How to Understand Addiction Treatment Success Rates: Evidence, Definitions & Reporting Standards
A treatment programme can report a high “success rate” while still leaving important scientific questions unanswered. Success may refer to abstinence, reduced substance use, treatment completion, retention, remission, overdose reduction, improved functioning or quality of life. These measures are not interchangeable.
studies were included in a 2025 systematic review examining how substance-use-disorder treatment outcomes are defined and measured. The review found substantial variation between studies. 1
For journalists, researchers and educators
This resource is designed to help readers interpret addiction-treatment outcome claims accurately. Original charts on this page may be cited or reproduced with attribution to DeBinge and a link to this resource. Primary research sources are listed in full below.
When reporting a treatment “success rate,” we recommend identifying the exact outcome, denominator, follow-up period, measurement method and handling of missing data rather than repeating a headline percentage without context.
There is no single scientifically sufficient “addiction treatment success rate”
In addiction research, the word success is not itself a standardised clinical endpoint.
A 2025 systematic review of 110 studies identified substantial variation in how treatment outcomes were defined and measured. Reported outcomes included abstinence, relapse, treatment retention, treatment dropout, adherence, readmission and quality of life. 1
Earlier work examining opioid-treatment trials reached a similar conclusion. Dennis and colleagues identified 21 specific outcome concepts, 53 subdomains and 118 different measurements across 60 clinical trials. 2
“Is an 80% success rate good?”
The first question should be:
“What exactly counted as success, among which patients, measured how, and at what point after treatment?”
Scope and methodology
This page is an evidence-synthesis and interpretation resource. It is not a new systematic review, does not rank treatment providers and does not claim that one addiction treatment is universally superior to another.
Priority is given to peer-reviewed systematic reviews, randomised clinical trials, consensus outcome projects and guidance from organisations including the U.S. Food and Drug Administration, National Institute on Alcohol Abuse and Alcoholism, National Institute on Drug Abuse Clinical Trials Network and Substance Abuse and Mental Health Services Administration.
Where an individual clinical trial is used as an example, the outcome definition and analysis population used by the investigators are identified.
A defensible treatment-outcome statement should identify at least five things
- Outcome: What was actually measured?
- Definition: What threshold counted as improvement, relapse or success?
- Denominator: Was the percentage calculated from everyone admitted, everyone starting treatment, treatment completers, or only people successfully followed up?
- Time: Was the result measured at discharge, three months, six months, one year or longer?
- Missing data: What happened to patients who could not be contacted, stopped treatment or missed assessments?
What common addiction-treatment outcome terms actually mean
Addiction research uses several distinct measures of treatment outcome. Even apparently familiar terms such as “relapse” and “retention” may be defined differently between studies.
For that reason, the operational definition is as important as the headline percentage.
| Outcome | What it usually measures | What should be reported | What it does not prove by itself |
|---|---|---|---|
| Abstinence | No use of a specified substance during a defined period. | Substance measured, abstinence window, self-report or toxicology, testing frequency and handling of missed tests. | Overall recovery, treatment retention, improved wellbeing or future relapse risk. |
| Relapse / return to use | A prespecified level or pattern of resumed substance use. | Exact threshold and time period used to define the event. | That every episode of substance use has identical clinical meaning. |
| Treatment completion | Completion of a planned treatment episode or programme. | Programme length, completion criteria and discharge rules. | Abstinence or continued recovery after leaving treatment. |
| Retention | Length of time a patient remains engaged in treatment. | Start point, allowable gaps, transfer rules and duration. | Clinical improvement by itself. |
| Adherence | Medication use, appointment attendance or participation in treatment. | Measurement method and threshold used to define adequate adherence. | That treatment itself caused improvement. |
| Remission | Reduction or absence of diagnostic symptoms for a defined period. | Diagnostic criteria and duration. | Full social, occupational and health recovery. |
| Recovery | A broader process involving substance use, symptoms, health, functioning and wellbeing. | Definition used and domains included. | A universally standardised single endpoint. |
| Overdose / mortality | Non-fatal overdose, fatal overdose or death during follow-up. | Cause definition, data source and completeness of follow-up. | All other benefits or harms associated with treatment. |
| Quality of life | Physical, psychological, social and functional wellbeing. | Validated instrument, baseline result and follow-up result. | Abstinence unless substance use is separately measured. |
Recovery is broader than abstinence alone
The National Institute on Alcohol Abuse and Alcoholism describes recovery from alcohol use disorder as a process involving remission from AUD symptoms together with cessation from heavy drinking over time. Its research framework also recognises improvements in physical health, mental health, relationships and quality of life. 6
SAMHSA similarly describes recovery through several dimensions, including health, home, purpose and community. 5
An international multidisciplinary consensus project published in 2024 recommended measuring domains including addictive behaviour, symptom burden, health-related quality of life, global functioning, psychosocial functioning, physical health and mental health or wellbeing. 7
Opioid use disorder research provides a useful example
A National Institute on Drug Abuse Clinical Trials Network consensus process began with 24 candidate outcome concepts and ultimately selected five core outcomes for opioid-use-disorder treatment research. 4
1. Patient global impression of improvement
A patient-reported assessment of whether the individual’s opioid-related problem has improved or worsened.
2. Non-fatal opioid overdose
Number of opioid-related overdose events occurring during a defined observation period.
3. Illicit or non-medical drug toxicology
A biological measure of substance use evaluated using a prespecified responder rule.
4. Duration of treatment
Continuous time enrolled in medication treatment.
5. Fatal opioid poisoning
A mortality endpoint measuring opioid-related death rather than relying only on surrogate outcomes.
It demonstrates that even expert consensus does not reduce opioid-treatment effectiveness to one generic “success rate.”
What the research data show: definitions can change the story
Figure 1. Addiction studies use many different outcome measures
Narváez-Camargo and colleagues included 110 studies in their 2025 methodological systematic review.
Among the outcome categories identified, 20 studies reported abstinence, 22 relapse, eight readmission, 24 dropout, 23 retention, 22 adherence or engagement, and 19 quality of life. A single study could report more than one outcome category. 1
| Outcome variable | Studies reporting outcome |
|---|---|
| Abstinence | 20 |
| Relapse | 22 |
| Readmission | 8 |
| Dropout | 24 |
| Retention | 23 |
| Adherence / engagement | 22 |
| Quality of life | 19 |
The review also identified substantial differences in the way individual outcomes were operationally defined.
“Dropout,” for example, could refer to failure to return for a subsequent treatment session in one study while another study used a substantially longer treatment-duration threshold. 1
Figure 2. X:BOT demonstrates why the denominator matters
The X:BOT randomised clinical trial compared extended-release injectable naltrexone with buprenorphine-naloxone for opioid-relapse prevention over 24 weeks. 3
Among everyone randomised, approximately 72% of participants assigned to extended-release naltrexone successfully initiated medication compared with approximately 94% assigned to buprenorphine-naloxone.
In the intention-to-treat population, a relapse event occurred in approximately 65% of participants assigned to extended-release naltrexone compared with approximately 57% assigned to buprenorphine-naloxone.
Among participants who successfully initiated the assigned treatment, relapse-event percentages were approximately 52% and 56% respectively. 3
| Outcome | XR-naltrexone | Buprenorphine-naloxone |
|---|---|---|
| Treatment induction | 72% | 94% |
| Relapse event — intention-to-treat | 65% | 57% |
| Relapse event — successfully inducted | 52% | 56% |
However, it is not the same question as an intention-to-treat analysis that includes everyone originally assigned to treatment.
X:BOT also demonstrates the importance of prespecified missing-data rules. The investigators described how missing urine samples and treatment discontinuation contributed to the relapse classification. 3
This illustrates why a published percentage cannot be interpreted fully without understanding the analysis rules behind it.
Figure 3. COMBINE demonstrates why the endpoint itself matters
The COMBINE Study enrolled 1,383 participants with alcohol dependence and evaluated combinations of medication, medical management and behavioural interventions. 8
The study reported multiple outcomes, including percentage of days abstinent, time to first heavy drinking day and a participant-level measure described as a “good clinical outcome.”
These measures are all potentially useful, but they are not interchangeable.
| Treatment condition | Percent days abstinent | Good clinical outcome |
|---|---|---|
| Placebo + medical management | 75.1% | 58% |
| Naltrexone + medical management | 80.6% | 74% |
| Placebo + CBI + medical management | 79.2% | 71% |
| Naltrexone + CBI + medical management | 77.1% | 74% |
The regulatory approach to alcohol-treatment outcomes is also evolving.
In February 2025, the U.S. Food and Drug Administration qualified a two-level reduction in alcohol risk-drinking level as a clinically meaningful endpoint that can be used in medication-development programmes for moderate-to-severe alcohol use disorder. 9
This provides another example of why abstinence alone should not automatically be treated as the only meaningful endpoint in every alcohol-use-disorder study.
Eight questions to ask before accepting any addiction-treatment success rate
The following framework can be used by journalists, researchers, clinicians, commissioners, patients and families when evaluating a treatment-outcome claim.
-
What is the numerator?
How many people actually met the stated outcome? -
What is the denominator?
Everyone admitted, everyone starting treatment, treatment completers, or only participants reached at follow-up? -
What exactly counted as success?
Abstinence, reduced use, remission, retention, programme completion, no overdose, improved functioning or a composite? -
How long were people followed?
A result at discharge is different from an outcome measured at six months or one year. -
How was the outcome verified?
Self-report, toxicology, medical records, independent assessment or a combination? -
How were missing data handled?
Were participants lost to follow-up excluded, counted as unsuccessful or handled using another prespecified rule? -
Who was studied?
Substance, severity, age, comorbidity, treatment history and treatment setting can affect comparability. -
What other clinically important outcomes were measured?
Retention, overdose, mortality, physical health, mental health, functioning and quality of life may all matter.
Common warning signs in treatment-outcome claims
| Claim pattern | Why interpretation is difficult | What stronger reporting would include |
|---|---|---|
| “90% success rate” with no definition | The endpoint cannot be independently interpreted. | Exact outcome definition, denominator and follow-up period. |
| Percentage calculated only among programme completers | People who disengaged early may disappear from the denominator. | Results for the original treatment cohort as well as completers. |
| No follow-up beyond treatment discharge | Completion is being measured rather than durability after treatment. | Clearly stated post-treatment follow-up periods. |
| Participants lost to follow-up are omitted | The reported percentage may be biased if those lost differ systematically from respondents. | Follow-up rate and explicit missing-data rules. |
| Different substance-use disorders combined into one rate | Different disorders may have different clinically meaningful endpoints. | Substance-specific reporting where possible. |
| “Verified” outcome with no measurement method | The reader cannot evaluate how the result was established. | Self-report method, toxicology schedule or record source. |
| No comparator or benchmark | Improvement before and after treatment does not by itself establish that treatment caused the improvement. | Appropriate comparison or cautious interpretation. |
| Safety outcomes not reported | Benefit cannot be interpreted fully without relevant harms. | Adverse events, overdose and mortality where appropriate. |
| One endpoint described as complete “recovery” | Recovery is multidimensional. | Substance use, health, functioning and wellbeing measures. |
A reporting template for addiction-treatment outcomes
Journalists and treatment providers can use the following structure to make outcome statistics easier to interpret.
“At [time point], [X/Y; percentage] of [defined denominator] met [operational outcome definition], measured using [measurement method].
Follow-up data were available for [percentage] of the original cohort. Participants with missing data were handled using [prespecified rule].
During the same observation period, treatment retention was [result], non-fatal overdoses were [number], deaths were [number], and [validated functioning or wellbeing measure] changed by [result].”
Why absolute numbers matter
A percentage without its numerator and denominator hides scale. An 80% result could mean eight of ten people or 800 of 1,000 people. Reporting both the absolute number and percentage gives the reader substantially more information.
Statistical uncertainty matters too
When treatments or populations are being compared, confidence intervals or another appropriate measure of uncertainty can help readers understand how precise an estimate is.
FDA guidance on multiple clinical endpoints also explains that endpoint selection and multiplicity can affect the likelihood of incorrect conclusions when analyses are not appropriately controlled. 10
Cross-study comparisons require caution
Treatment results are particularly difficult to compare directly when interventions use different induction procedures, operate in different settings, target different substances, include different patient populations or use different definitions of relapse and recovery.
The X:BOT trial is a useful example: treatment initiation requirements materially affected the intention-to-treat comparison, while the analysis among successfully inducted participants answered a different question. 3
Patient-important outcomes deserve space beside substance-use measures
Dennis and colleagues found that opioid trials frequently emphasised continued opioid use and treatment retention while patients described additional goals that were underrepresented in research. 2
The later international consensus project similarly incorporated functioning, quality of life, physical health and mental wellbeing alongside addictive behaviour and symptom burden. 7
For journalists, researchers and educators
Suggested citation
Jezierski R. How to Understand Addiction Treatment Success Rates: Evidence, Definitions & Reporting Standards. DeBinge. Evidence checked 24 September 2026.
Clinical review by Dr Fergus Law is currently pending.
Chart reuse
Journalists, researchers and educators may reproduce the original charts created for this page with attribution to “DeBinge — How to Understand Addiction Treatment Success Rates” and a link to this resource.
The cited original research remains the authoritative source for the underlying results.
Editorial and conflict-of-interest disclosure
DeBinge is a private addiction-treatment organisation and publishes patient and professional educational material.
This page does not present proprietary DeBinge treatment-outcome data, does not claim that DeBinge has a superior success rate, and does not use the studies cited here as proof of effectiveness for a specific DeBinge treatment programme.
Its purpose is to explain how addiction-treatment outcome claims can be interpreted more accurately.
Medical-review status
Prepared for medical review by Dr Fergus Law, BSc, MBChB, FRCPsych, Consultant Psychiatrist specialising in substance misuse and addiction.
Review status: Pending.
The wording should only be changed to “Medically reviewed by Dr Fergus Law” after he has reviewed and approved the final published version.
Data provenance
| Figure | Source | How the chart was created | Important limitation |
|---|---|---|---|
| Figure 1 | Narváez-Camargo et al. (2025) | Published outcome-category counts were transcribed and visualised as a bar chart. | Categories overlap because studies can report more than one outcome. |
| Figure 2 | Lee et al. (2018), X:BOT | Published treatment-initiation and relapse-event percentages were plotted side by side. | Treatment initiation and relapse are separate outcomes, and intention-to-treat and successfully inducted analyses answer different questions. |
| Figure 3 | Anton et al. (2006), COMBINE | Published percent-days-abstinent and good-clinical-outcome results were displayed on a common percentage scale. | The two measures represent different outcome constructs and should not be interpreted as interchangeable. |
Limitations of this guide
This resource is a structured educational synthesis rather than a prospectively registered systematic review.
The individual trials used as examples were selected because they illustrate important methodological issues and should not be interpreted as a comprehensive comparison of all addiction treatments.
Clinically meaningful outcomes can differ according to substance, severity, population, treatment setting and patient goals. Researchers conducting formal evidence synthesis should return to the original study protocol, statistical analysis plan and publication rather than extracting effect estimates from this educational page alone.
Version history
Version 1.0 — 24 September 2026
Substantive future revisions should update the evidence-check date and document material changes.
References
- Narváez-Camargo M, Lozano-Rojas O, Mancheño-Velasco C, Verdejo-García A. Substance Use Disorder Treatment Outcomes: Methodological Overview of Metrics and Criteria. International Journal of Methods in Psychiatric Research. 2025;34(2):e70027. DOI: 10.1002/mpr.70027
- Dennis BB, Sanger N, Bawor M, et al. A call for consensus in defining efficacy in clinical trials for opioid addiction: combined results from a systematic review and qualitative study. Trials. 2020;21:30. DOI: 10.1186/s13063-019-3995-y
- Lee JD, Nunes EV Jr, Novo P, et al. Comparative effectiveness of extended-release naltrexone versus buprenorphine-naloxone for opioid relapse prevention (X:BOT). Lancet. 2018;391(10118):309–318. DOI: 10.1016/S0140-6736(17)32812-X
- Kelly SM, Gryczynski J, Mitchell SG, et al. The opioid use disorder core outcomes set (OUD-COS) for treatment research: findings from a Delphi consensus study. Addiction. 2022. DOI: 10.1111/add.15875
- Substance Abuse and Mental Health Services Administration. About Recovery. SAMHSA
- Hagman BT, Falk D, Litten R, Koob GF. Defining Recovery From Alcohol Use Disorder: Development of an NIAAA Research Definition. American Journal of Psychiatry. 2022;179(11):807–813. DOI: 10.1176/appi.ajp.21090963
- Black N, Chung S, Tisdale C, et al. An International, Multidisciplinary Consensus Set of Patient-Centered Outcome Measures for Substance-Related and Addictive Disorders. Journal of Clinical Medicine. 2024;13(7):2154. DOI: 10.3390/jcm13072154
- Anton RF, O’Malley SS, Ciraulo DA, et al. Combined Pharmacotherapies and Behavioral Interventions for Alcohol Dependence: The COMBINE Study. JAMA. 2006;295(17):2003–2017. DOI: 10.1001/jama.295.17.2003
- U.S. Food and Drug Administration. FDA qualifies drug development tool to facilitate clinical trial research on alcohol use disorder. 14 February 2025. FDA
- U.S. Food and Drug Administration. Multiple Endpoints in Clinical Trials: Guidance for Industry. October 2022. FDA
- U.S. Food and Drug Administration. Opioid Use Disorder: Endpoints for Demonstrating Effectiveness of Drugs for Treatment — Guidance for Industry. October 2020. FDA
This resource should be reviewed periodically and updated when important outcome-reporting guidance or major evidence changes.