Sober Synthesis

8/2/2026

Jeff K.- Recovery Defined

Definitions, Criteria, and Correlates

Recovery from Substance Use Disorders

The term ‘recovery’ is one of the most common and useful words in any discussion about substance use disorders. However there is no widely agreed upon definition or criteria describing its parameters and applications. On an individual level it may suffice to recognize differences of opinion. On an operational level for purposes of research and clinical assessment ambiguity hampers ability to compare outcomes, methodologies, and theory.

This post focuses on a recent definition advanced by the National Institute on Alcohol Abuse and Alcoholism a branch of the NIH and a major source of research and information in the field. Examples of research studies are presented investigating the potential applications of the NIAAA definition.

Substance Use Disorders are a common source of deaths, morbidity, and economic cost in the US. Data from a 2025 survey conducted by SAMHSA estimates a past year occurrence affecting 48.4 million individuals with 28.7 million cases attributed to alcohol alone or with other addictive drugs. Direct and indirect economic cost in the US is estimated to be between $740 billion to 1 trillion dollars annually.

Only 19% of individuals with SUD receive treatment. There is an array of treatment options of widely varying setting, philosophy, and methodology. The three medications approved for treatment of AUD are grossly under-prescribed in both outpatient and inpatient settings. Basic withdrawal treatment is out of reach for many people with SUD who may not have insurance coverage forcing them to find other choices or ending up in emergency departments further driving up cost and use of resources.

The DSM-5 published by the American Psychiatric Association has been widely adapted as the reference standard and staging system for the diagnosis of substance use disorders. There are a total of 11 criteria which cover four domains related to substance use:

– Impaired control

– Social problems

– Risky use

– Physical dependance

Disease severity can be measured by the number of positive criteria met:

2-4 = mild, 4-5 = moderate, 6 or greater = severe. Severe disease is sometimes referred to as addiction.

Remission is added to the diagnosis when the individual is asymptomatic (0 criteria except for cravings) for greater than 3 continuous months. Between 3 – 12 months remission is considered early remission. Greater than 12 months symptom free the term sustained remission may be used.

Note that the criteria do not mention quantity or duration of substance use. It is therefore possible to be considered in remission while continuing some level of use of the primary or secondary substance.

It is also important to note that the DSM is a collection of diagnostic criteria, not a complete description of the condition itself. It has undergone multiple revisions since its inception in 1952. Revision is an ongoing process as new knowledge becomes available.

The NIAAA has published limits describing heavy high-risk drinking. These are the most commonly used and accepted in the US.

For men the limits are 5 or more drinks on any given day or 15 or more per week. For women limits are 4 or more drinks per day or 8 or more per week.

A US standard drink is defined as 14 grams of alcohol. The equivalent of one 12 oz 5% beer. Online calculators are available on the NIAAA website for conversion to other drinks and quantities.

Different limits for men and women reflect differing alcohol metabolic rates and average body composition

Recognizing the need for an operational measurable definition of recovery from Alcohol Use Disorder the NIAAA has developed a working definition. This was published along with a discussion of methodology used in the American Journal of Psychiatry, November 2022.

The definition is reproduced above and in references with links at the end of this post.

It contains three components:

– The individual should be asymptomatic as defined in the DSM-5 with zero symptoms other than craving.

– The individual should have ceased ‘heavy drinking’ as defined in the NIAAA limits.

– Improvements in psychosocial aspects such as physical and mental health, basic needs, other “dimensions of well being”

The first two components follow the most widely accepted criteria for remission and recovery from AUD. The last is more nuanced and may be considered a provisional statement open to clarification as new data becomes available. An individual may be considered ‘recovered’ if criteria are sustained over an unspecified time period.

In considering a definition of recovery several goals need to be kept in mind. It needs to be operational and represent a broad consensus opinion. It should have measurable components enabling comparative results between research studies. Clinical utility and acceptance by providers is essential enabling categorization and tracking of treatment goals.

A definition can serve as a focal point in discussions relating to overall philosophy, empirical testing, and economic aspects of AUD or SUD recovery.

This is not the first definition to be offered. Three of these are listed above. Critics have pointed out that while highly descriptive of the recovery process these definitions are vaguely worded and do not lend themselves to empirical testing.

This study was conducted in response to the NIAAA definition. The authors wanted to see if the definition was operational as a basis for a data driven investigation of recovery in an AUD population. It was published in September 2025.

Two locations were included. Subjects were included meeting DSM-5 criteria for AUD. The subjects were all recently, or soon planning to initiate recovery by their own choice and strategy. They were not guided by or provided with treatment from the researchers. The study was not designed to evaluate efficacy of treatment methods. A total of 422 subjects were included in the final data.

Data was obtained by interview and selected psychosocial testing at baseline, 3, 6, 9, and 12 months. Alcohol consumption quantity, frequency, number of drinking days, DSM symptoms, and measures of functional and mental health were included in the study.

This graph represents the primary findings of this study. Based on early results subjects were assigned to one of four categories. Remission here refers to asymptomatic status by DSM criteria. Drinking pattern was classified as either abstinent, low or high risk by NIAAA limits. Groups were defined as follows:

– Remission + Abstinence (R+A) green

– Remission + Low Risk Drinking (R+LRD) teal

– Remission + High Risk Drinking (R+HRD) blue

– Non – Remission (NR) gold

At each time interval percent of subjects in each class is indicated. The wavy lines represent transitions within and between groups. Markov modeling was used to infer hidden variables. Markov models are briefly explained below.

Note that percent of individuals abstinent to low risk increases throughout the study from 9% to 19% at 12 months. The non remission group decreased in size throughout the study with net transfer to one of the recovery groups. The abstinent group was the most stable throughout the study with 95% retention rate. Also note that transitions out of abstinence were to non remission rather than to one of the remission plus low or high risk drinking groups.

Technical Note: brief summary of hidden Markov modeling (MM) used in this study.

MM is a type of machine learning used to estimate a variable or state not directly seen by the user. There is an instructive YouTube video linked in the references.

In MM observed information is used to calculate probability of an unseen hidden value. Transitions between observed states such as weather shown in the example above are used to find the most probable mood your friend iwill be in just before you call him.

The probability of transition from one category to another during the 12 month study of AUD individuals attempting recovery is shown above. Note the bidirectional movement for each state with the exception of abstinent group which did not have significant transitions to one of the recovered drinking groups (ie moderation).

Likelihood of staying within a group is designated by the circular loops. Abstinent was the most stable group. Regression analysis indicates that those with greater disease severity were highly likely to be either abstinent or in non remission. Individuals with remission but remaining low or high risk drinking were most likely to have lower initial AUD severity.

This shows some of the characteristics at the start of the study. Color code is the same as the previous graph. From left to right in each box: non-remission NR – R+HRD – R+LRD – R+A

Note Those with highest pre-test drinks per day, AUD symptoms and consequences tended to remain either abstinent or non recovered. Those who received treatment had a high probability of remaining abstinent.

Symptoms for each category before and after the study period. Decrease in anxiety was observed for all three groups in remission. There near equal improvement for these measures between abstinence, R+ high or low drinking. This finding suggests that significant incremental gains are possible short of optimal total abstinence. A similar improvement was seen for depression symptoms in both abstinent and remission + drinking groups. Larger gains were seen for abstinent subjects.

Additional data from this study measuring subject ratings of overall quality of life. The top bars show effect size from pre to post for each group. Greatest increase was in abstinent and remission with low risk drinking groups (d = 0.73 and 0.75). As with mood and other measures positive effects were present for all three remission groups.

This is from an earlier study published in the journal Addiction, January 2019. This used retrospective data from project MATCH. A large multi center randomized study which assigned patients to one of three treatment modalities and followed them for one year.

The researchers recruited from a pool of MATCH participants and followed them to three years looking at outcomes without assigning any treatment modality. 805 individuals were included in the study.

This study was done prior to the NIAAA definition. DSM-3 classifications for AUD were used. The goal was to stratify groups based on level of function and drinking pattern. Quality and level of satisfaction in a number of psychosocial life domains was correlated with functional level and drinking pattern. Also assessed was employment status and use of marijuana or other drugs.

Based on early data and statistical modeling the best fit was found for a four profile model. As no significant differences in outcomes was found between abstinence and high functioning/infrequent light drinking these patterns were bundled together.

Four profiles and percentage in the population were identified:

– Low functioning / frequent heavy drinking (15.8%)

– Low functioning / occasional heavy drinking (16.1%)

– High functioning / occasional heavy drinking (16.9%)

– High functioning / occasional light drinking (51.2%)

The above graph shows the levels of positive responses in each category standardized as a probability function for each domain. Notable is the largest group fit the high functioning / occasional light drinking profile at 51.2%. This category included subjects reporting abstinence or near abstinence.

These graphs show results for two items assessed. Profiles are arranged into four blocks corresponding to profile 1 = low functioning / frequent heavy drinking (red) on the far left and profile 4 = high functioning / occasional non heavy drinking (yellow) on the far right.

For each profile results are shown for months 0, 3, 9, 15, and 30.

There are striking differences between high and low functioning groups with worsening depression levels by month 30 with both frequent and infrequent heavy drinking.

High functioning low heavy or occasional non heavy drinkers had decreased depression scores throughout from baseline to 30 months.

Social functioning scores again show no improvement by 30 months for the low functioning frequent or infrequent heavy drinking profiles. Scores improved for the occasional heavy to occasional light drinking groups with near identical results for both.

This post reviews a definition of recovery from Alcohol Use Disorder developed by the NIAAA. The definition includes two widely used criteria; the DSM-5 and the NIAAA limits for hazardous heavy drinking. The definition also includes psychosocial aspects of recovery.

A large longitudinal research study designed to test the utility of the definition at one year following attempts at recovery in an AUD population validates correlation of the definition to clinical outcomes.

The model encompasses outcomes outside of the traditional goal of total abstinence and finds significant gains for some individuals with low or infrequent heavy use patterns. It validates common observations that individuals at highest risk tend to either achieve near abstinence or fall into non recovery status.

Symptom remission with occasional asymptomatic alcohol use was found in individuals with baseline lower risk.

This suggests that goals short of total abstinence may be appropriate for some individuals with AUD,

Objective means of identifying lower to higher risk profiles at outset remain a challenge. Empirical validation of psychosocial measures useful for a testable definition will require further research in larger multicentric studies.

Thank you for your interest in this post. Sobersynthesis.com and this post are for information and educational purposes only. No commercial or institutional interest. This post should not be considered medical or professional advice. Images and data obtained from sources freely available on the World Wide Web.

Feedback and comments are always welcome.

Jeffk072261@gmail.com

Jeff Kay 07/2026

References

https://onlinelibrary.wiley.com/doi/10.1111/acer.70172

Kelly, J.F., Belisario, K.L. & MacKillop, J. (2025) Prevalence, predictors, correlates, and dynamic changes in the NIAAA-defined “recovery” definition. Alcohol: Clinical and Experimental Research, 49, 2579–2591. Available from: https://doi.org/10.1111/acer.70172

…………………………………………………………………

Acceptance of Non-Abstinence Goals by Addiction Professionals in the

United States, Alan K. Davis and Harold Rosenberg

Psychology of Addictive Behaviors 2013, Vol. 27, No. 4, 1102–1109

https://www.apa.org/pubs/journals/releases/adb-a0030563.pdf

……………………………………………………………………….

Defining Recovery From Alcohol Use Disorder:

Development of an NIAAA Research Definition

Brett T. Hagman, Ph.D., Daniel Falk, Ph.D., Raye Litten, Ph.D., George F. Koob, Ph.D.

Am J Psychiatry 179:11, November 2022

https://psychiatryonline.org/doi/10.1176/appi.ajp.21090963

…………………………………………………………………..