February 8, 2019
Students and faculty members at the UNC Gillings School of Global Public Health have proposed a new statistical method for developing tailored treatment data using data from mobile devices.
Study co-authors include Daniel Luckett, PhD, postdoctoral research associate in biostatistics; Anna Kahkoska, doctoral student in nutrition; Elizabeth Mayer-Davis, PhD, Cary C. Boshamer Distinguished Professor of nutrition and medicine; and Michael Kosorok, PhD, W.R. Kenan Jr. Distinguished Professor and chair of biostatistics.
Their paper, “Estimating Dynamic Treatment Regimes in Mobile Health Using V-Learning,” was published online Dec. 12, 2018, in the Journal of the American Statistical Association.
The researchers hope to continue their work by putting their algorithm into a mobile app that eventually can help people with Type 1 diabetes live healthier lives.
By using mobile health, or mHealth, technologies, patient status can be monitored in real time, and data produced by the mobile technologies have great potential to inform personalized health care decisions.
Potentially, mHealth may improve the lives of people with Type 1 diabetes, who must make decisions throughout the day about when to eat, exercise or take insulin, actions which interact in complex ways to affect a patient’s blood glucose levels. Recent technological advances have produced devices such as continuous glucose monitors, which track a patient’s blood glucose in real time, and accelerometers, which monitor a patient’s physical activity.
The Gillings School team set out to develop a statistical method that could be used in conjunction with data from continuous glucose monitors and accelerometers to help people with Type 1 diabetes better manage their disease. Using data from a previous study, they were able to show that their new method could learn strategies to help patients maintain stable glucose levels over time.
“We’re very excited about Daniel’s [Luckett’s] work,” said Mayer-Davis. “Development and application of methods to improve our understanding of variability across individuals and how to use this information to improve individual patient outcomes is a critical need in many chronic conditions.
Mayer-Davis said managing Type 1 diabetes requires maintaining healthy blood glucose concentrations all day, every day. That need compelled the type of novel approach Luckett devised.
“This is a fabulous example of how precision medicine approaches can be applied to subpopulations to improve health,” Mayer-Davis said. “In other words, this is an example of how a vision of precision public health can be realized.”
Contact the Gillings School of Global Public Health communications team at firstname.lastname@example.org.