Where obesity burden concentrates, what it means for the Medicaid-eligible population, and what it may cost — for any U.S. state. Decision support, not a predictive model.
Pick a state to load live data…
The Medicaid correction — three steps from headline to reality
The published state obesity rate is the wrong number for a Medicaid book. Three corrections fix it.
Addressable annual cost MEPS-based
Excess medical cost ≈ $1,861 per adult with obesity per year ($3,097 severe) — Cawley et al., JMCP 2021 (MEPS 2011–16, all-payer; likely conservative for Medicaid). Enter your covered lives above for your number.
GLP-1 budget-impact scenario illustrative
A simple "what if we treat some of them" model. All inputs are editable assumptions — a planning sketch, not a forecast. Reference points (2025–26): net GLP-1 cost ≈ $2,940/yr under Medicaid (BALANCE model) to $6,200–8,400 commercial after rebates; short-term medical offset ≈ 0% for obesity-only populations (offsets appear mainly in the obesity+diabetes subgroup).
Who carries the burden — disparities BRFSS
Obesity by race/ethnicity (latest available). Medicaid populations skew toward higher-prevalence groups — another reason the state average understates the Medicaid rate.
Priority counties
Above the state median on both need-intensity and volume — start here.
It compounds
Obesity & diabetes co-locate across counties.
County burden
Click a column header to sort. Index = standardized blend of obesity + diabetes + inactivity. Volume = est. adults with obesity.
State trend vs U.S.
Adult obesity prevalence (self-reported, BRFSS).
How this is calculated methodology
Calibrated Medicaid-eligible obesity, ranked
Same engine as the single-state view, run across the region. The headline is the published state rate; the calibrated column applies the Medicaid low-income proxy + the NHANES self-report correction (×1.16, low-income-specific). Click a column to sort.
What this says
From estimate to measurement — your own data paid engine
Everything above is a public-data proxy: it re-bases the state headline to the Medicaid-eligible group and corrects self-report bias, then shown as a range. It's built to rank and size the problem — not to budget against. Point the same engine at a plan's own enrollment/claims and the proxy disappears:
This demo (public data)
Calibrated estimate via FPL proxy + self-report correction · county prevalence from modeled PLACES estimates · illustrative cost & GLP-1 figures. Answers: "is there a problem worth my attention?"
Paid v2 engine (your members)
Measured prevalence — no proxy, no correction · exact spend & county counts on real members · GLP-1 scenario with the obesity+diabetes high-cost core. Runs inside your environment under a BAA; only aggregate rollups leave. Answers: "exactly how big, and what does it cost us?"
If the directional picture above maps to a real problem, the exact version is a scoped pilot on your own data. That's the conversation worth having.