As we move through 2026, our blog theme “Health in All Policies” continues to drive our public health discussions. This framework reminds us that well-being isn’t just about medicine or insurance. Instead, it is shaped by where we live, work, and age—our social, economic, and environmental conditions.
A study recently published ahead-of-print in Medical Care brings this concept to life. The paper, Diabetes Diagnosis Patterns in Medicaid: How State Policy, Managed Care, and Social Vulnerability Shape Detection, by Maria Alva, DPhil and colleagues, explores how these background factors influence type 2 diabetes (T2D) diagnoses for Medicaid beneficiaries.
The Data: Who is Being Diagnosed?
Analyzing Medicaid data on ~4.4 million enrollees from 2016 to 2021 across 11 states with complete data, the research team found that roughly 1 in 10 working-age adult beneficiaries had a recorded T2D diagnosis.
However, the numbers showed variation. Even after adjusting for demographics and clinical factors, diagnosed prevalence ranged from 6.5% to 13.0% depending on the state. The variations mapped closely to:
- Age and race/ethnicity
- County-level obesity rates
- Local policy contexts
The Barrier of Underdiagnosis
Disparities in diabetes rates aren’t just a reflection of biology; instead, they are heavily influenced by structural barriers. Many Medicaid enrollees face food insecurity and a shortage of doctors who accept their insurance. These obstacles delay primary and preventive care, leading to underdiagnosis.
In other words, a lower diagnosis rate in a specific area doesn’t always mean fewer people are sick. It sometimes means fewer people are getting tested.
How Insurance Models Affect Diagnosis
Policy decisions, such as physician payment structures, heavily influence diagnosis rates. Today, managed care (run by managed care organizations [MCOs]) is the dominant Medicaid model. Because states use these frameworks, they should have a stronger financial incentive to invest in preventive care and screening.
The study found that, unadjusted, 87% of beneficiaries with T2D were enrolled in capitated plans, compared to 85% of those without.
This makes sense when you look at the economics. Under a capitation model, insurance plans receive a risk-adjusted, lump-sum payment per enrollee. Because higher-risk diagnoses can increase the plan’s payment, these models have an incentive to accurately find and document chronic conditions.
However, as we can see in the portion of Fig. 2 below, that is not what they found when they adjusted for race/ethnicity and other confounders. Capitated plans had a lower predicted T2D diagnosis rate than non-capitated plans. On the other hand, comprehensive MCO plans had higher predicted probabililties than other MCOs.
Portion of Fig. 2, Posterior predictive distributions of type 2 diabetes diagnosis rates as a function of race/ethnicity and contextual factors. From Diabetes Diagnosis Patterns in Medicaid: How State Policy, Managed Care, and Social Vulnerability Shape Detection in Medicaid. Alva et al. Medical Care64(7):419-427, July 2026.
doi: 10.1097/MLR.0000000000002328. CMCO: Comprehensive Managed Care Organization.
The Outliers
People identified as “Other” race in this study included Asians and Pacific Islanders, Native Americans and Alaska Natives, and people with multiracial/ethnic backgrounds. They made up about 7% of the total sample, but had the highest prevalence: 14.3%. In the portions of Fig. 2 above and below, we can see that they have a much higher likelihood of T2D in some of the strata. The wider bars on either side of the point estimate are a result of the smaller sample sizes involved.
The Real Drivers: Poverty and Social Vulnerability
While the raw numbers show a difference by plan type, looking closer at the data reveals a deeper story. When controlling for other variables, the biggest differences in diagnosis rates weren’t defined by plan. Instead, they were driven by individual poverty and social vulnerability quartiles.
For example, the study noted patterns of the probability of diagnosis across Social Vulnerability Index (SVI) quartiles (see below). Those in the 4th quartile, the most disadvantaged, had lower probabilities of diagnosis—except among those in the “Other” race category.
This suggests that community-level disadvantage may be a barrier to obtaining necessary healthcare services, especially among Hispanic and Black enrollees (shown in green). Those enrollees show a clear pattern of increasing probability of diagnosis with each decrease in SVI quartile.
Portion of Fig. 2, Posterior predictive distributions of type 2 diabetes diagnosis rates as a function of race/ethnicity and contextual factors. From Diabetes Diagnosis Patterns in Medicaid: How State Policy, Managed Care, and Social Vulnerability Shape Detection in Medicaid. Alva et al. Medical Care64(7):419-427, July 2026.
doi: 10.1097/MLR.0000000000002328. Higher quartile indicates higher social vulnerability.
Moving Toward “Health in All Policies”
If we want to close these gaps, we cannot rely on the healthcare system alone. This study proves that “Health in All Policies” is a necessary call to action, not just a theoretical concept.
To systematically tackle chronic diseases like diabetes, policymakers must integrate health considerations into every sector, including:
- Housing: Ensuring safe, stable living environments
- Transportation: Creating reliable ways for people to get to clinics
- Nutrition: Improving access to healthy, affordable food
Furthermore, we must improve the quality of our health data. We need data that reflects the true burden of disease, not just the quirks of MCO coding practices. A coordinated, cross-sector approach is the best way to serve vulnerable communities and narrow the gap in chronic disease care.
For more insights on integrating health into policy-making, check out our other posts on Health in All Policies. Together, we can advocate for changes that improve health outcomes and ensure equitable healthcare access for everyone.
Lisa M. Lines