The National Institute on Drug Abuse Collaborative Cocaine Treatment Study Rationale and Methods Paul Crits-Christoph, PhD; Lynne Siqueland, PhD; Jack Blaine, MD; Arlene Frank, PhD; Lester Luborsky, PhD; Lisa Simon Onken, PhD; Larry Muenz, PhD; Michael E. Thase, MD; Roger D. Weiss, MD; David R. Gastfriend, MD; George Woody, MD; Jacques P. Barber, PhD; Stephen F. Butler, PhD; Dennis Daley, MSW; Sarah Bishop, MA; Lisa M. Najavits, PhD; Judy Lis, MSN; Delinda Mercer, MA; Margaret L. Griffin, PhD; Karla Moras, PhD; Aaron T. Beck, MD National Institute on Drug Abuse Collaborative Cocaine Treatment Study is a large, multisite psychotherapy clinical trial for outpatients who meet the DSM-IV criteria for cocaine dependence. For 480 randomized patients, the outcomes of 4 treatments are compared for an 18-month period. All treatments include group drug counseling. One The also adds cognitive adds individual drug therapy, one adds supportive-expressive psychodynamic therapy, and one counseling; one consists of group drug counseling alone. In addition, 2 specific interaction hypotheses, one involving psychiatric severity and the other involving degree of antisocial personality characteristics, are being tested. This article describes the main aims of the project, the background and rationale for the study design, the rationale for the choice of treatments and patient population, and a brief description of the research plan. Arch Gen Psychiatry. 1997;54:721-726 treatment The National Institute on Drug Abuse (NIDA) Collaborative Cocaine Treat¬ ment Study is a multisite clinical trial in¬ vestigating the efficacy of 4 treatments for outpatients with cocaine depen¬ dence. To our knowledge, it represents the largest clinical trial conducted to date on psychotherapy and counseling for cocaine dependence. This article de¬ scribes the rationale for the study design, which was developed in response to 2 NIDA Requests for Application for a Co¬ operative Agreement research pro¬ gram.1·2 The first phase of the project (conducted from October 1991 to July 1993) involved site selection, protocol development, and staff recruitment. The second phase (conducted from July 1993 to July 1994) involved training the staff to an acceptable level of competence, From the Department of Psychiatry, University of Pennsylvania (Drs Crits-Christoph, Siqueland, Luborsky, Woody, Barber, Moras, and Beck and Ms Mercer), and Philadelphia Veterans Hospital (Dr Woody), Philadelphia, Pa; the National Institute on Drug Abuse, Rockville, Md (Drs Blaine and Onken); Brookside Hospital, Nashua, NH (Drs Frank and Butler and Ms Bishop); Gaithersburg, Md (Dr Muenz); Western Psychiatric Institute and Clinic, University ofPittsburgh, Pittsburgh, Pa (Dr Thase, Mr Daley, and Ms Lis); and Harvard Medical School and McLean Hospital (Drs Weiss, Najavits, and Griffin) or Massachusetts General Hospital (Dr Gastfriend), Boston, Mass. A complete list of the members of the National Institute on Drug Abuse Collaborative Cocaine Treatment Study Group appears on page 722. field testing the assessment procedures and protocol treatments, and obtaining preliminary pilot data relevant to the hy¬ potheses of the study. The main clinical See also page 691 trial began in July 1994, and data collec¬ tion is scheduled to end in January 1998. This article describes the main aims of the study, the background for the project, and a description of the main el¬ ements of the study design. AIMS OF THE NIDA COLLABORATIVE COCAINE TREATMENT STUDY This study investigates the relative effi¬ cacy of 4 manual-guided treatments for outpatients who meet the DSM-IV crite¬ ria for cocaine dependence: (1) group drug counseling (GDC) alone, (2) indi¬ vidual cognitive therapy (CT) plus GDC, (3) individual supportive-expressive (SE) psychodynamic therapy plus GDC, and (4) individual drug counseling (IDC) plus GDC. The following main effects and interaction hypotheses are being tested. Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Pennsylvania User on 09/03/2013 Main Effects One of the primary aims of this study is to compare the short- and long-term efficacy and patient acceptance of CT plus GDC, SE psychodynamic therapy plus GDC, IDC plus GDC, and GDC alone. The comparison of the com¬ bined individual and group treatments with GDC alone addresses the practical question of whether it is sufficient to treat cocaine-dependent patients with GDC alone. The IDC plus GDC treatment also addresses a practical ques¬ tion: if GDC needs to be supplemented with individual treat¬ ment, is a professional psychotherapist needed or are the services of a substance abuse counselor equally effective? The IDC plus GDC treatment is also included to control for the extra time with an individual counselor that those patients in CT and SE psychodynamic therapy will receive. Interaction Hypotheses There are 2 hypotheses: (1) Patients with higher levels of psychiatric severity (concurrent psychiatric symp¬ toms) will show a better response to the psychotherapies (CT plus GDC and SE psychodynamic therapy plus GDC) compared with the drug counseling treatments (IDC plus GDC and GDC alone). (2) Patients with more antisocial personality characteristics will have a better re¬ sponse to CT plus GDC compared with SE psychody¬ namic therapy plus GDC, but patients without antiso¬ cial personality characteristics will have a better response SE psychodynamic CT plus GDC. to therapy plus GDC compared with BACKGROUND Developing and testing treatments for cocaine use disor¬ ders is a high national priority. The National Institute on Drug Abuse Household Survey 19943 indicates that the prevalence of frequent cocaine use has remained at 0.3% of the popu¬ lation since 1985. The effects of cocaine extend beyond the individual user and are wide ranging, especially in the ar¬ eas of drug-related violence and risk behaviors associated with the human immunodeficiency virus. The development of successful treatments for cocaine use disorders, there¬ fore, has the potential for a notable effect on the health of cocaine users; the larger social and medical problems as¬ sociated with its use could also be affected. Given the mag¬ nitude ofthe problem, the NIDA determined that a controlled clinical trial of the most promising psychotherapy and coun¬ seling approaches for cocaine dependence was warranted. Selection of the treatments for study was based on practi¬ cal considerations (eg, clinical relevance) and prior empiri¬ cal work, particularly the studies by Woody et al4·3 on psy¬ chotherapy and counseling approaches for opiate depen¬ dence. The treatments chosen for this study included IDC and GDC, SE psychodynamic therapy, and CT. Some form of IDC or GDC by a nondoctoral-level provider is the standard treatment offered to substance abusers in most clinical settings.6 Psychodynamically ori- Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Pennsylvania User on 09/03/2013 ented psychotherapy continues to be the most widely used professional treatment.7 In preparation for this project, NIDA staff conducted an informal survey of clinicians who work with substance abusers. This survey sug¬ gested that psychodynamic therapy was also frequently used in the community for the treatment of these pa¬ tients. Cognitive-behavioral therapies are also used com¬ monly in the community, with cognitive-behavioral re¬ lapse prevention strategies increasingly incorporated into the treatment of substance abusers in particular.8 Cog¬ nitive-behavioral and psychodynamic therapies not only represent 2 approaches that are commonly used but also serve as contrasting theoretical perspectives because of their distinctiveness from each other. It is not at all clear if these treatments are efficacious for cocaine dependence. In recent years, the results of sev¬ eral controlled trials testing these treatments, alone or in combination with psychopharmacological treatments for cocaine dependence, have been published; this literature has been reviewed elsewhere.gl0 Several studies have sup¬ ported the efficacy of cognitive-behavioral relapse preven¬ tion treatment for cocaine use disorders,11·12 particularly for delayed effects emerging at follow-up.13 Promising results using a behavioral therapy (contingency contracting and community reinforcement) have also been reported.M Less promising evidence concerning the use of interpersonal psy¬ chotherapy or psychodynamic therapy has been reported.15·16 The Role of Psychiatric Severity Rather than focus solely on main effect comparisons be¬ tween treatments in clinical trials, several researchers have suggested that it is more relevant to clinical practice to examine interaction effects, with an eye toward match¬ ing different treatment modalities to patient character¬ istics.17"19 One known modifier of response to psycho¬ therapy is level of psychiatric severity.20 Substancedependent patients with notable psychiatric symptoms may need a treatment that addresses these ongoing psy¬ chiatric problems that might be fueling the addictive pro¬ cess or making the substance use disorder more diffi¬ cult to treat. In contrast, substance-abusing patients with low levels of psychiatric symptomatology might find more intensive psychotherapeutic treatments to be less rel¬ evant. For them, standard drug counseling may suffice. The available literature tends to support these views. In an examination of outcomes in 6 rehabilitation pro¬ grams, McLellan et al21 found that patients with low lev¬ els of psychiatric severity improved in all treatment pro¬ grams, whereas patients with high levels of psychiatric severity showed virtually no improvement. Rounsaville et al22 found that higher psychiatric severity was associ¬ ated with poorer outcome in current functioning and psy¬ chosocial adjustment but not substance use outcome in a 2.5-year follow-up of persons addicted to opiates. Most relevant to this study, Woody et al23 found that opiatedependent patients with a high psychiatric severity achieved better psychiatric and drug use outcomes when they received psychotherapy in addition to IDC than when they received only the latter. With the use of an alcoholdependent sample, Kadden et al24 and Cooney et al25 found that patients with a higher psychiatric severity did bet- coping skills training group therapy than in inter¬ actional group therapy, but both treatments were equally effective for patients with low psychiatric severity. The previously described studies indicate consis¬ tently that patients with higher levels of psychiatric se¬ verity show greater improvement when their treatment includes professional psychotherapy. However, the ben¬ efits of this interactional model of treatment matching have not been studied prospectively in a cohort of cocaineter in dependent patients. The Role of Antisocial Personality Disorder Beutler18 reported that, in 8 of 9 studies, patients low on a trait measure of antisocial personality disorder did better in verbal psychotherapy compared with behavioral therapy. Several more recent studies found that patients with higher levels of traits associated with antisocial personality disor¬ der fared better in more directive treatments.24"27 The likelihood that only a few cocaine abusers are likely to manifest few traits associated with antisocial personal¬ ity disorder might explain the poor results from interper¬ sonal and psychodynamic treatments in controlled stud¬ ies. These findings serve as the basis for the hypothesis of the current project: patients who are more impulsive and who have more characteristics of antisocial personality dis¬ order will have better outcomes in cognitive-behavioral therapy, whereas less impulsive patients will have better outcomes in psychodynamic therapy. RESEARCH PLAN Patient Selection The selection criteria were designed to identify a represen¬ tative group of primary cocaine-dependent patients who would be suitable for outpatient treatment and whose per¬ sonal characteristics and life circumstances would not in¬ terfere with study participation. Patients aged 18 to 60 years are included if they have a principal current diagnosis of co¬ caine dependence or cocaine dependence in early partial re¬ mission (as determined by DSM-IV criteria). The principal diagnosis is established using a 0 to 8 severity rating scale adapted from the Anxiety Disorders Interview Schedule-Revised28 that reflects the diagnostician's evaluation of the subjective distress and functional impairment associated with the di¬ agnosis. In addition, patients must have used cocaine at least once in the 30 days before enrollment, have a current postal address and plan to live in the area for the next 2 years, be able to provide the name of at least 1 person who can gen¬ erally locate their whereabouts, and be able to understand and complete the assessment measures. Explicit written, in¬ formed consent was obtained from all patients. Patients are excluded if (1) they have a principal di¬ agnosis of alcohol or polysubstance abuse or dependence; (2) they have a diagnosis of current opioid dependence or opioid dependence in early partial remission (sustained full remission is acceptable); (3) there is evidence of dementia or another irreversible organic brain syndrome; (4) they have a psychotic disorder; (5) they have any history of or current suicide bipolar disorder; (6) they are at current imminent or homicide risk, require hospitalization or resi- Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Pennsylvania User on 09/03/2013 dential treatment, or are in a halfway house; (7) they are unwilling to discontinue a current psychotherapeutic treat¬ (8) they must continue to receive a psychotropic medication; (9) they have any life-threatening or unstable ment; medical illness or a medical illness that can create marked change in mental state; (10) incarceration is impending; (11) they were hospitalized more than 10 days of the past 30 days for treatment of cocaine use; (12) they are cur¬ rently legally mandated for treatment; or (13) they are more than 12 weeks pregnant. chiatric severity, and degree of antisocial personality characteristics. Sex, marital status, employment status, and mode of cocaine use were chosen because these vari¬ ables have been associated with outcome, retention, or both in prior studies6'30 or within our pilot data. Psychi¬ atric severity and degree of antisocial personality disor¬ der were chosen because they were specifically hypoth¬ esized to interact with treatment modality in relation to outcome. The urn randomization was separate for each site to ensure that treatments were balanced on the rel¬ evant factors within each site. Therapist and Drug Counselor Selection and For the chosen treatments to have the best chance pos¬ be provided sible of demonstrating efficacy, they under optimal conditions (ie, by therapists and drug coun¬ selors with specialized expertise and training). Details of therapist and counselor selection and training will be re¬ ported in separate articles. Briefly, all therapists and coun¬ selors were first selected from a pool of applicants. Train¬ ing then consisted of 4 weekend workshops plus supervised experience with 4 patients per therapist or counselor. Ratings of adherence and competence of treat¬ ment sessions by supervisors and independent raters were used to make final decisions about certifying each thera¬ pist or counselor as performing the treatment ad¬ equately enough to participate in the main trial. must Patient Screening and Randomization Procedures Patients are initially screened for study participation via tele¬ phone or on a walk-in basis to determine if the basic inclu¬ sion criteria are met. If the criteria are met, patients return for an enrollment evaluation, which includes a urine drug To be randomized, patients must complete the en¬ rollment evaluation, still meet the inclusion criteria, and complete an orientation phase that includes attendance and assessment requirements. The patient is required to attend 3 visits at the clinic within 14 days, including 1 group ses¬ sion at which another urine drug screen is collected, and 2 case management visits. Group counselors suggest attendance at local self-help groups, such as Cocaine or Al¬ coholics Anonymous; promote human immunodeficien¬ cy virus risk reduction; and address housing,job, or finan¬ cial needs. This orientation phase selects for treatment only those patients with enough motivation to attend at least a few sessions. Although it might ideally be of interest to test treatments on the full range of patients who contact a treat¬ ment program, it seems likely that psychotherapy treatments only have a chance of efficacy with patients sufficiently mo¬ tivated to attend a minimal number of orientation sessions. After satisfactory completion of the postorientation as¬ sessments, patients are randomized to treatment. Random assignment to treatment uses a special technique known as adaptive or "urn" randomization.29 Urn randomization ensures equivalence of the groups on preselected variables that may relate to outcome across treatments or that may interact with type of treatment to affect outcome. For this project, the following factors were in¬ cluded in the adaptive randomization algorithm: sex, mari¬ tal status, employment status, mode of cocaine use, psyscreen. Treatments Training Format. All 3 individual treatments have a 6-month ac¬ tive phase and a 3-month booster phase. The same ses¬ sion schedule is followed in all 3 individual treatments. During the first 3 months, individual sessions are held twice per week; during the next 3 months, sessions are held weekly. During the booster phase, 1 session is held each month. For the GDC modality, group sessions are held once a week for the 6 months of the active phase. During the booster phase, patients in the GDC alone treat¬ ment once group meet individually with the group counselor per month for a half hour visit. Cognitive Therapy. Cognitive therapy provided in this study follows a detailed manual.31 This treatment is based on the assumption that substance use disorders are re¬ lated to individuals' maladaptive beliefs and related thought processes. Cognitive therapy for substance abuse consists of 5 components: (1) collaboration, (2) case con¬ ceptualization, (3) structure, (4) socialization to the cog¬ nitive model, and (5) use of cognitive and behavioral tech¬ niques. Among the techniques used are Socratic questioning, advantages-disadvantages analysis, moni¬ toring of drug-related beliefs, activity monitoring and scheduling, behavioral experiments, and role playing. Psychodynamic Therapy. Brief SE psychodynamic therapy as conducted in this study follows the general SE psychodynamic therapy treatment manual32 and a more spe¬ cific variant of it developed for cocaine abusers.33 Accord¬ ing to this model, the problems associated with the use of SE cocaine and with its cessation are viewed in the context of understanding of the person's interpersonal and intra- an psychic functioning. The core conflictual relationship theme34 concept provides the framework for this understanding. The patient's core conflictual relationship theme, defenses, and views of self influence complicate the steps needed to stop using cocaine and to address the problems involved with co¬ caine dependence. The therapist uses supportive and inter¬ pretive techniques, particularly interpretations related to as¬ pects of patients' core conflictual relationship themes, that most interfere with the achievement of patients' goals. Individual Drug Counseling. This modality35 is based on a widely used approach to the treatment of drug addic¬ tion. It is time limited and focuses primarily on helping the patient achieve and maintain abstinence by encouraging behavioral changes, such as avoiding drug triggers, struc¬ turing one's life, and engaging in healthy behaviors (eg, ex- Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Pennsylvania User on 09/03/2013 ercise). Individual drug counseling is a staged approach with specific interventions keyed to the stages. In content and structure, it is consistent with the philosophy of the 12step approach, specifically that addiction is a disease that damages the person physically, emotionally, and spiritu¬ ally and that recovery is a gradual process. Participation in self-help groups is strongly encouraged. Group Drug Counseling. This treatment36 is designed to educate patients about the important concepts in ad¬ diction recovery, to strongly encourage participation in 12-step programs, and to provide a supportive group at¬ mosphere in which members can express feelings, dis¬ cuss problems, and learn to draw strength from one an¬ other. The group is progressive; the first 3-month phase is psychoeducational, and the second 3-month phase con¬ sists of open discussion with a focus on patients' help¬ ing each other solve problems in recovery. Assessments assessment battery was designed to capture mul¬ tiple domains: diagnoses, substance use patterns, psychopathological features or symptoms, quality of life, treat¬ ment services use, theory-specific measures of mediators The of outcome, treatment process, attrition, and motiva¬ tion for treatment. Most assessments are made at enroll¬ ment (or immediately following the orientation phase) and at monthly intervals during the active (months 1-6) and follow-up (months 9,12, 15, and 18) phases. In ad¬ dition, patients receive an assessment when they drop out of treatment or violate protocol. The primary outcome measure is the Drug Use Com¬ posite Scale of the Addiction Severity Index, a measure commonly used in substance abuse treatment with avail¬ able reliability and validity data.3740 This measure will be used to assess relative reduction in drug use during the active and follow-up phases of the protocol. Ob¬ served urine samples, collected once a week at group ses¬ sions, will be used as a secondary outcome measure. Urine samples will be used to assess time until remission (ie, 4 weeks of no drug use). In addition to the main efficacy measures, a pri¬ mary dependent variable of interest is retention in treat¬ ment. Days until a patient drops out of treatment or oth¬ erwise violates the protocol will be used as the measure of retention. Psychiatric severity is assessed via the psychiatric se¬ verity composite score on the Addiction Severity Index. The degree of antisocial personality characteristics is as¬ sessed using the Socialization Scale of the California Psy¬ chological Inventory. These measures have been used suc¬ cessfully in previous research on psychiatric severity21,23 and antisocial personality traits,24,25 respectively. Power Analysis and Sample Size Determination for a large, com¬ plex study like this one. Power will vary as a function of distribution of outcome measure, main effects vs inter¬ actional effects, and expected effect size. For the Addic¬ tion Severity Index, as the outcome measure analyzed via There is no single power calculation analysis of covariance, the power to detect main ef¬ fects is high (0.90-0.96) assuming 120 patients (30 per site) randomized into each of 4 treatments, 33% attri¬ tion (ie, lack of data), a type I error of 0.05, and effect sizes of 0.30 to 0.50 for our contrasts of interest. Analyzing each interaction separately (SE psycho¬ dynamic therapy vs CT at 2 levels of antisocial person¬ ality characteristics and SE psychodynamic therapy plus CT [pooled] vs IDC at 2 levels of psychiatric severity), we computed the size of the interaction effect that can be detected with a power of 0.80 and a Bonferronicorrected type I error of 0.025. As previously described, the sample size is 120 per arm of the study and there is 33% attrition. We found a Cohen F=0.55 for degree of antisocial personality characteristics and F=0.48 for the psychiatric severity interaction. Thus, with our pro¬ posed sample size, and correcting for attrition, the Bonferroni-corrected interaction tests are sensitive to mod¬ erate-sized effects. The statistical power for detecting differences in our other main outcome of interest, retention rates, was also examined. With the use of retention rates (in which drop¬ out is defined as no contact with a patient for 4 consecu¬ tive weeks) from our training phase of 90% at 1 month, 70% at 3 months, and 50% at 6 months and a power of 80%, an exponential function predicting dropout at time t months was calculated and used for power analysis for survival data. With the use of the method of Zelen and Dannemiller,41 the study's sample size was found to be adequate to detect a moderately small hazard ratio of 1.8 for each Bonferroni-corrected (2-tailed) pairwise com¬ parison of the 4 treatments, with a power of 0.80. an Data Analysis The initial analysis of the main treatment data will consist of analyses of the patterns of retention. This will be performed in a time-to-event analysis using the definition of retention previously given. Analyses of differences in efficacy using the Addiction Severity Index drug use composite score will be done via longitudinal regression approaches that allow for the use of all data collected. Additional cross-sectional analyses (analysis of covariance on 3- and 6-month outcomes) will be performed, as well as time-to-event analysis using the measure of days until abstinent determined from the urine data. Type of treatment, site, psychiatric severity, and degree of antisocial personality disorder will be the main independent variables for all of the previously described analyses. Statistically significant omnibus tests will be fol¬ lowed by post hoc comparisons ofindividual treatment mo¬ dalities, as appropriate. Several samples will be analyzed. An intent-to- pragmatic analysis that uses all data collected on pa¬ tients, regardless of whether the patient was a dropout or a protocol violator, will be conducted.42 Further analy¬ ses using only data available on patients up to dropout, treat protocol violation, or therapy completion will also be con¬ ducted. In this latter analysis, an attempt is made to dis¬ cern the effects of the therapies themselves, free of the effects of other treatments that patients may have re¬ ceived after dropping out of the program or otherwise violating protocol. Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Pennsylvania User on 09/03/2013 secondary analyses examining other out¬ come measures, mediators of outcome, and predictors of outcome will also be performed. The details of these analyses will be discussed in subsequent reports. Various SIGNIFICANCE knowledge, this study is the largest, most wellstudy of manual-guided psychotherapy and counseling for cocaine dependence yet conducted. The study is likely to yield practical information about the To our controlled of these patients, such as answers these ba¬ sic questions: Is professional psychotherapy a useful ad¬ dition to the standard drug-focused group counseling? Are there patient characteristics (eg, psychiatric sever¬ ity or degree of antisocial personality characteristics) that interact with type of treatment, thereby providing guid¬ ance to clinicians for making better matches between patients and treatments? The multisite nature of this project will allow for the examination of the extent to which answers to these questions are robust across different sites. treatment to Accepted for publication November 6, 1996. This study was supported in part by grants 1718- DA07090, U18-DA07663, U18-DA07673, U18-DA07693, and U18-DA07085from the National Institute on Drug Abuse, Rockville, Md; grant P50-MH-45178 and Career Develop¬ K05-DA00168from the Na¬ tional Institute of Mental Health, Rockville. We thank John Boren, PhD, National Institute on Drug Abuse, project officerfor this cooperative agreement. Reprints: Paul Crits-Christoph, PhD, Room 700,3600 Market St, Philadelphia, PA 19104. ment Awards K02-MH00756 and REFERENCES Maximizing the Efficacy of Psychotherapy and Drug Abuse Counseling Strategies in the Treatment ofCocaine Abusers. Rockville, Md: National Institute on Drug Abuse; 1991. US Dept of Health and Human Services publication DA-91-04. 2. Maximizing the Efficacy of Psychotherapy and Drug Abuse Counseling Strategies in the Treatment of Cocaine Abusers. Rockville, Md: National Institute on Drug Abuse; 1990. US Dept of Health and Human Services publication DA-90-04. 3. National Institute on Drug Abuse Household Survey 1994. Rockville, Md: National Institute on Drug Abuse, US Dept of Health and Human Services; 1994. 4. Woody GE, Luborsky L, McLellan AT, O'Brien CP, Beck AT, Blaine J, Herman I, Hole A. Psychotherapy for opiate addicts: does it help? Arch Gen Psychiatry. 1983; 1. 5. 6. 7. 8. 9. 10. 11. 40:639-645. Woody GE, McLellan AT, Luborsky L, O'Brien CP. Psychotherapy in community\x=req-\ based methadone programs: a validation study. Am J Psychiatry. 1995;152: 1302-1308. Means LB, Small M, Capone DM, Capone TJ, Condren R, Peterson M, Hayward B. Client demographics and outcome in outpatient cocaine treatment. Int J Addict. 1989;24:765-783. Jensen JP, Bergin AE, Greaves DW. The meaning of eclecticism: new survey and analysis of components. Professional Psychol: Res Pract. 1990;21:124-130. Rawson RA, ObertJL, McCann MJ, Marinelli-Casey P. Relapse prevention strategies in outpatient substance abuse treatment. Psychology Addict Behav. 1993;7:85-95. Crits-Christoph P, Siqueland L. Psychosocial treatment for drug abuse: selected review and recommendations for national health care. Arch Gen Psychiatry. 1996; 53:749-756. Najavits LM, Weiss RD. The role of psychotherapy in the treatment of substance use disorders. Harv Rev Psychiatry. 1994;2:84-96. Carroll KM, Rounsaville BJ, Gawin FH. A comparative trial of psychotherapies for ambulatory cocaine abusers: relapse prevention and interpersonal psychotherapy. Am J Drug Alcohol Abuse. 1991;17:229-247. 12. Carroll KM, Rounsaville BJ, Gordon LT, Nich C, Jatlow P, Bisighini RM, Gawin FH. Psychotherapy and pharmacotherapy for ambulatory cocaine abusers. Arch Gen Psychiatry. 1994;51:177-187. 13. Carroll KM, Rounsaville BJ, Nich C, Gordon LT, Wirtz PW, Gawin F. One-year follow-up of psychotherapy and pharmacotherapy for cocaine dependence. Arch Gen Psychiatry. 1994;51:989-997. 14. Higgins ST, Budney AJ, Bickel WK, Hughes JR, Foeg F, Badger G. Achieving cocaine abstinence with a behavioral approach. Am J Psychiatry. 1993;150:763769. 15. Rounsaville BJ, Glazer W, Wilber CH, Weissman MM, Kleber HD. Short-term interpersonal psychotherapy in methadone-maintained opiate addicts. Arch Gen 16. Psychiatry. 1983;40:629-636. Kang SY, Kleinman PH, Woody G, Millman RB, Todd TC, Kemp J, Lipton DS. Outcomes for cocaine abusers after once-a-week psychosocial therapy. Am J Psychiatry. 1991;148:630-635. 17. Shoham-Salomon V. Introduction to special section on client-therapy interaction research. J Consult Clin Psychol. 1991;59:203-204. 18. Beutler LE. Toward specific psychological therapies for specific conditions. J Consult Clin Psychol. 1979;47:882-897. 19. Beutler LE. Have all won and must all have prizes? revisiting Luborsky et al's verdict. J Consult Clin Psychol. 1991;59:226-232. 20. Luborsky L, Diguer L, Luborsky E, McLellan AT, Woody G, Alexander L. Psychological health-sickness (PHS) as a predictor of outcomes in dynamic and other psychotherapies. J Consult Clin Psychol. 1993;61:542-548. 21. McLellan AT, Luborsky L, Woody GE, O'Brien CP, Druley KA. Predicting response to alcohol and drug abuse treatments: role of psychiatric severity. Arch Gen Psychiatry. 1983;40:620-625. 22. Rounsaville BJ, Kosten TR, Weissman MM, Kleber HD. Prognostic significance of psychopathology in treated opiate addicts: a 2.5-year follow-up study. Arch Gen Psychiatry. 1986;43:739-745. 23. Woody GE, McLellan AT, Luborsky L, O'Brien CP, Blaine J, Fox S, Herman I, Beck AT. Psychiatric severity as a predictor of benefits from psychotherapy: the Penn-VA study. Am J Psychiatry. 1984;141:1172-1177. 24. Kadden RM, Cooney NL, Getter H, Litt MD. Matching alcoholics to coping skills of interactional therapies: posttreatment results. J Consult Clin Psychol. 1989; 57:698-704. 25. Cooney NL, Kadden RM, Litt MD, Getter H. Matching alcoholics to coping skills or interactional therapies: two-year follow-up results. J Consult Clin Psychol. 1991; 59:598-601. 26. Beutler LE, Engle D, Mohr D, Daldrup RJ, Bergan J, Meredith K, Merry W. Predictors of differential response to cognitive, experiential, and self-directed psychotherapeutic procedures. J Consult Clin Psychol. 1991;59:333-340. 27. Beutler LE, Machado P, Engle D, Mohr D. Differential patient and treatment maintenance among cognitive, experiential and self-directed psychotherapies. J Psychother Integr. 1993;3:15-31. 28. DiNardo PA, Barlow DH. Anxiety Disorders Interview Schedule-Revised (ADIS\x=req-\ R). Albany, NY: Phobia and Anxiety Disorders Clinic; 1988. 29. Wei LJ. An application of an urn model to the design of sequential controlled trials. J Am Stat Assoc. 1978;73:559-563. 30. Gainey RR, Wells EA, Hawkins JD, Catalano RF. Predicting treatment retention among cocaine users. Int J Addict. 1993;28:487-505. 31. Beck AT, Wright FD, Newman CF, Liese BS. Cognitive Therapy of Substance Abuse. New York, NY: Guilford Press; 1993. 32. Luborsky L. Principles ofPsychoanalytic Psychotherapy:A Manual for Supportive\x=req-\ Expressive Treatment. New York, NY: Basic Books Inc Publishers; 1984. 33. Mark D, Luborsky L. A Manual for the Use of Supportive-Expressive Psychotherapy in the Treatment of Cocaine Abuse. Philadelphia: Dept of Psychiatry, University of Pennsylvania; 1992. 34. Luborsky L. Core conflictual relationship themes (CCRT): a basic case formulation method. In: Eells T, ed. Handbook of Psychotherapy Case Formulation. New York, NY: Guilford Press. In press. 35. Mercer D, Woody G. Addiction Counseling. Philadelphia: Center for Studies of Addiction, University of Pennsylvania, Philadelphia Veterans Affairs Medical Center; 1992. 36. Mercer D, Carpenter G, Daley D, Patterson C, Volpicelli J. Addiction Recovery Manual. Philadelphia:Treatment Research Unit, University of Pennsylvania; 1994:2. 37. McLellan AT, Kushner H, Metzger D, Peters R, Smith I, Grissom G, Pettinati H, Argeriou M. The fifth edition of the Addiction Severity Index. J Subst Abuse Treat. 1992;9:199-213. 38. Kosten TR, Rounsaville BJ, Kleber HD. Concurrent validity of the Addiction Severity Index. J Nerv Ment Dis. 1983;17:606-610. 39. McLellan AT, Luborsky L, Cacciola J, Griffith JE. New data from the Addiction Severity Index: reliability and validity in three centers. J Nerv Ment Dis. 1985; 173:412-423. 40. Stoffelmayr BE, Mavis BE, Kasim RM. The longitudinal stability of the Addiction Severity Index. J Subst Abuse Treat. 1994;11:373-378. 41. Zelen M, Dannemiller MC. The robustness of life testing procedures derived from the exponential distribution. Technometrics. 1961;3:29-49. 42. Lavori PW. Clinical trials in psychiatry: should protocol deviation censor patient data? Neuropsychopharmacology. 1992;6:39-48. Downloaded From: http://archpsyc.jamanetwork.com/ by a University of Pennsylvania User on 09/03/2013