Assessing Niaaa's Operational Definition of Recovery from Alcohol Problems: A Latent Class Analysis of a Diverse Online Sample

Background: In 2020, the National Institute on Alcohol Abuse and Alcoholism (NIAAA) issued an operational definition of recovery from alcohol use disorder (AUD), which specifies that recovery is based on remission from DSM-5 AUD criteria (aside from the craving item) and cessation of heavy drinking (i.e., three drinks/day or seven/week or more for women and four drinks/day or 14/week or more for men). Thus far, has examined the extent to which people who self-identify as having resolved alcohol problems meet these criteria. This study aims to address this gap by examining rates and patterns of heavy drinking and AUD symptom endorsement in a diverse national sample of people in recovery from alcohol problems, inclusive of individuals in natural recovery (i.e., no utilization of treatment or self-help groups).

Methods: Participants were recruited online through Amazon MTurk as part of a larger recovery study. To be eligible, participants self-identified as having resolved a problem with alcohol (alone or with other drugs) for six months or more. The sample (n=385) included 44% women, 28% participants of color (8% Black, 8% other races, and 17% Latinx ethnicity), and 30% in natural recovery. After qualifying through an eligibility screening, participants completed a survey including questions about alcohol problem severity and treatment and recovery history; current drinking; and current and historical DSM-5 AUD criteria endorsement. After data cleaning, the research team conducted a latent class analysis to identify classes based on heavy drinking and AUD symptom endorsement and covariates associated with these classes.

Results: Each of the 10 AUD criteria (excluding craving) was endorsed by 3 to 8% of participants in the past three months, and 34% reported heavy drinking during this period even though all participants identified as having resolved their alcohol problems for at least six months. A two-class solution was the best fit for the data (-2LL=822.24; Entropy=.89; VLMR=8.36, p<.001). In class one (15% of sample; “high risk”), 45% of participants experienced at least one AUD criterion in the past three months, and one-third reported heavy drinking in this timeframe. The most commonly endorsed DSM-5 criteria were drinking more than intended, unsuccessfully trying to stop drinking, and spending a lot of time drinking or hungover. In contrast, class two (85% of sample; “low risk”) rarely (<5% of the time) endorsed any AUD criteria and reported no recent heavy drinking. Membership in the high-risk class was significantly associated with Black racial identity, being married, having been to jail, and full-time employment. Natural recovery was not significantly associated with class membership.

Conclusions/Implications: Approximately one-third of the sample did not meet the NIAAA recovery definition, due to lack of remission from AUD criteria and/or heavy drinking. In particular, the findings regarding heavy drinking offer provisional support to recent arguments in the literature suggesting that some individuals may persist in recovery despite continued heavy drinking. Future research should examine the pathways through which covariates such as race, marriage, and incarceration potentially influence risk group membership, with recovery capital being a proposed mechanism.