07 Jun ADA26: Machine Learning Predicts Type 2 Diabetes Risk Across 3 Million Adults: New Findings from Kaiser Permanente
MedicalResearch.com Interview with:
Dr. Rodriguez
Luis A. Rodriguez, PhD, MPH, RD
Research Scientist, Kaiser Permanente Northern California Division of Research
Assistant Professor, Department of Health System Sciences
Kaiser Permanente Bernard J. Tyson School of Medicine
Assistant Adjunct Professor, Department of Epidemiology & Biostatistics
University of California, San Francisco
ADA 2026 Poster Presentation: Machine-Learning Modeling for T2DM Prediction in over 3 Million Adults
American Diabetes Association 85th Scientific Sessions, June 2026
MedicalResearch.com: What is the background for this study? What are the risk factors used to develop the prediction model?
Response: Type 2 diabetes develops gradually over many years, often without clear warning signs. As a result, it can be difficult for health systems to identify which adults are most likely to benefit from prevention efforts before the disease develops.
In this study, we used electronic health record data from more than 3 million adults in Kaiser Permanente Northern California to develop a prediction model that estimates an individual’s risk of developing type 2 diabetes over 1, 3, and 10 years.
The model is based on information routinely collected during clinical care, including age, sex, race/ethnicity, body mass index, blood glucose levels, smoking, physical activity, medical and family history, and medication use. By combining these clinical, biological and behavioral factors, the model provides a more comprehensive assessment of diabetes risk than traditional screening approaches.
MedicalResearch.com: What are the main findings?
Response: We found that the model was highly accurate in identifying individuals at risk for developing type 2 diabetes across short-, medium-, and long-term time horizons. It showed strong performance in both discrimination and calibration, meaning it reliably distinguished between higher- and lower-risk individuals and closely matched predicted risk with observed outcomes.
Importantly, the model identified many high-risk individuals who might otherwise be missed using traditional screening methods — particularly younger adults or those without markedly elevated blood sugar levels. This highlights the value of using a broader set of routinely collected health data to improve early risk detection.
MedicalResearch.com: What should readers take away from your report?
Response: The key takeaway is that we may be able to do a much better job of preventing type 2 diabetes by using information that already exists in routine health care.
Rather than waiting until patients meet diagnostic thresholds, health systems could potentially identify individuals at high risk years earlier and target them for prevention. This approach would enable more efficient and scalable diabetes prevention efforts, helping ensure that limited resources — such as lifestyle intervention programs — reach the people who need them most.
MedicalResearch.com: What recommendations do you have for future research as a result of this study?
Response: The next step, after we finalize the external validation, is to move from prediction to action. Future research should focus on how to use these risk predictions to improve patient outcomes — for example, by testing strategies to increase engagement in diabetes prevention programs or to guide targeted use of preventive treatments.
It will also be important to evaluate how well this approach works in other health care systems and populations, and to assess its real-world impact on reducing diabetes incidence at the population level.
MedicalResearch.com: Is there anything else you would like to add? Any disclosures?
Response: A key strength of this study is that it was conducted within a large, integrated health care system, which creates a possible avenue for translating these findings into clinical practice.
This work reflects a broader effort to shift diabetes prevention from a reactive model to a proactive, population-based approach.
This study was funded by NIH/NIDDK. The authors have no other disclosures.
According to the National Institute of Diabetes and Digestive and Kidney Diseases, identifying individuals at risk for type 2 diabetes early and connecting them with proven lifestyle intervention programs can significantly reduce or delay the onset of the disease.
For more on diabetes prevention and metabolic health research, see MedicalResearch.com’s diabetes research coverage.
Disclaimer: The information on MedicalResearch.com is provided for educational purposes only, and is in no way intended to diagnose, cure, or treat any medical or other condition. Some links are sponsored. Products, services and providers are not warranted or endorsed by MedicalResearch.com or Eminent Domains Inc. Always seek the advice of your physician or other qualified health and ask your doctor any questions you may have regarding a medical condition. In addition to all other limitations and disclaimers in this agreement, service provider and its third party providers disclaim any liability or loss in connection with the content provided on this website.
Last Updated on June 7, 2026 by Marie Benz MD FAAD