Ernst D.B., Pettinati H.M., Donovan D.M. et al.
Annals of Family Medicine: 2008, 6(5), p. 435–440.
Unable to obtain a copy by clicking title? Try asking the author for a reprint by adapting this prepared e-mail or by writing to Dr Ernst at firstname.lastname@example.org. You could also try this alternative source.
In a programme intended to simulate primary care management of alcohol dependence, what made the difference to patients was how far the clinician maintained confident optimism and responded to the patient rather than strictly adhering to the treatment manual.
The purpose of this study was to examine the relationship between treatment outcomes and patient and clinician factors associated with a medical management intervention for the treatment of alcohol dependence. Intended to approximate a primary care approach, the intervention was developed for the COMBINE study, a randomised controlled trial combining two medications (naltrexone and acamprosate) with medical management, with or without specialist psychosocial alcohol therapy.
Overall results of the trial have previously been reported. This report focused on the medical management element of the treatments, examining links between outcomes during the trial and features of the patients and the clinicians and their behaviours, the latter based partly on a sample of audio-recorded sessions rated by observers. In particular, links were examined between
drinking and related problems
Abstinence from alcohol, amount of heavy drinking, and clinical improvement (ie, at most moderate drinking with no related problems).
during treatment and:
• the patients' attendance Number of medical management visits; total minutes in treatment. ;
• their alliance or therapeutic relationship with the clinician and satisfaction with treatment;
• and the clinician's adherence to and competence in delivering the manualised medical management protocol.
The analysis found that the more medical management visits a patient attended, but the less total time spent in those visits, the better the outcomes in terms of more drink-free days, reductions in heavy drinking, and a higher likelihood of clinical improvement. It seemed that patients who were doing poorly chose to or were able to attend less regularly, but needed more time at each visit. Patients who were more positive about their alliance with their clinician midway through treatment, or at the end expressed greater satisfaction with their treatment, were abstinent on significantly more days.
Two features of how the clinicians behaved were also associated with better patient outcomes. Those who according to observers most adequately conveyed professionalism, expertise, and confidence in the medical management protocol, yet who also deviated somewhat from strict adherence to that protocol, had patients who drank less often and less heavily than other patients. The dimension reflecting professionalism, expertise, and confidence was also related to overall clinical improvement. Observers commented that these clinicians seemed to convey optimism about recovery even in the face of discouraging news from the patient, and were prepared to respond to issues raised by the patient even if that meant departing from the manual. The authors interpreted these findings as indicating that "some flexibility in delivering medical management, based on good clinical judgment and in conjunction with optimism and hope for recovery, supports better outcomes with the intervention". Their broader conclusion was that medically trained clinicians with little specialised training in alcohol dependence treatment were able to deliver a brief and effective medication management intervention designed to be consistent with primary care practice.
commentary It is important to place these findings in the context of the overall trial (itself previously analysed by Drug and Alcohol Findings). For the trial 1383 alcohol dependent patients were randomly allocated to various abstinence-oriented treatments lasting 16 weeks. The most basic was placebo pills plus nine medical management appointments typically lasting under 20 minutes spent assessing, monitoring and feeding back the medical consequences of the patient's drinking, and promoting adherence to pharmacotherapy. Some of these patients were also allocated to active medications (naltrexone and/or acamprosate) and/or to a sophisticated programme of psychosocial therapy. The key question was how far the extra therapies improved on straightforward medical management with placebo pills. In essence, as long as medical management was bolstered by naltrexone, patients did about as well as they did with additional therapy. Even without active pills, during treatment, 58% of patients receiving basic care achieved a good clinical outcome.
Having discovered that fairly straightforward (if well structured and perhaps more intensive than normal) medical care could be associated with outcomes as good as after state-of-the-art therapy, the featured study set out to unravel the underlying processes. This it did by assessing how much of the difference in outcomes between patients could be accounted for by differences between patients, and how much differences between their clinicians. The analysis found that the patients' behaviour and perceptions of treatment were by far the most significant factors.
Despite features Several of the variables attributed to the patient (attendance; relationship with clinician; satisfaction) could have been influenced by the clinician they were assigned to. Conceivably, this would leave little room for the analysis to find any remaining link with the clinician's skills and approach. Making this less likely still was the fact that in this highly controlled trial, differences between clinicians would have largely been ironed out by the selection, training and supervision procedures, and by that fact that all the clinicians followed the same manual. of the study and analysis which would have obscured this, there was also a small but statistically significant link between the clinician and the outcomes. Patients had better drinking outcomes when their clinicians were characterised (not just with their successful patients, but across their entire caseloads) by unwavering, optimism-instilling confidence in the treatment, allied with flexibility in its application. Importantly, flexibility and optimism had to go together, consistent with the common sense understanding that it is not enough to have confidence in the form of blind faith in the treatment protocol, or to depart from it simply because you lack that faith. The former might lead to disregarding important clues from the patient, missing therapeutic opportunities, and to a damaged therapeutic relationship ('I'm not really being listened to'), the latter to the patient too lacking faith and to an incoherent approach which fails to offer a credible, consistently structured route to recovery.
In general psychotherapy these attributes are well recognised 'common factors' found to account for far more of what makes for good outcomes than the particular brand of therapy. In alcohol therapy too, as long as the approach is explicitly structured in a way which makes theoretical sense, it matters little what particular form it takes. Consistent with the featured study, what is important is being responsive enough to the patient to match your approach to their mood, personality and recovery preferences, even if that means departing from best practice distilled in state-of-the-art manuals.
This is not however a licence to abandon structure altogether. Coherence and structure are important ingredients of good therapy, offering a way to make sense of what may seem chaos and confusion so unstructured that there can be no such thing as a clear route, let alone a route out. For example, in one study of cocaine counselling, a moderate degree of fidelity to the treatment manual led to better outcomes than either following it very diligently or being relatively lax.
Though probably active in every therapeutic encounter, such influences might be more visible in some situations and populations than others. In the featured study, though very heavy Averaging 21 UK units most days. drinkers, the patients were more socially integrated and less severely dependent than some UK alcohol treatment caseloads. They could also only enter the study if they had achieved at least four days without drinking. All were prepared to accept referral to the study or responded to ads soliciting participants. In the cocaine study mentioned above, moderate adherence was most clearly beneficial when a patient had formed a relatively poor relationship with their therapist.
Drug and Alcohol Findings has published a series of articles dedicated to exploring of the impact of some of the common factors which might be important in the treatment of substance use problems.
Thanks for their comments on this entry in draft to Denise Ernst of the Center on Alcoholism, Substance Abuse, and Addiction at the University of New Mexico and to Petra Meier of the University of Sheffield. Commentators bear no responsibility for the text including the interpretations and any remaining errors.
Last revised 22 March 2009. First uploaded
REVIEW 2018 Cohesion in group therapy: a meta-analysis
DOCUMENT 2013 Sometimes best to break the rules