Persistence/recurrence of and remission from DSM-5 substance use disorders in the United States: Substance-specific and substance-aggregated correlates

Abstract

This study examined demographic and psychosocial correlates associated with persistence/recurrence of and remission from at least one of ten DSM-5 substance use disorders (SUDs) and three substance-specific SUDs (i.e., alcohol, cannabis, and prescription opioids). Data were collected from structured diagnostic interviews and national prevalence estimates were derived from the 2012–2013 National Epidemiologic Survey on Alcohol and Related Conditions. An estimated 25.4% of the U.S. population had at least one prior-to-past-year (prior) SUD. Among individuals with any prior SUDs, the prevalence of past-year substance use and DSM-5 symptomology was as follows: abstinence (14.2%), asymptomatic use (36.9%), symptomatic use (10.9%), and persistent/recurrent SUD (38.1%). Among individuals with prior SUDs, design-based multinomial logistic regression analysis revealed that young adulthood, higher educational attainment, higher personal income, never having been married, being divorced/separated/widowed, lack of lifetime substance use treatment, and stressful life events predicted significantly greater odds of past-year persistent/recurrent SUDs, relative to abstinence. In addition, remission from a prior tobacco use disorder decreased the probability of past-year persistent/recurrent SUD, relative to abstinence. Stressful life events were the only common correlates across the aggregation of all SUDs and each substance-specific SUD, but differences were found for specific stressful life events between drug classes. Nearly half (49%) of adults with prior DSM-5 SUDs continued to report past-year symptomatic substance use, while only one in seven individuals were abstinent. The findings suggest the value of examining remission associated with both substance-specific SUDs and aggregation of SUDs based on the shared and unique correlates of persistent/recurrent SUDs; this is especially true for stressful life events, which could be useful targets for enhancing clinical care and interventions.