THE BEST TREATMENTS
Projec Match one-year outcome results
The world’s most expensive research into the results of alcohol treatment is Project Match. It took eight years and spent about $27million trying to match alcoholic clients to best treatments… We analysed it to see if you can use its results in your own work
by ALEX GEORGAKIS MSc, March 1997
“The question ‘Does psychotherapy work?’ is virtually meaningless… the range of individual differences within standard diagnostic categories remains so diversified as to render meaningless any questions or statements about individuals who become so labelled… the question towards which all outcome research should ultimately be directed is the following: What treatment, by whom, is most effective for this individual with that specific problem, and under which set of circumstances?” (Paul, 1966)
The above quotation, written 30 years ago, has been echoed by many in the field of psychotherapy and addiction treatment. The idea that matching clients to treatment will significantly improve outcomes dates back to Aristotle for medicine and Jellinek for alcoholism. More recently, it has gained popularity in fields such as medicine, education, psychotherapy and addiction. It has intuitive appeal. And it has received promising support from numerous clinical trials.
In the alcoholism field, the “matching hypothesis” suggests that, instead of prescribing the same treatment to all alcoholics, we should assess clients’ individual characteristics and needs and offer a specified treatment which best matches those needs. The underlying rationale is that some treatments are more effective with certain types of clients than others.
In the US, the matching hypothesis was considered promising enough for the National Institute on Alcohol Abuse and Alcoholism to fund the biggest and probably the most scientifically rigorous clinical trial of psychotherapy ever conducted, with the final costs apparently exceeding $27million.
Project Match consisted of two independent but paralleled treatment-matching studies. One part of the study recruited clients from five outpatient settings (n=952), referred to as the outpatient group.
The other group consisted of clients who had completed at least seven days of inpatient or intensive day hospital treatment — not simply detoxification — and were receiving aftercare at one of five sites (n=774). These were referred to as the aftercare group. Altogether, 1,726 patients who met the study criteria were randomly allocated to one of three treatments.
Treatments were chosen using predefined criteria, such as potential for revealing matching effects, prior documentation of clinical effectiveness, and feasibility and practicability of implementation. Five treatments were shortlisted, three of which were chosen: cognitive behavioural coping skills therapy (CBT), motivational enhancement therapy (MET) and twelve-step facilitation (TSF).
CBT aims to teach alcoholics better ways of coping with problems in their lives; the idea is that this will affect their drinking behaviour, which is assumed to be “functionally related”. MET aims to increase motivation to change. Unlike CBT, it assumes that the alcoholic has the skills to recover successfully but lacks motivation.
TSF was selected mainly because of its widespread use, coupled by feasibility of implementation and its distinctiveness from the other two interventions. TSF’s goal is to help clients incorporate the belief system of Alcoholics Anonymous. It is important to note that TSF is not a test of the effectiveness of AA, only of a treatment which helps people to affiliate with AA and understand their condition as a spiritual and medical disease.
All three treatments aimed to help clients to achieve abstinence.
Despite the popular use of group therapy in the treatment of alcoholism, the investigators decided — mainly for practical and methodological reasons — to deliver all three treatments in individual sessions over a 12-week period. CBT and TSF consisted of 12 sessions of individual therapy. MET consisted of four sessions, once again delivered over a 12-week period (weeks 1, 2, 6 and 12).
80 therapists were certified to deliver the three treatments. Every attempt was made to ensure that therapists in each treatment condition were delivering the intended treatment, rather than a mixture of different treatments and techniques. In other words, all treatments were ‘manualised’.
They were delivered by therapists who believed in the efficacy of their particular treatment. All therapy sessions were videotaped. 25% of sessions were viewed by supervisors to ensure that therapists complied with the treatment manuals. And if therapists deviated from the model, extra supervision was given.
WHAT TYPE OF CLIENT DOES BETTER IN WHAT TYPE OF TREATMENT?
Project Match’s primary aim was to identify matching effects between client characteristics and treatment: what type of client does better in what type of treatment? Ten matching characteristics were selected for investigation, based on prior studies and theoretical rationales: severity of alcohol involvement, cognitive impairment, client conceptual level, “meaning seeking”, gender, motivational readiness to change, psychiatric severity, social support for drinking versus abstinence, sociopathy and alcoholism typology (that is, type A = low vulnerability and moderate problem severity, type B = high vulnerability and high problem severity).
Sixteen specific ‘predictions’ were generated and tested. For example, the investigators predicted that TSF treatment and CBT would do better with clients who are more severely dependent, compared with MET, which is less intensive. They predicted that cognitively impaired clients and clients who score high on ‘meaning seeking’ would have better outcomes with TSF compared with CBT and MET. In contrast, it was predicted that TSF would do worse than CBT with women and clients with high scores on psychiatric severity, and worse than MET with clients with high conceptual levels.
All patients who entered the trial were followed up three, six, nine, 12 and 15 months after entry to treatment. Information on their alcohol and drug use was verified by blood and urine tests and collaterals. The location rates were over 90% throughout the follow-up period.
The investigators chose two main drinking outcomes: % of days abstinent and number of drinks per drinking day. The first investigated frequency of drinking, the latter severity of drinking.
Result 1: All clients — aftercare and outpatient groups — showed dramatic reductions in drinking. These were sustained throughout the follow-up period.
Result 2: The aftercare group drank more before treatment than the outpatient group. But they showed greater improvements after treatment. For example, 35% were continuously abstinent versus 19% for the outpatient group.
Result 3: All three treatments did equally well, with no substantial differences in outcomes between them. Any outcome differences between treatments favoured TSF.
Result 4: With one notable exemption, the investigators found no matching effects. In other words, different types of clients had similar outcomes irrespective of the treatment offered.
Result 5: Outpatient clients without psychopathology — a global measure which combines symptoms of anxiety, affective, psychotic and personality disorders — did better if treated with TSF.
Result 6: There were no differences in outcomes between MET (four sessions) and TSF and CBT (12 sessions).
Result 7: In the aftercare group, the following characteristics were predictive of outcome: gender (women did better than men), severity of alcohol involvement (less dependent clients had better outcomes) and support for drinking (higher levels, worse outcomes). Also, towards the end of the follow-up, clients with higher psychiatric severity had fewer days abstinent.
Result 8: In the outpatient group, clients who were less motivated and had higher levels of support for drinking did worse.
The authors intend to publish additional papers which will shed light on a number of important areas and so narrow the scope of interpretations. They also intend to follow the outpatient sample three years after treatment.
Project Match has generated tremendous interest among researchers and clinicians on both sides of the Atlantic. The results were eagerly awaited and speculation was rife. For some, the results will be an anticlimax. Matching clients to treatment was seen as the way forward in improving alcohol treatment outcomes. Clearly, the results of the study challenge such notions. Some might criticise the methodology. Some might argue that it was over-ambitious. We can speculate that different treatments, different client attributes, different delivery systems — such as inpatient versus outpatient — might have produced the expected effects. But it is more fruitful to concentrate on what was found.
MOST FAVOURABLE RESULTS
For 12-Step practitioners, the results were clearly encouraging. This was the first real attempt to look at 12-Step treatment using such a complex and informative design. Any outcome differences found between treatments favoured TSF. And clients with no psychopathology did significantly better in TSF than CBT. And none of the mismatched predictions — for example, that women would do worse in TSF than CBT— were supported.
In the outpatient group, TSF clients had fewer drinking consequences at month nine, were more likely to be totally abstinent at month 15, and were sgnificantly more likely to remain continuously abstinent throughout the follow-up period (24% versus 15% for CBT and 14% for MET). One suspects that the intense concentration on abstinence found in the TSF treatment explains such differences.
Another important result is the relationship between intensity of treatment and outcomes. The finding that four sessions of MET were as effective as 12 of CBT and TSF suggests that MET might be more cost-effective, and a study is under way. But the authors of the report caution against premature conclusions. Clients’ experience of participating in one of the most rigorous and intense trials might have had a large impact on their drinking. For example, MET clients in the outpatient group would have received not only four therapy sessions but also three pre-treatment assessment sessions totalling eight hours, plus five face-to-face interviews, three of which were particularly long. Their families or significant others would also have been interviewed on four occasions.
In the aftercare arm, MET clients would have received all the above plus at least seven days of inpatient or intensive day hospital treatment. Being “forced” to look at your drinking and related behaviours in great detail on numerous occasions, even if it is not in the name of therapy, plus the knowledge that your family would be interviewed, might have had a large therapeutic impact.
Another interrelated finding was the outcome differences between the aftercare group, which had previously received inpatient or intense day hospital treatment, and the outpatient group. The aftercare group had better outcomes, despite their comparative disadvantages — such as being more dependent, less socially stable and having more previous treatments. Again we must proceed with caution.
Clients were not randomly allocated to inpatient versus outpatient treatment. The aftercare group consisted of clients who had completed at least seven days of treatment before receiving Project Match treatments. The potential bias introduced by excluding clients who dropped out might have had an impact on outcomes. The potential added value of an inpatient/day hospital treatment requires more investigation.
The apparent equivalence in outcomes between three very distinct treatments should not surprise those familiar with psychotherapy and addiction research. A popular explanation is that common therapeutic factors exist in very different therapies. It is these common factors which account for client changes.
All three treatments showed very promising outcomes.
Matching Alcoholism Treatments to Client Heterogeneity: Project Match Post-treatment Drinking Outcomes (1997) in Journal of Studies on Alcohol, 58, pages 7-29.
This article was first published in Addiction Today, March 1997.