Differential Selection of Obesity Treatments by Race/Ethnicity

Major advances in obesity treatment have emerged in the last decade, adding effective incretin-based injectables like semaglutide to existing treatments ranging from lifestyle modification to bariatric surgery. While the burden of obesity in the United States is high and disproportionately affects racial and ethnic minoritized individuals,^1,2 obesity interventions may be distributed inequitably. Insurance coverage is cited as a critical barrier to accessing novel expensive treatments, yet it is unknown whether inequities persist in this new treatment paradigm in the absence of insurance barriers.

What is the clinical question being addressed?

Do obesity treatments differ by race and ethnicity for Veterans with overweight or obesity?

What is the main finding?

Among 2.4 million Veterans with overweight or obesity, many minoritized populations were more likely to receive lifestyle interventions and less likely to receive more efficacious, expensive therapies, such as glucagon-like peptide-1 agonists.

All patients with ≥2 body mass index (BMI) values ≥27 kg/m² (Food and Drug Administration criterion for antiobesity medication eligibility) in the Veterans Health Administration (VA) Healthcare System from January 1, 2018, to December 31, 2022, and known race/ethnicity (<6% missing) were included in this retrospective cohort analysis. We compared baseline characteristics using chi-square (categorical variables) and analysis of variance (continuous variables), then evaluated the association between race/ethnicity (American Indian/Alaska Native [AI/AN], Asian/Native Hawaiian/Pacific Islander [NHOPI], non-Hispanic Black, Hispanic/Latino, non-Hispanic White) and receipt of obesity interventions with multivariable logistic regression. Three treatment categories were evaluated: completion of ≥1 visit in VA's comprehensive lifestyle intervention, “MOVE!”; ≥1 prescription for an antiobesity medication; and bariatric surgery. Medications included phentermine, phentermine-topiramate, orlistat, bupropion-naltrexone, liraglutide, and semaglutide (Ozempic and Wegovy). Bariatric surgery (defined by current procedural terminology and International Classification of Disease-10 codes) included adjustable gastric band, sleeve gastrectomy, and roux-en-y gastric bypass. Models were adjusted for age, sex, obesity class, diabetes diagnosis, Charlson comorbidity index, year of overweight/obesity diagnosis, and social deprivation index.³ MOVE! and medication prescription analyses were additionally adjusted for VA facility as a random intercept; bariatric surgery models did not converge with this term given the rarity of the outcome. We also conducted subanalyses of semaglutide prescription by race/ethnicity and a stratified analysis by sex. The Stanford Institutional Review Board and the VA Research and Development Committee approved this study. Requirement for informed consent was waived.

Among 4,722,568 VA patients with BMI data, 51% (n = 2,424,784) had a BMI >27 kg/m². Among this cohort with overweight or obesity (7.8% [n = 189,786] female, mean age of 63.6 ± 13.3, 71.4% [n = 1,730,895] White, 20.4% [n = 494,461] Black, 6.0% [n = 144,715] Hispanic/Latino, 1.5% [n = 37,142] Asian/NHOPI, 0.7% [n = 17,571] AI/AN), 38.3% (n = 941,467) had class I obesity (BMI 30-34.9), 18.1% (n = 440,097) had class II obesity (BMI 35-39.9 kg/m²), and 9.8% (n = 238,613) had class III obesity (BMI ≥40 kg/m²). Over one-third (38.2%, n = 933,715/2,424,784) had a diabetes diagnosis; 50.9% (n = 121,454/238,613) of those with class III obesity had diabetes. Black and AI/AN patients had the highest proportions of class III obesity, while Asian/NHOPI patients had the lowest (Figure 1A).

Figure 1.

Obesity Severity by Race/Ethnicity and the Likelihood of Obesity Treatment Across Modalities by Race/Ethnicity and Sex

(A) Obesity severity by race/ethnicity (B) likelihood of obesity treatment across modalities by race/ethnicity and sex. Models adjusted for age, sex, obesity class, diabetes, Charlson comorbidity index, year, and social deprivation index, with VA facility as random intercept for MOVE! and medications. AI/AN = American Indian/Alaska Native; MOVE! = comprehensive lifestyle intervention; NHOPI = Native Hawaiian or Pacific Islander.

Among patients with overweight/obesity (n = 2,424,784), 11.1% (n = 262,218) received lifestyle interventions, 8.0% (n = 193,723) received medication, and 0.1% (n = 2,820) underwent bariatric surgery. Lifestyle intervention use varied by race/ethnicity from 9.4% (n = 163,335/1,730,895) for White patients to 15.9% (n = 78,703/494,461) for Black patients (P < 0.001). Medication use varied from 7.9% among White patients (n = 136,142/1,730,895) to 8.9% among Hispanic/Latino and Asian/NHOPI patients (n = 13,144/144,715 and n = 3,315/37,142, respectively) (P < 0.001). Bariatric surgery use ranged between 0.1% and 0.2% across all race/ethnicity groups (P < 0.001).

Compared to White patients, Black and Hispanic/Latino patients of both sexes and male Asian/NHOPI patients received more lifestyle interventions (Figure 1B). Conversely, male Black and AI/AN patients were less likely to receive medications, though female Black patients and Hispanic/Latino patients of both sexes were more likely to receive medications. Black patients were also less likely to undergo bariatric surgery (AOR: 0.89 [95% CI: 0.81-0.97]), though this effect disappeared after stratification by sex.

Semaglutide was the most frequent medication used, comprising 84% of prescriptions. Semaglutide prescription rates varied across 130 VA facilities (0.4%-11.7%; P < 0.01). Black Veterans were less likely than White Veterans to receive semaglutide prescriptions even after adjustment for diabetes status (AOR: 0.78 [95% CI: 0.76-0.80]), with no differences for other race/ethnic groups.

In this analysis of contemporary obesity treatments among 2.4 million eligible patients in the largest U.S. integrated health care system, we identify opposing gradients of treatment across the spectrum of intervention options. To our knowledge, this represents the first national report of Food and Drug Administration-labeled antiobesity medication use to include semaglutide prescriptions, and the racial/ethnic and sex differences therein, among a real-world population of patients with overweight and obesity. In contrast to decades of cardiovascular studies showing undertreatment among minoritized individuals, we did not find consistent undertreatment across all obesity treatment modalities by race/ethnicity. Notably, despite the lack of insurance barriers in the VA population, several minoritized populations still received the more efficacious, expensive therapies disproportionately less than nonminoritized groups.

Gaps in insurance coverage have been cited in the non-VA population as a primary reason for decreased access to glucagon-like peptide-1 agonists.⁴ However, given Veterans' universal access to free or low-cost health care, our findings suggest that mechanisms influencing differential care are more complex than out-of-pocket cost of treatment. Other factors may include patient preferences, provider biases, access to specialty care, health care system barriers, availability of on-site bariatric surgery, and health-related social needs beyond those captured by the social deprivation index.

Our study has limitations. Though we adjusted for diabetes status, we cannot be certain whether the primary indication for semaglutide was obesity or diabetes. Veterans may have received treatment outside the VA, but particularly for medication use, we would expect this number to be low given lower medication costs within the VA. A small proportion (<2%) of patients received multiple treatment modalities, which are beyond the scope of this report. Despite adjustment, additional confounders may persist. While the VA population is racially and ethnically diverse, lower proportions of women and younger individuals compared to the general population may affect the generalizability of our findings.

Amid an obesity epidemic in the United States, we observe differential treatment among the highest-risk individuals, with certain racial and ethnic minorities likelier to receive lifestyle interventions and less likely to receive medications and surgery. These findings challenge the assumption that insurance gaps are the primary culprit for inequities in pharmacotherapy use. VA/DoD guidelines for the management of adult overweight and obesity do not address inequities in obesity treatment.⁵ Yet as regulations surrounding obesity interventions evolve, further investigation into the factors that influence selection of obesity treatment by race, ethnicity, and sex are critical to ensuring that all patients can access the most appropriate evidence-based therapies.

Funding support and author disclosures

Dr Yong is funded by the Department of Veterans Affairs Health Systems Research (1IK2HX002236-01A2). All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.