Prevalence, predictors, correlates, and dynamic changes in the NIAAA-defined "recovery" definition.

Background: Definitions of alcohol use disorder (AUD) "recovery" abound, but the National Institute on Alcohol Abuse and Alcoholism (NIAAA) was the first to offer an operational definition ("AUD remission with no heavy drinking"). Little is known about this definition or how it compares to others. Greater knowledge would inform its clinical and public health utility.

Design: Multisite, single-group, prospective study of individuals beginning a new AUD recovery attempt (N = 442) assessed at baseline, 3, 6, 9, and 12 months on AUD symptoms, alcohol use, and functioning/well-being. Hidden Markov models (HMMs) estimated NIAAA-defined recovery status and other recovery group statuses/transitions. Regressions tested group membership predictors; linear mixed models examined functioning/well-being.

Results: Participants were classified into four groups depending on remission status and level of alcohol use: (i) Remission + abstinence (R + A; n = 102); (ii) Remission + low-risk drinking (R + LRD; n=51); (iii) Remission + higher-risk drinking (R + HRD; n = 70); (iv) No remission (NR; n = 219). During follow-up, groups 1 + 2 (NIAAA definition) were quite prevalent (34.6%), evincing patterns of larger psychosocial improvements compared with NR and R + HRD. R + A and R + LRD, however, differed on baseline characteristics, with R + A, similar to NR, being more severe on AUD-related metrics, whereas both remitted but still drinking groups (R + LRD, R + HRD) were less severe. Of note, R + A was the most stably remitted group; R + LRD and R + HRD were less stable, moving increasingly into NR. Unlike R + HRD, who, similar to NR, exhibited persistent lower levels of self-esteem and happiness, NIAAA-defined groups were associated with widespread improvements.

Conclusion: The NIAAA recovery definition was consistent with empirical outcomes exhibited by the R + A and R + LRD profiles, with both showing large psychosocial improvements, but who may represent different AUD phenotypes/stages. Despite R + A exhibiting a pattern of historically greater clinical severity, they demonstrated more stable remission compared with both the NIAAA recovery definition-inclusive R + LRD, but especially R + HRD. Ongoing alcohol use appears to destabilize remission, with heavier use associated with less psychosocial improvement and AUD reinstatement.