Executive Summary
Residential rehabilitation occupies a small proportion of the UK addiction treatment landscape, yet it carries outsized significance for patients, families, commissioners, and the public imagination. It is often used as a last resort for people with severe dependence and complex needs, and is too easily judged against the wrong standard – as though it were a one-off, acute medical intervention that should produce durable outcomes from a single episode of care. Addiction recovery, however, typically unfolds over time and across multiple episodes of treatment and support – research suggests a median of two serious recovery attempts, with those most severely affected requiring considerably more over a period of years. Residential treatment is best understood as a time-limited but uniquely powerful window of opportunity, one that can stabilise patients, disrupt harmful environments, build recovery skills and recovery-oriented social connections, and actively shape the conditions that make sustained recovery more likely after discharge.
This paper reviews the evidence base relevant to residential rehabilitation, including major prospective cohort studies (NTORS, DATOS, DORIS, ATOS), a targeted systematic review of residential treatment studies, and treatment centre outcome evaluations spanning four decades. While randomised controlled trials are often impractical or unethical in this field, the available longitudinal evidence supports the conclusion that residential rehabilitation delivers meaningful benefits, especially when integrated into longer treatment pathways and continuing care. Notably, NTORS found that residential clients with the most severe and complex needs made the greatest treatment gains, while DORIS demonstrated that better outcomes for residential clients could not be explained by lower baseline severity. The most consistent findings across research traditions extend well beyond abstinence with improvements across mental health, social functioning, quality of life, and broader problem severity. Individual treatment centre studies, including a series from Castle Craig spanning 1999 to 2015, show that the overwhelming majority of patients who engage with residential treatment experience measurable improvement, including many who have not maintained complete abstinence.
The paper then examines what the research suggests actually drives positive outcomes. Process research by Moos, Kelly, and Morgenstern, alongside Vaillant’s longitudinal naturalistic observations, indicates that the quality of the treatment environment and the degree to which programmes facilitate recovery-supportive behaviours, coping skills, and social network change can matter more than individual patient characteristics or any single therapeutic technique. Mechanisms repeatedly associated with sustained recovery include changes in social networks, increased abstinence self-efficacy, strengthened coping and emotional regulation, engagement with mutual-help communities, and growth in meaning, purpose, and hope. These mechanisms align closely with the rationale for structured residential treatment and provide a more precise basis for programme design and evaluation than abstinence rates alone. At Castle Health, this evidence has directly informed clinical innovation. One example is ALAANA (Assertive Linkage to Alcoholics Anonymous and Narcotics Anonymous), a structured 12-Step Facilitation workshop designed to activate the recovery mechanisms the research identifies as most important.
Finally, the paper proposes a practical, standardised model for outcome evaluation that can be adopted across providers and calls on UK residential treatment providers to unite behind a collaborative, multi-site effectiveness study. The model recommends consistent baseline assessment, the use of validated public-domain instruments across key domains (quality of life, depression, self-efficacy, recovery capital, self-esteem, spirituality, and social network composition), and follow-up at 6 and 12 months post-discharge. A prospective longitudinal design, with all participating sites using the same instruments and protocols from admission through follow-up, would ensure the comparability and rigour that the field currently lacks. To improve credibility, studies should include non-completers wherever possible, employ transparent methods, and incorporate independent oversight. The aim is to generate honest, actionable evidence that supports clinical improvement, informs commissioning, and strengthens public confidence in residential treatment as part of an integrated system of care.
Key Messages
- Rehab is best evaluated as a time-limited but high-impact stage in a longer recovery journey, not as a one-off cure.
- The balance of evidence supports meaningful benefit from residential rehabilitation, including for severe and complex cases, particularly when linked to continuing care and aftercare.
- The most important outcomes extend beyond abstinence, including mental health, quality of life, functioning, and overall problem severity.
- Mechanisms of sustained recovery are increasingly clear, including social network change, self-efficacy, coping skills, mutual-help engagement, and meaning or purpose.
- The sector needs shared, standardised outcome measurement and benchmarking, built on validated instruments, baseline data, follow-up, inclusion of non-completers, and independent oversight.
- Castle Health is calling on UK residential providers to participate in a collaborative, multi-site effectiveness study using this shared framework – the first of its kind in the UK.
The Challenge
Residential rehabilitation – rehab – is just one small fraction of the addiction treatment landscape but it occupies a prominent place in popular imagination. For many, rehab signals the start (or restart) of a treatment and recovery journey. In the journey analogy, rehab can be seen as the first stage – it has been compared to a taxi that takes you to the train station.
In the United Kingdom, rehab also tends to be a treatment of last resort. The residents of a rehab suffer from more severe dependency than patients in community treatment settings. The people who enter rehab often do so after many frustrated attempts to quit or get help elsewhere. So when rehab leads to success, its defenders can say that it outperformed previous interventions. When it does not lead to lasting change, it is seen as no worse than earlier attempts.
But this framing is self-defeating. It justifies low expectations, restricts access, and reinforces the idea that rehab is only for the least hopeful cases. This can distort commissioning decisions, limit earlier intervention, and sideline residential treatment in the wider care system.
The marginalisation has at times been reinforced by official assessments. In a 2006 publicly funded review of alcohol treatment effectiveness, Raistrick, Heather, and Godfrey concluded that “12-Step residential treatment confers no added benefit compared with other forms of treatment and is less cost-effective than outpatient treatment.” National Institute for Health and Care Excellence (NICE) Guidelines adopted in 2007 and 2011 established high eligibility criteria for residential rehab on similar grounds, concluding it was not cost-effective compared to intensive outpatient models. However, these assessments assumed that suitable intensive outpatient services would be accessible – yet such services are still not widely available in the UK. The result has been a policy environment in which residential treatment is judged unfavourably against alternatives that do not, in practice, exist for most patients who need them.
It is important to recognise rehab’s potential and its limitations without resigning it to the margins. The real challenge is to ask what residential treatment uniquely offers, how it works, and how it can be used to best effect. When relapse occurs, it is often beyond the direct influence of the treatment provider, and so a successful outcome thus depends on what happens next in the patient’s recovery journey. Treatment provides a crucial window of opportunity. It exerts a positive influence in the short-term, providing space and time to learn what it takes to sustain long-term recovery.
Whether the treatment experience leads to lasting change depends largely on the broader context of a person’s life – including their support systems, self-efficacy (knowing how to act effectively) and sense of personal agency (the ability to choose to act). Framed this way, it’s understandable why many rehabs have not felt compelled to rigorously justify their effectiveness through long-term outcome tracking. If patients complete treatment feeling hopeful, stable and better equipped for recovery than when they arrived then less scrupulous providers can argue that they have fulfilled their role as being that crucial “taxi ride.”
They could even point to Professor Griffith Edwards’ observation that “within that longitudinal perspective, it is absurdly medico-centric to suppose that treatment efficacy is the single most important question. Treatment is more accurately perceived as being at best a timely nudge or whisper on a lifelong course.”
Still, acknowledging the limits of treatment does not let providers off the hook – it makes the case for clearer evidence and sharper questions about what works, and why. Understanding that rehab is only one part of a lifelong recovery journey should drive us to examine how that part can be made as effective as possible.
The question “does treatment work?” may be overly medico-centric, as Edwards suggests, but reframing it as “how does it work – when it does?” opens up rich avenues. Even if long-term outcomes hinge on factors beyond providers’ control, rehab services must still show that they are making the most of the window they do control.
This paper explores how residential rehab should be understood and evaluated, by looking at a broad range of scientific research that addresses efficacy, as well as the processes and conditions that enable recovery. Clarifying the factors that matter most helps to provide treatment centres with a guide to contributing more rigorously designed outcome studies to the field.
Some might question why this matters for a treatment modality that represents only around 2% of addiction treatment in the UK. But the significance of rehab outcomes extends well beyond its direct reach. Residential treatment offers a uniquely controlled environment in which to study the mechanisms of recovery – what changes in a person’s life when they get better, and why. Insights generated from high-quality rehab outcome studies can inform the design and improvement of standard community treatments that serve the vast majority of patients. Understanding what works in the most intensive treatment setting sharpens our understanding of what matters across all settings.
If rehab’s role is to provide a timely nudge on a lifelong course, then providers must demonstrate that they are delivering effective nudges – especially given the financial and personal costs involved. The central challenge is how best to use the space and time that rehab creates to activate key mechanisms of change and strengthen the personal and social support structures of long-term recovery.
Accountability
Since its relatively recent emergence, the addiction treatment field has always been expected to demonstrate its effectiveness using the same research frameworks applied to the rest of healthcare. The public, treatment providers, and funders often assume that rehab should follow the acute model of medical intervention i.e. an intensive treatment period leading to prolonged positive outcomes directly attributable to that intervention.
However, the efficacy of rehab cannot be measured as one would a specific cancer treatment or the potency of antibiotics against an infection. Addiction recovery involves complex behavioural, psychological, and social changes that unfold over months and years. There are no cures in the sense familiar from acute medical treatments. But this positive expectation isn’t entirely unreasonable because the ultimate goal of rehab is to facilitate stable recovery. Without this, the original symptoms will often re-occur and worsen.
The longitudinal reality underscores this point. Research by Dennis and colleagues (2005) and Kelly and colleagues (2019) has shown that the pathway to sustained remission typically involves multiple episodes of treatment and support – on average around 3-5 episodes over a period of years from first treatment contact. It is worth noting that the commonly cited mean of around five recovery attempts is skewed by a long tail of more severe cases; Kelly’s nationally representative study found the median number of serious recovery attempts is closer to 2, meaning that many individuals do achieve lasting recovery after just one or two well-delivered interventions. But the broader pattern is clear: for most people, recovery is a trajectory rather than an event, and any single treatment episode – however intensive – is best understood as one contribution to that longer journey. Evaluating a residential programme as though it were a one-off cure is therefore not just imprecise; it applies the wrong framework entirely.
This means rehab must demonstrate that it can meaningfully influence what happens next. The real test of effective rehab is whether it equips patients with the tools, insights, and foundations that enable them to navigate post-discharge challenges and maintain recovery in the real world. This necessitates the development of more precise and meaningful evaluation methods and a wider evidence lens that researches the active ingredients of treatment. Individual rehab providers must then satisfy the sceptics by collecting credible and multi-dimensional data about their own service’s outcomes.
Scientific Research
There is surprisingly little high-quality research that specifically sets out to examine residential rehab outcomes. The first explanation for this is that randomised controlled trials (RCTs) – the gold standard for measuring treatment effectiveness – are largely impractical in this field. RCTs would require randomly assigning addiction sufferers to residential rehab, community treatment, or no intervention, which would be unethical given that patients often have critical needs and clear clinical indicators for appropriate levels of care. The concept of equipoise – genuine uncertainty about which treatment is better – rarely exists in addiction treatment where we usually know residential care is more appropriate for severe cases.
Beyond ethical concerns, blinding isn’t possible as participants know their treatment setting. There is also the further obstacle that rehab is not a specific treatment method with standardised protocols – it is more a treatment setting, with multi-modal interventions that vary significantly between providers, making meaningful comparison difficult.
The failure to advance the RCT model within the addiction treatment field does not mean rigorous evaluation is impossible. However, it does require alternative designs such as matched cohort studies, longitudinal tracking, and mixed-methods research.
National Studies
Large-scale national cohort studies of addiction patients (while not designed specifically as rehab outcome studies) do provide crucial insights, demonstrating meaningful associations between residential treatment and positive outcomes for certain populations, particularly those with severe dependencies. Major prospective naturalistic cohort studies such as NTORS (UK), DATOS (USA), DORIS (Scotland), and ATOS (Australia) have followed individuals through treatment at regular intervals to determine treatment effectiveness.
This approach also faces significant challenges: patients often move between multiple services during follow-up, making it difficult to determine which intervention was most influential; individuals referred to different services may have varying baseline severity levels; and selection bias occurs when people choosing residential treatment differ systematically from those in community services. Despite these limitations, these four major studies collectively provide supportive evidence for residential rehabilitation’s effectiveness.
ATOS demonstrated that individuals achieving sustained recovery over an 11-year follow-up were more likely to have engaged in residential rehabilitation, particularly when combined with aftercare. NTORS found that nearly half of residential rehab patients remained abstinent from heroin at five-year follow-up, alongside substantial improvements in psychological health and criminal activity. Notably, NTORS reported that residential clients “presented with some of the most severe problems and complex needs, and these clients made some of the greatest treatment gains” – a finding that challenges the assumption that rehab’s relatively strong outcomes simply reflect an easier caseload. It is also worth noting that NTORS combined residential rehabs and NHS inpatient units into a single category; when the data was subsequently disaggregated, rehabs were found to have substantially outperformed the inpatient units, a distinction that was obscured in the published reports and may have contributed to the subsequent decline in funded referrals to the residential sector.
DATOS showed that longer retention in residential treatment was consistently associated with greater reductions in drug use, criminal activity, and unemployment. Even DORIS, despite finding overall low abstinence rates in Scotland’s treatment system, revealed that outcomes were better for those with longer or more intensive treatment approaches. Importantly, the DORIS researchers conducted multivariate analysis to test whether residential clients’ better outcomes could be explained by lower baseline severity; using the Severity of Dependence Scale, they found virtually no difference in dependence levels between those receiving residential rehabilitation and those treated in the community, indicating that rehabs were not simply selecting patients with less severe dependence.
Together, these studies suggest that residential rehabilitation plays a crucial role in achieving sustained recovery for individuals with complex addiction problems, particularly when integrated into longer treatment pathways with appropriate aftercare support.
Systematic Review (2019)
Turning to research that examines rehab outcomes directly, and across diverse populations and settings, a 2019 systematic review by De Andrade and colleagues – the first comprehensive systematic review specifically focused on residential rehabilitation – identified 23 admissible studies published between 2013 and 2018.
The review found moderate quality evidence that residential rehabilitation improves outcomes across multiple domains: substance use, mental health, social outcomes, and mortality. Notably, 17 of the 23 studies reported mental health outcomes including PTSD, depression, anxiety, and general mental health, with abstinence rates varying widely but generally showing positive results. The review concluded that best practice rehabilitation integrates mental health treatment and provides continuity of care post-discharge.
Treatment Centre Outcome Studies
While large-scale studies published in peer-reviewed journals tell one story, individual treatment centres often report markedly different – and more favourable – outcomes.
Some centres report success rates of 50–70%, significantly exceeding figures from the academic literature, with one UK provider stating that “96% of our clients… are still in recovery one year after treatment.” This contrasts with Miller, Walters, and Bennett’s 2001 composite review of both inpatient and outpatient treatment, which found just 24% achieved one-year continuous abstinence across all modalities. According to William White, “The primary cause of these discrepancies lies in the quality of the study methods and procedures upon which statements about effectiveness are made.”
Those who introduced the American “Minnesota Model” of rehab to the UK did so with a sense of enthusiasm and social mission, which perhaps discouraged rigorous scrutiny of its effectiveness. Centres attracted such a volume of referrals throughout the 1980s that there was little real pressure to establish whether the approach was producing the outcomes claimed. The earliest serious attempts at publishing validated treatment outcomes came from Hazelden in the 1970s and 1980s, reporting that 53-60% of patients remained abstinent one year after treatment. Reviewing these studies, Cook observed that despite methodological criticisms, “all patients appear to have gained some therapeutic benefit” – an early indication that the value of treatment extends well beyond abstinence rates alone. Around the same period, Clouds House in the UK presented similarly encouraging figures, with 54% of treatment completers achieving continuous abstinence from alcohol and mood-altering drugs, and an additional 14% maintaining abstinence following only a single relapse.
Despite methodological limitations – relying on self-reports and excluding early dropouts – these studies provided reassuring early evidence supporting 12-Step residential treatment. Cook concluded that the Minnesota Model appeared to produce outcomes equivalent to or better than those of other treatment programmes, whilst illustrating the need for comparison groups and longer follow-up.
In 1992, Clouds House commissioned an independent evaluation: Alex Georgiakis tracked 116 patients for 30 months post-treatment, with self-reports independently validated.
Remarkably, 61% were abstinent from all mood-altering substances at follow-up, with over half of these maintaining continuous abstinence since treatment. A further 5% were abstinent from their primary drug with no evidence of problematic substitution, and 5% were drinking within government health recommendations, while 29% were using drugs or drinking problematically. This was particularly significant given that a substantial number were socioeconomically disadvantaged with high levels of psychological distress and criminal histories – challenging assumptions that residential treatment only worked for stable, well-resourced individuals. The granularity of the Georgiakis findings also demonstrated the value of honest, nuanced reporting: a simple binary of “abstinent or not” would have obscured the fact that a significant proportion had achieved meaningful, if imperfect, improvement.
Castle Craig Hospital in Scotland – the flagship residential facility within the Castle Health group – has conducted a series of outcome studies since the 1990s, progressively evolving from internal methodology to validated instruments. The first significant evaluation was conducted by Hughes in 1999, examining 96 patients who had maintained employment during their period of alcohol dependency, following them up over an average 21-month period. With a 78% response rate, the results showed that 41% had achieved continuous abstinence throughout the follow-up period, with a further 19% classified as having good outcomes.
A more comprehensive study by McCann and Amos in 2000 examined all 206 patients who participated in Castle Craig’s Extended Care Unit over a two-year period, with follow-up averaging 429 days. Of those who completed the programme, 48% maintained continuous abstinence from all drugs and alcohol, with an additional 14% achieving good outcomes. Crucially, the study documented improvements well beyond abstinence: over 60% reported much improved physical health, over 50% reported much improved mental health, and nearly 70% reported much improved quality of life – particularly impressive given the severe physical and mental health difficulties these patients presented with at admission. These broader measures of improvement would become an increasingly important part of the story.
Dr George Christo’s 2007 evaluation study introduced the validated Christo Inventory Score (CISS) as a standardised outcome measure. Studying cocaine-dependent patients, Christo found that 60% achieved good outcomes despite high levels of dysfunction at intake, with abstinence rates rising from 46% among those completing detoxification alone to 66% among full programme completers – underscoring the value of treatment duration. A further 2015 Castle Craig study found that 73.4% of 89 respondents were totally abstinent at one-year follow-up, with 91.8% either abstinent or living with significantly reduced use.
When broader outcomes are considered – mental health, social functioning, overall problem severity – the 2015 Castle Craig study found that 91.8% of respondents had improved, 8% remained the same, and none had deteriorated. This pattern of overwhelming improvement, even among those who have not maintained complete abstinence, is consistent across comparable treatment centre studies and represents strong evidence that residential rehabilitation produces meaningful benefit for the great majority who engage with it. The answer to “does rehab work?” is yes, often; the more productive question is how much it works for each person, and over what time span.
Quasi-residential and community-based programmes, which evolved out of Clouds and Castle Craig, have shown similarly strong results. SHARP in Essex (a community programme based on the Clouds House model) achieved 72% completion rates with 79% of completers not re-engaging with any treatment services within the three-year pilot period. LEAP in Edinburgh found abstinence rates increased from 12% at baseline to 48% at four-year follow-up, rising to 60.7% among programme completers.
Despite their common heritage – with Castle Craig, Clouds House, SHARP and LEAP all rooted in the same therapeutic model – these individual studies exist in isolation, each using different methods and measures, and offer no basis for meaningful comparison between different approaches or populations.
What the field requires is comparative research across multiple centres using consistent methods. Rather than each provider designing isolated studies and reporting their own results, the sector needs common benchmarking to establish what works, for whom, and under what circumstances.
The American NAATP’s FoRSE programme is currently attempting to address this through centralised data collection among a large number of providers, but reporting difficulties and patchy participation appear to be limiting meaningful comparisons.
At Castle Health, we are attempting to address this gap directly. Castle Craig is currently planning a new series of outcome studies designed to move beyond the limitations of previous evaluations – adopting internationally recognised instruments, independent oversight, and the inclusion of non-completers. The rationale and proposed methodology for this approach are discussed in detail below.
The Treatment Environment: Moos and the Veterans Affairs Research
Important longitudinal outcome studies – especially those by Rudolf Moos and John Finney at the U.S. Department of Veterans Affairs – have helped identify the key therapeutic factors and programme characteristics that drive positive outcomes in residential addiction treatment. Thanks to the VA’s integrated healthcare system, Moos and his team were able to access long-term data and track outcomes for thousands of patients over periods of up to 16 years. Their research combined outcome data with robust process evaluation, offering a unique insight into what constitutes effective treatment.
In studies of over 3,000 individuals from 15 VA residential programmes, Moos found substantial improvements at 1-year, 2-year and 5-year follow-up. His comparative studies indicated that 12-step oriented residential programmes produced significantly better abstinence outcomes and 64% lower annual healthcare costs than cognitive-behavioural programmes.
Crucially, Moos showed that the quality of the treatment environment made a much bigger difference to outcomes than the individual characteristics of patients. In fact, the treatment setting accounted for seven to eight times more of the variance in results than factors such as a person’s background or condition at entry. This challenged the idea that success depends mainly on using the right technique or treatment “technology,” and instead highlighted that recovery is shaped by how well the person’s life situation fits with the support offered by the programme.
Moos was one of the first researchers to clearly separate two important factors in treatment: mediators – what happens during treatment that helps explain why it works – and moderators – the personal traits that influence whether it works for a given individual.
By focusing on what actually happens during the treatment process, he identified four key elements that consistently contributed to better outcomes: a supportive and structured environment; encouragement of healthy behaviours that replace substance use; clear abstinence-based expectations and strong peer role models; and the development of practical coping skills.
These elements closely mirror the design and philosophy of residential rehab programmes. The implication is that it is not just what treatment someone receives, but how it is delivered – and whether the environment actively supports change.
Vaillant’s Natural History of Recovery
George Vaillant’s decades-long work on the Harvard Grant Study and the Core City cohort, partly synthesised in The Natural History of Alcoholism, provides a uniquely valuable longitudinal perspective. His research followed men over periods of up to sixty years, tracking the natural course of alcohol dependence and the conditions under which stable recovery occurred. While Vaillant’s studies were not designed to evaluate residential programmes directly, their findings have profound implications for how we understand – and design – effective treatment.
Vaillant identified four factors that consistently predicted sustained recovery: experiencing the negative consequences of addiction, which creates readiness for change; finding a substitute dependency such as involvement in Alcoholics Anonymous or a similar recovery fellowship; developing new, supportive relationships that overcome the isolation of active addiction; and increased spirituality or a sense of purpose, meaning, and hope.
These factors align closely with the mechanisms identified by Moos and later process researchers. They also map directly onto the structure of quality residential rehabilitation. A well-run rehab removes someone from their drinking or using environment, immerses them in a recovery-focused peer group, teaches practical coping strategies, facilitates engagement with mutual-help fellowships, and connects them to ongoing aftercare support. In this way, residential treatment can operationalise the very conditions that Vaillant’s research identified as essential for lasting recovery.
It is worth noting the demographic limitations of Vaillant’s work: the Grant Study cohort consisted exclusively of white male Harvard graduates, though the parallel Core City cohort of inner-city Boston men provided an important socioeconomic counterbalance. Despite these constraints, the consistency of his findings across both groups – and their convergence with the process research that followed – strengthens rather than diminishes their relevance to treatment design.
12-Step Evidence
Not every rehab promotes a 12-Step approach, but the majority in the United Kingdom do integrate this method, and study after study has emerged demonstrating the effectiveness of 12-Step approaches to addiction recovery for those who engage with them. Researchers such as John Kelly conducted deeper analyses of the Project MATCH dataset, drawing new conclusions that went beyond the original findings, while a Cochrane review has provided systematic evidence supporting 12-Step facilitation as an effective treatment intervention.
Kelly’s subsequent process research identified specific mechanisms through which 12-Step participation works, with notable differences by gender. Women’s outcomes were found to be driven more by confidence in handling difficult emotions, while men’s were driven by confidence in navigating social situations without drinking. Age-related differences have also emerged: for young adults, 12-Step participation appears to help primarily by reducing substance-using contacts in their social networks, but is less effective at helping them form new recovery-supportive relationships – possibly because the average age at AA and NA meetings is around fifty, limiting access to same-aged recovery peers. Older adults, by contrast, benefit more fully from the social network mechanism in both directions. These findings have direct implications for programme design, suggesting that services working with younger patients should actively facilitate connections with same-aged peers in recovery rather than relying solely on general fellowship attendance.
Important process research by Morgenstern and colleagues has helped explain how 12-Step approaches work. Their 1997 study found that affiliation with Alcoholics Anonymous after treatment was related to maintenance of self-efficacy and motivation, as well as increased active coping efforts. This challenges common criticisms that 12-Step programmes leave people feeling helpless through concepts like “powerlessness” – instead showing that participation actually increases self-efficacy and enhances coping abilities.
The near universal integration of 12-Step approaches within residential programmes contributes to the growing evidence base supporting rehab, particularly given that residential facilities remain one of the few UK treatment settings where patients can access structured 12-Step facilitation. What was once viewed with scepticism by some has emerged as a significant therapeutic advantage with clear efficacy benefits.
Analysis
Critics raise legitimate concerns about the current rehab evidence base, noting that studies often have methodological flaws, or that their findings don’t always apply across different populations and settings. However, the cumulative evidence has been building a stronger case for rehab, and pointing the field in more productive directions.
Initially, researchers attempted to measure addiction treatment through a “technology model” – creating standardised therapy manuals and testing them like drugs in controlled trials. This approach seemed successful at first, with evidence-based treatments like cognitive behavioural therapy (CBT) and motivational interviewing showing promise. However, an unexpected pattern emerged which saw different treatments producing similar results. Whether someone received CBT, 12-step therapy, or motivational enhancement, outcomes were often the same. Even more surprising, when researchers tested whether treatments worked for the reasons they were supposed to (such as CBT improving coping skills), they often found the specific factors were not shown to have made the difference.
Critics argued the field was asking the wrong questions. Rather than trying to standardise treatment like medicine, they suggested focusing on broader questions: How do people actually change? What role do relationships, hope, meaning, and social support play? How does treatment fit into someone’s whole life context?
Today’s addiction research is moving away from the idea that treatment can be standardised like taking a course of medication. There’s growing recognition that recovery involves complex personal and social processes that cannot be easily controlled or replicated in a laboratory. The field is shifting towards understanding the nuanced reality of how change actually happens in people’s lives.
The implication for rehabs is that residential rehabilitation is unlikely ever to be definitively proved effective in the traditional medical sense. Twenty-five years ago, critics argued that the evidence base for residential treatment was very limited; they continue to make that argument today. However, rather than being a failure to conduct proper research it’s an evolution towards asking better questions. Rather than simply asking “does rehab work?”, we’re now exploring “how does it work, for whom, and under what circumstances?”
What has genuinely improved is our understanding of the healing factors (mechanisms) through which recovery happens. We now know that changes in social networks, self-efficacy, spiritual practices, and coping strategies are key drivers (mediators) of recovery. These are the same factors that Vaillant identified through decades of longitudinal observation, that Moos confirmed through process evaluation of treatment environments, and that Kelly and Morgenstern have demonstrated operate through 12-Step participation.
The convergence of evidence from such different research traditions – naturalistic long-term follow-up, programme evaluation, and clinical process research – is itself a powerful form of validation. Translating this knowledge into practice is not automatic, but it opens new opportunities for treatment centres to enhance their programmes. Rather than rigidly following therapy manuals, providers can focus on cultivating these active ingredients of change – shaping environments that support better interpersonal relations, build confidence, and reinforce purposeful living.
At Castle Health, this is already informing clinical design. In 2025, Castle Craig introduced ALAANA (Assertive Linkage to Alcoholics Anonymous and Narcotics Anonymous), a structured four-part 12-Step Facilitation workshop now integrated into the programme for all patients. Drawing on evidence-based TSF materials – particularly the American MAAEZ model, adapted to a UK context and updated in light of more recent research – ALAANA was designed specifically in response to the evidence reviewed in this paper: that structured linkage to mutual-help fellowships is among the strongest predictors of sustained recovery; that 12-Step participation activates key mechanisms including self-efficacy, social network change, and coping skills; and that the way these mechanisms operate varies by population. In Moos’s terms, it is an example of deliberately shaping the treatment environment so that patients understand not just what to do, but why it works – making the evidence real at the level of individual experience. ALAANA is one of several programme developments at Castle Health informed by this body of research, and the new outcome study methodology described below will allow us to evaluate whether such changes are producing the improvements we expect.
For rehabs, this means moving beyond simply tracking abstinence rates. Residential rehabilitation is not primarily a physical health intervention – it is a psychosocial treatment, and its outcomes should be measured accordingly, focusing on quality of life, psychological wellbeing, self-efficacy, and social functioning. The treatment centre outcome data reviewed above reveals a consistent and striking pattern: from Cook’s observation in the 1980s that “all patients appear to have gained some therapeutic benefit,” through the McCann and Amos findings of dramatically improved physical and mental health and quality of life, to the 2015 Castle Craig data showing 91.8% of respondents improved with none deteriorating – the evidence repeatedly shows that when broader measures of functioning are used, the picture is far more encouraging than abstinence figures alone suggest. The overwhelming majority of patients who engage with residential treatment experience measurable improvement in their overall condition. Providers should therefore monitor the processes we know matter: Are patients developing stronger social connections with people in recovery? Are they gaining confidence in their ability to cope with difficult situations? Are they finding meaning and purpose? These process measures, combined with longer-term outcome tracking, will provide a far richer understanding of their effectiveness than traditional medical-model evaluations ever could.
However, for this data to be credible and useful, providers must adopt rigorous methodological standards. This means using validated measurement tools that allow for comparison across services, ensuring independent oversight of outcome studies, and applying the “intent-to-treat” principle that includes all admitted patients – not just successful completers. Most importantly, rehabs need to collect comprehensive baseline data on the factors they intend to measure, since follow-up data is undermined when we don’t know exactly where patients started. When conducted properly, these studies serve dual purposes, providing honest accountability to stakeholders while generating insights that can genuinely improve treatment programmes.
Towards a Model for Outcome Evaluation
If the field is to move beyond isolated, incomparable studies towards meaningful benchmarking, individual treatment centres need a shared framework for outcome evaluation – one that is rigorous enough to be credible, practical enough to be adopted widely, and designed to measure the factors we now know matter most. What follows is a model that Castle Health has developed in collaboration with Dr Tim Leighton, whose doctoral research on change mechanisms in residential rehabilitation has informed both the instrument selection and study design. It is offered here as a potential template for the wider sector.
The starting point is baseline data. Too many outcome studies, even published ones, fail to give a clear picture of who enters treatment and their relevant characteristics. Without this, follow-up information is largely meaningless – we cannot measure change if we do not know where someone started. Baseline assessment should capture demographics (age, sex, ethnicity, relationship status, employment, education), substance use history (primary and secondary dependencies, severity of dependence, length of problem use, previous abstinence periods), prior help-seeking (including previous counselling, structured treatment, and engagement with mutual-help fellowships), and social network composition. Routine client records are often incomplete and may not capture the full range of information required; a dedicated baseline questionnaire administered separately will ensure consistency and completeness.
Beyond demographics and substance use history, baseline measurement should capture the specific outcomes known to mediate durable recovery. Research has increasingly shown that good drug and alcohol outcomes are supported or predicted by changes in factors that can be produced or strengthened during treatment and sustained afterwards. A robust example is abstinence self-efficacy – realistic confidence that abstinence can be maintained in challenging situations – a finding replicated across many studies, with indications that self-efficacy in different contexts (social situations, painful emotional states) may be more or less important for different groups. Another very robust finding is that durable recovery is associated with changing one’s social network, which both reduces active users in the network and increases the number of abstinent friends, which are independent predictors of good outcomes. Quality of life, depression, anxiety, self-esteem, recovery capital, and spiritual practices are all further outcomes that the literature identifies as significantly associated with sustained recovery.
To measure these factors economically, accurately, and consistently, we propose using validated, public-domain instruments that allow comparison across services. The proposed battery includes: the EUROHIS-QOL for quality of life; the PHQ-9 for depression; the Alcohol/Drug Abstinence Self-Efficacy Scale, which provides subscales for social situations, negative affect, physical discomfort, and craving responses; the Multi-Dimensional Inventory of Recovery Capital, preferred to shorter instruments because it provides separate scales for human, cultural, physical, and social capital; a validated self-esteem measure; and the Religious Background and Behaviour Questionnaire used in Project MATCH, which allows direct comparison with that dataset as analysed by Kelly and others. A bespoke social network assessment asks participants to identify up to five significant friends and score each on a four-point scale for their drug and alcohol use or recovery status, enabling changes in network composition to be tracked over time.
The study design itself must embody the methodological principles discussed throughout this paper. We recommend recruiting a cohort of at least 120 consenting patients from consecutive admissions, with follow-up at six and twelve months post-discharge (extendable thereafter). Crucially, the study must include non-completers to the maximum extent possible. The Clouds House study by Georgiakis found that late dropouts did at least as well as completers, whereas early dropouts did very poorly – a finding with important implications for retention strategies that would be invisible in a completers-only study. The same instruments should be administered at baseline, end of treatment, and at each follow-up point, enabling measurement of change across the treatment episode and beyond. Follow-up interviews can be conducted via secure video platforms to maximise participation, and oversight by an independent, qualified researcher will enhance the study’s credibility. The methods should be made completely transparent.
The value of this approach extends well beyond any single centre’s results. If adopted widely, the use of common instruments and methodology would enable genuinely meaningful comparison across providers for the first time. But this aspiration need not remain hypothetical.
From Individual Measurement to National Benchmarking
The outcome evaluation model described above is designed for individual centres, but its real potential is realised at scale. Castle Health is therefore calling on residential treatment providers across the UK to consider adopting this framework – or one built on the same principles – as the basis for a collaborative, multi-site effectiveness study.
The case for such a study is straightforward. UK addiction treatment outcome research has been characterised by single-site designs, inadequate sample sizes, inconsistent measurement, and limited follow-up. Each provider generates its own data using its own methods, and the result is a collection of isolated findings that cannot be meaningfully compared. We have no way of knowing whether differences in reported outcomes reflect genuine differences in treatment effectiveness, or merely differences in methodology, patient demographics, or reporting conventions. A multi-site study using shared instruments would resolve this at a stroke.
We propose a prospective longitudinal design in which all participating sites implement the validated measurement battery described above – the instruments measuring recovery capital, self-efficacy, social networks, quality of life, depression, self-esteem, and spiritual practices – at admission, end of treatment, and at 6- and 12-month follow-up intervals. Trained research staff independent of clinical teams would conduct assessments using standardised protocols across all sites, ensuring data quality and eliminating the bias associated with clinician-administered measurement. With a target of 15 to 20 participating sites and a combined cohort of sufficient scale, the study would have sufficient statistical power to detect meaningful differences between subgroups and to conduct the mediational analyses that single-centre studies cannot support. Comprehensive tracking systems and multiple contact methods would maximise follow-up retention, and intent-to-treat analysis would include all eligible admissions regardless of completion status.
The research questions such a study could answer are precisely the ones this paper has argued the field needs to address. What baseline characteristics predict treatment completion and sustained recovery? How do outcomes vary across sites and treatment modalities? Do increases in self-efficacy and recovery capital during treatment predict recovery at follow-up? Do social network changes mediate the relationship between treatment and sustained outcomes? Rather than being abstract academic questions, these have direct implications for programme design, treatment matching, and commissioning decisions.
Crucially, this would be a collaborative endeavour, not a competitive one. Data would be anonymised, oversight would be independent, and the methodology would be transparent. Participating providers would receive detailed feedback on their own outcomes benchmarked against the wider dataset – information that is valuable for service improvement regardless of where a centre sits in the overall distribution. The aim is not to rank providers but to generate the shared evidence that strengthens every provider’s position with commissioners, funders, regulators, and the public. Residential rehabilitation has been asked to justify itself for decades; a sector that can present credible, comparable, independently overseen outcome data from multiple sites running different models has an answer that no single provider, however rigorous, can offer alone.
Castle Health is committed to participating in and helping to convene such a study. We invite providers, researchers, and sector bodies with an interest in building this evidence base to make contact.
Conclusion
The critical questions have evolved from “does rehab work?” to “who benefits and why?” The evidence reliably shows that residential rehabilitation is effective, including for those with severe dependencies, particularly when integrated with appropriate aftercare. The large-scale national studies – most of which were underway or publishing key findings by the early 2000s – demonstrated this clearly, and while we have had only smaller studies since then, what has emerged remains supportive.
Furthermore, a crucial reframing has emerged from within the research field itself. In a landmark 1996 paper, McLellan, Woody, and colleagues argued that addiction treatment had been evaluated against inappropriate standards – judged by acute-care expectations of one-off cure rather than by the standards applied to other chronic conditions. When addiction treatment outcomes are compared with those for diabetes, hypertension, or asthma – in terms of treatment adherence, relapse rates, and the need for ongoing management – addiction treatment performs comparably well.
The recovery trajectory data reinforces this: if sustained remission typically emerges over multiple engagements across several years, then the appropriate question for any single treatment episode is not “did it cure?” but “did it meaningfully advance the patient’s recovery journey?” This perspective, later developed further in McLellan and colleagues’ influential 2000 JAMA paper on drug dependence as a chronic medical illness, does not lower the bar for rehab providers; rather, it sets the bar in the right place, demanding that treatment be evaluated for what it can realistically deliver within a continuing care framework, and that providers be held accountable for the quality and rigour of what happens during the treatment episode itself.
The language of a recent UK government self-assessment guide for residential rehab signals a notable shift in tone, and is a far cry from the era of the National Treatment Agency (NTA), when rehabs were systematically marginalised. The change of tone and direction in Scotland is even more promising. This renewed openness reflects both a growing recognition of 12-step-informed approaches and the inability of alternative treatment models to demonstrate superior outcomes. More importantly, we now understand the healing mechanisms that drive recovery – changes in social networks, self-efficacy, and coping strategies – knowledge that allows programmes to focus on what actually works rather than simply following therapy manuals.
Residential rehabilitation is, and will remain, an essential part of the treatment landscape. The evidence shows that it performs no worse than intensive outpatient alternatives, and when all costs are considered – including higher dropout rates in outpatient settings and the practical reality that most patients do not have access to intensive community programmes – cost differentials may not be as significant as assumed, particularly given that many charity-run residential facilities operate on remarkably economical budgets. Even as we improve our ability to match people to the right interventions, and as community-based alternatives evolve, there will always be a significant cohort for whom immersive, structured, 24/7 care is the most appropriate and effective option.
The evidence does enough to show that it works; our task now is to do more to understand how to make it work better, and identify those who need it most.
References
Banbury Lodge. (n.d.). Banbury Lodge: Alcohol & drug rehab centre in Oxfordshire. Retrieved July 21, 2025, from https://www.banburylodge.com/
Cook, C. C. H. (1988). The Minnesota Model in the management of drug and alcohol dependency: miracle, method or myth? Part II – evidence and conclusions. British Journal of Addiction, 83(7), 735–748.
De Andrade, D., Elphinston, R., Quinn, C., Allan, J., & Hides, L. (2019). The effectiveness of residential treatment services for individuals with substance use disorders: A systematic review. Drug and Alcohol Dependence, 201, 227–235.
Dennis, M. L., Scott, C. K., Funk, R., & Foss, M. A. (2005). The duration and correlates of addiction and treatment careers. Journal of Substance Abuse Treatment, 28(Suppl 1), S51–S62.
Edwards, G. (1987). The Treatment of Drinking Problems: A Guide for the Helping Professions (2nd ed.).
Gossop, M., Marsden, J., & Stewart, D. (2001). NTORS after five years: The National Treatment Outcome Research Study – changes in substance use, health and criminal behaviour during the five years after intake. National Addiction Centre.
Hughes, P. (1999). Outcome study of alcohol dependent patients at Castle Craig Hospital. Castle Craig Hospital.
Kelly, J. F., Greene, M. C., Bergman, B. G., White, W. L., & Hoeppner, B. B. (2019). How many recovery attempts does it take to successfully resolve an alcohol or drug problem? Estimates and correlates from a national study of recovering U.S. adults. Alcoholism: Clinical and Experimental Research, 43(7), 1533–1544.
Kelly, J. F., Humphreys, K., & Ferri, M. (2020). Alcoholics Anonymous and other 12-step facilitation programmes for alcohol use disorder. Cochrane Database of Systematic Reviews, 3, CD012880.
Kelly, J. F., Bergman, B. G., & Fallah-Sohy, N. (2018). Mechanisms of behavior change in 12-step approaches to recovery in young adults. Current Addiction Reports, 5(4), 516–527.
Kelly, J. F., Stout, R. L., Greene, M. C., & Slaymaker, V. (2014). Young adults, social networks, and addiction recovery: Post treatment changes in social ties and their role as a mediator of 12-step participation. PLoS ONE, 9(6), e100121.
Leighton, D. T. H. (2017). Inside the Black Box: An exploration of change mechanisms in drug and alcohol rehabilitation projects. PhD thesis. University of Bath, Department of Education.
McCartney, D. (2020, December 20). Is rehab effective or a waste of time and money? Recovery Review Blog. Retrieved July 21, 2025, from https://recoveryreview.blog/2020/12/20/is-rehab-effective-or-a-waste-of-time-and-money/
McCann, M. A., & Amos, P. (2000). Outcome study of patients treated in the Extended Care Unit at Castle Craig Hospital. Castle Craig Hospital.
McKeganey, N., Bloor, M., McIntosh, J., & Neale, J. (2006). Abstinence and drug misuse treatment: Results from the Drug Outcome Research in Scotland study. Drugs: Education, Prevention and Policy, 13(6), 537–550.
McKeganey, N. (2011, January). The effectiveness of residential rehabilitation services in the treatment for drug and alcohol problems (Unpublished report). University of Glasgow.
McLellan, A. T., Woody, G. E., Metzger, D., McKay, J., Alterman, A. I., & O’Brien, C. P. (1996). Evaluating the effectiveness of treatments for substance use disorders: reasonable expectations, appropriate comparisons. Milbank Quarterly, 74(1), 51–85.
McLellan, A. T., Lewis, D. C., O’Brien, C. P., & Kleber, H. D. (2000). Drug dependence, a chronic medical illness: implications for treatment, insurance, and outcomes evaluation. Journal of the American Medical Association, 284(13), 1689–1695.
Miller, W. R., Walters, S. T., & Bennett, M. E. (2001). How effective is alcoholism treatment in the United States? Journal of Studies on Alcohol, 62(2), 211–220.
Moos, R. & Finney, J. (2000). Inpatient substance abuse care and the outcome of later community residential and outpatient treatment. Addiction, 95(6), 833–846.
Morgenstern, J., Blanchard, K. A., Morgan, T. J., Labouvie, E., & Hayaki, J. (1997). Examining mechanisms of action in 12-step community outpatient treatment. Drug and Alcohol Dependence, 49(3), 233–242.
National Institute for Health and Care Excellence. (2007). Drug misuse: Psychosocial interventions (NICE Clinical Guideline 51). https://www.nice.org.uk/guidance/cg51
National Institute for Health and Care Excellence. (2011). Alcohol-use disorders: Diagnosis, assessment and management of harmful drinking and alcohol dependence (NICE Clinical Guideline 115). https://www.nice.org.uk/guidance/cg115
Office for Health Improvement and Disparities (OHID). (2025). Residential drug and alcohol treatment: self-assessment guidance. Department of Health and Social Care. Available at: https://www.gov.uk/government/publications/residential-drug-and-alcohol-treatment-self-assessment-toolkit/residential-drug-and-alcohol-treatment-self-assessment-guidance [Accessed 24 July 2025].
Raistrick, D., Heather, N., & Godfrey, C. (2006). Review of the effectiveness of treatment for alcohol problems. National Treatment Agency for Substance Misuse, University of York.
Strang, J., & Gossop, M. (Eds.). (2005). Heroin Addiction and ‘The British System’, Volume II: Treatment and Other Responses. Routledge.
Vaillant, G. E. (1983). The Natural History of Alcoholism: Causes, Patterns, and Paths to Recovery. Harvard University Press.
Vaillant, G. E. (2012). Triumphs of Experience: The Men of the Harvard Grant Study. Harvard University Press.
White, W. L., & Godley, M. D. (2005). Addiction treatment outcomes: Who and what can you believe? Counselor, 6(3), 52–55.
