DECODE, Analysis Group’s GenAI Model, Reveals Long-Term Benefits of Weight Loss on Cardiometabolic Conditions

September 26, 2025

Obesity is a chronic, multifactorial condition linked to a higher risk of more than 200 comorbidities, and substantial health care utilization and costs. Research is essential to understanding and treating obesity, but complex comorbidity interactions, a lack of long-term comprehensive data, and the need to evaluate a wide range of patient subpopulations present significant challenges. Leveraging the power of generative AI (GenAI), Analysis Group developed a comprehensive disease model to address those limits and help advance understanding of obesity and cardiometabolic conditions.

At the Professional Society for Health Economics and Outcomes Research (ISPOR) 2025, Analysis Group Managing Principal Eric Wu and Principal Jimmy Royer presented their Dynamic Evaluation of Cardiometabolic Obesity Disease (DECODE) model, a digital twin that was trained on large US electronic health record (EHR) and claims databases, and can simulate the progression of obesity and other related cardiometabolic conditions simultaneously.

In an application of DECODE to assess the long-term progression of obesity and its effects on patients with cardiometabolic conditions (and comorbidities such as type 2 diabetes, hypertension, sleep apnea, and osteoarthritis), researchers at Analysis Group found that a 10% weight loss over 10 years was associated with a 26% lower risk of heart failure and a 12% lower risk of bariatric surgery.

The success of the current use case encourages exploration of more GenAI applications developing holistic disease and treatment modeling. According to the team, the model “has shown great potential to help researchers better understand complex disease and comprehensive impacts of treatments.”