Click for next page ( 150


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 149
8 ISSUES OF ASSESSMENT, METHODOLOGY, AND RESEARCH DESIGN Despite the practical difficulties of research involving persons with alcohol-related problems, there have been some notable achievements in treatment evaluation in recent years. These accomplishments include conceptual advances, new measurement techniques, the growing sophistication of diagnostic procedures, and improved approaches to research design. In addition to reviewing the opportunities associated with these accomplishments, this chapter also discusses some of the major unresolved evaluation research Issues, such as who should be assessed, what variables should be measured, and the validity of current procedures. RECENT DEVELOPMENTS IN CONCEPTUALIZATION There are three important conceptual advances relevant to the current status and future conduct of research on the treatment of alcohol-related problems. These advances have guided the development of new assessment procedures that promise to provide a more sophisticated basis for measuring the underlying dynamics of alcohol dependence and its treatment. The first conceptual development is the description of alcohol dependence as a core syndrome within a broader range of alcohol-related problems. In recent years, research on the alcohol dependence syndrome (Edwards, 1986) has focused on a cluster of biological, psychological, and behavioral elements that are believed to be at the core of pathological alcohol-seeking behavior. With the development of new diagnostic criteria and assessment instruments to measure the presence and severity of alcohol dependence (Rounsaville et al., 1987; Robins et al., 1988), researchers are now in a much better position to evaluate specific treatment interventions that focus on such concomitants of dependence as craving, environmental precipitants, and biological vulnerability factors. A second development in conceptualization has resulted in a distinction between alcohol dependence-and alcohol-related disabilities. Whereas alcohol dependence is seen as a coherent syndrome with signs and symptoms that tend to occur together and mutually reinforce one another, alcohol-related disabilities are now considered to be a rather heterogeneous set of physical, psychological, and social impairments that occur independently of alcohol dependence. In research terms, this distinction means that the assessment of alcoholism should focus both on the underlying dependence and on the problems associated with it. It also means that alcohol-related problems that occur as a result of heavy drinking in the absence of dependence are also the proper targets of treatment prevention efforts. Finally, alcohol dependence is now being conceptualized and measured not as an all-or-none phenomenon but as a continuum that ranges in severity from mild to severe (Edwards, Arif, and Hodgson, 1981~. A third advance in conceptualization concerns new developments in the way treatment evaluation is viewed. Interest in treatment evaluation research has been increasing within the alcohol field over the past 20 years. The dominant approach in both clinical practice and treatment evaluation research has been to ask: Is treatment effective? Traditionally, research has attempted to discover the Main effects" of different treatments relative to the absence of treatment or exposure to some alternative treatment. A more sophisticated model of treatment evaluation, known as the Matching hypothesis," has emerged recently -149

OCR for page 149
in the fields of drug abuse, alcoholism, and psychotherapy (Luborsky and McLellan, 1981; Moos and Finney, 1985; VandenBos, 1986~. This model, which is discussed in detail in Chapter 11, places great demands on the skills, resources, and ingenuity of evaluation researchers. ADVANCES IN ASSESSMENT TECHNOLOGY Significant advances have been made in the technology of assessment, which have led to the development of new techniques for screening, diagnosis, and differential assessment. The research opportunities associated with advances in diagnosis and other assessment areas are discussed below. Screening is considered separately in Chapter 10. Diagnosis One of the most active areas of development since the last Institute of Medicine report (IOM, 1980) on research opportunities in the alcohol field has been diagnosis, including the tools for diagnostic assessment. The third edition of the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders (DSM-III) (American Psychiatric Association, 1980) was published the same year as the IOM report. DSM-III had a major impact on diagnosis in the mental health field (Spitzer, Williams, and Skodol, 1983) and fostered further development of structured diagnostic interviews. Together, ~operational" diagnostic criteria and structured interviews based on those criteria have had a profound effect on the way clinicians conceptualize and determine psychiatric diagnoses. For example, it is now clear that unstructured clinical examinations have major limitations. One problem is reliability among observers. The use of structured interviews has improved the reliability of diagnoses and other clinical assessments in the alcohol field. On the assumption that some types of information are best gathered by an interviewer, a related issue concerns the degree of clinical expertise the interviewer should possess. Recent research with non-clinically trained (lay) interviewers indicates that if the interview is highly structured, accurate diagnostic information can be gathered by interviewers with little~or no clinical experience (Coryell, Cloninger, and Reich, 1978; Helzer et al., 1981~. Some of the commonly employed examinations that are structured enough to be used by nonclinicians include the Diagnostic Interview Schedule (DIS) (Robins et al., 1981) and the Psychiatric Diagnostic Interview (PDI) (Othmer, Powell, and Penick, 19801. These instruments are essentially general diagnostic examinations. Three others that are designed specifically to gather information about substance use are the Composite International Diagnostic Interview, Substance Abuse Module (CIDI-SAM) (Robins et al., 1988), the Addiction Severity Index (ASI) (McLellan et al., 1980), and the Comprehensive Drinker Profile (Marlatt and Miller, 1984~. Related to the development of diagnostic instruments has been a series of ongoing programs of study within NIAAA, the American Psychiatric Association (APA), and the World Health Organization (WHO). These programs have led to recommended revisions in current diagnostic criteria in the Diagnostic and Statistical Manual (DSM-III-R) (APA, 1987), as well as in the International Classification of Diseases (ICD-10) (WHO, 1988~. They have also contributed to the development of new diagnostic instruments that are capable of classifying people reliably and quickly, according to the severity of their alcohol dependence (Robins et al., lam; Wing et al., in press). Moreover, advances in evaluation , -150

OCR for page 149
methodology (Meyer et al., 1985) have stimulated research in this area, and a number of assessment procedures have been developed to measure the severity of alcohol dependence (Stockwell, Murphy, and Hodgson, 1983; Skinner and Horn, 1984~. Research suggests that the severity of dependence predicts cravings for alcohol and failure to control drinking following relapse (Polich, Armor, and Braiker, 1981; Babor, Cooney, and Lauerman, 1987~. Although farther research is called for, it appears that the dependence syndrome construct has considerable promise for early detection, diagnosis, and treatment planning, at least in the European and North American countries in which most of the research has been conducted (Edwards, 1986~. Nevertheless, the concept of the alcohol dependence syndrome has been criticized from a number of points of view, and at present its scientific status as the basis for a worldwide definition of alcohol dependence requires further study. A more systematic program of research using cross-cultural comparative methods could provide useful information. To the extent that consistency is found across disparate samples of alcoholics from different cultures, the case for a core dependence syndrome would be strengthened, thereby facilitating epidemiological research, international communication, and comparisons across treatment evaluation studies. The following issues and questions offer opportunities for research on diagnosis: What is the construct validity of the dependence syndrome concept used as the basis for DSM-III-R and ICD-10? Research should be undertaken on the underlying assumptions of the alcohol dependence syndrome construct, particularly with respect to the relative separation of indicators of dependence from indicators of alcohol-related impairment. Because clinicians are using the same criteria to diagnose all substance abuse disorders, it is crucial to examine the commonalities and distinctions among psychoactive substances. To what extent do differences in severity of dependence represent differences in the abuse liability of different substances? Additional cross-cultural research would be useful to identify biobehavioral universals of alcohol dependence. Such research should be encouraged in the context of any program to revise international classification systems (e.g., the ICD-10~. What is the concordance between current (e.g., DSM-III) and proposed (e.g., DSM-III-R)~diagnostic systems, particularly in terms of interrater agreement and test-retest reliability? What is the internal and external validity of the new criteria proposed in DSM-III-R and ICD-10? Research should focus on predictive validity (assessing the clinical course of individuals diagnosed according to different diagnostic systems) and diagnostic stability (the persistence of diagnostic symptoms over different periods of the drinking career). Differential Assessment Differential assessment is the detailed evaluation of the patient's alcohol-use disorder in terms of etiology, presenting symptoms, substance-use patterns, alcohol-related problems, and other associated features. A variety of purposes can be served by the differential assessment. First, it can help to explain the underlying causes of a disorder. Second, data obtained by careful assessment can serge as both a baseline and a predictor of future behavior. Finally, differential assessment is crucial to individualized treatment planning, which is based on the identification of"matches" between patient characteristics or needs and specific treatment interventions. -151

OCR for page 149
There have been some notable advances in basic research that have begun to stimulate new approaches to assessment and patient placement. Laboratory studies of craving, psychological expectancies, and impaired control over drinking have helped to elucidate the natural bases of alcohol dependence (Meyer et al., 1985~. With the development of procedures to measure the severity of alcohol dependence, it may soon be possible to assign patients to different types and intensities of treatment, based on their vulnerability to readdiction and the suitability of different therapies for their treatment. The application of psychological learning theory to the study of relapse (see Chapter 3) has led to the development of specific assessment techniques that are designed to identify the patient's susceptibility to relapse and to minimize the effects of brief lapses (slips) (Annie and Davis, 1987~. One of the most promising areas of research related to the development of new assessment procedures is the identification of alcoholic subtypes. For example, there is some evidence to suggest that psychopathology and social instability are associated with poor prognosis, and these types; of patients may warrant more individualized treatment (Rounsaville et al., 1987~. In addition to structured diagnostic interviews (which allow reliable classification of psychopathology), new approaches to subtyping by personality variables and by other characteristics are being proposed (Cloninger, 1987~. Many of these advances have been facilitated by the use of standardized assessment procedures that were developed within other disciplines to measure behavior and symptoms often seen in alcoholics. These techniques include procedures to measure cognitive functioning, affective states, anxietr, depression, social skills, and drug use. Although these procedures appear to be suitable for use with alcoholics, there is also a need to develop measurement techniques that have been designed specifically for alcohol problems and that have been validated by using appropriate patient samples. A wide range of methods--including interviews, questionnaires, biological tests, performance tests, and clinical examination procedures--has been employed to gather clinical data on substance-use disorders. Clinical instruments vary along several dimensions (e.g., who gathers the information, how detailed it must be), but most interviews rely on the respondent as the main or sole informant. Self-administered instruments are also used; the assumption behind them is that if they are worded carefully enough and the amount of information needed is not lengthy, it is possible to bypass an interviewer altogether. Some commonly used questionnaires include the Michigan Alcoholism Screening Test (MAST) (Hedlund and Vieweg, 1984), the Alcohol Expectancy Questionnaire (Brown et al., 1980), the Alcohol Use Inventory (Wanberg, Horn, and Fisher, 1977), and the Alcohol Dependence Scale (Skinner and Horn, 1984~. In summary, there has been a burgeoning interest in the past decade in the development of questionnaires, interviews, performance tests, personality inventories, and biological tests that have vastly improved the scientist's ability to quantify virtually all of the relevant aspects of alcoholism as a multidimensional clinical disorder. Research instruments such as the MAST, the ASI, and the DIS have gained widespread acceptance among researchers and clinicians (Lettieri, Nelson, and Sayers, 1985~. All of these instruments generate reliable, standardized information that could be entered easily into a patient's hospital record to become part of the reporting statistics available to researchers, funding agencies, and health officials. Such data would be extremely useful as a basis for comparing programs and settings in terms of client characteristics and treatment effectiveness, in turn suggesting hypotheses for later clinical trials. -152

OCR for page 149
The following research initiatives offer the opportunity to build on recent advances in the technology of differential assessment: Appropriate emphasis should be placed on establishing the reliability and validity of various assessment procedures. Clients should be evaluated by using standardized, validated, reliable instruments. Efforts should be made to evaluate the clinical relevance of these instruments for program planning, patient placement, treatment matching, and outcome prediction. Emphasis should be given to measures of treatment readiness, motivation for treatment, alcohol- and drug-use patterns, seventr of dependence, alcohol-related disabilities, type and severity of psychopathology, social support network, and coping skills. In those instances in which instruments have been developed for use with nonalcoholic populations, efforts should be made to validate these measures on alcoholic patient samples. ~ The identification of psychiatric comorbidity as a common feature of the clinical picture of alcoholism implies that future research should investigate (a) the relative importance of differentiating alcoholics according to major comorbidities (e.g., antisocial personality, depression, anxiety disorder); and (b) treatment-matching hypotheses that combine new psychiatric assessment procedures with diagnosis-related therapeutic interventions. Given the importance of psychiatric interviewing for the development of accurate diagnostic information, the following questions should be pursued. What is the impact of varying degrees of structure in the interview instrument on the quality of clinical data obtained? How do interviews with and without cutoffs compare? How accurate are informants' reports? How much detail can respondents accurately provide about their disorders? Are inaccuracies biased? How do a patient's mental state and level of intoxication or withdrawal affect reporting? Research on the Reliability and Validity of Assessments Because patients in treatment for alcohol-related problems have often been found to be defensive, uncooperative, sociopathic, or cognitively impaired, there is reason to be skeptical about the accuracy of the information they provide about themselves for research and clinical purposes. This skepticism has generated a substantial amount of methodological research on the validity of verbal self-report methods; it has also stimulated the search for biological markers and other objective measures of alcohol consumption that are not as susceptible to response distortion (Babor, Stephens, and Marlatt, 1987; O'Farrell and Maisto, 1987~. Given the importance of accurate information about the patient's presenting symptoms, drinking history, and treatment outcome, research on the reliability and validity of assessment procedures should receive high priority. Research on Verbal Report Methods Verbal report techniques (e.g., interviews, tests, questionnaires) have become the most common method of obtaining clinical data on alcohol abuse and its modification through treatment. Recent research has raised questions about the extent to which this method provides reliable and valid information for research purposes. A review of methodological studies in the alcohol literature (Babor, Stephens, and Marlatt, 1987) shows that, although information obtained from alcoholics tends to be generally reliable and valid, there can be considerable variability in accuracy depending on the sensitivity of the information sought (e.g., arrest records), the specificity of the validation criteria (e.g., archival records, - 153

OCR for page 149
breathalyser readings, informants' reports), the personal characteristics of the respondents (e.g., sober versus intoxicated), the time window of the report (e.g., recent, lifetime), and the demand characteristics of the task situation (e.g., intake interview, research evaluation at follow-up). Verbal data obtained from alcoholic respondents are inherently neither valid nor invalid; they vary with the methodological sophistication of the data gatherer and the personal characteristics of the respondent. These facts suggest that an emphasis on demonstrating whether verbal report data are either valid or invalid is misplaced. A more important issue is how to improve the validity of these procedures. The variety of verbal report procedures used to assess drinking behavior and alcohol-related problems makes it difficult to evaluate and compare results across studies. Until recently, little attention was focused on the systematic evaluation of measurement procedures, assessment contexts, and response sets. Alcohol research would benefit from a new generation of methodological studies directed at procedures to enhance the validity of verbal report data and at evaluating new procedures that provide objective indicators~of recent alcohol consumption. The generic term for factors that might influence verbal report information in a research context is response effects. Although there has been some general research in this area (Bradburn, 1983), there has been little research focused specifically on substance use or on alcohol-related problems. Such studies are important in the alcohol field because research findings largely depend on the accurate reporting of alcohol problems by study participants. There are several opportunities for further research in this area. One is the impact of other types of response effects (e.g., accuracy of recall) on the precision of self-reports. There is evidence that the recall of health-related occurrences is inconsistent (Aneshensel et al., 1987) and that, even for such seemingly salient events as hospitalization, there may be underreporting after only one year (Cannel!, 1977~. Other research questions relate to the costs and sacrifices of various methods of data collection. Pertinent issues include the practical limits on the amount of information that can be obtained with self-report questionnaires and the use of telephone interviews to obtain diagnostic and other clinical data. The following opportunities may serve as useful guidelines for a research agenda on verbal report methods: More studies of assessment contexts and instructional sets are needed. For instance, preliminary results suggest that individual assessments yield more valid results than group assessments (Sobell and Sobell, 1981~; that telephone surveying and interviewing may be a reasonably reliable means of obtaining information on drug use (Frank, 1985~; that adolescents will give more valid self-reports in private settings (Gfroerer, 1985~; and that the use of computerized assessments may yield more valid data than other techniques (Skinner and Allen, 1983~. Individual difference factors (e.g., social desirability response sets) should be investigated because of their potential influence on validity. Until it can be demonstrated that social desirability scales (e.g., Crowne and Marlowe, 1960) actually measure response bias on alcohol-related assessments, however, their use should not be incorporated automatically into assessment batteries (cf. Bradburn, 1983~. Some invalid responses may be related to forgetting. In addition, there is a strong relationship between mental status and the validity of self-reports (Miller and Barasch, 1985), suggesting the need for careful assessment of organic impairment in respondents. -154

OCR for page 149
. Recently developed time-line follow-back (Sobell et al., 1986) and retrospective diary procedures should be evaluated to assess their ability to improve the reliability and validity of verbal reports by providing recall cues to respondents. Because no single measure of alcohol consumption is entirely valid, most researchers recommend the use of convergent lines of evidence to establish drinking status or to measure drinking behavior. Such analytic procedures as the multitrait-multimethod assessment matrix should be used to evaluate the discriminant and convergent validity of verbal report data. Different verbal report strategies for quantifying consumption (e.g., the quantity-frequency question, the diary method/time-line follow-back procedure) should be compared. The use of multiple biological indicators may have promise as a way of corroborating verbal reports, provided sensitivity problems can be overcome (Cushman et al., 1984~. Use of Biological Markers The presence of alcohol in blood, urine, sweat, or breath is strong evidence that drinking has occurred. Because the duration of abstinence and the amount of drinking following relapse are important criteria for treatment outcome (Babor, Cooney, and Lauerman, 1987), there is great interest in developing laboratory tests (markers) that detect alcohol consumption. An ideal marker would be one that reflected the mean blood alcohol level over several weeks, that did not depend on the presence of organ damage, and that returned to normal relatively quickly with abstinence (O'Farrell and Maisto, 1987~. The major problem with blood, breath, and urine alcohol tests is the relatively short half-life of alcohol after it is ingested. Alcohol is usually not present in any of these fluids hours after ingestion. Thus, a positive alcohol test indicates that drinking occurred that day, but it is not an indication of chronic alcohol use. In this regard, the results of a study by Orrego and colleagues (1979) are important for those interested in alcoholism treatment evaluation. The study was carried out as part of a clinical trial evaluating propylthiouracil treatment of alcoholic hepatitis, in which daily urine specimens were collected and analyzed for ethanol. . ~ ~ , Of the patients in this study who consistently denied drinking during a six-month period, 25 percent had positive urine specimens. Another 25 percent made claims of abstinence for one week that were refuted by urine tests during the six months. Because blood, breath, and urine alcohol tests are indicators of recent alcohol use only, one focus of current research is the development of methods to measure alcohol use over longer periods. Phillips (1984) studied a sweat patch that is applied to the skin and cannot be removed without leaving a mark. The patch collects sweat for seven days, which is then analyzed for ethanol. Results showed that 50 percent (11 of 22) of the subjects who returned for follow-up underreported the amount they drank. However, half of the normal volunteers failed to return to have their patches removed; consequently, lack of patient cooperation may present a problem with this method. Liver enzymes rise with excessive alcohol ingestion and therefore are useful in detecting drinking. Gamma glutamyl transpeptidase (GOT) increases with prolonged heavy drinking to a greater extent than other liver enzymes (Straw and Lieber, 1980~. However, GGT is not a specific marker, because liver injuries other than alcohol abuse also result in increased serum concentrations. Conceivably, isoenyme profiling techniques may make it possible to distinguish the GGT elevations caused by drinking from those associated with other factors. Recently, Irwin and coworkers (1988) reported that the parallel combination of the percentage of increase in GGT, aspartate aminotransferase (SGOT), and alanine -155

OCR for page 149
aminotransferase (SGPT) over posttreatment discharge values effectively distinguished recovering alcoholics who remained abstinent for three months from those who resumed drinking. Another marker involves the mean corpuscular volume (MCV) of red blood cells, which increases with heavy alcohol consumption independent of liver damage (Gheno, Magnabosco, and Mazzei, 1979~. Following the cessation of drinking, MCV returns very slowly to normal. Because an alcoholic may have been abstinent for a long period and still have an elevated MCV (Straw et al., 1979), the usefulness of MCV as a marker for relapse is limited. Laboraeor~y tests for such factors as GOT level and MCV reflect the consequences of alcoholism (i.e., tissue injury) rather than alcohol abuse. The tests of people who are abusing alcohol but have not yet developed organ damage will be normal, a consequence that explains the poor sensitivity of these tests. Other illnesses win also result in test abnormalities. Used by themselves, then, these tests do not have great sensitivity or specificity; however, the combinations used by Cushman and colleagues (1984) and Ryback and coworkers (1982) improve diagnostic accuracy. High-density lipoprotein cholesterol (HDLC) levels also increase during drinking episodes and return to control levels within one to two weeks after the cessation of drinking (Devenyi et al., 1981~. However, HDLC may not be suitable for monitoring drinking in patients with significant liver disease (cirrhosis or alcoholic hepatitis). Haskell, Camargo, and Williams (1984) reported that HDLC also falls significantly when moderate drinkers abstain and rises with the resumption of moderate drinking. Puddy and colleagues (1986) found similar results and, in addition, reported that apoliprotein A-I and A-II declined when moderate drinkers reduced their drinking from five drinks to one per day. Discriminant function analysis showed that a change in the serum apoliprotein A-II was the best of the markers compared by this group (apoliprotein A-I, HDLC, GOT, and MCV) for correctly classifying (with 96 percent accuracy) the category of drinking either five drinks or one. Platelet monoamine oxidase has been reported to be lower in alcoholics (Wiberg, Gottfries, and Oreland, 1977; Major and Murphy, 1978; Alexopoulos et al., 1981), but this finding was not replicated in a more recent study (Tabakoff et al., 1988~. The Tabakoff team also found that platelet adenylate cyclase activity, after stimulation with quanine nucleotide, cesium fluoride, or prostaglandin E, was significantly lower in alcoholics, although the basal platelet adenylate cyclase activity was the same in alcoholics and controls. Because the alcoholics in this study had been abstinent for one to four years, the utilizer of these markers for treatment outcome evaluation is uncertain. There is currently considerable interest in two markers: (1) the hemoglobin acetaldahyde adduct (A1) factor and (2) deglycosylated transferrin. In one study (Hoberman and Chiodo, 1982), hemoglobin At was significantly elevated in alcoholic patients with and without liver disease compared with normal subjects, but there was some overlap between the two groups and levels returned to normal with abstinence. The duration of abstinence required for hemoglobin At to return to normal needs to be defined. Stibler, Borg, and Joustra (1986) have shown that ~carbohydrate-deficient" transferrin is a specific marker for chronic drinking with a biologic half-life of approximately 17 days after cessation of alcohol ingestion. Behrens and coworkers (1988) have extended these observations on the specificity of "carbohydrate-deficient~ transferrin from Swedish patients to different racial groups in the United States. However, the sensitivity of this marker may be a problem. The Takase research team (1985) found serum desialotransferrin in only 60 percent of Japanese alcoholics. -156

OCR for page 149
At this time, with the exception of frequent alcohol tests, biological markers either are not sensitive enough or need further evaluation before they can be used alone to monitor treatment outcome. Markers can be of value as part of an assessment battery that includes self-reports, reports by collateral informants, and official records. The following opportunities for research on biological markers of treatment outcome should be pursued to improve the validity of treatment evaluation research: Researchers should develop and test markers that are sensitive to moderate and heavy alcohol consumption. Markers that return to normal with the cessation of drinking and show prompt elevation at the initiation of drinking would be particularly useful in treatment outcome studies. Another type of marker that should be developed and tested is one that is sensitive to cumulative ethanol intake but independent of alcohol-related tissue damage. Automation of Data Collection In recent years there has been an increasing trend toward the computerization of diagnostic interviews; that is, rather than using a paper-and-pencil instrument, software programs are written so that interview questions appear on a computer screen. The questions are read to the subject (alternatively, the subject may sit at the terminal and read them), and the response is keyed in, usually with a single number according to a prearranged code. There are numerous advantages to this technology, and it is likely to become increasingly popular. Additional advantages and savings are realized if computerized interviews can be self-administered by research subjects. These advantages go beyond economy. For example, there is evidence that a computerized, self-administered examination involving interaction with a machine rather than another human may foster a greater degree of respondent candor (Lucas et al., 1977; Duffy and Waterton, 1984~. Another possible advantage of the computer in clinical settings--perhaps one of the greatest potential advantages of automation--is in automating the collection of routine data as a research data base for ~quasi-experimental" or even true experimental designs. There have been some attempts to form an accumulating data base out of routine examinations, but typically these have depended on the willingness of busy personnel to transfer their clinical observations to computer-readable fonns, a labor-intensive rather than a labor-saving effort. It has been demonstrated that paper-and-pencil structured interviews can be used successfully as a means of routine data collection in clinical settings and can even be acceptable to clinicians in such a context (Helzer et al., 1981~. However, computerization does not merely make interviews and questionnaires more accessible and efficient by having structured interviews available for computer administration by lay or clinical interviewers. It also provides an electronic data base without additional labor. If patients can self-administer the interviews, an accumulating electronic data base would be provided at considerable savings of effort. There are a number of opportunities for research on data collection. Answers to the following questions would be valuable in deciding whether to pursue a policy of increased automation of diagnostic interview data: -157

OCR for page 149
What degree of clinical detail is ascertainable- by using a self-administered, computerized format? Are subjects more or less likely to remember past symptoms or to report sensitive data when queried by a computer? How well do structured interviews perform on a computer administration format? Given the opportunity, would clinicians and investigators find self-administered interview data useful? ASSESSMENT OF TREATMENT PROCESS, QUALITY, AND OUTCOME Advances in assessment techniques have contributed to a more accurate estimation of the relative contributions of client characteristics, therapeutic interventions, program settings, and environmental variables to the success or failure of treatment. Treatment Variables The relevance of treatment research to policy and clinical practice is directly related to the ability of such research to evaluate the costs and effectiveness of different treatment interventions as they relate to specific types of individuals. The complexity of the problem of matching individuals to treatments is related to the complexity of the treatment system in which problem drinkers are typically processed. This system consists of a complex array of treatment settings (e.g., inpatient, outpatient, residential), therapeutic approaches . . ~ . . (psychotherapy, AA, peer approaches), and treatment providers (e.g., psychotherapists, recovering alcoholics). Often, clients are exposed to several different treatment providers and therapeutic approaches in the same setting, after which they move through a series of settings during the course of different stages of treatment and recovery. At the program level of analysis where setting, staff, and therapeutic approaches interact, the mix of experiences may overwhelm the individual contributions of specific treatment variables or components. Specification of treatment is also needed in terms of the duration and quality of exposure. Anta-buse patients may have many treatment contacts, but the psychological impact of the contacts may be minimal. Alcoholics Anonymous (AA), on the other hand, may exert a profound effect on attitudes and behavior in a short time as a function of very intensive involvement of the alcoholic in the fellowship's group process. Recent developments in treatment specification are likely to affect the design of new procedures for the measurement of treatment process and quality (McCrady and Sher, 1985~. A number of instruments have been developed and validated for the purpose of studying characteristics of the treatment setting and process (Lettieri, Nelson, and Sayers, 1985~. Aspects of treatment specification include the intensity and duration of services received, the type of therapeutic modality the patient is given, the personality and skills of the therapist, the emergence of a "therapeutic alliances or partnership between patient and therapist, the characteristics of the setting, the competency of the program staff, the degree to which the treatment is delivered as ideally intended (its integrity), and the general atmosphere of the program as experienced by patients. These instruments include rating scales that are designed for trained observers to use to identify unique features of individual and group therapy, client rating forms that are used to assess the qualitative aspects of the services received, and facility inventories that provide a quantitative assessment of the structural properties of a program. -158

OCR for page 149
Outcome Variables Behavioral and social researchers (McLellan et al., 1983; Pattison, 1985) have argued that traditional outcome assessment methods are too global and imprecise to provide an adequate evaluation of client-treatment matches. This criticism is based on evidence that therapeutic change is multidimensional and that affecting different dimensions requires different types of treatment. There is a growing consensus (Emrick and Hansen, 1983) that, at a minimum, treatment outcome should be evaluated according to the following dimensions: drug and alcohol use, vocational adjustment, psychological problem severity, interpersonal relations, and criminal behavior. Another important trend in the specification of outcome variables has been the growing attention to process measures (Moos and Finney, 1985~. Process analysis focuses attention on the causal linkages between treatment components and dimensions of outcome. In general, little research has been devoted to the short-term impact of specific program components such as alcohol education, AA groups, and individual counseling. Short-term gains, such as the information acquired regarding the negative consequences of drinking, changes in attitude toward alcohol, and improvements in psychological adjustment, can be assessed quite readily by means of brief tests and rating scales at the termination of inpatient treatment or immediately following a prescribed period of outpatient treatment. At present, there is a strong need to apply these newly developed techniques of specifying treatment quality, process, and outcomes to identification of the active ingredients of traditional and experimental treatment interventions. lithe following areas should be explored in research on treatment process and outcome variables: What are the active ingredients of therapeutic interventions, including AA, cognitive-behavioral therapy, and so on? Do the various standard components of currently available multimodal rehabilitation programs (e.g., patient education, Twelve Steps groups) produce short-term changes in the patient's attitudes, motivation, and skills that predict long-term remission? Markers of compliance with treatment should be developed and tested to facilitate the evaluation of drug therapies. For example, riboflavin markers that can be placed in placebo and active medication can be used to detect compliance with drug therapy. Carbon disulfide breath analysis has a potential for evaluating compliance with disulfiram therapy. GROWING SOPHISTICATION IN METHODOLOGICAL APPROACHES In recent years, a variety of research approaches have been developed for use in treatment evaluation. As portrayed in Table 8-1, a basic distinction is often made between two types of research strategies: true experimental designs and quasi-experimental designs (Campbell and Stanley, 1963~. -159

OCR for page 149
Table 8-1. Methodological Approaches to Alcoholism Treatment Evaluation Research Research Design Populations Sampled From Descriptive/Quasi -Experimental Experimental One Facility Several Facilities All Facilities Before-After/ Outcome prediction studies Multisite, multiprogram random sample studies Multisite, multiprogram nonrandom sample studies Aggregate-level descriptive studies Preclinical studies Small-scale random assignment studies Multicenter clinical trials Experimental Approaches True experimental designs attempt to rule out all or most of the obvious confounding variables by randomly assigning subjects to experimental and control treatments. Preclinical studies are usually performed on small groups of patients (or animals) to establish a causal linkage between some treatment procedure (e.g., emetine) and a measure that should logically be associated with treatment outcome (e.g., conditioned aversion to alcohol). Small-scale random assignment studies operate at a broader level of analysis, by taking defined treatment modalities or programs and comparing them to no treatment, minimal treatment, or alternative treatments. Multicenter clinical trials represent the culmination of preclinical and small-scale studies. Quasi-Experimental Approaches In the alcohol field, true experimental designs, especially multicenter clinical trials, are difficult and expensive to carter out. This may explain why there have been numerous quasi-experimental studies in which researchers have explored treatment questions by weaving a net of circumstantial evidence around the observed associations between client characteristics and differential treatment response. The quasi-experiment is a design in which a variety of strategies other than random assignment are used to minimize the -160

OCR for page 149
influence of confounding variables. One approach has been to gather treatment and outcome data from a large number of patients who are being treated at different facilities or who have been exposed to different therapeutic modalities (multisite, multiprogram studies). Typically, the patients are not randomly assigned to these conditions, and the research is often designed to answer questions other than those dealing with treatment matching (e.g., Cronkite and Moos, 1980~. To control for the influence of major confounding variables, multivariate statistical procedures such as covariance analyses and partial correlation are used. Although the literature is not lacking in multisite, multiprogram prediction studies (Armor, Polich, and Stambul, 1978; Brown and Lyons, 1981; McLellan et al., 1983), it is important to note the limitations of previous research. First, most prediction studies have been limited to small samples drawn from a small number of unrepresentative treatment settings. Second, even when researchers have sampled patients across multiple settings, only a few studies have considered the characteristics of the setting as well as the quality and types of treatment received; even fewer have conducted extensive evaluations of patient characteristics. Primarily because of logistical problems and the costs of conducting this type of research, there has been an inverse relationship between the number of patients studied and the amount of data collected. Aggregate-Level Descriptive Studies In recent years, several attempts have been made to conduct general facility surveys as a means of generating descriptive statistics about the national configuration of alcoholism treatment services (Yahr, 1988~. These inventories include the National Drug and Alcoholism Treatment Utilization Survey (NDATUS), the American Hospital Association Annual Survey, and surveys conducted by the National Association of State Alcohol and Drug Abuse Directors and the National Association of Alcoholism Treatment Programs. The most comprehensive aggregate-level descriptive survey is the NDATUS, which has been conducted periodically by NIAAA since 1979. Survey forms for the NDATUS were distributed to all known treatment units in the United States in 1979, 1980, and 1982. A shorter suney, the National Alcoholism and Drug Abuse Program Inventory (NDAPI), was conducted in 1984 to define the universe of facilities that could later be used for more specialized sampling studies. These surveys have been designed to serve different functions, including resource planning and management on the state and federal levels; the generation of cross-sectional and longitudinal data on treatment utilization, staffing, funding, and length of stay; and the description of characteristics of programs, clients, and ownership of the facilities. In the case of the NDATUS, budgetary considerations have led to a drastic reduction in the amount of data collected; nevertheless, the abbreviated version (the NDAPI) sets the stage for more detailed surveys using random or stratified sampling techniques. The potential for weeding the more intensive, client-focused, multiprogram study designs to large-scale, representative sample surveys could constitute a significant advance in the treatment evaluation knowledge base. At present, the NDATUS is a brief questionnaire that requires the reporting unit to give its primary orientation (alcoholism, drug abuse, mixed), physical environment or setting (e.g., hospital, freestanding unit), population served (e.g., inner city, suburban), kinds of services provided (e.g., group therapy, education, aftercare), ownership characteristics, staffing patterns, and client census capacity. In addition, current patients are recorded according to race/ethnicity, age, and sex. Descriptive data obtained from the NDATUS - 161

OCR for page 149
applications have been extremely useful in describing characteristics of different treatment systems and identifying changes and trends over time (Yahr, 1988~. Although the NDATUS is lacking in most of the client and treatment variables now considered crucial to a proper understanding of treatment efficacy and efficiency, these data could be collected simultaneously from staff, administrators, and clients to provide a more comprehensive description of the services received by different types of clients. The final ingredient of such an approach is the measurement of short-term and long-term treatment gains by using periodic interviews with selected samples of patients at various intervals during and after treatment. Another approach, exemplified in the Epidemiologic Catchment Area Study (Helzer and Ptyzbeck, 1988), is to gather data about drinking behavior and treatment use from general published samples and compare these data with those collected from representative treatment settings. As shown in Table 8-1, the research findings emanating from preclinical, experimental, quasi-experimental, and descriptive research approaches should provide a comprehensive and complementary understanding of the many facets of alcohol problem treatment. Eventually, the active components of the most effective treatment should be identified, leading to a more systematic program of clinical trials that will provide a more definitive test of treatment efficacy, especially in relation to different client types, than has heretofore been possible (Finney and Moos, 1986~. Many opportunities exist for research employing complementary methodologic approaches. The considerations noted above suggest the following programmatic recommendations for funding agencies, research centers, and investigators: The new generation of assessment technologies should be employed to evaluate treatment efficacy by means of innovative research methods. Experimental, quasi-experimental, and naturalistic research designs each have their place in treatment evaluation. Descriptive studies of the treatment system at the aggregate level of analysis can provide valuable information about changes in the client population, variations in substance-use patterns, trends in program characteristics, and new approaches to treatment. Health services research should be encouraged as the first step toward an integrated program of descriptive and experimental research that will eventually lead to multicenter clinical trials of promising treatment interventions. Experimental research designs should be encouraged for alcoholism treatment evaluations whenever appropriate. When experimental designs are employed, there must be adequate specification of the client, program setting, therapist, and treatment variables that affect patient compliance and outcome. When experimental designs are not feasible because of ethical or practical constraints, the new generation of asessment technology should be employed to evaluate treatment efficacy by means of innovative, nonexperimental research methods. At times, program evaluation data have been used to support claims that one type of treatment is superior to a lack of treatment or to some comparison treatment. Success rates of 80 to 90 percent have been used in media advertisements to solicit patients to proprietary treatment programs. Outcome monitoring studies that do not meet acceptable criteria for making causal inferences have been used to justify requests for insurance reimbursement of unevaluated treatments. At present, there is a significant risk of -162

OCR for page 149
misinterpretation in the use of unsound research designs in studies sponsored by individuals or corporations that have a financial interest in the results. Whenever possible, evaluation research should be designed and conducted by investigators who are not likely to benefit from the result either materially or professionally. REFERENCES Alexopoulos, G. S., K Lieberman, I. Frances, and P. Stokes. Platelet MAO during the alcoholwithdrawalsyndrome. Am.J.Psych.136:1245-1250,1981. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders (DSM-IlI), ad ed. Washington, DC: APA, 1980. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders (DSM-III), ad ea., rev. Washington, DC: APA, 1987. Aneshensel, C. S., ~ L. Estrada, M. J. H. Hansen et al. Social psychological aspectsof reporting behavior: Lifetime depressive episode reports. J. Health Soc. Behav. 28:232-246; 1987. Annis, H. M., and C. S. Davis. Assessment of expectancies in alcohol dependent clients. In ~ Marlatt and D. Donovan, eds. Assessment of Addictive Behaviors. New York: Guilford Press, 1987. Armor, D. J., J. M. Polich, and H. B. Stambul. Alcoholism and Treatment. New York: John Wiley and Sons, 1978. Babor, T. F., N. L. Cooney, and R. J. Lauerman. The drug dependence syndrome concept as a psychological theory of relapse behavior: An empirical evaluation. Br. J. Addiction 82:393-405, 1987. Babor, T. F., R. S. Stephens, and G. ~ Marlatt. Verbal report methods in clinical research on alcoholism: Response bias and its minimalization. J. Stud. Alcohol 48~5~:410-424, 1987. Behrens, U. J., T. M. Worner, L. F. Braly et al. Carbohydrate-deficient transferrin, a marker for chronic alcohol consumption in different ethnic populations. Alcoholism: Clin. Exper. Res. 12:427-432, 1988. Bradburn, N. M. Response effects. Pp. 289-328 in P. E. Rossi and J. D. Wright, eds. Handbook of Survey Research. New York: Academic Press, 1983. Brown, J., and J. P. Lyons. A progressive diagnostic schema for alcoholism witevidence of clinical efficacy. Alcoholism: Clinical and Experimental Research 5~1~:17-25, 1981. Brown, S. A, M. S. Goldman, ~ Inn et al. Expectations of reinforcement from alcohol: Their domain and relation to drinking patterns. J. Consult. Clin. Psychol. 48:419-426, 1980. Campbell, D., and J. Stanley. Experimental and Quasi-Experimental Designs for Research. Chicago: Rand McNally, 1963. -163

OCR for page 149
Cannell, C. F. A Summary of Studies of Interviewing Methodology. USDHEW Publ. No. (HRA) 77-1343, Series II, No. 69. Washington, DC: U.S. Department of Health, Education, and Welfare, 1977. Cloninger, C. R. Neurogenetic adaptive mechanisms in alcoholism. Science 236:410-416, 1987. Cornell, W., C. R. Cloninger, and T. Reich. Clinical assessment: Use of nonphysician interviews. J. New. Ment. Dis. 166:599-606, 1978. Cronkite, R. C., and R. H. Moos. Determinants of the posttreatment functioning of alcoholic patients: A conceptual framework. J. Consult. Clin. Psychol. 48:305-316, 1980. Crowne, D. P., and D. Marlowe. A new scale of social desirability independent of psychopathology. J. Consult. Psychol. 24:349-354, 1960. Cushman, P., G. Jacobson, J. J. Barboriak, and ~ J. Anderson. Biochemical markers for alcoholism: Sensitivity problems. Alcohol Clin. Exp. Res. 8:253-257, 1984. Devenyi, P., G. M. Robinson, B. M. Kapur et al. High density lipoprotein cholesterol in male alcoholics with and without severe liver disease. Am. J. Med. 71:589-594, 1981. Dufty, J. C., and J. J. Waterton. Under-reporting of alcohol consumption in sample surveys: The effect of computer interviewing in fieldwork. Br. J. Addict. 79:303-308, 1984. Edwards, G. The alcohol dependence syndrome: A concept as stimulus to enquiry. Br. J. Addict. 81:171-183, 1986. Edwards, G., ~ Arif, and R. Hodgson. Nomenclature and classification of drug and alcohol related problems: A WHO memorandum. Bull. World Health Organization 59~2~:225-242, 1981. Emrick, C. D., and J. Hansen. Assertions regarding effectiveness of treatment for ~alcoholism: Fact or fantasy? Am. Psychol. 38:1078-1088, 1983. Finney, J. W., and R. H. Moos. Matching patients with treatments: Conceptual and methodological issues. J. Stud. Alcohol 47:122-134, 1986. Frank, B. Telephone surveying for drug abuse: Methodological issues and an application. Pp. 71-83 in B. ~ Rouse, N. J. Kozel, and L. G. Richards, eds. Self-Report Methods of Estimating Drug Use: Meeting Current Challenges to Validity. NIDA Research Monograph No. 57. Rockville, MD: National Institute for Drug Abuse, 1985. Gfroerer, J. Influence of privacy on self-reported drug abuse by youths. Pp. 22-30 in B. Rouse, N. J. Kozel, and L. G. Richards, eds. Self-Report Methods of Estimating Drug Use: Meeting Current Challenges to Validity. NIDA Research Monograph No. 57. Rockville, MD: National Institute for Drug Abuse, 1985. Gheno, G., V. Magnabosco, and G. Mazzei cronico. Min. Med. 3:1301-1306, 1979. -164 Macrocitosi ed anemia nell'alcoholismo

OCR for page 149
Haskell, W. C, C. Camargo, and P. T. Williams. The effect of cessation and resumption of moderate alcohol intake on serum high density lipoprotein subfractions: A controlled study. New Engl. J. Med. 310:805~10, 1984. Hedlund, J. L^, and B. W. Vieweg. The Michigan Alcoholism Screening Test (MAST): A comprehensive review. J. Operational Psychiat. lS:55 64, 1984. Helzer, J. E., and T. R. P~yzbeck. The co-occurrence of alcoholism with other psychiatric disorders in the general population and its impact on treatment. J. Stud. Alcohol 49:219-224, 19~. Helzer, I. A, L. N. Robins, J. L. Croughan et al. Renard Diagnostic Interview: Its reliability and procedural validity with physicians and lay interviewers. Arch. Gen. Psych. 38:393-398, 1981. Hoberman, H. D., and S. M. Chiodo. Elevation of the hemoglobin At fraction in alcoholism. Alcoholism: Clin. Exper. Research 6:260 266, 1982. Institute of Medicine, Division of Health Promotion and Disease Prevention. Alcoholism, Alcohol Abuse, and Related Problems: Opportunities for Research. Washington, DC: National Academy Press, 1980. I'win, M., S. Baird, T. Smith, and M. Schuckit. Use of laboratory tests to monitor heavy drinking by alcoholic men discharged from a treatment program. Am. J. Psych. 145:5, 1988. Lettieri, D. J., J. E. Nelson, and M. ~ Sayers, eds. Alcoholism Treatment Assessment Research Instruments. NIAAA Handbook Series, vol. 2. DHHS Publ. No. (ADM) 85-1380. Washington, DC: NIAAA, 1985. Luborsky, C, and ~ T. McLellan. Optimal matching of patients with types of psychotherapy: What is known and some designs for knowing more. Pp. 51-71 in E. Gottheil, ~ T. McLellan, and K ~ Druley, eds. Matching Patient Needs and Treatment Methods in Alcoholism and Drug Abuse. Springfield, IL: Charles C. Thomas, 1981. Lucas, R. W., P. J. Mullin, C. B. Luna et al. Psychiatrists and a computer as interrogators of patients with alcohol-related illnesses: A comparison. Br. J. Psychiatry 131:160-167, 1977. Major, ~ F., and D. Murphy. Platelet and plasma amine omdase activity in alcoholic individuals. Br. J. Psychiatry 132:548-554, 1978. Marlatt, G. A, and W. R. Miller. Comprehensive Drinker Profile. Odessa, FL: Psychological Assessment Resources, 1984. McCrady, B., and K Sher. Treatment variables. Pp.48-62 in B. S. McCrady, N. E. Noel, and T. D. Nirenberg, eds. Future Directions in Alcohol Abuse Treatment Research. Research Monograph No. 15. USDHHS Publ. No. (ADM)85-1322. Washington, DC: U.S. Department of Health and Human Se~vices9 1985. -165

OCR for page 149
McLellan, ~ T., L. Luborsky, F. Erdlen et al. The Addiction Severity Index Pp. 178-194 in E. Gottheil, ~ T. McLellan, and K D. Druley, eds. Substance Abuse and Psychiatric Illness. New York: Pergamon Press, 1980. McLellan, ~ T., L. Luborsky, G. E. Woody et al. Predicting response to alcohol and drug abuse treatments. Arch. Gen. Psychiatry 40:620-625, 1983. Meyer, R. E., T. F. Babor, M. Hesselbrock et al. New directions in the assessment of the alcoholic patient. In ~ H. Towle, ed. Proceedings: NIAAA-WHO Collaborating Center Designation Meeting and Alcohol Research Seminar. DHHS Publ. No. (ADM)85-1370. Washington, DC: U.S. Department of Health and Human Services, 1985. Miller, F., and ~ Barasch. The under-reporting of alcohol use: The role of organic mental syndromes. Drug Alcohol Depend. 15:347-351, 1985. Moos, R. H., and J. W. Finney. New directions in program evaluation: Implications for expanding the role of alcoholism researchers. Pp. 173-203 in B. S. McCrady, N. E. Noel, and T. D. Nirenberg, eds. Future Directions in Alcohol Abuse Treatment Research. Research Monograph No. 15. DHHS Publ. No. (ADM)85-1322. Washington, DC: U.S. Department of Health and Human Services, 1985. O'Farrell, T. J., and S. ~ Maisto. The utility of self-report and biological measures of alcohol consumption in alcoholism treatment outcome studies. Adv. Behav. Res. Ther. 9:91-125, 1987. Orrego, H., L. M. Blendis, J. E. Blake et al. Reliability of assessments of alcohol intake based on personal interviews in a liver clinic. Lancet 2:1354-1356, 1979. Othmer, E., B. J. Powell, and E. C. Penick. Psychiatric Diagnostic Interview (PDI), 9th rev. Kansas City, KS: University of Kansas Medical Center, 1980. Pattison, E. M. New directions in alcoholism treatment goals. Pp.7-26 in B. S. McCrady, N. E. Noel, and T. D. Nirenberg, eds. Future Directions in Alcohol Abuse Treatment Research. Research Monograph No. 15. DHHS Publ. No. (ADM)85-1322. Washington, DC: U.S. Department of Health and Human Services, 1985. Phillips, M. Sweat patch testing detects inaccurate self-reports of alcohol consumption. Alcoholism: Clin. Exper. Research 8:51-53, 1984. Polich, J. M., D. J. Armor, and H. B. Braiker. The Course of Alcoholism: Four Years After Treatment. New York: John Wiley and Sons, 1981. Puddy, I. B., J. Masarei, R. Vandongen, and L. J. Beilin. Serum apoliprotein A-II as a marker of change in alcohol intake in male drinkers. Alcohol and Alcoholism 21:375-383, 1986. Robins, L. N., J. E. Helzer, J. Croughan et al. National Institute of Mental Health Diagnostic Inte~view Schedule: Its history, characteristics, and validity. Arch. Gen. Psych. 38:381-389, 1981. -166

OCR for page 149
Robins, L. N., J. Wing, H. U. Wittchen et al. The Composite International Diagnostic Interview. Arch. Gen. Psych. 45:1069-1077, 1988. Rounsaville, B. J., Z S. Dolinsly, T. F. Babor et al. Psychopathology as a predictor of treatment outcome in alcoholics. Arch. Gen. Psych. 44:505-513, 1987. Ryback, R. S., M. J. Eckhardt, B. Felsher, and R. R. Rawlings. Biochemical and hematological correlates of alcoholism and liver disease. J. Am. Med. Assoc. 248:2261-2265, 1982. Saunders, S., O. G. Aesland, and M. Grant. AUDIT--The World Health Organization screening instrument for harmful and hazardous alcohol consumption. Br. J. Addiction, in press. Shaw, S., and C. S. Lieber. Mechanism of increased gamma glutamyl transpeptidase after chronic alcohol consumption: Hepatic microsomal induction rather than dietary imbalance. Substance and Alcohol Actions/Misuse 1:423-428, 1980. Shaw, S., T. M. Warner, M. F. Borysow et al. Detection of alcoholism relapse: Comparative diagnostic value of MVC, GGTp, and AANB. Alcoholism: Clin. Exper. Research 3:297-301, 1979. Skinner, H. A, and B. ~ Allen. Does the computer make a difference? Computerized versus face-to-face self-report assessments of alcohol, drug and tobacco use. J. Consult. Clin. Psychol. 51:267-275, 1983. Skinner, H. A, and J. L. Horn. Alcohol Dependence Scale (ADS) User's Guide. Toronto: Addiction Research Foundation, 1984. Sobell, L. C., and M. B. Sobell Effects of three interview factors on the validity of alcohol abusers' self-reports. Am. J. Drug Alcohol Abuse 8:225-237, 1981. Sobell, M. B., L. C. Sobell et al. The reliability of timeline method for assessing normal drinker college students' recent drinking history: Utility for alcohol research. Addict. Behav. 11:149-161, 1986. Spitzer, R. L., J. B. W. Williams, and ~ E. Skodol. International Perspectives on DSM-III. Washington, DC: American Psychiatric Press, 1983. Stibler, H., S. Borg, and M. Joustra. Microanion exchange chromatography of "carbohydrate deficient" transferrin (CDT) in serum in relation to alcohol consumption. Alcoholism: Clin. Ex~>er. Res. 10:535-544, 1986. Stockwell, T., D. Murphy, and R. Hodgson. The severity of alcohol dependence questionnaire: Its usefulness, reliability and validity. Br. J. Addict. 78:145-155, 1983. Tabakoff, B., P. L. Hoffman, K W. Lieberman et al. Differences in platelet enyme activity between alcoholics and nonalcoholics. New Engl. J. Med. 318:134-139, 1988. Takase, S., ~ Takada, M. Tsutsumi, and Y. Matsuda. Biochemical markers of chronic alcoholism. Alcohol 2:405-410, 1985. -167

OCR for page 149
VandenBos, G. R. Psychotherapy research: A special issue. Am. Psychol. 41:111-112, 1986. Wanberg, K W., J. ~ Horn, and F. M. Fisher. A differential model of alcoholism: The scales of the Alcohol Use Inventory. J. Stud. Alcohol 38:512-543, 1977. Wiberg, A., C. G. Gottfries, and To Oreland. Low platelet monamine oxidase acting in human alcoholics. Med. Biol. 55:181-186, 1977. World Health Organization. Draft revision of Chapter 5, Categories F~Fg9, Mental, behavioral, and developmental disorders: Clinical descriptions and diagnostic guidelines. In International Classification of Diseases, lOth ed. Geneva: WHO, 1988. Yahr, H. T. A national comparison of public and private sector alcoholism treatment delivery system characteristics. J. Stud. Alcohol 49:233-239, 1988. -168