A Microsoft Excel®-based model was developed to estimate the annual economic impact of obesity and overweight, encompassing both direct medical costs and indirect costs related to productivity loss across the non-farm civilian workforce. The analysis combined data from published reports and studies with original analysis of national databases.

Costs were presented at the national level, per employee with obesity and overweight, and for a hypothetical nationally representative employer with 10,000 employees across all industries as well as the seven major industries—Construction; Education & Health Services; Financial Activities; Government; Manufacturing; Professional & Business Services; and Transportation & Utilities—by combining different cost components associated with excess weight. All data were analyzed using R software version 4.3.2.

Person-level analyses were conducted using pooled adult sample records from the National Health Interview Survey (NHIS) and the Medical Expenditure Panel Survey (MEPS). These analyses compared health and economic outcomes for employed adults with obesity or overweight relative to their peers with a healthy weight (18.5 ≤ BMI < 25) [20]. The NHIS is an annual survey of a nationally representative sample of the U.S. civilian noninstitutionalized population. Data collected include self-reported demographics, health risk factors, and work-related variables (e.g., employment status, industry, missed workdays due to illness or injury, and days with reduced productivity due to illness or injury). The MEPS, a subset of NHIS respondents, includes additional survey questions and medical record extraction to gather data on healthcare utilization and expenditures.

We estimated the prevalence of obesity and overweight among adults by employment status, industry, and occupation using self-reported height and weight from the 2015-2018 combined NHIS files (n = 268,527). Self-reported height and weight can lead to systematic misclassification of BMI categories, and the inherent skewness of BMI distributions may result in underestimation of obesity prevalence [30].

A representative sample of employees from the US nonfarm civilian workforce, derived from the 2018-2021 combined MEPS files (n = 53,577), was used for this analysis. Due to the skewed nature of annual medical expenditures, with many individuals having zero expenditures and some having very large expenditures, we used a two-step process to estimate direct medical costs associated with overweight and obesity. This involved logistic regression and a generalized linear model with a log link applied to the combined dataset [3, 31]. The dependent variable was annual medical expenditures scaled to 2023 dollars using the medical component of the Consumer Price Index. Explanatory variables included body weight category (healthy, overweight, obesity), age group, gender, race, Hispanic ethnicity, marital status, industry, and year. The analysis excluded women who had pregnancy or childbirth during the year. Data limitations prevented identifying covered adult dependents of a person employed in a particular economic sector. Therefore, for modeling, we assume that excess medical costs associated with obesity and overweight among covered dependents is the same as the impact for the employed adult members. Research finds that married couples and domestic partners share many of the same disease risk factors and behaviors (e.g., smoking, diet, and physical activity level) [32, 33]. Furthermore, they generally have similar socioeconomic characteristics and share the same healthcare plan.

Obesity is linked to increased rates of work absenteeism (missed workdays) and presenteeism (reduced productivity while at work), as well as higher payments for disability and Workers’ Compensation programs [11, 13, 14, 34]. Employers may incur direct costs for absenteeism, such as wages paid to absent employees, worker replacement expenses (e.g., overtime pay or temporary worker hiring), and administrative costs. Indirect costs include diminished team productivity, safety concerns, lower service or product quality due to understaffing, and potential burnout or poor morale among employees covering for an absent or less productive coworker [35,36,37]. Overweight is weakly associated with increased absenteeism and was excluded from our cost calculations.

A 2023 study, examining MarketScan data on 719 482 employees with and without obesity between January 2015 and December 2019, estimates that obesity raises annual absenteeism costs by $891 (reaching $1036 for employees with class 3 obesity) relative to employees with healthy weight [11, 34]. Notably, these estimates consider only employee wages when calculating the productivity cost of missed workdays. Previous studies have suggested using wage multipliers to estimate the total economic cost of absenteeism from the employer’s perspective [13, 38,39,40]. This wage multiplier accounts not only for the immediate financial loss associated with absenteeism but also extends its scope to encompass broader repercussions. Our model employs a wage multiplier of 1.97 times the average wage to estimate productivity loss due to absenteeism [13]. Therefore, the $891 average annual cost of obesity-associated absenteeism based on employee wages translates to an average annual cost of $1755 per person with obesity for employers.

To estimate how costs associated with absenteeism vary by industry, we used Poisson regression analysis of pooled 2013–2018 NHIS data (n = 33,216), modeling annual workdays missed due to illness or injury while controlling for age group, race/ethnicity, gender, industry, and year. We then adjusted the national average to account for variation across industries in the number of obesity-related missed workdays and average earnings.

To estimate costs related to presenteeism, we used logistic regression analysis of pooled 2013-2018 NHIS data (n = 53 693), translating limited work capacity into a presenteeism measure. Among employed adults, 7.7% of NHIS respondents with obesity reported being “somewhat limited” in the kind and amount of work they could do due to physical, mental, or emotional problems. In comparison, 5.2% of employees with overweight and 5.0% with healthy weight reported similar limitations. After adjusting for demographics using logistic regression, obesity and overweight are associated with a 2.6 percentage point and 0.9 percentage point increase, respectively, in employees reporting limited work productivity. We use these percentages as a proxy for presenteeism (with results varying by industry).

For comparison, one study estimates diabetes-attributable presenteeism equates to a 6.6% decline in productivity [41]. A 2017 review of nine studies found that presenteeism costs per worker per year ranged from negative $776 to $2020 for overweight (midpoint = $622) and from $14 to $5304 for obesity (midpoint = $2659), relative to a healthy weight population and adjusted to 2023 dollars [14]. These studies used earnings as a proxy for the value of productivity, and when comparing these midpoint estimates to average earnings among the nonfarm civilian workforce, this equates to about a 4.4% decline in productivity associated with obesity and a 1.0% decline associated with overweight. A 2008 study among manufacturing employees (n = 341) reported a 4.2% health-related reduction in productivity among workers with BMI ≥ 35 (class 2 or class 3 obesity), with this loss being 1.2% higher than for other workers (BMI < 35) [42].

The impact of presenteeism measured using lost employee wages will underestimate the cost to employers, as employees’ diminished productivity can result in suboptimal work quality and output. As with modeling absenteeism, a wage multiplier accounts for the immediate financial loss associated with presenteeism and extends to the indirect costs to employers, including factors such as team cohesion, knowledge transfer, and project continuity [13, 43]. Our model employs a published wage multiplier of 1.54 times the average wage to calculate productivity costs due to presenteeism [13].

In the aforementioned study utilizing MarketScan data, obesity is associated with higher annual employer costs of $623 for short-term disability, $41 for long-term disability, and $112 for Workers’ Compensation Program payments per employee with obesity, relative to employees with a healthy weight with costs increasing with each higher BMI category [11, 34]. These findings serve as the foundation for our national estimates regarding the financial burden of disability and workers’ compensation attributable to obesity. Injury risk and associated costs vary across industries due to specific risk factors, demographic disparities, and economic influences. To model variation across industries in disability and workers’ compensation costs, we analyzed NHIS data from 2013 to 2018 using logistic regression to estimate the likelihood of injury claims among employees categorized by weight and industry while controlling for demographics and year of data collection. Further analysis of MEPS data, linked to NHIS, assessed variation in workers’ compensation payments per incident across industries. By combining industry-specific variations in injury risk and compensation costs per incident and applying them to average industry costs for disability and workers’ compensation payments, obesity-related expenses per worker were estimated by industry.

To model the impact of obesity and overweight across industry sectors, we incorporated industry-specific costs, accounting for variations in workforce demographics, earnings, employer insurance coverage, and occupational risks. Key national parameters and model assumptions are summarized in Supplementary Table A1.

To demonstrate the value of treating obesity, we utilized a published computer simulation model, the Disease Prevention & Treatment Microsimulation Model (DPTMM) [44,45,46,47,48,49,50]. Drawing on a combination of public datasets, published studies, and clinical trials, the model allows for detailed predictions of how changes in key biometric markers, such as body weight, blood pressure, cholesterol, and HbA1c, affect the future risk of obesity-related diseases, including type 2 diabetes, hypertension, coronary heart disease, heart attack, and stroke. This allows for the projection of treatment impacts on the reduction of disease incidence and severity and associated direct and indirect costs. The simulation used a constructed population file that is representative of the workforce and adult dependents by industry sector, with adults with obesity in NHIS 1:1 matched to adults with obesity in NHANES based on propensity scores calculated from factors including BMI, age, gender, race/ethnicity, and marital status [51]. NHIS contains data on industry and the presence of select diseases, while NHANES contains data on metabolic markers. The constructed population file is representative of the workforce and adult dependents, by industry sector. The simulation estimated annual clinical improvements and healthcare costs associated with achieving body weight loss of up to 5%, 10%, 15%, 20%, and 25% in the first year and maintaining the effect over the next four years. More information about the DPTMM can be found in the Supplementary.