The direct costs of substance use disorders (SUDs) in the United States are incurred primarily among the working-age population. During 2001 to 2020, more than 90% of deaths (865 732 of 932 573) caused by drug or alcohol poisoning and nearly 80% of nonfatal poisoning emergency department visits (20.1 million of 25.7 million) occurred among people aged 20 to 64 years.^1,2 Overall, 11% of US-employed adults (aged ≥18 years) reported having SUDs in 2020.³ More than half of employers with 50 or more employees surveyed in 2019 reported direct effects from the opioid overdose epidemic, such as impaired or decreased job performance.⁴ Previous estimates of SUD medical costs among US employees are well described but relied on workers’ self-reported health care utilization.⁵ Quantifying the medical cost of SUDs in the employer-sponsored insurance (ESI) population through analysis of actual financial transactions can highlight how SUDs affect workplaces and inform decision-making on the value of prevention strategies. This study aimed to estimate the annual attributable medical cost of SUDs in the ESI population from the health care payer perspective.

This economic evaluation did not require evaluation by an institutional review board or patient informed consent because all data were publicly available and no human participants were involved (45 CFR part 46). This study followed Consolidated Health Economic Evaluation Reporting Standards (CHEERS) reporting guidelines for economic evaluations.⁶ Merative MarketScan 2018 Commercial Claims and Encounters Databases were weighted to represent the non–Medicare eligible (aged <65 years) US ESI population (active employees, early retirees, and COBRA continuees plus dependents; Medicare supplemental plans not included). The databases report expenditures for inpatient services, outpatient services, and outpatient drugs from approximately 350 health insurance payers (large employers, health plans, and government and public organizations). Race and ethnicity are not reported in this database. By comparing total annual medical expenditures among similar enrollees with and without SUD diagnoses (referred to as control participants), this analysis aimed to estimate the cost that could be eliminated through prevention or successful treatment of SUD (referred to as the attributable cost). The main outcome measure was the annual attributable SUD medical cost in the ESI population overall and by substance type (eg, alcohol). The number of enrollees with an SUD diagnosis in the ESI population and the annual mean attributable cost of SUD diagnosis (any and by substance type) per affected enrollee are also reported. SUD diagnosis was defined by International Statistical Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) according to Clinical Classifications Software categories (alcohol-, cannabis-, hallucinogen-, inhalant-, opioid-, sedative-, stimulant-, and other substance-related disorders) on an inpatient or outpatient claim record (Table 1). Costs are reported in 2018 US dollars (unadjusted from data source); cost discounting was not applicable due to the 1-year analytic time horizon, which was selected to facilitate reporting in terms of annual cost, and the cost perspective was the health care payer.

Among 162 million 2018 ESI enrollees, 2.3 million (1.4%) had an SUD diagnosis (Table 2).⁹ The regression analysis sample included 210 225 individuals with an SUD diagnosis (121 357 [57.7%] male individuals; 68 325 [32.5%] aged 25-44 years) and 1 049 539 individuals with no SUD diagnosis (Table 3). This sample was used to model per-enrollee SUD costs, and the mean total annual expenditure was $26 051 (95% CI, $25 775-$26 327) among all enrollees with SUD diagnosis and $10 405 (95% CI, $10 331-$10 479) among all enrollees without SUD diagnosis. More than one-third (72 235 [34%]) of analysis sample enrollees with an SUD diagnosis had at least 2 comorbidities. The most common comorbidities among enrollees with an SUD diagnosis were depression (63 049 [30.0%]), hypertension (62 582 [29.8%]), and obesity (35 823 [17.0%]).

Regression modeling that compared enrollees with SUD diagnoses and control participants estimated the annual attributable mean cost of any SUD diagnosis (including multiple) was $15 640 (95% CI, $15 340-$15 940) per affected enrollee, and the total estimated medical cost in the ESI population was $35.3 billion (Table 4). Among ESI enrollees with an SUD diagnosis, alcohol-related disorders ($10.2 billion) and opioid-related disorders ($7.3 billion) were the most expensive. Any SUD diagnosis (including multiple) was associated with nonzero attributable per-enrollee costs in all assessed service categories (inpatient: $6502 [95% CI, $6333-$6671]; outpatient: $8628 [95% CI, $8442-$8814]; outpatient drugs: $1503 [95% CI, $1412-$1594]). Those results appear to be influenced by the prevalence of alcohol-related disorders and opioid-related disorders; more than half of enrollees with an SUD diagnosis had an alcohol-related disorder and nearly 30% had an opioid-related disorder (Table 2). Each of those SUD types was associated with relatively high mean per-affected enrollee, per-SUD type inpatient ($3988 [95% CI, $3837-$4140] and $3570 [95% CI, $3336-$3804], respectively), outpatient (95% CI, $4875 [95% CI, $4716-$5033] and $6280 [95% CI, $5976-$6585]), and overall per affected enrollee ($8939 [95% CI, $8680-$9197] and $11 871 [95% CI, $11 317-$12 424]) costs (Table 4).

This study’s primary contribution is quantifying the annual medical cost of SUDs in the ESI population based on recent financial transactions. In assessing total annual expenditures (inpatient, outpatient, outpatient drugs) among ESI enrollees with SUD, this analysis expands what is known about attributable costs in the ESI population compared with recent research that used hospital discharge data to analyze only inpatient costs.⁷ The inpatient SUD cost estimates from the present study of the ESI population are higher than comparable inpatient SUD costs among patients with all health care payers in the previous study. This is primarily because modeling here addressed not just the cost difference that was associated with SUDs among patients treated in hospitals (eg, costs for inpatients with and without SUD diagnoses being treated for the same heart condition), but also the cost difference for the entire hospital encounter if associated with SUDs based on comparing individuals with the same demographic characteristics, health plan type, and comorbidities, but different SUD diagnosis status. The $35.3 billion SUD medical cost represents a small fraction of total US personal health care expenditures paid by private insurance ($1.1 trillion in 2018).¹⁰ However, in this study, 1% of the ESI population had an SUD diagnosis compared with 11% of workers who self-report SUD,³ suggesting the medical cost that employers and their health insurance payers face is likely far higher than reported here. Employers can take action by developing workplace supported prevention, treatment, and recovery programs.¹¹

In this economic evaluation, we found that the estimated annual medical cost of SUDs in the non–Medicare eligible US ESI population was $15 640 per affected enrollee and $35.3 billion in total during 2018. The cost of strategies to support employees and their health insurance dependents to prevent and treat SUDs can be considered in terms of potentially offsetting the existing high medical cost of SUDs. Medical expenditures for SUDS represent the minimum direct cost that employers and health insurers face because not all people with SUDs have a diagnosis, and costs related to absenteeism, presenteeism, job retention, and mortality are not addressed.