Researcher granted more than $2 million to develop statistical models for analyzing data on physical activity

Indiana University School of Public Health-Bloomington (SPH-B) Associate Professor Roger Zoh—along with colleagues—are seeking answers to the following questions: What is the effect of physical activity or inactivity on the risk of developing chronic diseases and mortality when the available data might be error-prone? Can physical activity recommendations be more tailored; and if so, what factors should be used to obtain these recommendations?

To help address these concerns, Dr. Zoh was recently awarded a four-year National Institutes of Health (NIH) R01 Research Project Grant in the amount of $2,382,456. The funding will support development of statistical models to better estimate the impact of imprecisely measured physical activity and other modifiable behaviors on the risk of developing diabetes and other chronic diseases—as well as their impact on overall individual survival.

Dr. Zoh will collaborate with faculty from various institutions, including SPH-B, Columbia University, and the REGARDS (Reasons for Geographic and Racial Differences in Stroke) study at the University of Alabama at Birmingham, all connections which Dr. Zoh credits SPH-B Dean David Allison for facilitating.

Dr. Zoh has also worked with SPH-B Associate Professor Carmen Tekwe on a global outreach initiative to address challenges related to the collection and analysis of wearable devices or objective measures of physical activity in lower- or middle-income environments. He says that suboptimal analysis can result in misleading conclusions or a poor understanding of how physical activity and other modifiable behavior can truly benefit individuals with—or at risk of—developing chronic diseases such as diabetes and cardiovascular diseases.

"We have a saying in statistics: 'Garbage in, garbage out,'" says Dr. Zoh. "If the incoming data on physical activity and health outcomes is prone to error, or not as reliable as we think it should be, the conclusions we make will be potentially overly optimistic or too pessimistic."

An important aspect to consider in the creation of these statistical models is that individuals suffering from chronic diseases, such as type 2 diabetes, may have more restrictions when it comes to physical activity, making the accuracy of the measurement even more crucial. Unreliable or incorrect data can result from inconsistent usage of wearable devices, technological glitches, and devices being improperly calibrated to the individual at the time of wearing.

"We can make better-informed recommendations on physical activity regimens without individuals having to overdo or underdo it to improve their health status," says Dr. Zoh. "It is critical that we understand, as best as we can, the effect of physical activity on health and disease status."

For more stories about SPH-B students, faculty and staff making an impact locally and worldwide, visit  go.iu.edu/48bx.