05-13-2025
Right now, the frequency with which healthcare providers see patients to treat chronic diseases like diabetes doesn’t take into account key risk factors associated with the disease. Patients aren't sorted by risk level when deciding how often they should get care. Instead, doctors mostly decide based on a patient's current health, and different doctors have different opinions. This leads to uneven care. Some low-risk patients get many visits, while some high-risk patients barely get any. And patients with more disadvantages (lower income, less education, etc.) are often at higher risk but don't always get healthcare attention as often as they should.
This creates a big gap in healthcare access.
To fix this, Han Ye of Lehigh University; Ujjal Kumar Mukherjee of the University of Illinois Urbana-Champaign; and I suggest using a predictive system that better matches care to patient risk. Our new Journal of Operations Management research article shows that using patient socio-economic data, including things like income and education levels along with medical information, can help predict who is at higher risk for diabetes.
We developed a way for clinics to use these predictions to decide how to best allocate their limited resources, like doctor appointments, to the patients who need them most. This approach aims to provide more optimal use of healthcare resources. Our paper suggests that by using this data-driven approach to allocate healthcare encounters, clinics could potentially improve the overall health of their patient population with diabetes.
The research was conducted in close collaboration with the analytics team of a clinic. The providers, nurses and clinic management were concerned about how the population-level diabetes risk could be reduced, specifically for high-risk communities.
Implementing this approach would require operational changes in clinics. To improve diabetes care, hospitals would begin by predicting each patient’s future risk using models and estimating the expected health benefits — measured in Quality Adjusted Life Years (QALYs) — of providing care. Patients would then be ranked based on these expected benefits, and limited resources, such as consultation slots, would be allocated to those with the highest predicted gains. This ensures high-risk patients receive priority care, helping reduce overall diabetes risk in the population.
Dilip Chhajed is a clinical full professor of management in the Daniels School of Business, executive director of the IBE program, and associate dean for master’s and online programs. His areas of expertise are operations and supply chain management, with particular focus on health care operations and new product development.
