Click for next page ( 232


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 231
11 PAllt;NT-TREATMENT MATCHING AND OUTCOME IMPROVEMENT IN ALCOHOL REHABILITATION In the decade since the Institute of Medicine last reviewed the state of research on the treatment of alcohol problems (IOM, 1980), there has been a substantial amount of work on patient-treatment matching. Perhaps most important has been the recognition that matching patients and treatments is a sophisticated idea with as yet untapped possibilities for improving the effectiveness and efficiency of treatment. In particular, as a result of the work of Glaser (1980), Skinner (1981), and more recently Finney and Moos (1986), it has become accepted that (a) Matching studies are among the most conceptually, methodologically, and practically complicated of all forms of treatment evaluation research; and (b) appropriate studies of patient-treatment matching can be accomplished only after careful specification of patient and treatment characteristics. In addition to making conceptual progress, researchers have investigated basic patient characteristics that are generally predictive of outcome across a variety of treatments. Treatment methods have begun to be specified more clearly and to be applied in the manner specified. There has also been progress in the development and utilization of designs appropriate to the study of patient and treatment characteristics that have sufficient power to obsene real effects and in the use of statistical treatments that can adequately assess the complexity of various interactions. Advances in matching over the past decade have occurred primarily in methodological development, patient measurement, and treatment measurement. In the first part of this chapter, progress in each of these areas is summarized, and general suggestions for research opportunities are presented. Review articles by Miller and Hester (1986b), Annis (1987), and Longabaugh (1986) have been particularly helpful, and the reader is encouraged to use these sources for additional information and references. ADVANCES IN PATIENT-TREA1~MENT MATCHING RESEARCH, 1978-1988 The developments of the past decade provide a basis for an expectation of additional progress in the matching area. First, after a review of the concepts, methods, and results of existing studies, the committee suggests more focused designs and clearer evaluation questions tailored to the treatments being studied. In the second section of this chapter, the committee discusses potential areas of progress and identifies specific research opportunities. Advances in Research Methodology The past decade has seen the increased deployment of rigorous, experimental designs to assess treatment efficac y as well as the development of new and more sophisticated statistical procedures to analyze multiply determined outcomes. Many patient-treatment matching studies from the 1960s through the late 1970s were retrospective examinations of variables that had been generated at the time of admission to treatment from data taken from patients admitted to two or more different programs or -231

OCR for page 231
treatments. They were generally analyzed by using a series of t-tests, chi-squares, or simple correlations. Some of the recent progress in method development is due to the practice developed during the past 10 years of analyzing multiple outcome measures (e.g., amount, duration, and frequency of alcohol consumption; employment; social adjustment; medical services utilization) rather than abstinence or nonabstinence alone when evaluating alcoholism treatment efficacy. This practice has led to the recognition that outcome is not unidimensional and that many aspects of the patient's posttreatment condition (e.g., employment situation or family, psychiatric, and medical status) can have a direct bearing on the likelihood of relapse. The earlier concentration by clinicians and researchers on a unidimensional outcome tended to obscure the complexity of the treatment process and the factors that accounted for treatment failure. The more recent development of measuring multiple outcomes has encouraged researchers to give more thought to determining reasonable goals and expectations from venous types of treatments and deciding what types of patients could be expected to show change in each. This trend has resulted in more rigorous designs with better specification of treatments and the patients who should undertake them, as well as testable hypotheses regarding outcome. Another development that has facilitated matching and is also related to the practice of evaluating multiple outcome measures is the use of multivariate statistics. Current reports rarely contain only two-by-two chi-square distributions of abstinent and nonabstinent frequencies in two treatment programs. It is now recognized (Skinner, 1981; Finney and Moos, 1986) that the number and complexity of the interactions among patient pretreatment characteristics, during-treatment factors, and posttreatment environmental factors cannot be characterized without the sophisticated use of multivariate statistical procedures. The introduction of new statistical procedures has also increased the rigor of measurement, the sophistication of the designs used, and the specificity of the hypotheses tested. Much progress in our understanding of matching has come from the use of more rigorous methodologies and experimental designs. This usage should be encouraged. However, because randomized controlled trials are not appropriate or possible for all evaluation and treatment matching studies in all clinical settings, it is necessary to develop a range of rigorous, quasi-experimental designs. There have been few evaluations of patient treatment matching strategies outside of public or institutional settings that treat mainly lower socioeconomic strata patients. Given the importance of social supports and stability in determining outcome across a varieW of treatments, it may be that some of the conclusions reached to date apply to only a limited segment of the patient population. In addition, what some researchers have seen as resistance to evaluation on the part of clinics and clinicians may be the result of ethical and financial pressures. Many private clinics that might have liked to participate in evaluations may have been prevented because of design limitations. Three matching strategies that have been developed within recent years are discussed below. 1. Ewing (1977) proposed a Cafeteria matching strategy in which patients are offered several alternatives and are permitted to select among them. The major limitation of this strategy is that the attractiveness of a treatment is assessed along with its efficacy, and attractiveness has been shown to be a major factor in determining retention in treatment and acceptance of treatment goals (Luborslgr et al., 1984; Miller, 1985~. -232

OCR for page 231
- ~ 2. Feedback designs employ early findings to generate testable hypotheses that in turn generate self~orrecting further hypotheses (Glaser, 1980; McLellan et al., 1980, 1983a,b). Patients are evaluated pre- and posttreatment within a multiprogram network, with no modification of standard patient-program assignments. Data are analyze, and patient-program Statistical hunches are developed on the basis of retrospective data. In the second stage of the project, patients are assigned, based on the launches, to the programs that may be best for them. Data are again analyzed, and ~matched" patients are compared with ~mismatched. patients. Limitations include the influence of the self-selection process, the need for adequate variability in program characteristics, and the need for a relatively stable treatment network. 3. Experiments can be designed to test the addition of a critical treatment element to the usual treatment. This strategy can be effective in cases in which there is a minimum standard treatment available. Because no patient is denied the standard treatment and because participation in this type of study can lead to extra care, these designs are often well accepted by both patients and treatment staff. The limitations of this design concern the sample sizes that can reasonably be generated and the types of treatments for which this design can be used. - All three of these designs may allow matching research to be done in settings, treatments, and populations that have not previously been studied. They may also permit the research findings generated thus far to be more easily translated into clinical practice. The following are opportunities for research on methodology development: Euang's cafeteria matching strategy should be reexamined. Statistical models of clinical decision making should be employed. The clinical decision-making process of senior, experienced-clinicians can be modeled mathematically by using discriminant function or path analyses. Studies should be performed to examine the data that are used in making treatment assignment decisions (Miller and Hester, 1986b). Experiments designed to test the addition of a critical element to the usual treatment should be attempted. Studies should be performed using feedback designs. Advances in Patient Measurement During the past decade there has been greater recognition of the full range of problems commonly seen among alcoholics who present for treatment. As discussed in Chapter 8, a number of measurement instruments have been developed to assess these problems, both at the time of admission, when such instruments serve as predictors of treatment outcome, and at follow-up points, when they function as outcome criteria. The social, psychological, and economic problems of alcohol-dependent patients may contribute to relapse and therefore should be targets for intervention during alcohol treatment (Finney and Moos, 1986~. This realization has broadened the scope and complexity of treatment evaluations. Several groups have used developments in patient measurement to search for generic patient variables that may be broadly predictive of outcome across several different types and settings of treatment. This rapidly progressing area of research has been reviewed by Longabaugh (1986), Miller and Hester (1986b), and Annis (1987~. Four types of -233

OCR for page 231
measurement appear to be generally predictive of treatment outcome as evidenced by replication in more than one treatment setting and in more than one type of patient population: 1. Social stabilitY/socia1 supports. Fewer supports result in worse treatment response generally but especially in the case of outpatient treatment. This factor appears to be a major generic predictor (McLellan et al., 1983a; Longabaugh and Beattie, 1985~. 2. Psychiatric diagnosis including severity/number, duration, and intensity of symptoms. Greater severity indicates generally worse treatment response, especially for outpatient treatments. This measure of overall impairment, irrespective of diagnosis, has been a good predictor across a variety of outcome domains and patient populations (McLellan et al., 1983a,b; Cooney et al., 1987; Rounsaville et al., 1987; Babor et al., 1988~. Specific diagnoses (e.g., major or intermittent depressive disorders, panic attacks) have been the symptom patterns usually associated with greater overall severity of psychiatric illness among alcohol-abusing patient populations. These diagnostic entities have received a great deal of attention as predictors of outcome. The differentiation of primary and secondary psychiatric disorders in alcoholics has been emphasized by Schuckit (1984) as a predictor of treatment outcome. As new and more specific treatments develop for psychiatric disorders, these treatments are being provided to patients with concomitant alcohol problems. 3. Severity of alcohol use/severity of alcohol dependence syndrome. Greater severity means worse treatment response generally and, especially for outpatients, more rapid return to drinking and drinking problems. This variable has probably been the best predictor of posttreatment alcohol consumption and alcohol-related problems, although it has been somewhat less effective as a predictor of adjustment in other areas such as employment, medical health, or family relations (Orford, Openheimer, and Edwards, 1976; Polich, Armor, and Braiker, 1980; Lyons et al., 1982; Babor et al., 1988~. 4. Antisocial personality (ASP) disorder. The presence of ASP is generally indicative of poor treatment response across all modalities with the possible exception of enforced disulfiram treatment (Schuckit, 1984; Stabenau and Hesselbrock, 1984; Hesselbrock, Meyer, and Keener, 1985; Powell et al., 1985~. Other patient variables have been related to treatment outcome in general, but they have not been examined extensively. These variables include a family history of alcoholism, particularly in a f~rst-degree relative (Schuckit, 1985) and cognitive or information-processing problems (Walker et al., 1983~. There is also a large literature on the use of personality tests such as the Minnesota Multiphasic Personality Inventory (MMPI) to develop topologies of clients that would be differentially responsive to treatment; however, thus far no group within these topologies has emerged that has a clear relationship to treatment response. New and promising approaches to the subtyping of patients based on combinations of personality traits, drinking history, and familial variables have been developed by Morey, Skinner, and Blashfield (1984), Zucker (1987), and Cloninger (1987~. To the extent that these topologies are found in the future actually to represent homogeneous groups of alcoholics, it will be more efficient to match treatments to patient types defined by multiple characteristics than to patient types defined by individual traits or characteristics. Although it is encouraging that there has been a great deal of replication of these generic "patient predictor findings," all of which appear to be conceptually sensible and potentially useful clinically, more research should be done on the following areas of patient measurement: -234

OCR for page 231
Parametric work is needed to determine the strength of these variables across a variety of patient mixes and treatment types. The extent to which robust predictors cova~y, especially in selected segments of the population, should be examined. For example, it is possible that a high level of alcohol dependency, ASP, and a history of familial alcoholism may occur together, especially in chronic male alcoholics. The American Psychiatric Assomation's revised edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R) differs in major ways from DSM-III, and it will be important to determine whether the differences in classification affect patient outcome across a range of treatments. With the exception of ASP, there has been very little evidence that the presence of a specific psychiatric disorder is predictive of treatment outcome. Global symptom severity has been more predictive than specific diagnosis. Specific treatments must be developed to address the specific needs of certain diagnostic groups (e.g., depressed alcoholics, alcoholics with panic disorder) before matching can become optimal. Additional work is needed to better define and describe the dimensions of ASP (including impulsiveness, learning disorder, childhood aggression, relationship problems, and criminal activity) to determine the particular factors responsible for the observed resistance to treatment for alcohol problems. Motivation for change is widely viewed as an important predictor of response tO treatment, but there has been little indication that this quality can be measured either reliably or validly. A clear definition of motivation and a means of measuring it would be valuable. Work reported by Prochaska and DiClemente (1983) that describes a method for measuring "readiness for change" should be extended. ADVANCES IN TREATMENT MEASUREMENT New treatment models have been developed during the past several years, including the relapse prevention model (Gorski and Miller, 1982; Marlatt and Gordon, 1985) and brief interventions (Sanchez-Craig and Walker, 1982; Sanchez-Craig, 1984; Miller, 1985~. There have also been wider use of standard treatments in different settings and experimentation with outpatient detoxification and rehabilitation (Longabaugh et al., 1983; Hayashida et al., 1989~. In addition, manual-directed relapse prevention efforts (Marlatt and Gordon, 1985) and self-help treatments (Sanchez-Craig, 1984; Heather, 1986; Miller, 1986) have been developed and implemented. These new models are welcome developments for a variety of reasons. The manuals are an important source of training in theory and practice for clinicians and promote the standardized delivery of treatment. They are also valuable for researchers, enabling the derivation of better measures of treatment goals and processes. Although there has been progress in treatment process standardization, other areas still require work. New treatments are needed for specific segments of the population such as the cocaine-alcohol-dependent patient, the antisocial personality alcoholic, the alcoholic schizophrenic, and the medical patient with an alcohol-related disorder (e.g., alcohol cardiomyopathy, alcohol hypertension). . These new treatments should be conceptually related to the specific problems of the target populations and should use manuals for guidance tO ensure the accuracy of treatment application. It should be noted that patient-treatment matching can occur only when there are clearly different treatments. Many treatments are minor variants in practice. There is no evidence at this time that any single process or approach will work for every patient, and newer methods may succeed where older ones have not. As much attention should be paid to -235

OCR for page 231
developing instruments that measure characteristics of treatments as has been paid to the development of instruments that measure characteristics of patients. Although it is frequently said that patient characteristics account for more variability in outcome than do treatment characteristics, this perception could be a function of the lack of available treatment measures (Glaser, 1980~. The development of treatment manuals for specific therapies has promoted cnterion-based measures of the extent to which treatments are actually delivered as intended, the quantity of treatment provided to a patient, and ratings of its quality. Moreover, evaluation research using these manual~erived process measures has begun (Luborsly et al., 1984~. Instruments are now needed to assess the nature and amount of various treatment components across a range of treatment programs and across various stages of the treatment process. This assessment would provide some indication of which training components are most effective as well as suggestions about how to train therapists and improve the delivery of these components. The lack of such instruments has handicapped the study of matching and the accurate evaluation of treatment efficacy. The following questions represent opportunities for research on treatment measurement: What is treatment supposed to do? Many insurance companies and treatment funding agencies have reduced to 28 days the amount of time for which they will reimburse alcohol rehabilitation (formerly, they reimbursed from 90 to 120 days). Have the goals of treatment changed to reflect these time constraints? Are the goals of rehabilitation clearly different from detoxification, or do they complement each other? Have patients reached the established goals of treatment by the time they are discharged or released? There are few studies evaluating progress during the course of treatment (as opposed to outcome measures on the completion of treatment). How much entreatments does it take to produce a given level of change for a specified number of patients? It is necessary to develop a more quantified estimate of treatment effects so that more informed decisions can be made regarding how much treatment is necessary and in turn how much treatment should be compensated by insurance companies. Does the achievement of during-treatment goals relate to posttreatment success? The-work by Finney, Moos, and Mewborn (1980) showing the importance of the posttreatment environment has led to new studies. However, even this excellent research lacks the specificity of measurement necessary to determine if the "treatments has been applied in the intended manner and to an adequate extent. What are the "active ingredients" of treatment? Treatment programs offer many services or interventions. In a study by McLachlan and Stein (1982), patients received a standard treatment package of medical consultation, disulfiram or calcium carbimide, group psychotherapy, education, relaxation training, nutritional counseling, physiotherapy, physical training/exercise, and individual planning for an alcohol-free life-style. Are all of these necessary? Are they all associated with outcome? What are the essential or Active ingredients? Investigators have been attempting to match people to treatment settings (inpatient and outpatient) or treatment programs without knowing which factors are responsible for observed patient changes. -236

OCR for page 231
RESEARCH ON PATIENT-TREATMENT MATCHING: PERSPECTIVES AND OPPORTUNITIES Given the advances of the last decade in the areas of methodology development and patient measurement, what can be expected in the future in the area of matching research? The analysis of more discrete and better defined stages of the rehabilitation process offers the greatest Dotential for progress in refining the process of treatment selection and the provision or specific treatment components during rehabilitation. It can also help to improve the services offered in the posttreatment environment. Just as each of these stages of the rehabilitation process takes place in a different context and requires the patient to set different goals, the clinical possibilities for patient-treatment matching and the potential for matching research will also be different. cat ~ ~ ~ c' . In the remainder of this section the committee examines the potential for matching research in each of four areas that correspond to typical treatment situations: (1) matching before treatment starts, (2) matching at the initiation of treatment, (3) matching during the treatment process, and (4) matching following the rehabilitation intervention. Critical commentary is provided regarding the matching research conducted to date, along with suggestions for future methods to be applied and examples of specific types of studies that could be performed. Matching Before Treatment Starts: Special Populations and Self-Selection Ideally, the outcome evaluation researcher would like to examine the effects of a representative treatment program, technique, or modality on randomly selected patient samples that proportionally represent the total patient population. However, the population of ~alcohol-problemed" individuals is not completely represented by the patients who present for treatment. Moreover, because the nature of the treatment program, including its location, cost, referral network, charter, and preferred modalities, will determine in large part the types of patients who seek treatment, the sample of patients evaluated at a specific treatment program will not be representative of the population of treatment-seeking individuals. Therefore, all outcome studies done to date reflect imperfectly applied treatments in nonrepresentative, generally self-selected patient samples; most often these samples comprise the most severely impaired patients. Designers of treatment programs have recognized and utilized patient self-selection in their marketing strategies and in their clinical attempts to develop programs tailored to the individual needs of the patient. This process, which can be termed "solicited self-selection," is sensible in that it is not likely that the effects of even conceptually identical and comparably applied treatments would be similar, for example, for both a sample of older, male, lower-socioeconomic, chronic alcoholic veterans treated in a Veterans Administration hospital and a sample of adolescent, middle-class girls referred to treatment at a private facility. It is apparent that a form of matching takes place prior to the initiation of treatment through the process of specialization as well as the selective marketing and referral of seemingly appropriate patients. This type of marketing has encouraged the development of special programs for special populations such as adolescents, Native Americans, women, abused women, adult children of alcoholics (ACOAs), homeless men, and many others. These specialized programs are by far the most extensive matching work carried out in this country, yet little more than descriptive information is available about outcomes. -237

OCR for page 231
Special populations are important for patient-treatment matching (particularly at the level of treatment entry), but they have been difficult to study. It is not easy to discern who constitutes a special population. In a recent seminal report by Saxe and coworkers (1983), the special populations discussed are elderly people, adolescents, women, blacks, Hispanics, Native Americans, drunk drivers, and public inebriates/skid row alcoholics. The problems associated with evaluating treatment options for these groups are numerous and obvious. The three major definitional problems are that (1) these groups have substantial internal variability (e.g., Hispanics include Puerto Ricans, Mexicans, Cubans, etch; (2) these groups are not mutually exclusive (e.g., ~ member of the black special population may be elderly, Hispanic, a skid row woman, and convicted of drunk driving); and (3) they are not exhaustive (i.e. more recent designations have added pregnant, physically handicapped, and psychiatrically ill alcoholics to the groups already mentioned). The matching research that has been done in this area has attempted to determine whether differential outcomes are seen among these groups when they are treated in the same program. To date, the modest amount of research that has been completed has shown no clear indication that these group designations are associated with different outcomes among those patients who have entered treatment (Westermeyer, 1982; Blum, 1987; Blume, 1986~. It is quite clear, however, that members of these groups do not enter available "mainstream" treatments in proportions that are representative of the alcohol problems within those groups. For this reason, the major efforts in this area related to patient-treatment matching have been in the development of tailored programs designed and operated by and for selected special population groups. The goal of these efforts has been to attract more alcohol-impaired individuals from these groups into treatment. Although there has been a marked increase in the number of programs available, it is not yet clear that proportionally more members of these groups have been attracted to treatment or that greater proportions of special populations enter special programs rather than traditional programs. When these programs have offered attractions that were specifically directed toward their target populations (e.g., child care for women's programs, special access for handicapped programs), it is not clear whether or to what extent the actual treatment provided within these programs differed from more mainstream types of treatment or whether they were associated with differential opportunities. These facts of clinical life do not imply that research in this area is impossible. Matching or the prediction of outcome studies can be performed simultaneously with ongoing treatments as long as the questions addressed and the methodologies employed are suitable in the treatment context. The following questions represent opportunities for research on programs designed with specialized segments of the patient population in mind: Do patients with the "right" patient profile stay longer, show more improvement, and remain improved longer than patients with the "wrong" profile? Is greater demographic and socioeconomic homogeneity among patients associated with better retention in a specific treatment? How much and what types of diversity can a patient population tolerate and still maintain cohesiveness? This is a particularly relevant question for those treating the increasing numbers of cocaine and alcohol abuse patients who present for treatment at traditional ~alcohol-only" programs. Can these patients be treated along with alcohol-only patients? Are treatment goals and methods compatible for these two types of patients? Can women be treated as effectively in mixed male and female settings as in specialized women's facilities? Similar questions could be asked with respect to adolescent alcoholics as well as other significant subgroups in the total patient population. -238

OCR for page 231
Do similar treatment strategies and components (e.g., group therapy, education, Alcoholics Anonymous, disuliram) have qualitatively similar effects on outcome among programs with different patient profiles? Matching at the Initiation of Treatment: Levels of Treatment Intensity There are several different levels of treatment intensity offered in most areas: (a) advice/self-help; (b) brief interventions (usually fewer than five counselor appointments lasting about a week); (c) outpatient care (usually one to three hours per day, three to five days per week); (d) partial hospitalization (usually six to eight hours per day); (e) residential inpatient (nonmedical inpatient setting); and (f) medical inpatient (in a specialized unit in a general or psychiatric hospital). There now exist a body of valid, replicated data on patient factors in treatment matching and enough understanding of the various levels of treatment that the development of more carefully staged designs is possible. Staged or hierarchical designs refer to the tailoring of matching hypotheses to the treatment goals and patient populations appropriate for different levels of treatment structure (e.g., no treatment, brief treatment, outpatient, inpatient). New work in this area should build on conclusions from previous studies that are sensible and have been replicated. The work of many investigators has been reviewed by Sanchez-Craig (1984) and by Miller and Hester (1986a). These efforts indicate that individuals with less severe and shorter periods of problem drinking, better social supports, and fewer medical or psychological problems can improve without intensive treatment. The following are opportunities for research on matching at treatment initiation: Studies of matching to different levels of treatment intensity should attempt to select patients with approximately the same level of problems and social supports. The results of such studies might then permit a better understanding of the patient factors within the clinically appropriate group that are associated with outcomes from each level of treatment intensity. Studies of matching within each level of treatment intensity are needed to evaluate the treatment components (e.g., group therapy, individual therapy, medication, education) and patient characteristics in the clinically appropriate group that are associated with favorable outcome. Models of patient assignment to different levels of treatment intensity (e.g., Hoffman et al., 1987) should be evaluated in a series of controlled trials at various sites and with various segments of the patient population. Matching During the Treatment Process: Role of Treatment Components There is a fairly discrete set of treatment components that is provided, or at least offered, to most alcohol-dependent patients in treatment, regardless of the treatment modality or setting. These components include (a) group therapy (usually focused on issues of treatment need and denial); (b) individual therapy (usually personal counseling on relationship problems and crises); (c) alcohol or substance abuse education; (d) attendance at Alcoholics Anonymous (AA) meetings; and (e) antidipsotropic medications (usually disulfiram). -239

OCR for page 231
Although some research has been conducted to examine matching in treatment settings (e.g., inpatient versus outpatient) and among programs, there has been little matching work done on treatment components (e.g., medication, therapy, education) and even less on therapist or counselor technique. Studies in these areas are potentially important in that the failure to find evidence of differences in efficacy as a result of matching to different programs or settings may be due to the similarity of the therapeutic methods employed in these treatment venues. As has already been emphasized, there is a need for more detailed measurement of the treatment process, as well as a need to identify the active ingredients of treatment. If the active ingredients are identified at the process level and if they are applied fairly similarly across treatment settings and programs, then it would not be surprising if there were not much evidence of differential outcomes from Patient-setting or Patient-program matching. The following are opportunities for research on matching during treatment: Random patient assignment methods in controlled experimental trials can be used most effectively used in studies within a treatment setting or program to investigate the value of different combinations of treatment components or the addition of a specific component to the usual treatment. The study by Woody et al. (1984, 1985) of psychotherapy as an adjunct to standard counseling is an example of an approach that can provide clear data on the value of specific treatment components. Each of the standard treatment components now used in rehabilitation programs (education, the Twelve Steps, group therapy, etc.) should be evaluated for its contribution to outcome by comparison with programs that have all other aspects of the treatment except the target component. Matching Following the Rehabilitation Intervention: Role of the Posttreatment Environment The work of Finney, Moos, and Mewborn (1980) has shown that the posttreatment environment can profoundly influence the overall outcome for treated alcohol abusers. In the past, the posttreatment environment of patients depended on the patient's personal resources because most programs concentrated on primary treatment and the funds needed to develop individually tailored posttreatment programs were not available. Continuing treatment options offered to patients who had completed primary care were generally restricted to AA meetings and possibly a weekly or monthly continuing care meeting at the primary care site. Because of recently shortened periods of reimbursed care for primary rehabilitation and some new financial incentives to provide outpatient treatment, clinical programs are now devoting more time to the development of posttreatment continuing care programs and have attempted to bring the family of a patient into the continuing treatment process. The availability of these services provides an opportunity for patient-treatment matching research following the period of primary rehabilitation. The following are opportunities for research on patient-treatment matching following primary rehabilitation: Comparative studies of AA treatment, relapse prevention, individual therapy, or family therapy following the completion of primary rehabilitation might be initiated in a -240

OCR for page 231
variety of treatment settings and patient populations. Do these interventions add anything beyond primary treatment? What types of patients benefit most from each of these treatments? Idealtr, this research should involve parametric studies that investigate the optimum duration and intensity of treatments and should include measures of cost -effectiveness. Comparative studies of different posttreatment environments (e.g., halfway houses, quarter-way houses, family treatment centers) should be conducted to determine the overall efficacy of these environments and the types of patients best suited to them. Comparative studies of family treatments that are independent of the rest of the patient's treatment program could be used to evaluate the contributions of various forms of family education to the posttreatment adjustment of the patient. For example, during the course of the patient's rehabilitation, families could be assigned to Al-Anon, family therapy, alcohol education, or individual counseling. It should then be determined whether these interventions add anything beyond primary treatment for the affected patient and whether these different approaches can be matched with specific types of families. PATIENT-TREATMENT MATCHING: SOME CONCLUSIONS AND RESEARCH RECOMMENDATIONS Work to date on patient-treatment matching leads to three conclusions, which are recapitulated here. 1. Patient factors appear to be more predictive of outcome from treatment than are treatment process factors. Techniques for measuring patient characteristics have shown major development in breadth, reliability, and validity over the past decade. In contrast, treatment processes have been almost unstudied, and there are no available instruments for reliable and valid treatment measurement. The broader range of treatments now available and under development may reveal more potent treatment process factors if treatment is actually provided in an appropriate manner and for an adequate length of time. 2. Of the patient variables that have been studied, psychosocial factors have been shown to be the most important predictors of outcome for different treatment intensities (e.g., inpatient, partial hospitalization, outpatient care). Patients with better social and economic support and fewer psychiatric problems do well in most treatments and seem to benefit equally from inpatient or outpatient interventions. Lower socioeconomic strata patients and those having more serious psychiatric problems do less well in treatment generally and fare particularly poorly in outpatient care. Such patient factors as the severity of alcohol dependence, family history of alcoholism, and presence of antisocial personality disorder have been generally predictive of poorer outcomes front all treatments but are not differentially predictive of response to specific treatments. 3. There have been very few studies in which patients were matched to different treatment components (e.g., group therapy, individual therapy, medication, relapse prevention) Within a given level of treatment intensity. There are at this time no clear predictors of differential outcomes from any of these components. Work on patient-treatment matching offers the potential for significant, practical advances. To achieve this potential, the committee makes the three following recommendations: 1. There is a need for a more specific focusing of matching questions (Longabaugh, 1986; Annis, 1987~. Efforts should be made to study well-defined treatments that have clear therapeutic goals for specific segments of the patient population. Matching studies -241

OCR for page 231
that employ appropriate designs should be considered at each level of the rehabilitation process. At the levels of referral to the treatment program (primary care) and posttreatment environment (aftercare), experimental designs that employ random patient assignment and nonexperimental designs such as a cafeteria approach (Ewing, 1977) or a feedback system (Glaser, 1980) should be considered. When assigning patients to treatment components or treatment providers within a specific program or environment, experimental designs with random patient assignment are preferable to evaluate the differential efficacy of components of approximately equal attractiveness and comparable intensity. 2. More innovative interventions are needed, as well as programs designed to address specific treatment problems of different groups in the population (e.g., the psychiatrically ill alcoholic, the antisocial alcoholic, the cocaine- and alcohol~ependent patient). Similarly, there is a need to continue evaluation and patient-treatment matching work with recently developed treatments for problem drinkers (Sanchez-Craig, 1984; Miller 1985), for relapse prevention (Gorski and Miller, 1982; Marlatt and Gordon, 1985), and for community reinforcement (Azrin et al., 1982~. As discussed in Skinner (1981), it is difficult to study the optimum matching of patients and treatments when there is so little variability in the philosophy, duration, or basic therapeutic components of most treatments. 3. Reliable, valid, practical, and generalizable instruments are needed to measure the types, amounts, and duration of alcohol treatment interventions a patient receives during the course of rehabilitation. These measurements are important both for training therapists and for evaluating of treatment efficacy. If treatments are not applied in an appropriate manner, then it is unreasonable to think that they will work. We do not always know whether a specific intervention (e.g., group therapy for denial), much less a multi-senrice treatment program, is practiced in the manner originally intended. We do not always know the extent to which different individuals in a single treatment receive the same types, amounts, or duration of treatment components. The often repeated claim that patient factors account for more outcome variation than treatment factors may be simply a function of the unavailability of treatment measurement instruments or the close association between certain client characteristics (e.g., age, marital status, antisocial personality) and the posttreatment environments to which these individuals typically return. The ability to characterize a treatment intervention or program as well as it is now possible to characterize patients should substantially enhance our ability to predict outcomes and assign (match) patients to optimal treatments. REFERENCES Annis, H. M. Effective treatment for drug and alcohol problems: What do we know? Paper presented at the annual meeting of the Institute of Medicine, National Academy of Sciences, Washington, DC, October 21, 1987. Azrin, N. H., R. W. Sisson, R. Meyers et al. Alcoholism treatment by disulfiram and community reinforcement therapy. J. Behav. Ther. Exp. Psychiatry 13:105-112, 1982. Babor, T. F., Z. S. Dolinsky, B. J. Rounsaville et al. Unitary versus multidimensional models of alcoholism treatment outcome: An empirical study. J. Stud. Alcohol 49:167-177, 1988. Blum, R. W. Adolescent substance abuse: Diagnostic and treatment issues. Pediatric Clinics of North America 34:523-537, 1987. Blume, S. B. Women and alcohol: A review. J. Am. Med. Assoc. 256:1467-1469, 1986. -242

OCR for page 231
Brandsma, J. M., M. C. Maultsby, and R. J. Welsh. The outpatient treatment of alcoholism: A review and comparative study. Baltimore, MD: University Park Press, 1980. Cloninger, C. R. Neurogenetic adaptive mechanisms in alcoholism. Science 236:410-416, 1987. Cooney, N. L., R. M. Kadden, M. D. Litt et al. Alcoholism aftercare treatment matching: Posttreatment results. Paper presented at the Fourth International Conference on Treatment of Addictive Behaviors, Bergen, Norway, 1987. Ewing, J. ~ Matching therapy and patients: The cafeteria plan. Br. J. Addict. 72:13-18, - 1977. Finned, J. W., and R. H. Moos. Matching patients with treatments: Conceptual and methodological issues. J. Stud. Alcohol 7:122-134, 1986. . . . Finney, J. W., R. H. Moos, and C. R. Mewborn. Post-treatment experiences and treatment outcome of alcoholic patients sur months and two years after hospitalization. J. Consult. Clin. Psychol. 48:17-29, 1980. Glaser, F. B. Anybody got a match? Treatment research and the matching hypothesis. In G. Edwards and M. Grant, eds. Alcoholism Treatment in Transition. London: Crown Helm, 1980. Gorski, T. T., and M. Miller. Counseling for Relapse Prevention. Independence, MO: Herald House-Independence Press, 1982. Hayashida, M., A. I. Alterman, ~ T. McLellan et al. Comparative effectiveness and costs of inpatient and outpatient medical alcohol detoxification. N. Engl. J. Med. 320~6~:358-365, 1989. Heather, N. Change without therapists: The use of self-help manuals by problem drinkers. Pp. 331~359 in W. R. Miller and N. Heather, eds. Treating Addictive Behaviors: Processes of Change. New York: Plenum Press, 1986. Hesselbrock, M. N., R. E. Meyer, and J. J. Keener. Psychopathology in hospitalized alcoholics. Arch. Gen. Psychiatry 42:1050-1055, 1985. Hoffmann, N. G., F. Ninonuevo, J. Mozey et al. Comparison of court-referred DWI arrestees with other outpatients in substance abuse treatment. J. Stud. Alcohol 48:591-594, 1987. Institute of Medicine (IOM). Alcoholism, Alcohol Abuse, and Related Problems: Opportunities for Research. Washington, DC: National Academy Press, 1980. Longabaugh, R. The matching hypothesis: Theoretical and empirical status. Paper presented at the Annual Meeting of the American Psychological Association, 1986. -243

OCR for page 231
Longabaugh, R., and M. Beattie. Optimizing the cost effectiveness of treatments for alcohol abusers. Pp. 104-136 in B. McGrady, N. E. Noel, and T. D. Nirenberg, eds. Future Directions in Alcohol Abuse Treatment Research. Washington, DC: NtAAA, 1985. Longabaugh, R., B. McCrady, E. Fink et al. Cost effectiveness of alcoholism treatment in partial versus inpatient settings: Sur month outcomes. J. Stud. Alcohol 44:1049-1071, 1983. Luborsly, L., A. T. McLellan, G. E. Woody et al. Therapist success and its determinants. Arch. Gen. Psychiatry 81:123-130, 1984. Lyons, J. P., J. Welte, J. Brown et al. Variation in alcoholism treatment orientation: Differential impact upon specific subpopulations. Alcoholism Clin. Exp. Res. 6:333-343, 1982. Marlatt, G. A, and J. R. Gordon. Relapse Prevention. New York: Guilford Press, 1985. McLachlan, J. F. C., and R. L. Stein. Evaluation of a day clinic for alcoholics. J. Stud. Alcohol 43:262-272, 1982. McLellan, ~ T., K ~ Druley, C. P. O'Brien et al. Matching substance abuse patients to appropriate treatments. A conceptual and methodological approach. Drug Alcohol Depend. 5~3~:189-193, 1980. McLellan; A; T., L. Luborsly, G. E. Woody et al. Increased effectiveness of substance abuse treatment: A prospective study of patient-treatment Matching. J. New. Ment. Dis. 171~10~:597-605, 1983a. McLellan, ~ T., L. Luborsly, G. E. Woody et al. Predicting response to alcohol and drug abuse treatments: Role of psychiatric severity. Arch. Gen. Psychiatry 40:620-625, 1983b. Miller, W. R. Motivation for treatment: A review with special emphasis on alcoholism. Psychol. Bull. 98:84-107, 1985. Miller, W. R. Haunted by the Zeitgeist: Reflections on contrasting treatment goals and conceptions of alcoholism in Europe and the United States. In T. F. Babor, ed. Alcohol and Culture: Comparative Perspectives from Europe and America. New York: New York Academy of Sciences, 1986. Miller, W. R., and R. K Hester. Inpatient alcoholism treatment: Who benefits? Am. Psychol. 41:794-805, 1986a. Miller, W. R., and R. K Hester. Matching problem drinkers with optimal treatments. Pp.175-203 in W. R. Miller and N. Heather, eds. Treating Addictive Behaviors: Processes of Change. New York: Plenum Press, 1986b. Morey, L. C., H. ~ Skinner, and R. K Blashfield. A typology of alcohol abusers: Correlates and implications. J. Abnormal Psychol. 93:408-417, 1984. -244

OCR for page 231
Orford, J., E. Openheimer, and G. Edwards. Abstinence or control: The outcome for excessive drinkers two years after consultation. Behav. Res. Ther. 14:409-418, 1976. Polich, J. M., D. J. Armor, and H. B. Braiker. Patterns of alcoholism over four years. J. Stud. Alcohol 41:397416, 1980. Powell, B. I., E. C Penick, M. R. Read et al. Comparison of three outpatient treatment interventions: A twelve-month follow-up of men alcoholics. J. Stud. Alcohol 46:309-312, 1985. Prochaska, J. O., and C. C. DiClemente. Stages and processes of self-change of smoking: Toward an integrated model of change. J. Clin. Consult. Psychol. 51:390-395, 1983. Rounsaville, B. J., Z S. Dolinsly, T. F. Babor et al. Psychopathology as a predictor of treatment outcome in alcoholism. Arch. Gen. Psychiatry 44:505-513, 1987. Sanchez-Craig, M. Therapist's Manual for Secondary Prevention of Alcohol Problems: Procedures for Teaching Moderate Drinking and Abstinence. Toronto: Addiction Research Foundation, 1984. Sanchez-Craig, M., and K Walker. Teaching coping skills to chronic alcoholics in a coeducational halfway house. I. Assessment of programme effects. Br. J. Addict. 77:35-50, l9g2. Saxe, L^, D. Dougherty, K Esty et al. Health Technology Case Study 22: The Effectiveness and Costs of Alcoholism Treatment. Washington, DC: U.S. Congress, Office of Technology Assessment, 1983. Schuckit, M. ~ Subjective responses to alcohol in sons of alcoholics and control subjects. Arch. Gen. Psychiatry 41:879-884, 1984. Schuckit, M. ~ The clinical implications of primary diagnostic groups among alcoholics. Arch. Gen. Psychiatry 42:1043-1049, 1985. Skinner, H. ~ Different strokes for different folks. Differential treatment for alcohol abuse. Pp. 349-367 in R. F. Meyer, T. F. Babor, B. C. Glueck et al., eds. Evaluation of the Alcoholic: Implications for Research, Theory and Treatment. NIAAA Research Monograph No. 5. USDHHS Publ. No. (ADM)81-1033. Washington, DC: NIAAA, 1981. Stabenau, J. R., and V. Hesselbrock. Psychopathology in alcoholics and their families and vulnerability to alcoholism: A review and new findings. Pp. 108-132 in S. M. Mirin, ed. Substance Abuse and Psychopathology. Washington, DC: American Psychiatric Press, 1984. Walker, R. D., D. M. Donovan, D. R. Kivlahan et al. Length of stay, neuropsychological performance and aftercare: Influences on alcohol treatment outcome. J. Consult. Clin. Psychol. 51:900-911,1983. Westermeyer, J. Ethnic factors in treatment. Pp. 709-717 in E. M. Pattison, and E. Kaufman, eds. The American Encyclopedic Handbook of Alcoholism. New York: Gardner Press, 1982. -245

OCR for page 231
Woody, G. E.; ~ T. McLellan, L. Luborsky et al. Psychiatric severity as a predictor of benefits from psychotherapy. Am. J. Psychiatry 141:1171-1177,1984. Woody, G. E., ~ T. McLellan, L. Luborsly et al. Sociopathy and psychotherapy outcome. Arch. Gen. Psychiatry 42:1081-1086, 1985. Zucker, R. The four alcoholisms: A developmental--account of the etiological process. In P. C. Rivers, ed. Alcohol and Addictive Behavior. Lincoln, NE: University of Nebraska Press, 1987. -246