithm can also take into account the passage of time in each patient’s experience, including every visit, prescription, and diagnostic test. The model input for drugs is based on their active ingredients.

Using causal inference, researchers categorized the active drug and placebo patient groups that would be found in a clinical trial.  The model tracked patients for two years, and compared their disease status at that end point to whether or not they took medications, which drugs they took, and when they started the regimen.  

"With causal inference, we can address the problem of having multiple treatments. We don't answer whether drug A or drug B works for this disease or not, but figure out which treatment will have the better performance," said Zhang, who leads the Artificial Intelligence in Medicine Lab and is a core faculty member in the Translational Data Analytics Institute at Ohio State.

The group hypothesized that the model would identify drugs that could lower the risk for heart failure and stroke in coronary heart disease patients.

The model yielded nine drugs considered likely to provide those therapeutic benefits, three of which are currently in use. This means that the analysis identified six candidates for drug repurposing.

The analysis indicated that a diabetes medication called metformin, as well as a drug called escitalopram that’s used to treat depression and anxiety, could lower the risk of heart failure and stroke in the model patient population.

The results show the potential for machine learning and AI tools to accelerate drug repurposing, saving researchers both time and money.

"This work shows how artificial intelligence can be used to 'test' a drug on a patient, and speed up hypothesis generation and potentially speed up a clinical trial. But we will never replace the physician - drug decisions will always be made by clinicians,” said Zhang.

While this study focuses on proposed repurposing of drugs to prevent heart failure and stroke in patients with coronary artery disease, researchers said that the framework could be applied to other conditions as well.

"My motivation is applying this, along with other experts, to find drugs for diseases without any current treatment. This is very flexible, and we can adjust case-by-case," Zhang said. "The general model could be applied to any disease if you can define the disease outcome.”