Alcohol Use Disorder: Recovery, Outcomes, and the Long Road Forward

A comprehensive knowledge base article synthesizing expert panel discussion for clinicians, researchers, and people affected by AUD

Overview — Recovery Is Real and Common

Here is the most important thing to know about alcohol use disorder (AUD) and recovery: most people with AUD do get better over time. This is not wishful thinking. It is what large-scale population data show.

Using the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC-III), researchers found that among individuals with prior-year AUD, only 34.2% had persistent AUD — meaning roughly two-thirds had moved into some form of improved status ^[1]. Approximately 70% of persons with AUD improve without formal interventions, and fewer than 25% ever use alcohol-focused services ^[2] (Note: this specific figure could not be independently verified against the source abstract — the underlying study supports the general finding but the exact number should be confirmed before publication). Recovery is not rare. It is not reserved for people who "hit bottom" or find the right program. It is the statistical norm across the population.

At the same time, relapse is common — and it is not failure. AUD is a chronic, relapsing condition for many people, much like diabetes or hypertension. Relapse is part of the natural history of the disorder, not evidence that a person is beyond help or that prior treatment was wasted. One study found that among people who relapsed after sustained remission, the average time in remission before relapse was 3.6 years, with a range extending up to 23 years ^[3] (Note: this specific figure could not be independently verified against the source abstract — the underlying study supports the general finding but the exact number should be confirmed before publication). People can and do recover after multiple relapses. The trajectory is rarely a straight line.

This article holds both truths at once: recovery is real and common, and the path is often nonlinear, shaped by severity, social resources, structural barriers, and time.

Terminology — Recovery, Remission, Abstinence

Precision in language matters here — both for research and for the conversations clinicians have with patients and families. These terms are not interchangeable.

Recovery is the broadest term. The NIAAA recently formalized an operational definition requiring two components: (1) remission from DSM-5 AUD symptoms (except craving) and (2) cessation of heavy drinking — critically, not requiring full abstinence ^[4]. This is a paradigm shift from older, abstinence-only frameworks. Non-abstinent recovery — drinking at low-risk levels without meeting AUD criteria — is explicitly recognized as a legitimate outcome.

Remission is a DSM-5 clinical term with two time-based specifiers:
- Early remission: 3 to 12 months without meeting AUD criteria (except craving)
- Sustained remission: 12 months or longer without meeting criteria

Abstinence means no alcohol use at all. It is one valid recovery pathway, but not the only one.

Controlled drinking (also called non-abstinent recovery or low-risk drinking) refers to drinking below heavy-use thresholds without meeting AUD diagnostic criteria. NESARC-III data show that 17.9% of individuals with prior-year AUD achieved asymptomatic low-risk drinking — a non-abstinent recovery state — while 16.0% were fully abstinent ^[1]. Non-abstinent recovery is not a consolation prize. It is a documented, common outcome.

Lapse refers to a single episode of drinking after a period of abstinence or controlled use. Relapse refers to a sustained return to problematic drinking. These are not the same thing, and treating them as equivalent causes real harm — both clinically and in how people in recovery talk to themselves after a slip.

WHO Risk Drinking Level reductions represent an emerging continuous outcome framework. Rather than asking only "did you drink or not?", this approach measures movement between risk categories (high, medium, low, abstinent) as clinically meaningful progress — even without full abstinence. It changes how we count success.

Realistic Trajectories — What the Data Show

What does recovery actually look like over time? The honest answer is: variable, nonlinear, and more hopeful than clinical settings often suggest.

Cross-sectional snapshot (NESARC-III): Among individuals with prior-year AUD, the distribution of outcomes was as follows — 34.2% persistent AUD, 21.5% asymptomatic high-risk drinkers, 17.9% asymptomatic low-risk drinkers (non-abstinent recovery), and 16.0% abstinent ^[1]. Roughly one-third were still struggling; two-thirds had improved in some way.

Treatment dose and 3-year outcomes: Project MATCH data show that attending all 12 CBT or TSF sessions was associated with significantly fewer heavy drinking days and alcohol-related consequences at all post-treatment time points, including 3 years post-treatment, compared to attending 0–2 sessions ^[5]. What happens during treatment has durable effects — but only if the dose is adequate.

8-year community outcomes: In one of the longest follow-up studies in the corpus, Humphreys et al. found that AA attendance in the first 3 years predicted remission, lower depression, and higher quality relationships with friends and partners at 8 years ^[6]. Critically, inpatient treatment days showed no independent relationship to 8-year remission or psychosocial outcomes. Community-based, sustained support outperformed acute clinical intervention on long-term outcomes.

Help-seeking multiplies recovery odds: NESARC Wave 1 data show that only one-quarter of individuals with alcohol dependence ever sought help. But help-seeking increased the likelihood of any recovery (hazard rate ratio = 2.38) and abstinent recovery (HRR = 4.01). Adding 12-Step participation to formal treatment nearly doubled recovery odds compared to formal treatment alone ^[7].

Recovery is rarely linear. Long-term relapse — after years of sustained remission — is real. The most prevalent and potent risk factor identified in one study was not craving or neurophysiology, but change in recovery vigilance ^[3]. Vigilance is a social and psychological construct. It is modifiable. And it requires ongoing attention, not just early-recovery effort.

Lapse vs. Relapse — A Distinction That Matters

One of the most clinically consequential distinctions in AUD recovery is the difference between a lapse and a relapse — and the field has not always made this distinction clearly enough.

A lapse is a single episode of drinking. A relapse is a sustained return to problematic drinking patterns. Most lapses do not become relapses. Treating every slip as catastrophic failure — in clinical framing or in a patient's own self-talk — can actually increase the risk of full relapse by triggering shame, hopelessness, and disengagement from support.

The corpus does not contain a dedicated lapse-vs-relapse study, which the expert panel flagged as a genuine gap. However, the data on residual symptoms are instructive: even one residual DSM-5 AUD symptom at 6 months post-treatment tripled the odds of a subsequent "slip" at 12 months (OR: 3.7, 95% CI: 1.5–9.0) in older adults ^[8]. This argues for ongoing monitoring after treatment ends — not because a slip is inevitable, but because residual symptom burden is a real and measurable risk signal.

For clinicians: a lapse is a clinical event that warrants engagement, not judgment. For people in recovery: one drink does not undo years of work. The question after a lapse is not "did I fail?" but "what do I do next?"

Pathway Pluralism — Multiple Roads to Recovery

There is no single correct path to recovery from AUD. The evidence supports multiple pathways, and the most important clinical task is often matching the right pathway to the right person at the right time — not insisting on one approach for everyone.

Abstinence-focused pathways include 12-Step programs (AA, TSF), residential treatment, and medication-assisted treatment (MAT) aimed at supporting full abstinence. These pathways have the strongest evidence base for severe AUD. Among those achieving abstinent recovery in NESARC-III, 43.2% had received formal treatment — compared to only 12.3% of non-abstinent recoverers ^[1]. This suggests that abstinence-focused, treatment-supported pathways are particularly important for higher-severity presentations.

Moderation-focused pathways include Moderation Management, controlled drinking programs, and harm reduction approaches. These are not appropriate for everyone — particularly not for those with severe AUD or medical contraindications to any alcohol use — but they are legitimate and evidence-supported for lower-severity presentations. The NESARC-III data confirm that non-abstinent recovery is a common real-world outcome ^[1].

Harm reduction pathways include managed drinking approaches and the Sinclair Method (targeted naltrexone use to reduce drinking rather than achieve abstinence). These approaches recognize that any reduction in alcohol-related harm is clinically meaningful, even without full abstinence.

Natural recovery — improvement without any formal treatment — is the most common pathway statistically ^[2] (Note: this specific figure could not be independently verified against the source abstract — the underlying study supports the general finding but the exact number should be confirmed before publication). However, it is concentrated among lower-severity presentations. It should not be used as a rationale for withholding treatment from those who need it.

The key clinical principle: pathway pluralism is not relativism. Not every pathway is equally appropriate for every person. Severity, comorbidities, social context, and patient preference all inform the match.

WHO Risk Drinking Levels as Outcomes

There is a meaningful shift underway in how clinical trials and treatment programs measure success ^[4]. Traditionally, the primary outcome in AUD treatment research was binary: abstinent or not. This framework misses a great deal of clinically meaningful change.

The WHO Risk Drinking Level framework categorizes drinking by risk level — high, medium, low, and abstinent — and measures movement between categories as a meaningful outcome. A person who moves from high-risk to medium-risk drinking has achieved something real, even if they have not achieved abstinence.

This matters for several reasons. First, it aligns with how the NIAAA recovery definition works — recognizing non-abstinent outcomes as legitimate ^[4]. Second, it broadens how trials count success, potentially revealing treatment benefits that binary abstinence outcomes would miss. Third, it validates the experience of people in recovery who are drinking less and living better, even if they have not stopped entirely.

Reductions in WHO Risk Drinking Level correlate with improvements in AUD diagnostic status and reduced criterion burden. This is not a lowering of standards — it is a more accurate measurement of what treatment actually accomplishes for a heterogeneous population.

Maintenance Pharmacotherapy — How Long?

The expert panel identified a significant gap in the document corpus: almost no evidence on pharmacotherapy duration, discontinuation criteria, or re-treatment after relapse was present in the reviewed documents [noted by Dr. Addiction, Round 3]. This is a genuine limitation of this knowledge base, and clinicians should consult ASAM guidelines and NIAAA clinical resources for pharmacotherapy-specific guidance.

What the corpus does support is the chronic disease framing that should govern pharmacotherapy decisions. AUD is not an acute condition that resolves after a treatment episode. Residual symptoms at 6 months independently predict relapse at 12 months ^[8]. Negative emotionality remains a dynamic risk factor throughout the first year of recovery ^[9]. These findings argue for sustained clinical engagement — including pharmacotherapy where indicated — rather than discontinuation after initial stabilization.

The general clinical consensus (not from this corpus, but from ASAM and NIAAA guidelines — noted here as general knowledge) supports:
- Naltrexone: minimum 6–12 months; some patients benefit from longer-term maintenance
- Acamprosate: similar duration considerations
- Disulfiram: requires sustained adherence to maintain deterrent effect

The question of when to discontinue pharmacotherapy and how to manage re-treatment after relapse cannot be answered from this document corpus. This is an honest gap.

Recovery Capital — The Infrastructure of Recovery

Recovery capital is one of the most important and actionable concepts in this field. Developed by Granfield and Cloud, the framework describes the aggregate of personal and social resources that support and sustain recovery. It includes four domains: social capital (relationships, social networks), human capital (education, employment, health), financial capital (economic resources), and community capital (access to recovery-supportive environments and services).

Recovery capital is not just a theoretical construct. It is measurable and it predicts outcomes.

• Total recovery capital scores were associated with significantly greater odds of meeting NIAAA recovery criteria (OR = 1.61, p = .001) ^[10]
• Individual recovery capital domains were differentially predictive depending on how long someone had been in recovery — meaning the resources that matter most shift over time ^[10]
• During the COVID-19 pandemic, recovery capital showed consistent protective effects against relapse for both women (aOR 0.90) and men (aOR 0.93) ^[10]

The duration-specificity finding is clinically important. What a person needs at six months of recovery is not what they need at six years. Early recovery may require intensive social support and structured programming; sustained recovery may require meaningful roles, purpose, and community contribution. Qualitative research supports this arc: receiving support was crucial in early recovery, while providing support was important for sustained recovery ^[11].

Community-level factors also matter. Lower health insurance rates, greater income inequality, lower education, and lower income at the community level were each independently associated with lower-functioning recovery profiles three years post-treatment ^[12]. Recovery capital is not just individual — it is structural. You cannot randomize someone out of their zip code.

Recovery Community Organizations

Recovery Community Organizations (RCOs) are peer-led, community-based organizations that provide recovery support services outside of clinical settings. They are recognized by SAMHSA as a key component of a Recovery Oriented System of Care (ROSC). Examples include Faces & Voices of Recovery and Young People in Recovery.

RCOs serve populations that clinical systems routinely fail to reach — including communities of color, people with limited insurance, and those who never engage with formal treatment. Qualitative research from Detroit's Black community documents that Black Americans used blended pathways including religion and spirituality that mainstream recovery systems largely ignore ^[11]. RCOs are often better positioned to meet people in these pathways than clinical programs.

However, the expert panel was honest about a significant limitation: no controlled studies exist for recovery community centers specifically ^[13]. The evidence base for RCOs as distinct organizational entities — their structure, funding models, outcomes, and reach — is largely absent from this corpus. This is a genuine gap, not a reason to dismiss RCOs, but a call for better research.

Peer Recovery Support

Peer recovery support specialists (PRSS) — people with lived experience of AUD who are trained and certified to support others in recovery — represent one of the most promising and rapidly expanding components of the recovery support system.

The systematic review by Day et al. found that peer-delivered interventions were associated with decreased relapse rates, reduced re-admission, increased engagement, and increased social support ^[13]. These are not marginal findings. They represent an emerging evidence base for a model that extends recovery support well beyond the clinical contact window.

Medicaid reimbursement for peer recovery support services has expanded significantly in recent years, creating a sustainable funding pathway for this workforce. The RC-Link trial ^[14] is testing peer recovery coaching initiated during inpatient hospitalization, with predicted mechanisms of change being social support and self-efficacy. This is exactly the bridge between clinical settings and community infrastructure that the field has needed — but outcome data from this trial are not yet available in this corpus.

The theoretical mechanism is sound: peer specialists provide social support, model recovery identity, and maintain connection during the high-risk post-discharge period when clinical contact drops off. The outcome data, while promising, need longer follow-up studies to fully establish effectiveness.

Sober Living and Recovery Housing

Recovery housing — including the Oxford House model and transitional recovery housing programs — provides a structured, substance-free living environment that supports early and sustained recovery. It is a concrete form of recovery capital: stable, recovery-supportive housing is both a resource in itself and a platform for building other resources.

The Day et al. systematic review found that recovery housing studies showed positive results, with significant differences from standard care ^[13] (Note: this specific figure could not be independently verified against the source abstract — the underlying study supports the general finding but the exact number should be confirmed before publication). The evidence base is modest but consistently positive on housing stability and recovery outcomes.

Recovery housing is particularly important for people leaving residential treatment or incarceration, who face the highest relapse risk in the immediate post-discharge period. The structural logic is straightforward: returning to an environment saturated with drinking cues and social pressure to drink undermines even the most motivated recovery effort. Recovery housing changes the environment, not just the individual.

Re-Treatment After Relapse

Relapse is not the end of the recovery story. It is a clinical event that calls for re-engagement, not abandonment.

The chronic disease framing is essential here. We do not tell a person with diabetes that their insulin is "not working" because their blood sugar rose after a period of good control. We adjust the treatment plan. The same logic applies to AUD. Relapse after a period of remission is expected in the natural history of a chronic condition — and re-treatment is the appropriate clinical response.

The evidence supports this framing. Among people who relapsed after sustained remission, the mean years in remission before relapse was 3.6 years ^[3] (Note: this specific figure could not be independently verified against the source abstract — the underlying study supports the general finding but the exact number should be confirmed before publication). Those years of remission were not wasted. They represent real health gains, reduced harm, and a foundation for re-engagement.

Integrated multidisciplinary care for alcohol-associated liver disease — a high-severity, high-stakes population — produced significant reductions in active severe AUD (from 85.2% to 51.9%, p<0.001) and reduced emergency department utilization ^[15] (Note: this specific figure could not be independently verified against the source abstract — the underlying study supports the general finding but the exact number should be confirmed before publication). Even in the most medically complex presentations, re-engagement with treatment produces meaningful outcomes.

The clinical message: re-treatment after relapse is not a sign of treatment failure. It is the expected practice of chronic disease management.

Quality of Life and Functional Recovery

Recovery is not only about drinking. It is about living.

The field has increasingly shifted from asking "did you drink?" to asking "how are you functioning?" — employment, housing stability, family relationships, physical health, mental health, and sense of purpose. This shift from "recovery from" to "wellness in" recovery reflects a more complete understanding of what people actually need and want from the recovery process.

The 8-year data from Humphreys et al. are instructive here: AA attendance in the first 3 years predicted not just remission but lower depression and higher quality relationships with friends and partners at 8 years ^[6]. Recovery, when sustained, improves the whole person — not just the drinking behavior.

Recovery capital domains — employment, housing, relationships, community connection — are both outcomes of recovery and inputs to it ^[10]. This bidirectionality matters: investing in functional recovery (helping someone get stable housing, reconnect with family, find meaningful work) is not separate from treating AUD. It is treating AUD, through the lens of chronic disease management.

Identity and Long-Term Recovery

One of the most important — and least studied — dimensions of sustained recovery is identity transformation. Research by Best, Granfield, and others suggests that long-term recovery often involves a fundamental shift in how a person sees themselves: from "person with a drinking problem" to "person in recovery" to, eventually, simply a person living a meaningful life.

This identity shift is not cosmetic. It may be one of the primary mechanisms by which recovery is sustained over years and decades. The qualitative research in this corpus gestures at this: facilitators of recovery shift by stage, with receiving support being crucial in early recovery and providing support being important for sustained recovery ^[11]. The transition from recipient to contributor is an identity transition as much as a behavioral one.

The most prevalent and potent risk factor for long-term relapse — change in recovery vigilance ^[3] — may itself be a proxy for identity. When a person's recovery identity weakens, vigilance drops. When vigilance drops, relapse risk rises. The mechanism connecting identity to outcome is theoretically compelling but empirically underexplored in this corpus.

The expert panel was unanimous: the identity transformation literature is a significant gap in the current evidence base. This is an area where qualitative and mixed-methods research is essential, and where the corpus is thin.

Mortality — Recovery Reduces Death Risk

AUD is a life-shortening condition. It is associated with liver disease, cardiovascular disease, cancer, injury, and suicide. Sustained recovery reduces all-cause mortality — this is the clinical urgency underlying everything in this article.

The corpus does not contain dedicated mortality outcome data for AUD recovery populations — a gap flagged explicitly by the epidemiologist on the panel. However, the integrated care data are suggestive: multidisciplinary care for alcohol-associated liver disease produced significant reductions in active severe AUD and emergency department utilization ^[15], outcomes that carry direct mortality implications in a population with high short-term death risk.

The absence of mortality data in this corpus is an honest limitation. Clinicians counseling patients with AUD and significant medical comorbidities — liver disease, cardiovascular disease, pancreatitis — should draw on mortality data from other sources (NIAAA, WHO, and published cohort studies) to communicate treatment urgency. The evidence that recovery improves survival is strong in the broader literature, even if it is not represented in this specific document corpus.

Racial and Structural Disparities in Recovery

Recovery is not equally accessible to everyone. The evidence documents significant disparities that are structural, not individual.

NESARC-III data show that odds of achieving recovery were lower among non-Hispanic Black individuals relative to persistent AUD, while women and married individuals had greater odds of recovery ^[1]. Whites are proportionately more likely to utilize alcohol-focused services than Black and Hispanic individuals ^[2].

Qualitative research from Detroit's Black community illuminates the mechanisms: delayed recovery initiation due to systemic barriers, chronic early-recovery relapse cycles coupled with inadequate support, and reliance on blended pathways including religion and spirituality that mainstream recovery systems largely ignore ^[11]. These are not individual failures. They are structural failures — racism operating at personal, interpersonal, and societal levels.

Community-level social determinants compound individual-level disadvantage. Lower health insurance rates, greater income inequality, and lower community income were each independently associated with lower-functioning recovery profiles three years post-treatment ^[12]. A person cannot recover their way out of structural poverty or systemic racism. Recovery support systems must be designed with this reality in mind.

Evidence Gaps — What This Knowledge Base Cannot Answer

Honest acknowledgment of what we do not know is as important as what we do. The expert panel identified the following gaps:

Long-term population-level trajectory data. The corpus contains no nationally representative longitudinal study tracking recovery trajectories — including relapse, re-entry, and sustained remission — across a decade or more. NESARC-III is cross-sectional [corpus-gap]. Project MATCH provides 3-year follow-up but from a treatment sample ^[5]. We cannot answer from this corpus what recovery looks like at 10 or 20 years in a representative population.

Pharmacotherapy evidence. The corpus contains almost no evidence on medication-assisted treatment — naltrexone, acamprosate, disulfiram — their role in maintenance, optimal duration, or re-treatment protocols after relapse. This is a profound gap for a chronic disease management framework.

Lapse-to-relapse transition. The corpus does not contain studies specifically examining what distinguishes a single heavy drinking episode from full relapse, or what interventions interrupt the lapse-to-relapse progression. This is arguably the most clinically actionable unanswered question in sustained recovery management.

Recovery community organization outcomes. No controlled studies exist for recovery community centers specifically ^[13]. The RCO evidence base is largely absent from this corpus.

Identity transformation mechanisms. The corpus does not address how recovery identity forms and sustains over time, or whether identity shift is itself a mechanism of long-term maintenance. This is a qualitative and mixed-methods research gap.

Racial disparity mechanisms and trajectories. The corpus documents disparities ^{[1] [11]} but contains no quantitative longitudinal data on whether and how racial gaps in recovery outcomes close or widen over time.

Severity-stratified natural recovery rates. The corpus does not provide clean, severity-stratified natural recovery rates. The inference that natural recovery is concentrated among lower-severity presentations is well-supported but not directly demonstrated ^[1].

Summary — What to Take Forward

Recovery from AUD is real, common, and achievable through multiple pathways. The majority of people with AUD improve over time — many without formal treatment, though treatment substantially increases the odds of recovery, particularly for severe presentations ^[2] (Note: this specific figure could not be independently verified against the source abstract — the underlying study supports the general finding but the exact number should be confirmed before publication).

Recovery is not a single event. It is a long-term process shaped by severity, social resources, structural context, and time. Relapse is part of that process for many people — not a sign of failure, but a signal for re-engagement [corpus-gap].

Recovery capital — the aggregate of social, human, financial, and community resources — is the most consistently documented modifiable predictor of sustained recovery across multiple study designs and populations [corpus-gap]. Investing in recovery capital is not separate from treating AUD. It is treating AUD.

Community-based, long-duration support — including mutual aid, peer recovery support, and recovery housing — predicts long-term outcomes in ways that acute clinical intervention alone does not ^{[6] [13]}. The chronic disease model demands sustained engagement, not discharge after stabilization.

And finally: recovery is not just the absence of drinking. It is the presence of a meaningful life. The goal is not sobriety as an end in itself, but health, connection, purpose, and the full human experience that AUD can take away — and that recovery can restore.

This article was produced by synthesizing a multi-expert panel discussion grounded in verified research documents. All citations reflect papers cited in that discourse. Gaps in the evidence base are noted explicitly throughout. Clinicians should supplement this resource with current ASAM guidelines, NIAAA clinical resources, and emerging pharmacotherapy literature not represented in this corpus.

Verified References

• ^[8] Behrendt, Silke, Kuerbis, Alexis, Braun-Michl, Barbara et al. (2021). "Residual alcohol use disorder symptoms after treatment predict long-term drinking outcomes in seniors with DSM-5 alcohol use disorder.". Alcohol Clin Exp Res. DOI: 10.1111/acer.14722 [abstract-verified: yes]
• ^[14] Byrne, Kaileigh A, Pericot-Valverde, Irene, Ross, Lesley A et al. (2025). "The Peer Recovery Coaching Linkage (RC-link) intervention study: Protocol for a randomized controlled trial for alcohol use disorder recovery in the inpatient hospital setting.". Contemp Clin Trials. DOI: 10.1016/j.cct.2025.108097 [abstract-verified: partial]
• ^[7] Dawson, Deborah A, Grant, Bridget F, Stinson, Frederick S et al. (2006). "Estimating the effect of help-seeking on achieving recovery from alcohol dependence.". Addiction. DOI: 10.1111/j.1360-0443.2006.01433.x [abstract-verified: yes]
• ^[13] Day, Ed, Pechey, Laura Charlotte, Roscoe, Suzie et al. (2025). "Recovery support services as part of the continuum of care for alcohol or drug use disorders.". Addiction. DOI: 10.1111/add.16751 [abstract-verified: partial]
• ^[1] Fan, Amy Z, Chou, Sanchen Patricia, Zhang, Haitao et al. (2019). "Prevalence and Correlates of Past-Year Recovery From DSM-5 Alcohol Use Disorder: Results From National Epidemiologic Survey on Alcohol and Related Conditions-III.". Alcohol Clin Exp Res. DOI: 10.1111/acer.14192 [abstract-verified: yes]
• ^[9] Garber, Molly L, Belisario, Kyla L, Doggett, Amanda et al. (2026). "Relations between impulsivity and drinking in adults with alcohol use disorder during recovery: A longitudinal observational cohort study.". Alcohol Clin Exp Res (Hoboken). DOI: 10.1111/acer.70250 [abstract-verified: partial]
• ^[10] Gilbert, Paul A, Soweid, Loulwa, Kersten, Sarah K et al. (2021). "Maintaining recovery from alcohol use disorder during the COVID-19 pandemic: The importance of recovery capital.". Drug Alcohol Depend. DOI: 10.1016/j.drugalcdep.2021.109142 [abstract-verified: partial]
• ^[4] Hagman, Brett T, Falk, Daniel, Litten, Raye et al. (2022). "Defining Recovery From Alcohol Use Disorder: Development of an NIAAA Research Definition.". Am J Psychiatry. DOI: 10.1176/appi.ajp.21090963 [abstract-verified: yes]
• ^[6] Humphreys, K, Moos, R H, Cohen, C (1997). "Social and community resources and long-term recovery from treated and untreated alcoholism.". J Stud Alcohol. DOI: 10.15288/jsa.1997.58.231 [abstract-verified: yes]
• ^[3] Kelly, John F, Klein, Morgan, Zeng, Katherine et al. (2025). "Long-term relapse: markers, mechanisms, and implications for disease management in alcohol use disorder.". Front Public Health. DOI: 10.3389/fpubh.2025.1706192 [abstract-verified: partial]
• ^[10] LaBarre, Charles, Bowen, Elizabeth, Knapp, Kyler S (2025). "Recovery Capital Is Associated with a Greater Likelihood of NIAAA-Defined Recovery for Adults with Prior Alcohol Problems.". Subst Use Misuse. DOI: 10.1080/10826084.2025.2478598 [abstract-verified: yes]
• ^[5] Pfund, Rory A, Hallgren, Kevin A, Maisto, Stephen A et al. (2021). "Dose of psychotherapy and long-term recovery outcomes: An examination of attendance patterns in alcohol use disorder treatment.". J Consult Clin Psychol. DOI: 10.1037/ccp0000703 [abstract-verified: partial]
• ^[15] Sengupta, Shreya, Anand, Akhil, Yang, Qijun et al. (2025). "The impact of integrated care on clinical outcomes in patients with alcohol-associated liver disease: Early outcomes from a multidisciplinary clinic.". Hepatol Commun. DOI: 10.1097/hc9.0000000000000603 [abstract-verified: partial]
• ^[12] Swan, Julia E, Aldridge, Arnie, Joseph, Verlin et al. (2021). "Individual and Community Social Determinants of Health and Recovery from Alcohol Use Disorder Three Years following Treatment.". J Psychoactive Drugs. DOI: 10.1080/02791072.2021.1986243 [abstract-verified: partial]
• ^[2] Jalie A Tucker, Susan D Chandler, Katie Witkiewitz (2020). "Epidemiology of Recovery From Alcohol Use Disorder.". Alcohol research : current reviews. DOI: 10.1093/ije/30.3.427 [abstract-verified: partial]
• ^[11] Vose-O'Neal, Adam, Christmas, Shanesha, Alfaro, Karen A et al. (2025). "Understanding pathways to recovery from alcohol use disorder in a Black community.". Front Public Health. DOI: 10.3389/fpubh.2025.1537059 [abstract-verified: yes]

Replacement Resolution Audit

Each REPLACE verdict from the adjudication pass was resolved by re-querying the indexed fulltext corpus and selecting the highest-scoring paper that the Level 3 verifier confirmed supports the claim.

• ^[16] → NO REPLACEMENT FOUND (considered 3 candidates; none verified)
• ^[16] → ^[17] (verifier: partial; score 0.71). Title: How recovery definitions vary by service use pathway: Findings from a national survey of adults.
• ^[18] → NO REPLACEMENT FOUND (considered 3 candidates; none verified)
• ^[18] → ^[17] (verifier: partial; score 0.71). Title: How recovery definitions vary by service use pathway: Findings from a national survey of adults.
• ^[3] → NO REPLACEMENT FOUND (considered 5 candidates; none verified)
• ^[3] → NO REPLACEMENT FOUND (considered 5 candidates; none verified)
• ^[19] → ^[5] (verifier: partial; score 0.70). Title: Participant Demographic and Baseline Drinking Factors Can Predict Alcohol Use Disorder Pharmacotherapy Clinical Trial Co
• ^[10] → NO REPLACEMENT FOUND (considered 4 candidates; none verified)
• ^[20] → ^[10] (verifier: partial; score 0.61). Title: Recovery Capital Is Associated with a Greater Likelihood of NIAAA-Defined Recovery for Adults with Prior Alcohol Problem
• ^[15] → ^[11] (verifier: yes; score 0.53). Title: The prevalence of subthreshold psychiatric symptoms and associations with alcohol and substance use disorders: from a na
• ^[21] → ^[15] (verifier: yes; score 0.69). Title: Understanding pathways to recovery from alcohol use disorder in a Black community.