A real-world analysis of more than 1 million newly diagnosed type 2 diabetes patients shows how BMI-based treatment patterns may place a disproportionate burden on AAPI patients.
Key Takeaways
- Nearly 1 in 4 newly diagnosed AAPI patients with type 2 diabetes had a BMI below 25 at diagnosis (more than twice the rate seen in the non-AAPI population — 24% vs. 9%)
- Normal/low BMI patients were seen to start treatment after diagnosis less often, and used GLP-1s at lower rates than higher-BMI patients
- Within the AAPI population, lower BMI patients initiated treatment at a significantly lower rate than higher-BMI AAPI patients (63% vs. 79%)
- Lower BMI AAPI patients were prescribed a GLP-1 within the first year after diagnosis at roughly half the rate of lower BMI non-AAPI patients (7% vs. 14%)
- Lower BMI AAPI patients received their first treatment an average of ~26 days later than lower BMI non-AAPI patients
Type 2 diabetes (T2D) is often recognized and treated through a clinical lens shaped by body weight. For many Asian American and Pacific Islander (AAPI) patients, that lens can miss a key part of the risk profile. AAPI populations are known to develop insulin resistance, beta-cell dysfunction, and T2D at lower BMI values than Western reference populations. While BMI is understood to be an imperfect tool in clinical practice, it remains one of the most utilized visible signals in identifying metabolic risk, guiding treatment intensity, and shaping access to newer therapies.
To better understand how BMI relates to real-world treatment patterns among newly diagnosed T2D patients, we used Marmot, Komodo’s healthcare-native AI platform powered by the Healthcare Map. The analysis included over one million adult patients (18-65) with available race and BMI data between 2021-2025. Patients were stratified by race (AAPI vs. non-AAPI) and BMI at diagnosis: below 25 vs. 25 and above. Patients with missing race data were excluded. BMI was defined using the measurement closest to the diagnosis date within 30 days before or after diagnosis. Treatment initiation was defined as receipt of at least one diabetes medication within 365 days after diagnosis.
Here’s what we found:
Lower BMI (defined as normal or low, i.e., not high) was more than twice as common among newly diagnosed AAPI patients as it was among non-AAPI patients.
- Nearly 1 in 4 (24%) newly diagnosed AAPI patients had a BMI below 25, compared with 9% among non-AAPI patients.
Lower BMI AAPI patients initiated treatment at a substantially lower rate than higher-BMI AAPI patients.
- Within one year of diagnosis, 63% of lower BMI AAPI patients received at least one diabetes medication. Among higher-BMI AAPI patients, the rate was 79%. Medications included metformin, GLP-1 receptor agonists, SGLT2 inhibitors, insulin, and older oral agents.
- The same pattern appeared among non-AAPI patients. In that population, 64% of lower BMI patients initiated treatment within one year, compared with 81% of higher-BMI patients.
Lower BMI AAPI patients were prescribed GLP-1s at roughly half the rate of lower BMI non-AAPI patients. Metformin remained the dominant first treatment across all groups.
- Among treated normal/low BMI patients, 7% of AAPI patients were prescribed a GLP-1 during the first year after diagnosis, compared with 14% of non-AAPI patients. Similarly, GLP-1s were seen used as a first treatment in 3% vs. 8%.
- GLP-1 use within the first year also tracked with BMI. Treated higher-BMI AAPI patients used GLP-1s at nearly three times the rate of treated normal/low BMI AAPI patients: 20% vs. 7%.
- Metformin remained the dominant first treatment among treated normal/low BMI AAPI patients, at 79%, followed by older oral agents at 15%. Insulin use showed a separate cross-racial gap: treated normal/low BMI AAPI patients used insulin at 10% vs. 32% among their non-AAPI counterparts at first treatment — and at less than half the rate in the first year after diagnosis, 17% vs. 43%.
Lower BMI patients moved into treatment more slowly.
- Lower BMI AAPI patients received their first treatment about 26 days after diagnosis compared to non-AAPI patients with lower BMI (195 days vs. 169 days).
These findings point to a BMI-linked treatment gap with disproportionate implications for AAPI patients, a population where low BMI is more common in T2D. Normal/low BMI patients were less likely to start treatment after diagnosis, waited longer for first therapy, and used GLP-1s at lower rates than higher-BMI patients. Because nearly one-quarter of newly diagnosed AAPI patients had a BMI below 25, these treatment patterns may affect AAPI patients at greater scale. While this analysis does not establish causality or clinical appropriateness, it does identify a measurable blind spot in how type 2 diabetes care reaches patients whose metabolic risk appears at lower BMI.
Real-world data helps reveal gaps that traditional research can miss — especially when disease presentation, treatment patterns, and patient populations do not fit standard assumptions. For AAPI patients with type 2 diabetes, the findings show why large-scale, longitudinal data matters: it can surface where care pathways diverge, where treatment signals may be calibrated too narrowly, and where further investigation is needed. With Marmot, powered by Komodo’s Healthcare Map, stakeholders can move from isolated observations to population-level evidence that supports more precise, equitable care.
About Marmot: This analysis was conducted using Komodo Health’s Marmot, the healthcare-native AI analytics platform powered by the Healthcare Map and built to deliver transparent, reproducible insights at scale. By tracing treatment patterns across more than 1 million newly diagnosed type 2 diabetes patients, Marmot helps surface where real-world care diverges across populations — turning complex patient journeys into evidence that Komodo can use to identify gaps, characterize disparities, and support more equitable healthcare decisions.
To see more articles like this, follow Komodo Health on LinkedIn, YouTube, or X, and visit our Resources Hub.
