The Princeton University Department of Operations Research and Financial Engineering (ORFE) and ODH, Inc., a health technology company providing data aggregation and analytics solutions, have entered into a joint research project to develop new machine learning techniques that will enable health plans to assess and prioritize the mental and social health factors underlying their members’ conditions, and propose appropriate interventions.
In the health care setting, machine learning provides systems with the ability to scan huge volumes of data and use pattern recognition to make predictions about future outcomes.
Data scientists, researchers and mathematicians from ODH and ORFE will collaborate to develop a methodology that identifies hidden patterns from complex medical information using data-driven techniques. For example, today, a patient whose symptoms include jaundice, swollen abdomen and nausea may receive a primary diagnosis of liver failure. The technique to be developed by the researchers will identify additional, secondary factors which may be overlooked by the patient’s physicians and health plan, such as the patient’s alcohol addiction and related social challenges, which are likely to trigger a visit to a hospital, emergency room or other high-cost setting.
Additionally, the technique will prioritize these secondary factors, enabling care coordinators to identify which ones are likely to respond most immediately to intervention. The technique will also recommend to care coordinators the most effective interventions—such as referring patients to substance abuse treatment programs, educating them about the importance of medication adherence, or arranging transportation to a provider’s office—with the goal of improving outcomes.
“The healthcare industry is just starting to come to grips with the potential of machine learning,” Michael Jarjour, president and CEO of ODH, Inc., said. “We see a huge opportunity to push the boundaries by improving machine learning methodologies so that we can better identify underlying behavioral and social factors contributing to individuals’ health conditions and target interventions accordingly. We are thrilled to collaborate with Princeton on this venture. The University’s expertise in artificial intelligence and data science will be a key driver of our research.”
“The next generation of machine learning technologies for healthcare must confront unique technical challenges arising from this domain,” Samory K. Kpotufe, assistant professor of operations research and financial engineering, Princeton University, said in a statement. “Predictive tools have to be simple enough to yield interpretable health recommendations but must also achieve high accuracy. So, there’s a tension between accuracy and interpretability that has to be addressed, along with other practical constraints, such as, computational tractability, environmental and seasonal changes, and patients’ privacy. Given the expertise of ODH in health care technology, we are looking forward to a fruitful collaboration. Confronting the challenges of machine learning for healthcare is bound to generate a rich set of research questions with potential impact beyond the healthcare domain.”