Higher BMI, higher medical costs: BMI ~37 adds $3,800/year - 2024

Page by Claude AI - April 2026

Annual extra medical costs at BMI 35 and BMI 40+

Bottom line: The best causal (instrumental-variables) estimate is that a U.S. adult with class II obesity (BMI 35–39.9) incurs roughly $3,000 more per year in total medical costs than a normal-weight adult, and class III obesity (BMI ≥40) incurs roughly $5,850 more per year — about 2× and nearly 4× higher than their normal-weight peers' total spending, respectively (Cawley, Biener, Meyerhoefer et al., JMCP 2021; 2017 USD). Naive (non-IV) regressions put the class III figure closer to $3,100/yr (Ward et al., PLOS ONE 2021; 2019 USD). Costs rise sharply and nonlinearly above BMI 35, with inpatient spending being the largest driver. Adjusting for medical-care inflation, the 2024-dollar equivalents are roughly $3,800 (class II) and $7,400 (class III).

Headline per-person figures by BMI class

The single most authoritative BMI-class-specific breakdown comes from Cawley, Biener, Meyerhoefer, Ding, Zvenyach, Smolarz & Ramasamy, J Manag Care Spec Pharm 27(3):354–366 (2021), using MEPS 2001–2016, N=63,508 adults, with a two-part instrumental-variables model (child's BMI as instrument). Reference group is normal weight (BMI 18.5–<25). All figures in 2017 USD.

BMI band	Incremental annual cost vs. normal weight (IV)	% increase	~2024 USD (CPI-U)
Class 1 (30–<35)	$1,713 (90% CI 1,343–2,084)	+68%	≈ $2,180
Class 2 (35–<40)	$3,005 (2,245–3,766)	+120%	≈ $3,820
Class 3 (≥40)	$5,850 (4,004–7,696)	+234%	≈ $7,440
All obesity (≥30)	$2,505 (1,909–3,102)	+100%	≈ $3,200
Per 1-unit BMI (linear)	+$201/yr	—	≈ $255

The paper does not publish a split between BMI 40–44.9 and ≥45, but shows that inpatient spending alone rises +178% for class 1 and +924% for class 3 vs. normal weight — the clearest published evidence of sharp nonlinear acceleration above BMI 40. Biener, Cawley & Meyerhoefer's 2017 JGIM commentary similarly describes the cost curve as flat through overweight, rising slowly in class I, then "rising rapidly in the range of obese class II (35 ≤ BMI < 40) and especially obese class III (40 ≤ BMI)."

Time trend from the same 2021 paper (2017 USD, IV): class III cost rose from $5,395 (2001–05) → $5,957 (2006–10) → $6,312 (2011–16), ~17% growth in 15 years on top of inflation.

How these numbers compare across methods

Study	Year / dollar year	Method	Key per-person estimate
Finkelstein et al., Health Affairs 2009	2006 data / 2008 USD	Two-part OLS, obese vs. non-obese	+$1,429/yr (Medicare +$1,723, Medicaid +$1,021, private +$1,140); aggregate $147B
Cawley & Meyerhoefer, J Health Econ 2012	2000–05 MEPS / 2005 USD	IV vs. OLS; obese vs. non-obese	OLS +$656 vs. IV +$2,741 — IV ≈ 4.2× OLS (women $3,613; men n.s.)
Biener, Cawley & Meyerhoefer, JGIM 2017	2006–13 / 2013 USD	IV, obese vs. non-obese	+$3,429/yr; 93.6% borne by third-party payers
*Cawley et al., JMCP* 2021**	2001–16 / 2017 USD	IV, by BMI class vs. normal	Class 2 $3,005; Class 3 $5,850 (see table above)
*Ward, Bleich, Long & Gortmaker, PLOS ONE* 2021**	2011–16 / 2019 USD	Two-part model (not IV), BMI self-report-corrected via NHANES	Obesity overall +$1,861; severe obesity +$3,097; +$253 per 1-unit BMI above 30
CDC "Fast Facts" (current)	Cites Ward 2021 / 2019 USD	—	~$173B aggregate; per-person figures as above
Thorpe & Joski, JAMA Netw Open 2024	Projection / 2024 USD	Extrapolation of Cawley 2021	Projected >$385B aggregate obesity spending in 2024
JMCP systematic review (Bjornson et al. 2025)	Synthesis	—	Cites Cawley 2021 figures as current benchmark; >$261B aggregate

Why IV estimates run 2–4× higher than OLS. Self-reported weight in MEPS contains classical measurement error that attenuates OLS toward zero, and weight is endogenous to unobserved health, SES, and care access. Instrumenting adult BMI with a biological child's BMI purges both biases. Cawley & Meyerhoefer (2012) found a 4.2× gap; subsequent IV studies land in a similar range. A 2024 Health Economics methods paper (Biener, Meyerhoefer & Cawley) shows the measurement error is actually non-classical (mean-reverting — heavier people under-report more) but confirms the IV-OLS ordering and approximate magnitudes hold. Takeaway: treat OLS-based estimates like Ward 2021 and Finkelstein 2009 as lower bounds; treat Cawley 2021 IV estimates as the best causal benchmark.

Payer breakdown and service drivers

From Cawley et al. 2021 (obesity vs. normal weight, per-person annual increment, 2017 USD):- Public insurance +$2,877, private insurance +$2,058, out-of-pocket +$229 — public payers bear disproportionately more of the marginal cost. Third-party payers cover 88.5% of the total obesity-attributable cost.- By service: inpatient +$1,088 (+290%), outpatient +$787 (+67%), prescription drugs (excluding anti-obesity meds) +$917 (+187%). Inpatient is the biggest absolute driver and accelerates most with class III.

Finkelstein's older breakdown (2006 MEPS/MCBS) attributes 12.9% of private, 11.8% of Medicaid, and 8.5% of Medicare spending to obesity; Medicare + Medicaid finance ~41–50% of the total obesity cost burden.

Lifetime cost estimates

Finkelstein, Trogdon, Brown et al. (2008, Obesity 16:1843–48) calculated discounted lifetime medical costs attributable to obesity for a 20-year-old, ranging from ~$5,300 (Black women, Grade I) to ~$29,500 (White women, Grade II/III) in 2006 USD. In current dollars that Grade II/III lifetime increment is roughly $45,000. No recent (2020s) peer-reviewed lifetime update incorporating current prevalence of severe obesity has been published.

GLP-1 era and 2023–2025 updates

No peer-reviewed study has yet republished per-BMI-class cost estimates using post-2022 MEPS data that would capture real-world GLP-1 uptake. The 2025 JMCP systematic review (Bjornson et al.) continues to cite Cawley 2021 as the current benchmark. CBO's 2024 analysis of Medicare coverage of anti-obesity medications assumed the baseline obesity-attributable medical cost structure from Cawley 2021 and Ward 2021, and concluded that at current GLP-1 pricing the drugs would increase federal spending over 10 years — the offsetting medical savings (estimated using these same per-BMI-class figures) do not yet outweigh drug acquisition costs. Thorpe & Joski (JAMA Netw Open Dec 2024) project the aggregate direct medical cost of adult obesity exceeded $385 billion in 2024, extrapolating Cawley 2021's per-person increments by overall health-spending growth.

Key caveats worth flagging

Cawley-program estimates require a biological child in the household, so the sample is ~24–65 years old; costs in elderly adults (where absolute medical spending is highest) are likely underestimated.
Self-reported BMI under-classifies obesity, biasing all MEPS-based estimates downward.
No published BMI 40–44.9 vs. ≥45 split exists from the major U.S. studies; the evidence for steep acceleration above BMI 40 comes from the BMI² term, the inpatient-specific increments (+924% at class III), and a UK cohort (Bojke et al., Diab Obes Metab 2024) showing costs rise monotonically across successive BMI strata.
The 2021 Cawley et al. JMCP paper was funded by Novo Nordisk; the methodology and MEPS data are public and consistent with the authors' earlier non-industry-funded work, but note the disclosure.

Takeaway for quick citation

For a single defensible number, cite Cawley, Biener, Meyerhoefer et al., JMCP 2021: a U.S. adult with BMI 35–39.9 incurs roughly $3,000/yr (2017 USD; ~$3,800 in 2024 USD) and one with BMI ≥40 incurs roughly $5,850/yr (2017 USD; ~$7,400 in 2024 USD) in extra total medical costs compared to a normal-weight adult — with the class III increment accelerating sharply because inpatient admissions rise nearly tenfold. Naive (non-IV) MEPS regressions cut these figures roughly in half and should be treated as lower bounds.