Collaboration On
AI-powered Patient Safety Research Flags Drug Side Effects
A recent capstone project by an interdisciplinary team of graduate students from Carnegie Mellon University and the University of Pittsburgh School of Medicine employed artificial intelligence to examine health care data for potentially dangerous drug interactions.
The students mined association rules in medication data using a method often applied to e-commerce data. Instead of suggesting similar products for customers to buy together, they generated the rules from the data to guide the discovery of different medication combinations that may lead to adverse outcomes, said Rema Padman, Trustees Professor of Management Science and Healthcare Informatics in the Heinz College and faculty advisor of the capstone project.
Alan Scheller-Wolf, Richard M. Cyert Professor of Operations Management at the Tepper School, and Ari Lightman, digital media and marketing professor at Heinz College, facilitated access to this data at the University of Pittsburgh. Scheller-Wolf and Lightman are both faculty leads at the Initiative for Patient Safety Research (IPSR), which was established to address the persistent problem of medical errors, the third leading cause of death in the U.S.
Their involvement was instrumental in enabling the capstone team to identify and analyze adverse drug events, reactions, or medication errors in millions of records from more than 600,000 patients collected over three years.
“The dream is to solve these really big, complicated problems in society, like adverse medical events, but you’re not going to have one person solve it, or even one university,” said Scheller-Wolf. “We need a team, and collaborating across Oakland with Pitt has a huge advantage.”
The capstone project team, consisting of four Carnegie Mellon master’s degree students, used AI and machine learning to review databases at the University of Pittsburgh.
Using association rule mining, a method typically applied to e-commerce data, the students identified medication combinations that might lead to adverse outcomes. Padman noted that these methods help sift through vast amounts of data to extract useful information.
The students also worked with Richard D. Boyce, associate professor of biomedical Informatics at the University of Pittsburgh, to secure access to the data, including Pitt’s Medication Error Avoidance at Regional Scale (MEARs) database and the U.S. Food and Drug Administration’s Adverse Event Reporting System database (FAERS).
The capstone team conducted their research through an open-source approach using a highly secure virtual workbench and data science tools. They demonstrated their approach by focusing on patients taking colchicine, used to treat gout, and looking for interactions with other medications. They identified frequent pairings, such as colchicine with metoprolol, a high blood pressure medication that could worsen gout.
Scheller-Wolf noted that optimizing business solutions could incentivize health care companies to adopt these methods.
“It’s a big problem, but if we could do something about it even marginally, we could make life better for a lot of people,” he said. “We could save a whole slew of money that we could then use to do something good rather than trying to undo something bad, so that’s a big deal.”
Despite the fragmented nature of the health care industry, research like this can help build momentum for investment in patient safety.
The collaboration between the Tepper School of Business, the Heinz College, and the University of Pittsburgh exemplifies an interdisciplinary approach to tackle complex health care challenges. By combining expertise in operations management, digital media, data analytics, and biomedical informatics, Carnegie Mellon encourages students to develop innovative solutions to critical patient safety issues. This project highlights the importance of teamwork across different fields, demonstrating how collective efforts can lead to significant advancements in health care technology and patient safety.
