County-Level Industrial Job Profiles and Substance Use/Misuse/Use Disorders Among U.S. Working-Aged Individuals: A Latent Profile Analysis

Background and Purpose: Industry sectors differ in work environments and norms related to substance use. While higher employment in certain industries may have both direct and indirect effects—through social interaction and community spillovers—on residents’ substance use, less is known about how county-level labor market composition influences these outcomes. This study aims to identify distinct county groups based on their industry-specific job shares, examine their sociodemographic and substance-related characteristics, and assess associations between county group membership and substance use/misuse/use disorder risks among working-aged individuals in the United States.

Methods: We utilized a nationally representative sample of working-aged individuals (15-64) in the United States (n≈273,900) from the restricted-use 2015–2019 National Survey on Drug Use and Health (NSDUH), linked with county-level industry-specific job shares and contextual data from sources, including the U.S. Census Bureau, CDC, and the U.S. Drug Enforcement Administration. Outcomes included past-month substance use/misuse and past-year substance use disorder, separately for alcohol, marijuana, illicit drugs, and prescription drugs. Our analytic strategy involved three steps: (1) latent profile analysis of job share distributions across 3,143 U.S. counties to identify county-level industry composition profiles; (2) comparison of county-level sociodemographic and substance-related characteristics across the identified profiles; and (3) multivariate logistic regression testing associations between county class membership and individuals’ substance use outcomes, adjusting for individual, county, and state-level covariates. Estimates were adjusted for NSDUH’s complex sampling design to ensure generalizability.

Results: We identified seven distinct county groups: Class 1 (“Manufacturing”), Class 2 (“Agriculture & Natural Resources”), Class 3 (“Healthcare & Trades”), Class 4 (“Sales”), Class 5 (“Extractions”), Class 6 (“Hospitality & Entertainment”), and Class 7 (“Knowledge & High-Risk Services”). County groups differed in sociodemographic and substance-related policy characteristics. For example, Class 2 and Class 5—mainly in the West and South—had the highest proportion of Hispanic population (31.0%), the lowest rates of high school attainment (80.5%), and less active substance-related policies and activities. Class 6—mostly in the West—had the highest rate of single-parent households (36.3%) and marijuana legalization (73.3% for medical use and 33.3% for recreational use). Class 7 had the highest African American population (12.3%), high school attainment (90.9%), and median household income ($75,718) as well as the highest rates of having prescription drug monitoring laws. Logistic regressions showed that individuals residing in Class 2 and Class 5 had higher odds of binge drinking and alcohol use disorders compared to residents in Class 1. Residents in Class 4 and Class 6 were at greater risk for marijuana use and past-year marijuana use disorder. Class 7 residents had elevated risks for use/misuse/use disorders involving marijuana, illicit drugs, and prescription drugs.

Conclusions and Implications: Distinct labor market compositions not only reflect variation in sociodemographic and policy contexts but also correspond to varying risks for substance use and related disorders. These findings highlight the importance of tailoring prevention and treatment strategies to local labor market contexts to enhance their effectiveness and promote more equitable health outcomes across diverse U.S. counties.