Faculty Spotlight: Dr. James Henderson - Improving healthcare through the power of data

Data is more than just numbers. When used effectively, it can be a powerful tool for improving healthcare for both patients and the people who provide it. Researchers like Dr. James Henderson, assistant research scientist in the Division of General Medicine, use statistics and machine learning to inform and evaluate interventions that enhance healthcare delivery, support clinicians, and identify opportunities to deliver high-value care. 

Dr. Henderson’s research focuses on innovations in healthcare organization, particularly how health systems can leverage data from electronic health records (EHR) to deliver quality care while reducing unnecessary burden on clinicians. 

“My goal is to fill evidence gaps and find out what works, for who, and when, so patients get quality, evidence-based care without unnecessary interventions. I also want to ensure that everyone providing care, from physicians and advanced practice providers, to nurses, medical assistants, and non-clinical staff, find their work meaningful and can maintain a healthy work-life balance.”

Finding purpose through data and discovery

Dr. Henderson’s career as a statistician was driven by a desire to understand how data can explain the world around us. Describing himself as “deeply curious,” he finds fulfillment in tackling difficult problems such as the ones in health services research and healthcare delivery design. 

A defining moment in his career came when Dr. Henderson joined the Michigan Program on Value Enhancement (MPrOVE), a transitional research unit dedicated to improving the value of healthcare at Michigan Medicine and Beyond. Supported by the Institute for Healthcare Policy and Innovation and the Clinical Quality Department, MPrOVE introduced him to a community or researchers committed to turning evidence into action. “I caught the bug, so to speak, but I’m glad I did.”

Turning data into actionable improvements

In his research, Dr. Henderson calls on large-scale data sources like EHR and administrative datasets to answer important clinical questions. These resources provide valuable real-world evidence about how healthcare is delivered in everyday settings and allow researchers to examine trends at a population level, whether national or regional through insurance claims or local using Michigan Medicine ER data. 

“Because these datasets are routinely used to support care delivery, billing, and payment, they allow us to answer questions retrospectively,” he said.”At the same time, those same characteristics create challenges and limitations that researchers need to consider.”

One example was when Dr. Henderson and other collaborators used a combination of administrative data, including insurance claims, to estimate how routine blood tests used to assess Vitamin D levels changed across the U.S. for privately insured patients and ones through Veterans Health Administration. This was following a national campaign to avoid this practice. They were also able to compare the impact of the campaign to a more direct policy shift in Ontario, Canada. 

However, as healthcare continues to evolve and new technologies, treatments, and models of care present opportunities to improve patient outcomes, there can also be challenges that require thoughtful implementation. In another study, Dr. Henderson and his colleagues used local EHR data to assess the effectiveness of health system-based weight management treatments for patients with obesity. Their findings showed these management treatments are effective, when used, at promoting weight loss among patients with obesity, but they were used infrequently so the data presented them as having a minimal impact at a population level.

More recently, Dr. Henderson has studied how healthcare systems can effectively organize care for people with obesity and weight-related conditions around the use of highly effective GLP-1 receptor agonists and making sure they are used appropriately. In collaboration with Drs. Dina Griauzde, Cassie Turner, and other colleagues, he compared the efficacy of these medications when prescribed through the dedicated weight navigation program at Michigan Medicine, by people who have specialty training in obesity medicine, to those who were prescribed elsewhere. 

They found these medications were more effective in a setting where there is additional support for population health management and more time to have in-depth conversations with patients about treatment goals, expectations, and long-term management than a primary care visit can provide.

“I think the challenge is how we incorporate popular and highly effective medications like this into care,” said Dr. Henderson.“ How do we design not just the prescribing process, but all the things that go around it to make sure they stay effective. It’s especially important with something like weight loss medication because obesity lies at the root cause of a lot of preventable illnesses and morbidity. If we get this right, we can really impact the health of our patients and our communities.” 

Beyond evaluating clinical interventions, Dr. Henderson has also focused on translating data-driven insights into operational improvements within Michigan Medicine. Working with the MPrOVE team, he led the development and implementation of a machine learning model designed to predict the appropriateness of outpatient surgical cases for ambulatory surgical centers.

Implementing the model substantially reduced provider review burden while maintaining high rates of appropriate scheduling at ambulatory surgery centers. Today, it has enabled a reorganization of screening workflows so clinician review is targeted to cases where it provides the greatest value and is used within the health system to assess thousands of outpatient surgical cases each year.

Using variation in care to solve complex challenges

Some of Dr. Henderson’s work examines variation in care, or the differences in how clinicians and health teams diagnose, treat, and manage the same medical conditions. One area that particularly excites him is the use of generative AI and related models.

“I’m super excited about how these models have taught us to represent data through foundational models in ways that are useful for answering questions we don’t know how to ask yet,” he said. “I think this will greatly enhance our ability to provide context for understanding care decisions and help reduce unwarranted variation, while also allowing us to learn about what works best for patients.”

Working together to improve care

For Dr. Henderson, the most rewarding part of his work is collaborating with colleagues across the University of Michigan Health system who share a commitment to advancing healthcare. From close collaborators in General Medicine, to others in the department and across the university, he is inspired to help them by getting to the right question, choosing the right approach, and leveraging data to produce meaningful insights while recognizing what is possible and where gaps remain.

“Every day I get to engage with smart, dedicated people who are trying to make healthcare better,” he said. “Getting to do this work alongside them is what continues to inspire me.”

See more of Dr. Henderson’s research.