Trends in Use of Obesity Medications and Metabolic and Bariatric Surgery Among All of Us Participants From 2003 to 2023

ABSTRACT

Objective

This study aimed to evaluate the trends in antiobesity medication (AOM) and metabolic and bariatric surgery (MBS) use relative to the 2013 American Medical Association (AMA) declaration of obesity as a chronic disease and semaglutide approval in 2021 in the United States.

Methods

We employed a repeated cross‐sectional design to analyze national electronic health records (EHR) of US adults (≥ 18 years) with obesity (BMI ≥ 30 kg/m²) or overweight (BMI ≥ 27 kg/m²) with ≥ 1 comorbidity for AOM and those with BMI ≥ 35 kg/m² for MBS, using the All of Us dataset. We used interrupted time‐series models to evaluate trends from January 1, 2003, to October 1, 2023.

Results

AOM rates increased by 0.31%/year (95% CI, 0.25%, 0.37%) after 2013 and further increased by 1.42%/year (95% CI, 1.16%, 1.68%) after 2021. MBS use declined by 0.10%/year (95% CI, −0.15%, −0.04%) before 2013. Semaglutide approval was not associated with MBS use decline after 2021. After 2013, patients with BMI ≥ 40 kg/m² experienced a larger increase in AOM use (0.50% [95% CI, 0.40%, 0.61%]) vs. < 40 kg/m² (0.25%/year [95% CI, 0.20%, 0.30%]) and stable MBS use.

Conclusions

The declaration increased AOM use and, to a lesser extent, MBS use. Semaglutide approval increased AOM use but had no impact on MBS use.

Study Importance

• What is already known?

  • ○
    The utilization of antiobesity medications (AOMs) and metabolic and bariatric surgery (MBS) in the United States remains low, despite their effectiveness in treating obesity.
  • ○
    Only 1.0% of eligible patients in the US received AOMs, while less than 1% received MBS.
  • ○
    Recent studies in the US have shown a rapid increase in the use of AOMs, particularly the GLP‐1 receptor agonists, while MBS use declines or remains stable.
• What does this study add?

  • ○
    The 2013 declaration of obesity as a chronic disease changed the trend in the use of obesity treatments, as evidenced by a significant increase in the use of AOMs and a stable trend in MBS use after 2013.
  • ○
    The approval of semaglutide in 2021 further impacted obesity treatment use, including an accelerated increase in AOM use and an insignificant impact on MBS use, despite an observed decline.
• How might these results change the direction of research or focus of clinical practice?

  • ○
    Our study establishes how policy change and clinical innovation impact the use of obesity treatments over the years. Given the impact of the 2013 declaration and semaglutide approval in 2021, future studies may further explore how other changes such as insurance coverage impact the trends in AOM and MBS use, as a means of assessing the effect of policy change on health care use and health outcomes.

1. Introduction

Obesity is a major public health problem in the United States (US), with about 40% of US adults having obesity and 10% having severe obesity in 2023 [1]. Recent approval of newer antiobesity medications (AOMs) has transformed obesity pharmacotherapy. Compared to earlier medications, long‐lasting glucagon‐like peptide‐1 (GLP‐1), glucose‐dependent insulinotropic polypeptide (GIP), and GLP‐1 dual receptor agonists (RAs) provide superior effectiveness in weight loss [2]. Nevertheless, metabolic and bariatric surgery (MBS) still offers the greatest weight loss [3]. Despite their effectiveness, utilization of AOMs and MBS remains low in the US [4, 5]. From 2015 to 2022, only 1.0% of eligible patients in the US received AOMs [4], while less than 1% received MBS between 2013 and 2018 [5].

In 2013, the American Medical Association (AMA) declared obesity a chronic disease that requires medical attention. This was done to promote changes in patients' and clinicians' perceptions and the clinical and policy landscape in obesity management [6]. Following the declaration and the Food and Drug Administration (FDA) approval of semaglutide in 2021 for obesity, the use of AOMs might be expected to increase. However, there is limited information available on use trends [7, 8]. These two events might also be expected to impact MBS. Specifically, the declaration could have increased the use of MBS, while the approval of semaglutide might have reduced its use if patients were to choose medications over surgery [9, 10]. Lastly, it is not known whether these trends are impacted by the severity of obesity.

It is important to evaluate the impact of policy and clinical innovation on treatment utilization and health outcomes. Significant events, such as the 2013 declaration of obesity and approval of effective GLP‐1/GIP RAs, could support clinical guidelines, increase prescribing by physicians, and alter health plan coverage, all leading to increased uptake of medication. They could also have unintended consequences, such as widening health disparities or supplanting the use of effective treatments such as MBS. Our study will provide evidence to support future policy development, clinical innovations, and population‐based interventions to improve obesity treatment in the US.

We therefore conducted an interrupted time‐series analysis of AOM and MBS use among US adults (≥ 18 years) from 2003 to 2023, using the All of Us (AoU) database. We hypothesized that following the 2013 declaration, receipt of both AOM prescriptions and MBS would increase, and that following approval of semaglutide in 2021, receipt of AOM prescriptions would increase while MBS would decline. We also hypothesized that these trends would be amplified for patients with severe obesity.

2. Methods

2.1. Study Design and Data Source

We conducted a repeated cross‐sectional study, constructing annual cohorts of eligible patients based on their body mass index (BMI), using the deidentified Controlled Tier Dataset (v8) from AoU Research, a program supported by the National Institutes of Health (NIH) to enroll and collect health data on individuals who reflect the diversity of the US population [11]. We used the AoU dataset because it offers access to multiple years of electronic health record (EHR) data across a diverse, nationwide population with detailed information on disease conditions and treatments. The data collected included health conditions, medications, demographics, and laboratory results from EHR, anthropometrics, and surveys. The study was exempt from the Institutional Review Board's approval for human subjects since the data were deidentified.

2.2. Study Population

We captured all eligible AoU adult patients with BMI measurements in EHR between January 1, 2003, and October 1, 2023. For each year, the study sample (denominator) included those with at least one BMI measurement recorded. For the AOM analysis, we captured patients with obesity (BMI ≥ 30 kg/m²) or overweight (BMI ≥ 27 kg/m²) with ≥ 1 obesity‐related comorbidity, consistent with FDA‐approved indications for AOM use [12]. For the MBS analysis, the eligible sample was limited to patients with BMI ≥ 35 kg/m² to align with clinical guidelines [3]. The obesity‐related comorbidities included type 2 diabetes (T2D), dyslipidemia, hypertension, obstructive sleep apnea, stroke, atrial fibrillation, coronary artery disease, and heart failure, identified using the Current Procedural Terminology (CPT) and the International Classification of Diseases (ICD), Ninth and Tenth Revision codes (Table S1). Patients with T2D were captured using the ICD codes or eMERGE algorithm [13], with the initial diagnosis date being the date when a patient met either of (1) an ICD code with diabetes medication, (2) an ICD code with abnormal glucose (fasting blood glucose ≥ 126 mg/dL, glycated hemoglobin (A1C) ≥ 6.5% or random glucose ≥ 200 mg/dL), (3) two ICD codes and an outpatient insulin prescription, (4) diabetes medication with abnormal glucose, or (5) outpatient insulin preceded by diabetes medication criteria (Table S1). Pregnant women and patients with cancer were excluded from the study (Figure 1).

FIGURE 1.

Identification of eligible patients for inclusion. AOM, antiobesity medication; EHR, electronic health records; MBS, metabolic and bariatric surgery.

2.3. Outcome Measurement

We captured the FDA‐approved AOMs, including phentermine‐topiramate, naltrexone‐bupropion, orlistat, phentermine, injectable semaglutide (Wegovy), and liraglutide (Saxenda), under the brand names for obesity. Similarly, prescriptions for tirzepatide, semaglutide, and liraglutide approved for T2D were captured as off‐label use for obesity if the patient with obesity had no T2D at the time of the prescription (Table S2) [14]. For MBS, we identified the first surgical procedure using the Healthcare Common Procedure Coding System (HCPCS), CPT‐4, ICD‐9, and ICD‐10 codes (Table S1) [15, 16]. The receipt of AOM prescriptions and MBS was captured as categorical (yes or no) variables. Demographic characteristics considered included age (< 40 years and ≥ 40 years), based on higher risk of obesity among people aged 40 and above [17], and gender (male and female). Patients were grouped into two BMI categories (< 40 kg/m² and ≥ 40 kg/m² with/without ≥ 1 comorbidity), based on the Centers for Disease Control and Prevention BMI categories for adults [18].

2.4. Statistical Analysis

A segmented interrupted time‐series (ITS) regression was used to assess the trends in AOM prescription and MBS receipt before and after policy changes and pharmacotherapy approvals. Segmented ITS regression is a quasi‐experimental method that estimates the effect of an intervention by comparing the observed trends in the post‐intervention period to the counterfactual if the pre‐period trends had continued [19]. It relies crucially on the assumption that there was no other co‐incident policy change and that the counterfactual in the post‐period can be appropriately modeled by projecting forward the pre‐period trend [20]. We used linear regression to estimate a potential level, i.e., instantaneous change in the outcome after intervention, or slope change, i.e., change in the linear rate of increase or decrease from pre‐ to post‐period [21].

Overall, the segmented ITS regression equation is represented as follows:

$$Y_{\mathrm{t}}={\beta }_0+{{\beta }_1}^{*}\text{Time}+{{\beta }_2}^{*}\text{Intervention}+{{\beta }_3}^{*}\text{Post}\_\text{Intervention}+e_{\mathrm{t}}$$

where Y _t represents the percentage of patients receiving AOMs or MBS in year “t”; β ₀ estimates the baseline level; β ₁ estimates the percentage change per year before the intervention (baseline trend); β ₂ represents the percentage level change per year after the intervention; β ₃ estimates the percentage change in the trend after the intervention compared with the pre‐intervention trend; and e _t is the error term. We note that β ₁ + β ₃ gives the post‐intervention slope.

We aggregated the study data by year because annual aggregation over longer intervals yields more comparable denominators, produces more stable rates, and reduces noise due to autocorrelation and seasonality, which can induce bias in slope estimates [22]. To assess the effect of the AMA declaration on AOMs and MBS, we defined the pre‐period to be the 10 years (2003–2012) prior to the declaration and then assessed changes between 2013 and 2020 (post‐declaration) and between 2021 and 2023 (post‐semaglutide approval). Furthermore, we assessed the trends in AOMs excluding approved GLP‐1/GIP RAs. Finally, we conducted subgroup analyses to assess whether effects of the declaration varied by BMI (< 40 kg/m² vs. ≥ 40 kg/m²), age (< 40 vs. ≥ 40 years), gender (male vs. female), race/ethnicity (Black, Hispanic, other races [Asians, American Indians, Alaska Natives, Native Hawaiians, and other Pacific Islanders] vs. White), and race in two groups (non‐White vs. White).

We used the Durbin‐Watson test to check for potential autocorrelation and the Newey‐West standard errors to correct autocorrelation where detected. All analyses were conducted in the R statistical package, v4.4.2, using a two‐tailed significance level of 0.05.

3. Results

We identified 208,099 patients for the AOM analysis with a mean age of 59.3 years and a mean BMI of 36.7 kg/m²; 64.2% were female, 51.4% were White, 21.3% were Black, and 20.2% were Hispanic (Table 1). Likewise, we identified 107,351 patients for MBS analysis with a mean age of 57.4 years and a mean BMI of 41.6 kg/m²; 71.3% were female, 48.8% were White, 24.6% were Black, and 19.9% were Hispanic.

TABLE 1.

Characteristics of patients with or without an AOM prescription or MBS procedure.

Variables	AOM analysis sample (n = 208,099)	MBS analytic sample (n = 107,351)
	n (%)	n (%)
Age	59.3 ± 15.0 years	57.4 ± 14.7 years
BMI	36.7 ± 6.9 kg/m2	41.6 ± 6.9 kg/m2
Age group
18–44 years	52,477 (25.2)	30,288 (28.2)
45–64 years	80,409 (38.6)	44,243 (41.2)
65 + years	75,213 (36.1)	32,820 (30.6)
Gender
Female	133,632 (64.2)	76,496 (71.3)
Male	73,985 (35.6)	30,636 (28.5)
Others	482 (0.2)	219 (0.2)
Race
Black	44,313 (21.3)	26,428 (24.6)
White	106,907 (51.4)	52,390 (48.8)
Hispanic	41,945 (20.2)	21,379 (19.9)
Others	11,347 (5.5)	5387 (5.0)
Unknown	3587 (1.7)	1767 (1.6)

Abbreviations: AOM, antiobesity medication; MBS, metabolic and bariatric surgery.

3.1. Trends in AOM Prescriptions in Relation to the 2013 Declaration and Approval of Semaglutide in 2021

Following the 2013 declaration, receipt of AOM prescriptions increased by 0.31%/year (95% CI, 0.25%, 0.37%) and further increased by 1.42%/year (95% CI, 1.16%, 1.68%) after semaglutide approval, with instantaneous changes of 0.56% (95% CI, 0.18%, 0.94%) and 0.81% (95% CI, 0.33%, 1.29%) after the declaration and semaglutide approval, respectively (Figure 2). Excluding GLP‐1/GIP RAs, AOM prescriptions declined by 0.07%/year (95% CI, −0.11%, −0.03%) before the declaration, increased by only 0.10%/year (95% CI, 0.05%, 0.16%) after the declaration, and declined after 2021 by 0.08%/year (95% CI, −0.32%, 0.16%) (Figure S1).

FIGURE 2.

Trends in the receipt of AOM prescriptions relative to the 2013 declaration of obesity and approval of semaglutide in 2021. [Color figure can be viewed at wileyonlinelibrary.com]

3.2. Trends in MBS in Relation to the 2013 Declaration and Approval of GLP‐1/GIP RAs

Before 2013, MBS use declined by 0.10%/year (95% CI, −0.15%, −0.04%). After the 2013 declaration, MBS use slightly increased by 0.01%/year but then declined by 0.09%/year (95% CI, −0.44%, 0.27%) after semaglutide approval in 2021 (Figure 3).

FIGURE 3.

Trends in MBS use relative to the 2013 declaration of obesity and approval of semaglutide in 2021. [Color figure can be viewed at wileyonlinelibrary.com]

3.3. Trends in AOM Prescription and MBS Receipt by Demographics

After the declaration, the increase in receipt of AOM prescriptions was greater for patients with BMI ≥ 40 kg/m² vs. < 40 kg/m² (0.50% [95% CI, 0.40%, 0.61%] vs. 0.25%/year [95% CI, 0.20%, 0.30%]). After 2021, receipt of AOM prescriptions rose faster among those with BMI ≥ 40 kg/m² vs. < 40 kg/m² (2.26% [95% CI, 1.78%, 2.75%] vs. 1.18%/year [95% CI, 0.95%, 1.42%]) (Figure 4). MBS use declined significantly before the declaration among patients with BMI ≥ 40 kg/m² (−0.20% [95% CI, −0.32%, −0.08%]) vs. < 40 kg/m² (−0.01%/year [95% CI, −0.04%, 0.02%]). After the declaration, MBS use remained stable among the two categories, with those with BMI ≥ 40 kg/m² slightly increased by 0.03%/year. Patients with BMI ≥ 40 kg/m² experienced a higher decline in MBS use after 2021 by 0.15%/year compared to 0.03%/year among those with < 40 kg/m² (Figure 5). The receipt of AOM prescriptions increased faster among those aged ≥ 40 years vs. < 40 years after 2013 (0.33% [95% CI, 0.26%, 0.39%] vs. 0.25%/year [95% CI, 0.17%, 0.32%]) and after 2021 (1.47% [95% CI, 1.18%, 1.77%] vs. 1.17%/year [95% CI, 0.84%, 1.51%]) (Figure S2). After the declaration, MBS use remained flat among those aged ≥ 40 years but increased among those aged < 40 years at 0.09%/year (95% CI, −0.12%, 0.29%) (Figure S3).

FIGURE 4.

Trends in the receipt of AOM prescriptions, stratified by obesity severity (BMI ≥ 40 kg/m² vs. < 40 kg/m²). [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 5.

Trends in MBS use, stratified by obesity severity (BMI ≥ 40 kg/m² vs. < 40 kg/m²). [Color figure can be viewed at wileyonlinelibrary.com]

The AOM rate among females increased faster than among males after 2013 (0.35% [95% CI, 0.28%, 0.42%] vs. 0.23%/year [95% CI, 0.19%, 0.28%]) and after 2021 (1.59% [95% CI, 1.26%, 1.92%] vs. 1.07%/year [95% CI, 0.86%, 1.27%]) (Figure S4). However, there was little evidence of gender differences in MBS use before and after the declaration (Figure S5). Across all racial groups, AOM use increased significantly after the declaration and further increased after semaglutide approval, with no significant differences among racial groups (Figure S6). After the declaration and semaglutide approval, MBS use across all racial groups remained stable with no significant difference between groups (Figure S7).

4. Discussion

This study examined trends in the receipt of AOM prescriptions and MBS in relation to the 2013 declaration of obesity and subsequent approval of semaglutide in 2021. We found that prior to the 2013 declaration, AOM prescriptions were flat, while MBS was declining. As expected, following the declaration, receipt of AOM prescriptions increased significantly, and MBS stabilized. The approval of semaglutide in 2021 accelerated the upward trend in AOM prescriptions, but we did not see the hypothesized decline in MBS. Lastly, we found that both trends were exaggerated among patients with severe obesity.

The low uptake of AOMs before the declaration likely reflects the high cost of therapy, its modest effectiveness, limited insurance coverage, and patient perceptions and willingness to take medications for obesity [12, 23]. Furthermore, many clinicians did not consider obesity to be a chronic disease that requires medication but as a behavioral issue or a matter of personal responsibility [24, 25]. These attitudes likely contributed to low rates of diagnosis, counseling, prescriptions of AOMs, and delayed initiation of weight management plans [26, 27].

Our findings suggest that the declaration successfully addressed some of these issues, such as provider bias, patient perception, and insurance coverage. By reframing obesity as a chronic, relapsing condition requiring evidence‐based, long‐term management, the declaration likely prompted clinicians to rethink their approach to it [28]. In addition, the declaration might have promoted a standard criterion for health insurance coverage, especially among patients with severe obesity [29]. However, the introduction of newer AOMs, characterized by greater efficacy and lower adverse events, explained the lion's share of the increase in AOMs [30, 31], with prescriptions more than doubling from 2022 to 2023 [8].

Although our study reported a decline in MBS use before 2013, other studies reported uneven trends in MBS prior to 2013. For instance, Johnson and colleagues reported an overall increase in the number of bariatric procedures before 2013 [32], encasing periods of decline from 2004 to 2007 and again from 2009 to 2012 [32]. Others reported a decline of more than 50% in MBS use from 2008 to 2011 [33], aligning with our findings prior to 2013. While advances in surgical technique and expanded insurance coverage likely supported expansion [34], concerns over risk, changing clinical guidelines, referral hesitancy, and growing interest in AOMs likely exerted downward pressure [8, 23]. Stringent processes involving insurance coverage and preauthorization could also contribute to the decline.

The use of MBS stopped declining between 2013 and 2021, agreeing with the report given in the 2025 guideline of the Obesity Society [35]. This observation could be due to changes in perceptions of surgery, referral patterns, and stigma [35]. Following semaglutide approval in 2021, MBS began to decline again, suggesting a possible substitution effect, with patients choosing effective AOMs over MBS [36]. This observation, which comports with our hypothesis, should be interpreted cautiously because it was not statistically significant. Further research with longer post‐2021 data is required to definitively assess this relationship. This is important, because patients lose more weight with MBS, and if they opt for medication instead, they may lose out on valuable benefits. Nevertheless, the increase in AOM prescriptions dwarfed the decline in MBS, suggesting that most patients now opting for AOMs never considered MBS.

Following the declaration, patients with BMI ≥ 40 kg/m² had a significantly greater increase in AOM prescriptions, showing a targeted adoption of pharmacotherapy among patients with more severe obesity [12, 37]. Similarly, MBS trends after the declaration increased slightly faster for those with severe obesity, following a decline before the declaration in both BMI categories. This difference may reflect the greater disease burden in heavier patients and the established evidence that MBS is the most effective and durable treatment [38]. The observations by age group also suggest that older patients prefer AOMs [39], while younger adults are more proactive in seeking MBS, which is safer and offers fast and long‐lasting benefits [5].

The significant increase in the AOM prescriptions, especially in the post‐2013 era, signifies improved awareness of obesity treatment, patients' preference for a less invasive option, providers' confidence in AOMs, and fewer perceived barriers. Meanwhile, the stable rates in MBS use during this time raise concerns. Although advancements in MBS have improved safety [40], reduced cost [41], and expanded insurance coverage [42], these changes have not resulted in increased MBS use, suggesting that prior assumptions about barriers to MBS may not be accurate. Instead, limited awareness of these advancements, patient hesitancy, or increased confidence in newer AOMs may be contributing to the shift. Further research should explore the contribution of these factors to the flat trend in MBS, especially among those with severe obesity. These people are likely to have distinct health priorities, concerns about treatment risks, or preferences for nonpharmacologic approaches, which can influence their engagement with available therapies [43].

A key strength of the study was its duration, spanning from 2003 to 2023, offering a long‐term view of the evolving patterns in obesity management and allowing us to examine inflection around the 2013 declaration by the AMA and approval of semaglutide in 2021. While the percentages of AOM and MBS receipt are small, they represent meaningful changes in terms of the underlying population. These are meaningful changes over the baseline rates which were very low prior to the declaration. For example, a change of 0.5%–1% in treatment rates corresponds to hundreds to thousands of patients receiving the treatments. Hence, the rates, slopes, and changes observed in our study translate into important changes in obesity treatment. Providing this context is essential because the narrow y‐axis scales can make visual changes more pronounced. Although ITS regression helps to identify changes following interventions, our estimates could be impacted by unmeasured contemporaneous factors that occurred during a similar time as the AMA declaration and semaglutide approval. For instance, the availability of additional effective AOMs could have resulted in more AOM prescriptions. In addition, Medicaid expansion under the Affordable Care Act (ACA) after 2014 could have increased uptake of MBS [34]. Another limitation is that the denominator population (patients identified as having obesity) changed over time, depending on whether patients had an encounter in which their BMI values were recorded. This variation could affect the estimated rates of obesity treatment receipt even if the absolute number of treated patients remained stable. In addition, shifts in patient characteristics, such as age, health status, or frequency of health care encounters over time, may have affected treatment eligibility and opportunities for diagnosis, introducing selection bias. Furthermore, the AoU data consists of volunteer participants and oversamples underrepresented racial or ethnic groups. Hence, the distribution of race/ethnicity does not reflect that of the US population, which may limit generalizability. Nevertheless, the AoU research program offers access to many years of EHR data across diverse, nationwide volunteers, including detailed prescriptions and procedure records necessary to identify changes in therapeutic use. Also, we captured AOM use from EHR without considering AOMs obtained outside traditional health care settings such as retail pharmacy or telehealth services. Hence, some degree of underreporting is possible. Nevertheless, the use of AoU and the internal consistency of trends over time support the robustness of our findings. In addition, we defined obesity using BMI, which is not always accurate, but it is the most common marker in clinical settings. Lastly, we captured only the tirzepatide brand for T2D in the analysis, as the brand for obesity was not available in the AoU dataset at the time of the study.

5. Conclusion

The 2013 declaration was associated with increased receipt of AOM prescriptions, and the approval of semaglutide in 2021 accelerated that increase. Meanwhile, the declaration stopped the decline in MBS use, which increased slightly after 2013. However, the approval of semaglutide did not definitively impact the use of MBS after 2021. These trends were accentuated among those with severe obesity. These findings highlight the impact of policy and the introduction of new drugs on obesity treatment, with new drugs having a greater impact.

Author Contributions

Concept and design: Olajide A. Adekunle, Michael B. Rothberg, and Phuc Le. Acquisition, analysis, or interpretation of data: Olajide A. Adekunle, Dev Yash Gupta, Ha T. Tran, and Yihua Yue. Drafting of the manuscript: Olajide A. Adekunle. Critical review of the manuscript for important intellectual content: Michael B. Rothberg, Phuc Le, Christopher Boyer, and Hamlet Gasoyan. Statistical analysis: Olajide A. Adekunle and Christopher Boyer. Supervision: Michael B. Rothberg, Christopher Boyer, and Phuc Le.

Funding

The authors have nothing to report.

Conflicts of Interest

Dr. Rothberg reported receiving consulting fees from the Blue Cross Blue Shield Association outside the submitted work. The remaining authors declare no conflicts of interest.

Supporting information

Table S1: Diagnosis and procedure codes.

Table S2: Names and doses of AOM.

Table S3: Rates and 95% confidence intervals for the receipt of AOM prescriptions and MBS relative to the 2013 declaration of obesity and approval of semaglutide in 2021.

Table S4: Rates and 95% confidence intervals for the receipt of AOM prescriptions (after excluding approved GLP‐1/GIP RAs)

Table S5: Rate and 95% confidence intervals for the receipt of AOM prescriptions, stratified by obesity severity (BMI ≥ 40 kg/m² vs. < 40 kg/m²)

Table S6: Rates and 95% confidence intervals of MBS use, stratified by obesity severity (BMI ≥ 40 kg/m² vs. < 40 kg/m²)

Table S7: Rate and 95% confidence intervals for the receipt of AOM prescriptions, stratified by age (< 40 years vs. ≥ 40 years)

Table S8: Rate and 95% confidence intervals for the receipt of MBS, stratified by age (< 40 years vs. ≥ 40 years)

Table S9: Rate and 95% confidence intervals for the receipt of AOM prescriptions, stratified by sex (female vs. male)

Table S10: Rate and 95% confidence intervals for the receipt of MBS, stratified by sex (female vs. male)

Table S11: Rate and 95% confidence intervals for the receipt of AOM prescriptions, stratified by race/ethnicity (Black, Hispanics, or others vs. White)

Table S12: Rate and 95% confidence intervals for the receipt of MBS, stratified by race/ethnicity (Black, Hispanics, or others vs. White)

Table S13: Rate and 95% confidence intervals for the receipt of AOM prescriptions, stratified by race in 2 groups (non‐White vs. White)

Table S14: Rate and 95% confidence intervals for the receipt of MBS, stratified by race in 2 groups (non‐White vs. White)

Table S15: Model estimates and 95% confidence intervals for the aggregate trends

Table S16: Model estimates and 95% confidence intervals for trends stratified by BMI, age, sex, and race/ethnicity

Table S17: p values for racial/ethnicity differences in trend slopes for AOM and MBS.

Figure S1: Trends in the receipt of AOM prescriptions relative to the 2013 declaration of obesity and approval of semaglutide in 2021, while assuming the absence of all GLP‐1/GIP receptor agonists.

Figure S2: Trends in the receipt of AOM prescriptions, stratified by age (≥ 40 years vs. < 40 years)

Figure S3: Trends in MBS use, stratified by age (≥ 40 years vs. < 40 years)

Figure S4: Trends in the receipt of AOM prescriptions, stratified by gender (female vs. male)

Figure S5: Trends in MBS use, stratified by gender (female vs. male).

Figure S6: Trends in the receipt of AOM prescriptions, stratified by race/ethnicity (Black, Hispanic, or others vs. White).

Figure S7: Trends in MBS use, stratified by race/ethnicity (Black, Hispanic, or others vs. White).

Figure S8: Trends in the receipt of AOM prescriptions, stratified by race (non‐White vs. White).

Figure S9: Trends in MBS use, stratified by race (non‐White vs. White).

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.