Patient-Treatment Matching Fails to Show Promise for Alcoholism Treatment
Kenneth A. Lucas, CSAC, CADAC
In many areas of medicine, matching patients to treatments is widely practiced. For instance, patients with a diagnosis of cancer will almost certainly be matched to surgery, radiation or chemotherapy. Likewise, patients with severe diabetes will be matched to treatments that call for weight loss, exercise, restriction of sugar intake and the possible administration of insulin.
But how about those with the diagnosis of alcoholism? Those in the field of alcoholism treatment understand that alcoholics differ widely. For instance, some have cognitive impairment, some have poor coping skills, while others have high anxiety rates, poor socialization, psychiatric severity and low social support. So wouldnt it make sense to take a tip from more conventional medicine and match an alcoholic to treatments based on their particular characteristics in order to improve treatment outcomes and better utilize scarce resources?
Armed with studies showing that some treatment approaches were more effective than others for patients with certain characteristics, the National Institute on Alcohol Abuse and Alcoholism (NIAAA) initiated Project MATCH in late 1989. Project MATCH (Matching Alcoholism Treatment to Client Heterogeneity) was designed to build on earlier studies but with a large sample size, it was far more ambitious. The goal of Project MATCH was to learn whether different types of alcoholics respond selectively to particular treatments. Specifically, the study tested 16 patient-treatment combinations that appeared promising based on earlier experimental studies.Structure of Project MATCH A total of 1,726 patients were recruited at treatment facilities throughout the United States, making this the largest clinical trial of psychotherapies undertaken to date. A fourth of the patients were women, while 15% were from minority populations. There were two parallel arms representing the two major venues of treatment for alcoholic patients: an "outpatient" arm, with patients recruited directly from the community, and an "aftercare" arm, consisting of patients who had just completed an inpatient or intensive day hospital treatment.
Procedures were the same in both study arms. Participants were first assessed, using interviews and tests, to obtain information on demographic characteristics, personality, drinking behavior, factors predisposing to alcohol problems, the personal and medical effects of their drinking, and alcohol treatment history. Both groups involved identical assessment methods, treatment procedures within and across programs, follow-up evaluations and analytic techniques. The two-group design allowed exploration of possible differences in matching among patients recruited from different settings.
Ten patient characteristics were studied, mostly based on promising leads from earlier, more limited, studies. They were severity of alcohol involvement, cognitive impairment, conceptual level, gender, meaning seeking, motivation, psychiatric severity, social support for drinking versus abstinence, anti-social tendencies and alcoholic typology.Treatments All 1,726 patients were randomly assigned to one of three treatments: Twelve-Step Facilitation (TSF), Cognitive-Behavioral Therapy (CBT), or Motivational Enhancement Therapy (MET). These treatments were selected because they showed potential for matching, promising outcomes and utility in clinical situations.
TSFThis consisted of 12 weekly sessions in which the therapist encouraged patients to attend and become involved in the traditional fellowship activities of Alcoholics Anonymous (AA) and to introduce the first five of 12 steps. Involvement in AA included finding a sponsor, attending meetings regularly and reading AA literature. TSF was an approach designed specifically for Project MATCH. Although grounded in the 12-Step principles, it was a professionally delivered, individual therapy different from the usual peer-organized AA meetings and was not intended to duplicate or substitute for traditional AA.
CBTThis consisted of skills, taught and coached by a therapist, to enable patients to cope with situations and emotional states known to precipitate relapse. Patients, in 12 weekly sessions, practiced drink-refusal skills, learned to manage negative moods, and learned to cope with urges to drink.
METThis consisted of therapists using techniques of motivational psychology and, rather than training the patients in particular skills, encouraged individuals to consider their situation and the effect of alcohol on their life, develop a plan to stop drinking, and implement the plan. MET consisted of four sessions over the course of 12 weeks.
Procedures for administering each of these three treatments were carefully described in detailed manuals developed for each treatment. Carefully trained and supervised professionals delivered the treatments in individual therapy sessions. All therapy sessions were videotaped (with patient permission) and 25% were randomly selected for monitoring by supervisors to ensure that the therapy was conducted as intended.Results The findings challenged the notion that patient-treatment matching is a prerequisite for effective alcoholism treatment. The NIAAA trial confirmed only one of the hypothesized patient-treatment matches. There was a significant match on psychiatric severity with TSF among the "outpatients." Patients with few or no psychological problems had significantly more abstinent days with TSF than with CBT. The investigators concluded that patient-treatment matching, as exemplified by the 16 combinations of characteristics and treatments studied in Project MATCH, adds little to enhance treatment outcomes. In addition to the knowledge gained about matching, the trial also demonstrated that compared to their status before treatment, drinking and negative consequences declined regardless of which of the three treatments participants received. Test participants also decreased use of other drugs, were less depressed and improved their liver function. These improvements were maintained through the 12 months following treatment. Overall, gender or ethnic difference did not affect treatment outcome.
Commenting about the research, NIAAA director Enoch Gordis, MD said that, "The major finding from Project MATCHthat matching patients to treatments added little benefit to treatment resultswas a surprise to clinical investigators and to service providers alike. However, it certainly is not the first time that reasonable hypotheses, when rigorously investigated in a large population, failed to yield an expected result.