‘Dangerous data’: drinking after dependence

Send email for updates


Effectiveness bank home page. Opens new windowHot topic

Below is one of our selection of Hot Topics, important issues which sometimes generate heated debate over the facts or their interpretation. Click the GO button or the Blue title to trigger a customised search for relevant Findings logo documents. Links to other documents. Hover over for notes. Click to highlight passage referred to. Unfold extra text Unfold supplementary text

Copy title and link | Comment/query | Other hot topics |

‘Dangerous data’: drinking after dependence


First cracked in 1960s London, the orthodoxy that abstinence is the only feasible treatment goal for ‘alcoholics’ seemed shattered in 1973 by evidence that even physically dependent patients could learn to drink in moderation. Controversy was fierce, reaching the US Congress, TV networks and the courts. Explore the history and contested research behind an issue facing every dependent drinker starting treatment.

 1 Why the heat?

Your cholesterol is high and your doctor says, “No butter, no cheese, no cholesterol-raising foods – full stop.” You plead, “Can’t I just cut down and take some tablets?” The doctor is unmoved: “If you want me to help, do as I say, otherwise you are clearly not serious about protecting your health. Come back after you have a heart attack – then maybe you’ll see it my way.”

Contents

 1 Why the heat?
The controlled drinking issue and why it aroused such passion.

Milestones

 2 A gentlemanly start
In early 1960s London, psychiatrist D.L. Davies opened up the first telling crack in the abstinence-only consensus.

 3 The gloves come off; the Rand report
Support for the feasibility of controlled drinking in a 1976 US report was likened to “playing Russian roulette with the lives of human beings”.

 4 The Sobells decisively crack the abstinence consensus
Published in 1973, findings from the hard-to-explain-away solidity of a randomised trial showed successful treatment of US patients need not conform to abstinence-only orthodoxy.

 5 Sobells in the firing line
The heat is turned up in a bitter and decades-long controversy over the Sobells’ findings.

Contemporary evidence

 6 Evidence accumulates and expert opinion converges
The heat dies down as it becomes clearer that neither controlled drinking nor abstinence has a definitive advantage as a treatment goal, and as evidence emerges about what types of patients do better with either strategy. Major UK and US trials confirm the feasibility of controlled drinking as a treatment destination, especially for less severely dependent and vulnerable patients with social support for moderation.

 7 Who takes the decision and how?
With both objectives on the care-planning table, shared decision-making between client and patient has become the recommended way to decide.

 8 What do the authorities say?
Guidance from Britain’s National Institute for Health and Care Excellence and from the Department of Health’s National Treatment Agency for Substance Misuse accepts controlled drinking as a goal for less severely dependent patients or when abstinence is rejected.

Not so long ago that was effectively the stance dependent drinkers could expect to face. At issue was not just what patients should be advised, but whether they should be denied treatment until revelation or deterioration impressed on them the need to stop drinking altogether and forever. Moderation tempted but was seen as merely a steep and slippery slope to excess; abstinence was the only safe ground.

Challenges to this orthodoxy generated the most bitter and prolonged controversy ever seen in substance use treatment. Decades behind us now and with the doors closed on some of the fiercest debates, still the impacts and underlying drivers remain. Far from receding into a box labelled ‘pointless debates’, prioritising abstinence as a treatment objective returned to policy prominence in the UK from 2008 as a component of influential visions of ‘recovery’.

In academia the heat died down somewhat as controlled-drinking objectives accumulated research support, but was stoked again when in 2020 the prestigious vehicle of a Cochrane review was interpreted by its authors as vindicating what in the USA and perhaps too internationally are the main structured routes to abstinence – approaches based on Alcoholics Anonymous’s 12 steps. “Let’s not turn back the clock,” warned a commentary on this interpretation from the lead author of the book Controlled Drinking first published in 1981. Professor Nick Heather was concerned that “an exclusive focus on abstinence in treatment and the use of [Alcoholics Anonymous/12-step facilitation treatment] as the sole means to achieve it, which the Cochrane review … is likely to encourage, ignores decades of progress in broadening and articulating the treatment response to [alcohol use disorder], together with the findings of recent rigorous research, and is therefore retrogressive.”

Why such heat over a seemingly innocuous decision between patient and clinician on which form of remission to go for? And why does it persist, despite evidence that the health benefits of reducing drinking from very high to high levels can be greater than reductions from lesser levels to zero? In part heat was generated by concerns that allowing controlled drinking would let ‘alcoholics’ (presumed unable to stop drinking once they start) off the hook of non-drinking and set them up to fail, perhaps fatally. On the other side was the concern that while insisting on abstinence did nothing to improve outcomes, it did limit treatment to the minority whose problems were so severe they were prepared to countenance a life without drink. Underlying these views were incompatible visions of dependence as a distinct disorder characterised by inevitable loss of control, or one end of a continuum of behaviour which even at its most extreme can be replaced by moderation if the circumstances are sufficiently supportive.

In his response (free source at the time of writing) to the Rand report (of which more below), Ron Roizen, a long-time student of the history and sociology of problem drinking in the USA, took us deeper into why its findings on the feasibility of post-treatment controlled drinking were seen as so “dangerous”. Dr Roizen drew a distinction between the truth of a theory and its utility, stressing that mistaken theories may still be thought useful. As promulgated then and now, the “classical” disease theory of alcoholism is a special case of this duality, since “acceptance of that theory is itself the essence of alcoholism treatment” – ‘treatment’ actually consists of convincing the patient they are atypical and different from normal drinkers, that they are an ‘alcoholic’ to the core and for ever, constitutionally unable to touch a drop without descending into a destructive bender. From this perspective, “ ‘abstinence’ is not solely a measure of the patient’s improvement, but a sacred and essential element in the ‘treatment’ process … a sign that the model of alcoholism has been accepted by the patient”.

Further reading

For more see this detailed and freely available US account from Ron Roizen. It dealt more than adequately with the history of the controversy, but was unable to fully take on board later developments arising from early British findings. A much cited book offers a British perspective on the debate (turn to chapter four) and the evidence. On the evidence see also this Effectiveness Bank analysis of a recent UK study (the background notes are particularly informative) and this review.

Recognising that enables us to understand visceral reactions to challenges to an abstinence-based understanding of recovery. The perceived utility of the theory as a treatment tool capable of rescuing the otherwise un-rescuable holds only insofar as “therapists can present the theory honestly and openly without fear of contradiction … Without the ability … to create a genuine conviction in the classical disease concept of alcoholism, the theory, its treatment implications, and its authority and legitimacy dissolve.” From this point of view, “an attack on the abstinence criterion is an attack on the classical disease concept of alcoholism … And undercutting that truth is only done at great peril because the embracing of that truth proves to be the most successful treatment known for the condition.”

But ‘undercut’ it was; how, and what the reaction was, is the story told in this hot topic. It began in earnest in the early 1960s in south London with the first research-driven crack in the abstinence consensus, later to become a gaping wound in the USA. Here we re-examine the major milestones along this journey, distinguished by the bitterest conflicts ever to mar scientific discourse on addiction treatment. So many commentaries and studies have been devoted to the issue that this contribution cannot claim to be comprehensive: see the documents recommended in the “Further reading” panel to fill in the gaps.

 2 A gentlemanly start

D.L. Davies: opened up the first telling crack in the abstinence-only consensus

D.L. Davies: opened up the first telling crack in the abstinence-only consensus.

Though it flared hottest in the USA, the controversy dates back (1 2) to a 1962 report by British physician and psychiatrist David Lewis Davies (referred to almost universally as D.L. Davies) on seven patients discharged before 1955 from south London’s Maudsley hospital. Followed up in 1961, though previously “severely addicted” these men (they were all men) were said after discharge to have sustained controlled drinking. They were very much the minority of a total of 93 patients – but given their status on admission, that they existed at all was considered remarkable.

Davies’ report started by restating the views of the time: due to presumed “irreversible” changes after years of heavy drinking, “Among those who treat alcoholics there is … wide agreement that these patients will never again be able to drink ‘normally’.” Yet it seemed the seven had, and for between seven and eleven years, associated with major changes in working or domestic lives which divorced them from constant contact with drink, or resolved the troubles which had helped generate and sustain dependence.

Witness to convention’s grip, the report ended by endorsing the orthodoxy its findings challenged: “It is not denied that the majority of alcohol addicts are incapable of achieving ‘normal drinking’. All patients should be told to aim at total abstinence.” Davies did, however, see his findings as giving the lie to the aphorism, ‘once an alcoholic, always an alcoholic’. Approaches not “constrained” by this view could achieve “complete cures”: bolstered by radical changes in their lives (aided in these cases by two to five months in hospital), severely dependent drinkers whose drinking has social roots need not spend their lives in deepening dependence or teetering on its edge if they touched a drink. It was this use of the findings to, even if only modestly, deny the absoluteness of the abstinence-only mandate which distinguished Davies’ report from earlier accounts. The effect was to generate a considerable response, but one limited to professional circles and “dismissive” rather than outraged.

Even before Davies’ report abstinence had been tilted from its pedestal by observations that moderation was possible after dependent drinking, and that considering the individual’s entire life circumstances, abstinence was not always the key to overall improvement or the preferred drinking outcome. The year Davies’ report was published also saw publication of the classic description of ‘The abstinent alcoholic’ – formerly dependent drinkers who have sustained abstinence but are nevertheless unhappy, unfulfilled and/or nervously hanging on. In the account from Connecticut alcohol clinics in 1950s USA, they were the majority among the non-drinkers. With Davies’ report in the vanguard, by the mid-60s the pieces of the jigsaw had been amalgamated into a systematic and comprehensive challenge to physiological determinism in the construction of ‘alcoholism’ and the associated elevation of abstinence to the sole acceptable treatment goal and yardstick of success.

Findings contested

Some of the 18 responses to Davies’ report published by the same journal put their fingers on the key practical issue: how to absolutely extinguish ‘dangerous’ chinks of hope that post-dependence moderation might be possible, when research showed these chinks existed, and were not even extremely rare. Conceding that Davies’ seven were truly alcoholic would it was feared, “cloud the clarity of our definition of alcoholism and our exactness of treatment, for [the few exceptions] occasionally make liars out of us when we tell a patient he is an alcoholic and can never drink again.” Redefining these exceptions as “pseudo-alcoholics … is the only way we can stick to our guns that a true alcoholic can never drink again. If we ever concede this dictum, we, as therapists, and our alcoholic patients, will be lost.”

Exceptions make liars of us when we tell a patient he can never drink again … we will be lost

Ironically, these doom-laden concerns came from a doctor who well before Davies’ report had himself documented the phenomenon of non-abstinent recovery among what he had then unambiguously termed “alcoholics”. Faced with the implications spelt out by Davies, his focus turned to how conceding that some people do manage non-abstinent recovery might undermine an unswerving abstinence-only stance in treatment, leading him to reverse-engineer his definition of alcoholism to exclude these ‘exceptions’. Similar concerns were to be expressed ( below) over a decade later in response to a US report which reinforced Davies’ contentions. In both cases, the impulse was to suppress or redefine findings if they contradicted the theory, the opposite to the way science is meant to develop by contradictory findings forcing theories to be amended or abandoned.

Replying to the comments on his article, D.L. Davies made the point that far from denying the value of abstinence, all seven patients had first ‘reset’ their relationships with alcohol by a period without drinking at all, and that he remained of the view that “The patient who challenges the doctor’s advice to become a teetotaller, on the basis of what I have reported, should be told that the evidence suggests that only a minority may recover completely, and that he … would be wiser to assume that he will not be in that group.” Doubts that some of the seven were truly ‘alcoholic’ were countered by pointing out that this diagnosis had been made before it was known how the patients would fare after discharge, and that “In practice, one feels satisfied that a man is addicted to alcohol when he has tried to break off his use of the drug, which in some way is proving harmful, and has failed.”

Griffith Edwards: claimed his “mentor” had been deceived

Griffith Edwards: claimed his “mentor” had been deceived.

Aided by access to Davies’ records and fresh data from the same seven patients, another 23 years passed before the most robust critique was published, and it came from within the institution Davies had led. It started by recognising the magnitude of Davies’ intervention. His successor at the Maudsley and the associated Institute of Psychiatry in south London lauded Davies as “a pioneer who made a daring exploration of what was at the time virtually forbidden territory”. Later he explained that “Davies was questioning not just a medical consensus, but the central and hallowed organising idea of the American alcoholism movement.” These comments came from the prestigious figure of the late Griffith Edwards, but a sharp edge embellished his homage to his predecessor and “mentor”.

That edge had become apparent in 1979 when the journal Edwards started editing the year before published an interview with Davies. The interviewer – probably Edwards himself – told Davies of a personal encounter at the Maudsley with one of his seven patients. Contrary to the impression given to Davies in 1961, he had confessed to “drinking like a fish the whole time” and threatening to “bash the living daylights” out of his wife if she told the truth to Davies’ follow-up worker, to whom she had confirmed his more positive account. Significantly as it later transpired, Professor Davies also confessed to something: “I never regarded myself … as a research worker.”

The encounter prompted Edwards to re-examine such records as remained and to interview all six surviving patients plus relatives or carers. Conducted in 1983, the study was published in 1985. Having died in 1982, Davies could not challenge findings which cast doubt on whether some of the patients truly were severely dependent at the time of admission and whether most really had sustained ‘normal’ drinking after discharge. How starkly different was the picture gained from similar sources just over two decades before can be appreciated by the notes on ‘Case 2’. In 1961 Davies had seen a success story: “Drinks 1–2 pints of an evening but no spirits. Never drunk.” In 1983, Edwards saw a “catastrophic outcome”: “Heavy drinking recommenced not later than 1955; much subsequent morbidity culminated in 1975 with Wernicke-Korsakoff syndrome about 1973.”

In 1994 Professor Edwards revisited his critique of Davies. His account implied that “Case 2” was the patient who had prompted the new follow-up. Yet in respect of this critical patient, mysteries remained, such as how he had thought he could successfully deny drinking problems when according to Edwards, during the relevant period he had been seen for those problems by one of Davies’ own clinical team. In 1994 Edwards said Case 2 had been discharged in 1954, while earlier he had said it had been 1950.

On methodology too, there was a gaping discrepancy between Davies’ account of his study and that reconstructed by Professor Edwards. In his 1979 interview Davies had explained: “Every one of [the seven patients] was personally seen. Certainly Edgar Myers saw them and … I’m sure I saw them all myself.” In his 1962 report itself, Davies had mentioned another source – a female psychiatric social worker who “made a personal visit to their homes and in some cases to their place of work [and] made specific inquiry (from relatives as well as from the patient) about the drinking history since discharge.” After re-examining surviving records, Edwards concluded that Davies had relied largely on the “reports written by the follow-up social worker”, who in only one case had seen the patient himself. If in fact Davies had much more information, it might partly account for the discrepancies in the pictures given of how the patients had fared.

Adjudicating between these accounts is beyond our resources, and perhaps impossible at this remove from the events. Professor Edwards was, however, in no doubt. In 1994 he remained confident that his check on the patients had revealed that Davies’ account was “substantially inaccurate”. A research-naive clinician “had been substantially misled” by “intentionally unreliable witnesses,” which his (in retrospect) flawed methodology was not up to exposing. Be that as it may, later not-so-flawed work described below was to come to the same conclusions as Davies.

In a book published in 2003, Edwards later embraced normal drinking as a goal for many patients, but still maintained that (emphasis added) “abstinence is the only feasible objective” for those with a fully developed history of dependence. Among his criteria for identifying who should attempt which objective were those ( below) trialled by the Sobells in the USA.

 3 The gloves come off; the Rand report

The Davies episode was gentlemanly in conduct and limited to professional circles, but the following decade bitter disputes originating with US research overflowed across newspaper headlines and TV networks, in one case spawning legal proceedings.

A trusted mainstream source became the target of one of the hottest disputes. Funded by the US federal government’s National Institute on Alcohol Abuse and Alcoholism, in 1976 the Rand Corporation had published a report on new government alcoholism treatment centres. Rand was and remains a respected non-profit research centre known for its “empirical, nonpartisan, independent analysis”, whose origins lie in strategically important work for the US military.

David Armor: first author of the Rand report criticised for disseminating “dangerous” information

David Armor: first author of the Rand report criticised for disseminating “dangerous” information.

Using routine records collected at intake and six months later plus a follow-up survey of patients 18 months after intake, Rand’s researchers had found that fairly complete remission was the norm (70% were substantially improved on a range of measures), that most patients achieved this without altogether stopping drinking, and that at follow-up about as many had been drinking ‘normally’ (moderately at levels “far below what could be described as alcoholic drinking”) over the past six months as had sustained abstinence. Among “definite alcoholics”, between six to 18 months relapse was no more common if former patients had been drinking normally (16% had) than if they had sustained long-term abstinence. For the researchers, these were the crunch findings: “We cannot overemphasize the import of these findings … it appears that some alcoholics do return to normal drinking with no greater likelihood of relapse than alcoholics who choose permanent abstention.”

The study had its weaknesses, such as very incomplete six-month follow-up data routinely collected by the treatment centres (at the best centres, still only about a third of patients) and a better but still low follow-up rate of 62% in the researchers’ 18-month follow-up, which drew its sample from just eight of the 44 centres. Several checks reassured the researchers that their samples remained reasonably representative of all patients within the study’s remit. Nevertheless, sampling limitations were among the reasons why they did not claim to be able to track relapse as such, just to compare relapse rates between different groups of clients. But when that comparison was between post-treatment abstainers versus moderate drinkers, it was enough to endow the study with landmark status.

‘Dangerous’ data

Unlike Davies’ seven “exceptions”, Rand had found moderation almost as common and enduring a mode of recovery among alcohol treatment patients as abstinence, findings difficult to dismiss and a more substantial challenge to the hegemony of abstinence. Aware of the storm their findings might provoke, the authors disavowed any intention of recommending remitted alcoholics resume drinking. Still the storm broke, overflowing from scientific circles to numerous newspaper reports and editorials generally reaffirming the prudence of maintaining the abstinence standard. Deploying a metaphor which echoed over the years, holding out the prospect of controlled drinking was likened to “playing Russian roulette with the lives of human beings”.

Speaking the year after their report was published, Rand’s authors highlighted as “the most serious” of the critiques “at least from the viewpoint of scientific freedom” a strand also seen ( above) in the response to Davies’ paper – the willingness to suppress unwelcome findings. “Rather than denying the reality that some alcoholics are able to resume normal drinking, the thrust of these criticisms was directed instead at disseminating such ‘dangerous’ information to the public. The essence of this position seems to be that even the suggestion that abstinence may not be absolutely necessary for every alcoholic is so dangerous as to warrant suppression.” For David Armor and his team, the ‘danger’ lay in the opposite direction – of inflexibility about treatment goals deterring “a large segment of the alcoholic population who currently do not seek help from traditionally based therapies”.

Extended follow-up shows who manages controlled drinking best

In 1980 Rand conducted a four-year follow-up of the same patients, enabling a more extended assessment of whether moderate drinking could be sustained. This time information was obtained on 85% of the targeted sample. Based on the last six months of the four years, 54% were classified as “problem drinkers” and 46% in remission, constituting 28% who had not drunk at all and 18% “drinking without problems”. Two-and-a-half years later, 30% who at 18 months had been “long-term abstainers” (ie, for at least six months) had relapsed compared to 53% of short-term abstainers. In the middle were the 41% of non-problem drinkers who relapsed; when relevant variables were analysed together, they were not significantly more likely than the long-term abstainers to have resumed problem drinking.

Though overall abstinence showed no significant signs of being the superior anti-relapse strategy, it did show such signs for a subgroup of the most dependent patients. Those who were older (40 and over on admission) and also more severely dependent were less likely to relapse after post-treatment abstinence than after non-problem drinking, while the reverse was the case for those younger and less severely dependent. Patients who were severely dependent on entry to the study (specifically, during the month before) were after treatment far less likely to become moderate drinkers than abstainers, the breakdown was about even for lesser degrees of dependence, while patients who did not register symptoms of dependence at the start of study were much more likely to be continuing to drink moderately than abstaining. It was the proportion drinking moderately which changed most as dependence levels increased, becoming far less common (12%) after severe dependence than after the absence of dependence symptoms on entering treatment (45%). So to simplify, the answer to whether treated alcohol-dependent patients can in the long-term sustain moderate drinking was, ‘Yes they can and it is not uncommon, especially when dependence is less severe and entrenched.’

Rand’s authors themselves felt the most important implication of their findings was that “the key ingredient in remission may be a client’s decision to seek and remain in treatment rather than the specific nature of the treatment received” – an insight revisited decades later after another major US study – the Project MATCH trial.

 4 Linda and Mark Sobell: groundbreaking trial led to accusations of fraud

Linda and Mark Sobell: controversy peaked over their groundbreaking trial

Linda and Mark Sobell: groundbreaking trial led to rejected accusations of fraud.

One reason why Rand’ researchers expected their findings to be controversial was the reaction to a US trial conducted in 1970 by husband and wife team Mark and Linda Sobell, the key results from which had been published in 1973. There was also a direct connection: Linda Sobell supplied data for an appendix to the Rand report.

Rand’s report came from a respected mainstream source, but in 1970 the Sobells were at or towards the start of their careers in psychology research (1 2). The young couple devised an audacious and for the time methodologically advanced experiment which, unlike the studies from D.L. Davies and Rand, tested controlled drinking using the gold-standard methodology of a randomised trial – the format which most reliably isolates an intervention as the cause of the findings rather than differences between participants assigned to it versus a comparator.

At Patton State Hospital in California they recruited 70 men voluntarily admitted for the treatment of alcohol dependence, all classified as “Gamma alcoholics”, meaning they suffered physical withdrawal symptoms when they stopped drinking and had lost control over their consumption. These 70 were the test bed for a therapy programme which in suitable patients aimed for controlled drinking, in the context of a study capable of detecting with a high degree of certainty whether – as alleged for Davies’ patients – they were ‘pulling the wool’ over researchers’ eyes. It was perhaps these strengths which helped propel the Sobells episode to the peak of the controversy and generate its enduring status as the most telling of the early challenges to abstinence-only orthodoxy.

Promising first-year follow-up findings

The first step in the study was to divide the 70 patients into 30 considered suitable to aim for abstinence and 40 for controlled drinking, the latter chosen principally on the basis that patients had sought this objective, shown in the past they could manage it, and had a supportive environment to return to on discharge. Half of each set were then allocated at random to the hospital’s normal abstinence-oriented treatment, forming control groups against whom to benchmark the impact of radical programme (free source at time of writing) the other half were assigned to.

Design of the Sobells’ study

The new programme was geared to the goal – abstinence or controlled drinking – for which the patients had been considered most suitable. It entailed allowing them to drink during treatment in a mocked-up bar, electric shocks when they drank at all or exceeded controlled-drinking standards, showing videos of how they looked when drunk, and training in how to manage or avoid what for the patient were situations conducive to drinking or over-consumption. Of the resulting four groups of patients, the key group were the 20 considered suitable to aim for controlled drinking and whose treatment was geared to this objective, represented bottom left in the figure above. Compared to the rest, their progress would show whether allowing ‘alcoholics’ to pursue treatment goals which involved continuing to drink would prove to be the disaster orthodoxy predicted.

Published in 1973 in the journal Behaviour Research and Therapy, first-year follow-up results turned any expectation of disaster on its head. Over the year after discharge from hospital, on 70% of days the 20 patients considered suitable for and trained in controlled drinking neither drank heavily nor were consigned to hospital or prison due to their drinking – the best results of all the patients, and achieved despite controlled drinking on around a quarter of days. The 70% ‘good days’ figure was halved to 35% among patients who though similarly suitable to try for controlled drinking, had been left to the clinic’s conventional abstinence-oriented treatment; instead they spent half their days “drunk”. Neither was there any sign of a fading in effect. Over the second half of the follow-up year, at 73% versus 32% the gap between controlled-drinking patients and their controls was slightly greater than in the first six months, and this at a time when all the most successful set of patients drank at some stage. Patients also allocated to the new programme, but aimed at the abstinence goal for which they were considered suitable, did almost as well as those trained in controlled drinking, but nearly always by not drinking at all.

These results were not due to extremes dominating the averages; the disparities were similar when expressed as proportions of patients (rather than of days) who spent most of their time not drinking heavily – 85% of those trained in controlled drinking versus 32% of their controls. At 87%, patients allocated to the new programme but geared to abstinence also outperformed their controls, of whom barely more than a quarter (27%) managed mostly to avoid heavy drinking.

Rejection of US orthodoxy need not condemn patients to the progressive deterioration predicted for untreated ‘alcoholics’

How did the Sobells see their findings? Of “major importance” they said was that the patients “trained in and encouraged to practice controlled drinking” had demonstrated “their successful capacity to acquire and sustain such behavior”. Significantly in the light of later developments ( below), the researchers relied for this judgement not on the success of these patients in approximating societal drinking norms, but their success relative to others in the same study: no matter how limited it might have been, if their progress was greater than that of their comparators, the programme was at least a step in the right direction.

The results seemed a clear vindication of an intervention based on seeing addiction as a learnt behaviour and of the judicious allocation of even physically dependent patients to learn how to control and moderate their drinking. Controlled-drinking patients had been selected partly because of their “sincere dissatisfaction with [Alcoholics Anonymous] and with traditional treatment modalities”; the study showed this rejection of US orthodoxy need not condemn them to the progressive deterioration predicted for untreated ‘alcoholics’.

There were, however, gaps in the design of the research. No set of patients considered suitable for abstinence was instead offered training in moderation – perhaps an ethically impossible conjunction. Neither was any patient considered suitable for a controlled-drinking objective instead trained in abstinence using the Sobells’ radical programme. It left open the possibility that patients who wanted to go for controlled drinking, and seemed as if they could manage this, would have done better if trained instead in abstinence. The greatest methodological concern (acknowledged by the authors) was that the interviewer who gathered the follow-up data knew which group the interviewee had been allocated to, so was in a position to deliberately or inadvertently steer their responses to advantage the new treatment programme. But perhaps the biggest gap was that a follow-up limited to a year left critics free to reassure themselves that no matter how promising in the short-term, the performance of the controlled-drinking patients could not be sustained. Subsequent follow-ups addressed this uncertainty.

Promise sustained into second and third years

In 1976 and 1978 results were published for the second and third years after the patients had been discharged from hospital, enabling an assessment of whether what orthodoxy predicted would be a disaster had simply been delayed. If anything, the reverse was the case. The controlled-drinking trainees whom abstinence-only advocates would have considered most at risk continued on key measures to do best, and even better than during the first year.

During the second year the same data-gathering methodology as in the first yielded similar results. Up from 70% the year before, on 85% of days patients considered suitable for and trained in controlled drinking neither drank heavily nor were in hospital or prison due to their drinking – still the best results of all the patients, achieved while they drank in a controlled manner on nearly a quarter of days, and significantly superior to the 42% among their controls. As an attempt to assess holistic recovery, a new outcome measure combined (with equal weightings) drinking status, how informants saw the patients’ adjustment to relationships and stressful situations, and patients’ satisfaction with their occupational/vocational status. Patients trained in controlled drinking also excelled on this measure, suggesting their recovery was the most broad-based of the four sets of patients and significantly broader than among their controls.

For the third year the methodology changed, as did the researchers. Addressing accusations of bias on the part of the Sobells, a new set of researchers not associated with their work conducted the study, and their interviewers and the team who rated the interviews were not told which group patients belonged to, correcting the main methodological weakness of the Sobells’ follow-ups. Led by Glenn Caddy, two of the three new researchers had also effectively checked the Sobells’ second-year findings, again with assessors ‘blind’ to which set patients belonged to.

Of the 70 original participants, 62 were located and 49 of the surviving 58 were interviewed. The controlled-drinking patients had continued to show improved outcomes, though participants missing from the data may have somewhat altered the picture. Though just three of the 20 had abstained throughout the third year, the 13 patients who could be assessed had avoided heavy drinking on 95% of days, including 29% in the form of controlled drinking. Corresponding figures for their controls offered conventional abstinence-oriented treatment were 75% and 35%. Days of heavy drinking (“drunk days”) were 5% versus 25%. The controls too seemed to have improved relative to previous years, but the gap in heavy-drinking days between them and patients trained in controlled drinking remained substantial and statistically significant.

 5 Sobells in the firing line

Just as with Davies’ research at the Maudsley ( above), the most robust challenge to the Sobells’ findings came on the back of new data from the patients themselves in the form of another follow-up by a sceptical research team. Published in Science magazine in 1982, it cast doubt on the validity of the earlier findings – so much so that via the journalist Philip Boffey, critique co-author Irving Maltzman assured readers of the New York Times that “Beyond any reasonable doubt, it’s fraud.” Another co-author told reporters their findings cast “grave doubt on the scientific integrity of the original research”. On the spectrum of academic criticism, the gravest of accusations was being levelled at the Sobells, one with its origins in serendipitous links between key players panel below.

Genesis of a campaign

Published in 2008, in his book Alcoholism: its Treatments and Mistreatments Dr Irving Maltzman continued the campaign which had begun over three decades before with his doubts over the Sobells’ findings. For lovers of serendipity, the book included a fascinating account of the coincidences and connections which led to the campaign that occupied his later life.

At the time of the Sobells’ study Dr Maltzman was chair of the psychology department at the University of California in Los Angeles. His research focused on human psychophysiology, far from the treatment of alcohol dependence, but his position required him to assess a research proposal on alcohol treatment from a young assistant professor in the department. The idea was that psychology students would escort ex-military “alcoholics” being treated at a nearby hospital to a local bar and train them to control their drinking, using the methods described by Mark and Linda Sobell in a 1972 monograph which presented initial findings from their study.

Irving Maltzman: for over three decades campaigned against the Sobells’ study

Irving Maltzman: for over three decades campaigned against the Sobells’ study.

With no axe to grind over controlled drinking, understandably his concern was rather over throwing students in the deep end of the flood-tempting conjunction between ex-military ‘alcoholics’ and bars: “I did not think that young undergraduate students could effectively assume such a responsibility.” But a former student of his did have an axe to grind. Mary Pendery had experienced a drinking problem and sought help at the abstinence-oriented mutual aid meetings of Alcoholics Anonymous, but insisted she was no “ideological teetotaler”.

Dr Pendery was then a clinical psychologist heading the alcoholism treatment unit at the San Diego hospital for ex-military personnel, and Maltzman phoned for her opinion on the Sobells’ controlled-drinking approach. He recalls her averring she “had never seen an alcoholic learn to drink in a controlled fashion” and suggesting they visit Patton hospital to talk to the Sobells. Coincidentally, another ex-student of Maltzman’s, Jack Fox, was chief clinical psychologist at the hospital. Though the Sobells had moved on, he still suggested Maltzman and Pendery come to see the treatment facilities (including the bar) used in the study. On that visit “Jack and other members of the clinical psychology staff, who also happened to be former PhD students of mine, urged us to follow up on the patients … Jack stated that he saw many of the patients returning to the hospital for treatment following relapse.” Thus started a campaign based on the undoubted fact that many patients subject to the Sobells’ radical new programme did relapse, but which consistently ignored the fact that more did so who had been left to the hospitals’s conventional abstinence-oriented treatment.

Mary Pendery: contested the reality of post-alcoholism controlled drinking

Mary Pendery: contested the reality of post-alcoholism controlled drinking.

Irving Maltzman’s methodology-focused campaign against what he saw as fraud continued at least up to a few years before he passed away in 2015. Mary Pendery’s passionate objections to controlled drinking as a treatment aim may well have been sustained until her murder in 1994 by George Sie Rega, whom she came to know while he was being treated at the ‘abstinence-only’ alcoholism unit in San Diego where she worked (1 2). It seems they became lovers before Dr Pendery left for Wyoming; from there, later she recontacted him and he joined her. Reportedly in a “in a drunken rage” and “deep in alcoholic relapse”, he shot her before turning the gun on himself – a tragedy which has been laid at the door of “America’s strange love-and-death dance with addiction as a disease – supposedly set off irreversibly, irretrievably, irremediably by any consumption of a substance to which one has ever been addicted, and only remedied by perpetual abstinence”.

Five to nine years after the patients had been treated at Patton hospital, Maltzman together with Jolyon West and lead author Mary Pendery ( panel above) managed to contact and interview all but one of the 19 survivors from among the key 20 in the Sobell study – the ones judged appropriate for and trained in controlled drinking. One task for the interviews was to identify records confirming how the patients had fared, reducing reliance on memory and honesty. Though information was given on the patients up to “the end of 1981”, the report focused on contrasting the picture it pieced together of their drinking in the first and third years after discharge with the picture given for the same periods by the Sobells and by Glen Caddy.

Over those three years the critics identified just one of the 20 patients as having “succeeded at controlled drinking”. What was meant by this was unclear, and the report eludes a direct check against the articles it purported to contradict because it is primarily a narrative of adverse events for each individual (such as heavy drinking, hospitalisations and illness), without these being classified as the Sobells and Caddy had into proportions of days abstinent, moderately drinking, drunk, or institutionalised due to drinking. That adverse events happened was not at issue, especially in the six months after discharge, when even according to the Sobells on a third of days patients allocated to the controlled-drinking programme drank heavily or were in hospital or jail due to drinking. In comparison to the Sobells’ methodology, highlighting the ‘bad days’ and inviting readers to judge the patients’ progress on these was bound to present the more unpalatable (but also more unrepresentative) picture the critics were looking for.

Bullets lack a benchmark

The fundamental weakness of Pendery’s refutation study was that while it did partially document the progress of the key patients, it did not benchmark this against the other groups. In their reply, Mark and Linda Sobell actually admitted to many more hospitalisations for these patients than their critics had uncovered – but the ‘bad days’ spent in heavy drinking or institutionalised due to drinking were far fewer than among comparable patients allocated to conventional abstinence-oriented treatment. All the adverse events Pendery and colleagues reported could be accepted without affecting the Sobells’ core contention: that though the controlled-drinking patients (all with a poor prognosis) were not uniformly successful in overcoming their dependence, they did far better than their conventionally treated comparators.

The fundamental weakness was that it did not benchmark the key patients against their controls

Reasons given by Mary Pendery and colleagues for not reporting on control patients (even though they had also been followed up) would have needed to have been taken into account in adjusting and interpreting a comparison with the controlled-drinking patients, but seem well short of a justification for not making the comparison in the first place. Perhaps for the critics the clinching consideration was that “we are addressing the question of whether controlled drinking is itself a desirable treatment goal, not the question of whether the patients directed toward that goal fared better or worse than a control group that all agree fared badly” – reasoning which neglected the reality that no treatment or treatment goal is desirable in isolation, but only relative to the alternatives open to the patients. This same fundamental point was made by the most thorough of the investigations into the Sobells’ work. The Dickens inquiry (of which more below) argued that “drawing inferences from [these data] with respect to treatment effectiveness demands a comparison. Science, the activity, would have demanded such a comparison even though Science, the magazine, did not.”

That comment was made in respect of the most important of the available outcomes – patient deaths; it turned out to be one where lack of a comparison was clearly critical. Among the evidence cited by Pendery and colleagues were four “alcohol-related deaths” up to the end of their follow-up period in 1981 among patients trained in controlled drinking. But further investigation by the Sobells showed that the death record was actually worse among patients assigned to traditional abstinence-based treatment. The four allocated to training in controlled drinking had died on average about nine years after participating in the Sobells’ study, three for clearly alcohol-related reasons. But six comparison patients had died, on average about seven-and-a-half years post-participation, four clearly due to their drinking. If anything, training in controlled drinking had prevented early alcohol-related deaths.

The comparison did not stop CBS’s 60 Minutes TV programme broadcasting in March 1983 a tour of the graves of the four men, who according to the SAGE Encyclopedia of Alcohol were represented as having been killed by an irresponsible experiment: “60 Minutes failed to mention the six men who had died after undergoing the standard abstinence-based treatment or the poor outcomes that abstinence-based treatment programs have in general.”

One volley hits home

Where the Sobells had to bow to criticism was in respect of their assertion that “follow-up interviews were regularly conducted every 3–4 weeks” and that (free source at time of writing) “Subjects … were contacted for follow-up every 3–4 weeks for a period of 2 yr”. In fact, the schedule was not maintained. It was not a trivial point: at each contact patients were asked for a detailed drinking record since their last contact, enabling the calculation of the proportion of days of controlled drinking, abstinence, or heavy drinking on which the outcomes of the study largely rested; the shorter the recall period, presumably the more accurate the data.

One of the investigations into the study found (1 2) that over the Sobells’ two-year follow-up contacts averaged about 15. It was far more than the “four or fewer times” Mary Pendery said “most” patients had been contacted, but well below the scheduled 24. Critics complained that the Sobells must have known they had not kept to schedule, yet had not disclosed this in their publications – an omission which formed the prime basis for alleging fraud in the form of fabrication of missing data, one majored on in an article, book and news report.

The mystery remains why, if the intention was fraud, the Sobells revealed in their second-year follow-up article (free source at time of writing) that “Five of the 69 subjects found were extremely difficult to locate for follow-up. Final data for these subjects was completed long after their designated follow-up intervals had expired” – a public admission that the follow-up schedule was not always achieved.

No fraud, say investigators

Though the criticism was valid, investigations of the follow-up frequency issue uncovered (free source at the time of writing) error and carelessness, not fraud, and none said shortcomings invalidated the findings: “There was no convincing evidence that your admittedly erroneous estimates of the frequency of follow-up contacts significantly affected the conclusions drawn from this research.” It might be added that 15 follow-up contacts over two years would still make the study’s tracking of post-treatment drinking one of the most detailed ever conducted.

The judgement quoted in the previous paragraph came from the ethics committee of the American Psychological Association, which mounted one of the four inquiries into the Sobells’ scientific integrity. Similar territory was adjudicated on in a court case initiated by former study participants. The inquiries’ methods and findings have been endorsed by the Sobells (1, free source at the time of writing; 2), attacked by critics (1 2), and usefully reviewed by the independent voice of Ron Roizen.

First and most thorough it seems was an investigation by the Sobells’ then employers, the Addiction Research Foundation in Canada. Known after its chair as the Dickens Committee inquiry, it was conducted by four eminent academics, one of whom was at the time Canada’s minister of state for science and technology. Their report was delivered in 1982.

The following year the Sobells received the verdict of an investigator sent by the science committee of the US Congress to check for evidence of fraud. Hard on its heels, in 1984 came the report of an investigation overseen by a five-member panel consisting of senior staff from the US government’s drug and alcohol, health, and health research departments, known after the panel’s head as the Trachtenberg report. The final inquiry was mounted by the ethics committee of the American Psychological Association, which reported in 1984. Its instigation has variously been described as a complaint by the Sobells that their critics’ accusations had damaged their reputations, a similar complaint in which the Sobells had been joined by the lead author of the third-year follow-up report, and a complaint in the opposite direction (free source at the time of writing) from Dr Irving Maltzman about the Sobells’ work.

Last to come to a conclusion, and one of several legal proceedings, was a lawsuit against the state of California mounted by some of the participants in the Sobells’ study along with their relatives, alleging that the study had led to arrests, pain, public humiliation and four deaths, and that data had been “negligently or intentionally misrepresented and falsified”. Launched in 1983, it was dismissed (free source at the time of writing) by a judge in 1987.

Not a verdict on scientific validity

All the investigations focused on allegations of scientific fraud, most seriously in the form of intentional fabrication of evidence and a cover-up of this fabrication. None came near endorsing the charges. Errors, ambiguities, incomplete descriptions of methodology – these there were, but such shortcomings are common in substance use evaluation research, and not usually seen as indicative of fraud. As the Sobells saw it, in this case the heat had been turned up from routine and probably inconsequential shortcomings to far-from-routine fraud, because the notion that “chronic dependent alcoholics could successfully control their drinking was a terribly threatening idea at the time to alcoholism treatment personnel,” a threat which led critics spearheaded by Drs Maltzman and Pendery “to unleash this attack on this study as a way of disabusing anyone from believing that such ‘controlled drinking’ is possible”.

However, to conclude there was no fraud is not the same as declaring that methodologically, all was well. Regardless of the motivations of the accusers and the innocence of the accused, the possibility would remain that inadequacies in the research rendered the findings unreliable, and therefore also the interpretations and conclusions based on these findings. None of the investigations were geared to examining the methodology of the study as such, only in so far as it might pertain to allegations of intentional wrongdoing. Ironically, on this count the follow-up on which the main critics rested much of their case helped by largely validating the Sobells’ records and published findings.

When closely compared, findings presented by the critics as “in marked contrast to the favorable controlled drinking outcomes reported by the Sobells and Caddy et alwere found to offer more confirmation than refutation, lending the weight of an independent follow-up to the original study. Another close comparison was conducted by the Dickens committee, which found that on the important issue of re-hospitalisation due to alcohol-related causes, “for each person … the Pendery et al. and the Sobells’ data coincided”. Rather than the data itself diverging wildly, the divergence was largely in how it was interpreted, itself dependent on what the interpreter wished to portray, and whether the contrast was made with the poorer performance of the era’s conventional treatment, in particular as applied to the controls in the Sobells’ study.

In his essay on the controversy, Ron Roizen explained how the same data can signify the opposite to observers on different sides of the abstinence-only divide. Evidence that many potential patients reject abstinence may be seen as reinforcing the need to suppress alternative goals to drive patients down the abstinence channel – or the opposing need make those alternatives more widely available to attract more problem drinkers into treatment. The abstinence camp would see only a complete break from drinking as a success and incremental gains as failures presaging relapse, while those same gains could look like success to commentators prepared to laud moderation-based recovery. For Roizen, “So distant and mutually unintelligible were the orientations of these two approaches that they could raise the specters of bad faith and fraud between the camps.”

The Sobells themselves were not above some re-interpretation of results consistent with a desire to present their controlled-drinking participants in the best light. In both their first- and second-year (free source at time of writing) follow-up reports, days spent in hospital due to drinking were presented as a negative outcome which meant the patient was not “functioning well” and had been admitted for “alcohol-related health problems (usually for detoxification)”, implying serious and prolonged relapse. Faced with their critics’ stress on hospitalisations as a signal of controlled drinking’s failure, and first-year data showing these events were more common among patients trained in controlled drinking, the Sobells re-interpreted hospital admission as a positive strategy for preventing serious relapse, in line with their programme’s “encouragement to intervene early when drinking got out of hand”. Their reasoning relied on the short duration of most admissions, but the same reasoning was not applied to comparison patients, and it was unclear whether short stays actually were therapeutic or (for example) devoted to patching up alcohol-related injuries. This ambiguity should not, however, obscure the fact that the comparison patients ‘compensated’ for their lower first-year hospitalisation tally with more days of alcohol-related imprisonment, surely an unambiguously bad outcome.

 6 Evidence accumulates and expert opinion converges

Where have these milestones in controlled-drinking research led us? Though controversy is far from extinguished, in 2005 a commentator felt “the professional debate [about controlled drinking] in the field does not … elicit as much passion as it once did”. Among the reasons was “increased consensus that abstinence remains the preferred, safer outcome for individuals with alcohol dependency,” allied with a realistic acceptance of “a harm-reduction strategy emphasizing the outcome of reduced harm and improved psychosocial function as an alternative to a sole focus on abstinence”.

The Sobells’ later writings exemplified these trends. In 1995 their editorial for the Addiction journal revisited the debate sparked over two decades earlier by their findings. Eight responses in the same edition signified its continuing vitality. The Sobells accepted that “Recoveries of individuals who have been severely dependent on alcohol predominantly involve abstinence,” but queried whether this is because the need for abstinence is inherent to recovery from severe dependence. Instead they argued that adverse life circumstances such as poor social support and employment prospects tend to accompany more severe dependence, and these may be what obstruct reduction-based recovery. Beyond this minority for whom abstinence is best suited, they rehearsed the long-standing contention that reducing alcohol-related harm across an entire population demands acceptance of use-reduction goals because many prospective patients (especially those less or non-dependent) who want to tackle their drinking simply will not accept interventions predicated on abstinence.

In 2011 the Sobells returned to controlled drinking in another Addiction editorial. They noted that in the interim evidence had accumulated (see this example) that across the full spectrum of alcohol use disorders, including people who have never been in treatment, “low-risk drinking outcomes occur and are common”. Due they felt to staffing by formerly dependent drinkers steeped in the philosophy of Alcoholics Anonymous, US treatment services had largely ignored this evidence, deterring patients who might have sought treatment if controlled drinking had been on the table.

In support of its arguments the first editorial had cited a 1984 report (free source at the time of writing) on a Canadian trial which had randomly allocated problem drinkers to treatment expressly aiming either for abstinence or for moderation. As part of their treatment moderation-oriented patients were offered training in controlled drinking, the sole difference between the regimens. Most patients seemed to be drinking heavily enough to meet criteria for dependence but as yet were not (severely affected by their drinking.

Told their allocation during the first session of counselling, 23 of the 35 allocated to an abstinence goal either found it unacceptable or expressed reservations, but were not allowed by the study to switch to moderation. In contrast, just five of the 35 allocated to moderation rejected that endeavour; on ethical grounds, they were allowed to switch to an abstinence goal. During and at the end of treatment, goal allocation had generally not significantly affected drinking reductions. In the six months after treatment had ended, whatever goal had been impressed on them, most patients in the end had chosen to drink moderately, usually without reporting serious consequences; just 7% allocated to abstinence had actually achieved this goal.

Reviews explore who does best with what goal

So far the cited research has pitted abstinence as a treatment goal against moderation/controlled drinking, and shown that even for severely dependent patients, neither universally sweeps up the outcome prizes. Several reviews have confirmed that conclusion, but also helped answer the more nuanced and practice-relevant questions of what types of patients do best with either goal, and under what circumstances. Answers are important because without strong evidence about whether the kind of patient facing them can sustain controlled drinking, treatment staff may be unwilling to advocate non-abstinence goals, preferring the less professionally risky ground of an abstinence-based approach, but also running the countervailing risk of drinkers who could benefit from help being deterred from seeking or accepting it. The closer we get to answers about who does best with what goal in what circumstances, the less reason there will be to universally insist on abstinence, helping to douse the passion in the debate and open up treatment to a more diverse caseload.

The most comprehensive and recent of the reviews was completed in 2020. Its overall verdict was honours even: results from the studies did not unequivocally favour either abstinence-based or reduced-drinking goals/approaches. Neither was a controlled-drinking goal more effective among women than men. More surprisingly, the evidence also failed to support accepted wisdom that abstinence would become increasingly suitable as drinking disorder diagnoses became more severe. However, other kinds of analyses might reverse this conclusion, and other dimensions of severity might prove more relevant (1 2). The reviewers did find evidence that outcomes from controlled-drinking goals benefited from treatment geared to those goals.

Earlier reviews concurred on overall results, but sometimes differed on who does best with which goal (1 2 3). Moderation goals have been judged most appropriate for patients whose dependence is less embedded or severe, who face more severe medical and psychological risks from continuing to drink, believe these goals are feasible generally or for them in particular and are adamant about their choice, are employed, psychologically and socially stable, younger, and female. However, neither alone nor in combination are such indicators sufficiently closely associated with successful controlled drinking to be able to securely identify whether a particular individual should be channelled down this route. The findings of these reviews are expanded on in the supplementary text; click to unfold Unfold supplementary text.

Additional to reviews (and sometimes unable to be included in them or not within their remit), results from three of the largest and most sophisticated alcohol trials ever seen have recently been explored for their relevance to the controlled-drinking issue – especially the feasibility of controlled drinking as an outcome after treatment, findings with implications for the advisability of choosing this goal at the start. Findings from UKATT (along with other UK studies) and from the US Project MATCH and COMBINE trials are explored below. Among their many thousands of patients, collectively they reinforce the feasibility of remission in the form of moderation or controlled drinking.

UK studies offer little reason to insist on abstinence

In the seminal Sobells’ study and the Canadian trial described above, patients had been allocated by the researchers to aim for abstinence versus non-abstinence. Outside a research context too, sometimes goals are imposed by the treatment service or by the clinician, but in the UK the more usual situation is that together with their clinicians, patients choose their goals – the route whose consequences were documented in the early accounts of D.L. Davies and the Rand corporation. Since the first of these appeared in the early 1960s, what have we learnt about how patients fare if they themselves opt for moderation versus abstinence as an initial treatment goal?

For the UK the most nationally representative answer emerged from the UKATT trial of psychosocial therapy for 742 patients seeking treatment for alcohol problems at specialist services in England and Wales. Implemented in the late ’90s, it remains Britain’s largest alcohol treatment trial. Its main and rather disappointing findings were that eight sessions of an intensive and comprehensive therapy based on enlisting the patient’s social network were not shown to have been more effective overall than just three sessions of a more basic motivational approach. Dashed too were expectations that that each of the two approaches would differentially benefit certain types of patients (1 2).

But along the way UKATT did illuminate the controlled-drinking issue by way of secondary analyses of which patients preferred which goal and how they fared after expressing this initial preference. An initial report documented differences at study entry between those who (according to the judgement of treatment staff screening patients for the study) were probably aiming versus not aiming for abstinence. The caseload was fairly evenly split, 54% aiming for abstinence, 46% not. Before treatment, abstinence-aiming clients were generally drinking more intensely, experiencing greater drink-related and other problems, and were more socially isolated. Based on what was known about them before treatment, once overlapping influences were taken into account, around 70% of the sample could correctly be classified as aiming versus not aiming for abstinence, with abstinence the more likely choice among those who: were women; drinking more heavily but on fewer days; had detoxified immediately before entering the study (a signal of severe physical dependence); had a social network less encouraging or accepting of drinking, meaning that abstinence would be a less socially isolating choice than among the kinds of patients who preferred to continue to drink; lacked social support in general; and had recently experienced relatively severe alcohol-related problems. Strongest of these predictors were sex, drinking pattern, pre-treatment detoxification, and lack of social support for drinking. While these were the predictors that survived the elimination of overlapping variables, compared to the remainder, clients opting for abstinence were also relatively motivated and confident in their attempts to resist drinking, more likely to be out of employment, had more positive expectations about the effects of drinking, and suffered greater mental and physical health problems. As opposed to alcohol-related problems, a measure specifically of dependence bore no significant relationship to choice of drinking goal.

Outcomes at 3 and 12 months for patients who started treatment aiming for abstinence or for other goals

Having made their initial choice, how well did these patients overcome their drinking problems? A further analysis revealed that regardless of this choice, by 12 months after the UKATT therapies had ended patients were doing about equally well in terms of reducing drinking and its unwelcome consequences. Even among those who at first wanted to stop drinking altogether, more later substantially ameliorated their drink-related problems while continuing to drink than did so by abstaining chart.

Before UKATT there seem to have been four relevant British studies, all conducted at NHS inpatient alcohol treatment units, and all but one in Liverpool, including two at the same unit. Like UKATT, the four studies found that goal choice was meaningful, in the sense that successful outcomes generally took a corresponding form. The three Liverpool studies also agreed that overall success rates in eliminating risky drinking were similar whether or not abstinence was chosen. The remaining study found that opting for abstinence was more likely to be followed by non-problem drinking, but did not report whether lesser degrees of improvement were also more likely than among the remaining patients.

Large-scale US studies validate non-abstinent recovery

UKATT was inspired by another very large alcohol treatment trial, the US Project MATCH trial. Its primary aim was to compare the effectiveness of 12 weeks of different forms of psychosocial therapy for different types of patients, but led by Dr Katie Witkiewitz of the University of New Mexico, analysts plumbed its extensive databank for evidence on the controlled drinking issue (1 2 3). Her team focused on patients who had not just emerged from intensive inpatient or day programmes, known as the ‘outpatient’ arm of the trial. For these patients, MATCH’s therapies were effectively standalone treatments. The question asked was not so much whether low-risk drinking (interpreted in some analyses as allowing occasional heavy drinking) was a viable treatment goal, but whether it had proved a viable treatment destination, prefigured by a similar drinking pattern during treatment.

Abstinence should be demoted from its status as the gold standard of recovery

The same research stable reached similar conclusions after conducting a the same kind of analysis of data from COMBINE, another large-scale US alcohol treatment trial. The implications were that while in the US context, abstinence (or near abstinence) is the most common basis for broad-based recovery from alcohol dependence, nevertheless a substantial minority who continue to drink – and sometimes even to drink in excess of national guidelines – do about as well in terms of their psychological health, social functioning, and quality of life. Conversely, some who are virtually abstinent do poorly on the same measures. It led the analysts to argue that abstinence should be demoted from its status as the gold standard of recovery or at least an essential ingredient, to be seen instead as one of several bases for recovery, and sometimes a basis for non-recovery. For more on these studies unfold Unfold supplementary text the supplementary text.

Pooled analysis from US and UK trials probes who does best with which objective

In 2017 Dr Witkiewitz and colleagues published two analyses based on pooled data from the three large-scale trials described individually above under the headings (click to highlight) “UK studies offer little reason to insist on abstinence” and “Large-scale US studies validate non-abstinent recovery”. UKATT, Project MATCH, and COMBINE offered unrivalled detail on a pooled sample of 3,851 patients treated for drinking problems in the UK and the USA at 27 clinical centres, with sufficient commonalities across the studies to make pooling feasible.

The first analysis (free source at time of writing) found the sample divided most neatly into seven categories based on their patterns of drinking during the (first) 12 weeks of treatment. Some 41% of patients were best classified as virtual abstainers throughout treatment and 17% as low-risk drinkers. The remaining five classes to some degree featured heavy drinking.

Based on their drinking and wider welfare nine months after the trials’ treatments had ended, the analysts said “providers could inform patients that one heavy drinking episode itself is not predictive of long-term failure and that returning to abstinence or low-risk drinking following heavy drinking is predictive of better long-term outcomes”. However, they could not say which types of patients might best be encouraged to opt for abstinence versus low-risk drinking as a treatment objective. That question was addressed by another analysis (alternative source at time of writing) of the same pooled sample, which linked characteristics of the patients at the start of the trials to which of the seven drinking patterns they were most closely aligned to during treatment. The implications of the findings were that low-risk drinking was best managed by patients who leading up to treatment were relatively less severely dependent and/or drinking less, whose social networks were less packed with heavy drinkers, and who suffered less from negative moods and feelings akin to depression. Surprisingly out of the frame were the patient’s drinking goals. For more on this key study unfold Unfold supplementary text the supplementary text.

Don’t throw out the abstinence baby

Concern to rebalance the dominant emphasis on abstinence with recognition of the viability of alternatives should not blind us to the fact that selecting abstinence as a goal is often prognostic of the best treatment outcomes. In the studies included in the previous section, total or near abstinence was the most common basis for broad-based recovery, perhaps partly reflecting US treatment culture. This does not necessarily mean that pressuring the unwilling to accept an abstinence objective will improve their outcomes. Especially in societies such as the USA where abstinence is seen as the gold standard, selecting this objective may not be an active ingredient in itself, but simply a sign of commitment to overcoming your problems with drink which in other cultures or circumstances might take the form of a commitment to controlling one’s drinking. That said, abstinence clearly remains a viable and often (for patient and service) preferable ambition, sometimes in studies favoured by most patients.

Already the reviews cited above have shown the continuing salience of abstinence. This section describes a few sample studies to offer a taste of the kind of research incorporated in the reviews, showing that abstinence remains a goal associated with desired drinking outcomes, sometimes more closely than other goals. An important caution is that ‘association’ does not necessarily mean ‘causation’. When (as was usually the case) patients perhaps together with services chose their goals, other factors associated with this choice which the studies could not take into account may have been the active ingredients in differential outcomes, not the actual choice itself.

Among the studies the reviews uncovered was an analysis of relevant findings from the large US COMBINE alcohol treatment trial of medical care allied with pharmacotherapy. In a previous section (click to highlight) we looked at the prevalence and stability of non-abstinent recoveries three years after patients had been allocated to their treatments. Published in 2013, a second analysis examined the fate of patients with different drinking goals as they entered COMBINE’s treatments.

At the start 25% of participants were aiming for controlled drinking and 37% complete abstinence; most of the rest were also broadly aiming for abstinence, but perhaps not for ever, completely, or in every circumstance. The more dependent a patient was at the start of treatment, the more likely they were to opt for an abstinence goal, but this and other such differences were statistically ‘evened out’ in subsequent analyses to profile the relation of goal choice to drinking during treatment. During the 16 weeks of treatment patients aiming for total abstinence did actually spend more days not drinking, but when they did drink, drank more heavily. Non-problem drinking/abstinence was more likely among patients aiming for total abstinence than those aiming for moderation, but especially when medical care had been supplemented by psychosocial therapy, the gap was minor – – about 76% v. 72%. Though overall around two-thirds of patients aiming for moderation achieved remission, somewhat more did so when abstinence was firmly on the patient’s agenda. However, these results emerged after abstinence-oriented treatments; the gap favouring an abstinence goal would probably have been narrower and perhaps non-existent had non-abstinence aiming patients been offered a treatment adapted to their objectives.

Other recent studies to find abstinence the preferable aim include one from 2018 on alcohol treatment in Switzerland. Patients who as far was known persisted throughout with a controlled-drinking as opposed to an abstinence objective were more likely to have scored as less severely problematic drinkers at entry to treatment. After these initial indicators of problem severity and other factors had been taken into account, both in terms of drinking reductions and attainment of non-hazardous drinking or abstinence, 12 months after leaving treatment drinking outcomes were significantly better among patients who throughout treatment had focused on attaining abstinence as opposed to controlled drinking. However, three-quarters of those aiming for controlled drinking had actually reduced their drinking – possibly, the authors admitted, to degrees which achieved their own ambitions, even if not those set by the study or by national guidelines. Nearly 4 in 10 patients changed their drinking goal between the start and end of treatment, mainly from abstinence to controlled drinking (31%) rather than the reverse (12%).

Another recent study to find abstinence-aiming patients do better was a US trial of acamprosate prescribing for alcohol dependence in family doctor practices. In 2016 it reported that over the 12 weeks of treatment, patients who endorsed an abstinence goal on average reduced their heavy-drinking days by far more than those who did not. This study also exemplified another frequent finding: that reduced but continued drinking was a very common goal. Had non-commitment to abstinence barred them from treatment, most of the study’s participants would have lost the chance to engage in programmes during which on average they roughly halved their heavy-drinking days.

In 2020 abstinence emerged – but not convincingly – as probably the objective most closely associated with low-risk drinking in a US study of patients treated for conjoint substance use disorders and post-traumatic stress by the nation’s service for ex-military personnel. They were asked at the start of treatment whether they preferred not to use their primary substance (which for almost all was alcohol) at all or to reduce use. The main analysis took into account intensity of substance use, age, gender, alcohol dependence severity, and whether drinkers were also dependent on other drugs, and used all the available data to estimate missing outcomes at the end of treatment. It revealed that the 20 patients who chose abstinence were four times more likely than the 19 who chose use-reduction to be abstinent or drinking within the USA’s low-risk drinking guidelines. However, this difference was far from statistically significant, so could have been due to chance sampling variation. Statistically significant differences emerged only when other factors were not taken into account, meaning that what seemed like a big advantage due to choosing abstinence could instead have been due to this choice being associated with what really were the influential factors, such as intensity of use or dependence before treatment. Unplanned analyses of this kind are also considered unreliable because they can capitalise on different ways of analysing the data until one comes up with the desired results. In this case too, the analysis abandoned the recommended method for estimating missing data in favour of one which did not use all the information to hand. All this means that the most secure conclusion is that there was no reliable evidence that aiming for abstinence produced the safest drinking patterns, but that was the way the data was tending given the limited set of influences the study was able to adjust for. Again, this study revealed how many – nearly half – the patients might have refused treatment if an attempt had been made to insist on abstinence. In that light the researchers felt their results were not sufficiently in favour of an abstinence goal to limit or delay treatment of patients adamantly opposed to that goal, provided that low-risk substance-use goals are explicitly targeted and there are no medical contraindications to continued drinking.

A broader perspective

Treatment trials tend to miss or exclude deeply marginalised populations, and in societies where only a minority of people experiencing drinking problems enter treatment, their participants are unrepresentative of the totality of recovery. Dipping into research on the marginalised niche and on the broader population offers some context for what has so far been an account based on treatment and treatment-trial populations.

Generally controlled drinking is itself controlled by the patient, but in some programmes for marginalised populations the service does the controlling by administering set doses of alcoholic beverages at set times. Sometimes offered to inpatients as a way of ameliorating withdrawal, of greater interest in the current context are programmes which administer alcohol to people in the community as a therapeutic or harm reduction strategy.

A 2018 review of both types of programmes found that those based in the community integrate social support with patient care, including referrals or access to counselling and addiction treatment, primary health care, meals, shelter-style supportive or permanent harm-reduction housing, and entertainment or group activities. The services are geared to the needs of their typically indigent participants, who tend to be chronically homeless or lacking stable housing, with severe alcohol use disorders, and unable or unwilling to participate in abstinence-based housing or treatment – the kind of people who would often have been excluded from alcohol treatment trials. Positive outcomes included fewer encounters with the police, improved personal hygiene, and uptake of other health and social support. Even among these populations, a more conventional harm-reduction approach not involving the supply of alcohol – but also not predicated on abstinence – can generate benefits, including reduced drinking, drunkenness and alcohol-related harm.

Focusing on treatment populations risks giving a distorted impression of the bigger picture of how people normally emerge from problem drinking, especially in countries where abstinence is the favoured treatment goal. Published in 2000 (free source at the time of writing), Mark and Linda Sobell teamed up with a colleague to review studies which shed light on “natural recovery” from substance use problems, meaning people whose recoveries were not attributable to treatment or self-help groups.

Three-quarters of the studies they found had been conducted in North America. In the current context their key finding was that “across all studies two-fifths (40.3%) of alcohol recoveries involved low-risk drinking, suggesting that such drinking is a common route to recovery among naturally recovered alcohol abusers”. Note however that it was still a lower figure than the 60% of recoveries based on abstinence. The review covered substance use problems in general, but since three-quarters of the studies were of problem drinking, the overall finding that typical and average recovery durations were over six years suggests that in most cases recovery from problem drinking was stable and enduring, aided (according to the problem users themslves) most commonly by social support or transition to new (and presumably more recovery-friendly) social circles.

 7 Is shared decision-making the way forward?

The history of the controlled drinking controversy makes us face not just what decision is made about treatment goals, but also who makes that decision and how – where the power lies. If controlled drinking is to emerge from the cupboard to lie on the treatment-planning table alongside abstinence, how is the decision to be made about which to go for, and by whom?

Shared decision-making between patient and clinician has been recommended, but in turn raises the question about how to engineer this in a meaningful rather than tokenistic manner. Helpfully a Dutch study showed that shared decision-making can be systematised rather than left to the uncertain initiative of the clinician. In relation to life in general, one result was that patients felt more able to make their own decisions, more in control, and less submissive – possibly portending a more stable shift away from a dependent mind-set than could be achieved by less explicit shared decision-making. However, the lack of studies leaves it an open question whether systematic shared decision-making improves drinking outcomes.

In practice, reliance on clinicians to help navigate through “confusing and unpredictable” treatment pathways may undermine the ethos of shared decision-making, tipping the scales of power towards clinicians. The UK study which identified this deficiency called for treatment pathways to be more ‘patient-friendly’ and for health care professionals to support and build ‘self-efficacy’ among patients – their belief in their own power to succeed rather than ceding that power to the clinician.

Complicating shared decision-making, among 20 seemingly severely dependent patients interviewed at alcohol treatment services in London, the same study found that definitions of alcohol dependence varied widely – for some depending on volume or strength of alcohol consumed, for others indicating a need (as opposed to a want) to keep drinking. In turn these variations influenced their views about what and who treatment was for. Many saw ‘cutting down’ as an important step towards achieving abstinence and/or regaining control of drinking, while their practitioners tended to support cutting down only to the extent that it was a step towards abstinence, not a goal in itself. From the patients’ perspectives, moderation could be both a means to an end and the end in itself, contrary to the mainstream clinical view of abstinence and moderation as mutually exclusive goals.

As this study suggested, practitioners in several countries are known to be less inclined to support moderation as a treatment goal among patients with more severe drinking problems, though more would support it as an intermediate step towards abstinence. Surveyed in 1999/2000, two-thirds of the leaders of British substance use services fully endorsed the acceptability of controlled drinking as an intermediate outcome for non-dependent alcohol ‘abusing’ clients and only slightly fewer as a final outcome, but the corresponding figures for physically and psychologically dependent clients were 42% and 29%. However, absolute dogmatism was relatively rare; even when rejection was at its maximum, only 23% saw controlled drinking as a “completely unacceptable” final goal for dependent drinkers, and 60% of services made such treatment available to their dependent clients.

In 2011 similar questions were put to US addiction clinicians. For people whose alcohol problems fell short of dependence, over half saw non-abstinent drinking as an acceptable intermediate (58%) and final (51%) goal, but far fewer did so for clients who were dependent (respectively 28% and 16%). As well as the severity of problem drinking, evaluations of the acceptability of non-abstinence goals were informed by patients’ health problems (86% deemed these ‘very important’), number of previous treatment episodes (70%), presence of mental health disorders (68%), age (67%) and emotional stability (65%). Of those who saw non-abstinence as an unacceptable treatment goal, 4 in 10 did so partly because “It would send the wrong message to clients” – echoing the concern evident since controlled drinking spiked as a controversy that allowing the possibility of moderation would undermine commitment to abstinence as the sole way forward. The researchers concluded that overall “individuals with alcohol and drug problems who avoid treatment because they are ambivalent about abstinence should know that – depending on the severity of their condition, the finality of their non-abstinence goal, and their drug of choice – their interest in moderating their consumption will be acceptable to many clinicians, especially those working in outpatient and independent practice settings”.

When similar questions were also put to experienced addiction therapists in Poland, they too were much more likely to accept reduced drinking as a final treatment goal for the less severe diagnosis of alcohol abuse (77%) than for dependence (36%). Again, a common reason for rejecting this goal (expressed by around half the therapists who did reject it) was that it would send the wrong message to clients.

What are the consequences of a mismatch between the treatment goals of a service and its staff and those of their patients? Possibly, it seems, less successful treatment. Published in 2016, a Swedish study investigated this issue at two non-residential services, one requiring an abstinence goal, the other accepting of low-risk drinking. Their patient intakes roughly reflected what research suggests makes patients suitable for these differing goals, the abstinence-based service admitting relatively more male and older patients who were drinking more intensely and had lacked socioeconomic resources, signalled by fewer years of schooling. Regardless of the service they attended, at the start of treatment about half the patients were aiming for abstinence (compared to other patients, they were older and had been drinking more heavily and for longer) and a quarter for moderation. When patients’ goals matched that of the service, two-and-a-half years later 94% were abstinent or drinking at low-risk levels compared to just 63% when the goals had clashed. Overall, just over half the patients who wanted to drink at low-risk levels at the start of treatment were doing so when followed up.

 8 Contemporary guidance

While their patients may be ambivalent about preferring abstinence as a treatment goal, UK practitioners who follow official guidance should not be. Alcohol treatment services in the UK are unambiguously advised by the National Institute for Health and Care Excellence (NICE) to guide drinkers at the more severe end of the spectrum of alcohol use disorders towards abstinence and to favour moderation lower down the scale, yet without ever ruling out non-abstinent goals if working with these is required to engage the drinker in treatment:

In the initial assessment in specialist alcohol services of all people who misuse alcohol, agree the goal of treatment with the service user. Abstinence is the appropriate goal for most people with alcohol dependence, and people who misuse alcohol and have significant psychiatric or physical comorbidity (for example, depression or alcohol-related liver disease). When a service user prefers a goal of moderation but there are considerable risks, advise strongly that abstinence is most appropriate, but do not refuse treatment to service users who do not agree to a goal of abstinence.

For harmful drinking or mild dependence, without significant comorbidity, and if there is adequate social support, consider a moderate level of drinking as the goal of treatment unless the service user prefers abstinence or there are other reasons for advising abstinence.

For people with severe alcohol dependence, or those who misuse alcohol and have significant psychiatric or physical comorbidity, but who are unwilling to consider a goal of abstinence or engage in structured treatment, consider a harm reduction programme of care. However, ultimately the service user should be encouraged to aim for a goal of abstinence.

Before NICE had pronounced on treatment goals, in 2006 the Department of Health and what was its National Treatment Agency for Substance Misuse had issued guidance for England which promoted a similar strategy. It stressed that goal choice should not exclude drinkers from support or treatment, but did see abstinence as “the preferred goal for many problem drinkers with moderate to severe levels of alcohol dependence, particularly … whose organs have already been severely damaged through alcohol use, and perhaps for those who have previously attempted to moderate … without success”. Even for these drinkers, it continued, if abstinence is not acceptable, moderation is better than nothing, and may lead to abstinence. On how the decision should be made, in relation to care planning in general the guidance saw patient choice as not only an entitlement, but a strategy which improves the chances that the chosen treatment will succeed because “it has been selected and committed to by the individual”.

If severity of dependence is an influence on the chances of sustaining controlled drinking, the 2013 revision to the US diagnostic framework for mental disorders decisively removed what before seemed a clear dividing line in considerations of who should aim for abstinence. The American Psychiatric Association had previously classified alcohol abuse and alcohol dependence as distinct disorders in its Diagnostic and Statistical Manual of Mental Disorders, providing a line across which to favour abstinence. However, in 2013 it integrated these diagnoses under the single designation “alcohol use disorder”, acknowledging that these drinking problems do not divide in two, but range unbroken across a spectrum of experiences and symptoms from mild to severe. This re-imagining of drinking problems as a continuum throws up new questions about whether there can be a concrete point at which abstinence becomes the preferable or only acceptable treatment goal, even if research supports the salience of the severity dimension.

1981 edition of watershed book on controlled drinking featured a foreword by D.L. Davies, initiator of the controversy the book explored

1981 edition of watershed book on controlled drinking featured a foreword by D.L. Davies, initiator of the controversy the book explored

Notwithstanding this change, still valid are the conclusions of the British textbook Problem Drinking published in 1997, which authoritatively summed up the evidence: Research shows that no matter how physically dependent, moderation is for some feasible, especially when there are sufficient supports in the patient’s life, but the more severe the dependence, the more likely abstinence is to be the suitable strategy. The 2020 review described above cast doubt on severity of dependence as a factor, but not with sufficient weight to warrant deleting this part of the conclusions reached by Professors Nick Heather and Ian Robertson, whose earlier watershed book Controlled Drinking had drawn together the strands as they stood in 1981. Then as now, neither abstinence nor controlled-drinking objectives could be said to have been proved preferable for dependent drinkers. Controlled Drinking’s original 1981 edition featured a foreword by D.L. Davies, reminding us of the origins ( section above) of the controversy the book explored illustration.

This is how Drug and Alcohol Findings summed up the evidence: “Treatment programmes for dependent drinkers should not be predicated on either abstinence or controlled drinking goals but offer both. Nor does the literature offer much support for requiring or imposing goals in the face of the patient’s wishes. In general it seems that (perhaps especially after a little time in treatment) patients themselves gravitate towards what for them are feasible and suitable goals, without services having to risk alienating them by insisting on a currently unfavoured goal.”

Thanks for their comments on an earlier version of this essay to Dr Peter Rice, then Consultant Addiction Psychiatrist at the Tayside Alcohol Problems Service in Scotland. Thanks for comments on this version to: David J. Armor, lead author of the first Rand report, and now at George Mason University in the USA; Ray Hodgson, who commented on the controversy at the time and was at the time of his retirement Research Director at Alcohol Research UK; and to Ron Roizen, who also commented on the controversy at the time and is now an independent scholar known for his contributions to the history and sociology of drinking and the response to drinking in the USA. Commentators bear no responsibility for the text including the interpretations and any remaining errors.

Last revised 26 March 2021. First uploaded 06 March 2011

Comment/query
Give us your feedback on the site (two-minute survey)
Open Effectiveness Bank home page
Add your name to the mailing list to be alerted to new studies and other site updates


Top 10 most closely related documents on this site. For more try a subject or free text search

STUDY 2015 Navigating the alcohol treatment pathway: A qualitative study from the service users’ perspective

REVIEW 2012 BAP updated guidelines: evidence-based guidelines for the pharmacological management of substance abuse, harmful use, addiction and comorbidity: recommendations from BAP

REVIEW 2010 Assessment and management of cannabis use disorders in primary care

STUDY 2011 Randomized controlled trial of a brief intervention for unhealthy alcohol use in hospitalized Taiwanese men

DOCUMENT 2012 Quality standard for drug use disorders

STUDY 1992 Disulfiram treatment of alcoholism

REVIEW 2012 Acamprosate for alcohol dependence: a sex-specific meta-analysis based on individual patient data

STUDY 2015 The impact of a Housing First randomized controlled trial on substance use problems among homeless individuals with mental illness

STUDY 2013 Alcohol assessment and feedback by email for university students: main findings from a randomised controlled trial

REVIEW 2018 A meta-analysis of the association between patients’ early perception of treatment credibility and their posttreatment outcomes