Research: AI instrument may help spot Kind 2 diabetes traits within the U.S.


The work examining risk factors down to the county level could lead to tailored public health strategies

BUFFALO, NY – A new study from the University of Buffalo reports the benefits of using artificial intelligence to better understand type 2 diabetes in the United States.

The study describes how machine learning – a subset of AI where computers act intelligently without being explicitly programmed – can help study the prevalence of the disease, which affects more than 34 million Americans, and future trends detect.

The work was led by Zia Ahmed, a senior scientist and associate research professor at the UB RENEW Institute. It was published in Nature’s Scientific Reports on March 26th.

Ahmed says the prevalence of type 2 diabetes varies significantly in the United States due to wide-ranging socio-economic and lifestyle risk factors.

The study was based on data published in the Centers for Disease Control and Prevention (CDC) US Diabetes Monitoring System and the CDC’s Behavioral Risk Factor Surveillance System. Additional data such as six risk factors – access to higher education, poverty, obesity, physical inactivity, access to exercise areas like public parks, and access to healthy food – come from the US Census Bureau’s population estimation program.

The machine learning program used by the research team – a geographically weighted random forest model – outperforms existing methods, according to Ahmed.

A better understanding of the differences in these risk factors could help in intervention and treatment approaches to reduce or prevent type 2 diabetes, says Ahmed. He adds that from a policy perspective, the study results could lead to tailored and more effective prevention strategies, which is crucial given the projected increase in diabetes.

“Zia has extensive training, research and oversight experience in geospatial science in the fields of agriculture, health and the environment. His current research interest is explainable artificial intelligence (XAI) to explore the spatial heterogeneity of the local contribution to prediction. His knowledge and skills in advanced data techniques and machine learning impacts various areas of focus at RENEW, including environmental, genome and health, ”said Amit Goyal, SUNY Distinguished Professor and founding director of UB’s RENEW Institute. “Zia is an excellent mentor and works well with students at all levels. In addition to the students listed here, he has worked with many other students since his time at UB RENEW. “

Ahmed has over 20 years of environmental modeling and data analysis experience. The areas of expertise include data mining; geographic information systems, remote / proximal detection and geostatistics; linear / non-linear model, mixed effects model, multivariate statistics and machine learning; and database management.