Methadone

INTRODUCTION

Methadone maintenance treatment is considered by many to be the treatment of choice for opiate dependence and has been termed "best practice" for this disorder. Methadone is among the most heavily researched drugs, and the field has consensually agreed that treatment retention, absence of illicit drug use while in treatment, full-time employment, and reduced and/or eliminated criminal activity are the essential outcome variables for this treatment (1-4). Because the methadone treatment outcome literature consistently reports that patients who are detoxified from methadone have relapse rates ranging from 50% to 90% 1 year after treatment (5-8), the methadone treatment literature often recommends that patients remain on methadone for an indefinite period of time (1). This means that outcome variables for methadone maintenance treatment essentially are those that occur within treatment. (There is, however, clinical consensus that treatment duration, dosage, and ancillary care services should be individualized and determined according to the needs of the patient.)

A variety of medical, social, and psychological variables have been studied as predictors of the outcome variables listed above. Although these studies often are contradictory, length of stay is consistently related to subsequent treatment outcome using a variety of outcome indices. That is, patients tend to improve the longer they remain in treatment (1,6).

Some studies have used basic psychological tests to search for outcome predictors. However, psychological tests have not fared well in predicting treatment dropouts among substance abuse programs. Neither scores on the Minnesota Multiphasic Personality Inventory-2 (MMPI) clinical scales (9) nor on the Millon Clinical Multiaxial Inventory (MCMI) scales (10) differentiated dropouts from program completers from an inpatient drug abuse (i.e., opiates and cocaine) rehabilitation program. The MMPI-2 substance abuse scales (MAC-R, AAS, APS) have been negatively related to length of stay in a VA substance abuse (alcohol and drugs) residential program (11) as well as to several other correlates of substance abuse but have not been used with methadone maintenance treatment outcome variables (12).

Studies have used the Negative Treatment Indicators (TRT) scale from the MMPI-2 (13) with mixed results. TRT scores did not predict differences between psychiatric patients who received an irregular discharge from inpatient treatment and patients who received a regular discharge (14) and were not related to length of stay in a VA residential treatment program (11).

Treatment dropout was studied among 65 perinatal, female substance abusers in one of two housing conditions, either living in the community or in a program-sponsored transitional housing unit. All subjects were assessed with the TRT scale. Treatment consisted of manualized counseling sessions three times a week for 20 weeks. These patients were also paid for their participation. Results indicated that the TRT scale scores did not predict treatment dropout. Perhaps the strong need for housing in this group, combined with the need to avoid jail and/or the loss of their children could have been mitigating factors that may have superceded any treatment resistance. In addition, it is possible that some patients might be unwilling to change (i.e., elevated TRT scores) but not withdraw from the program (15). Hence, this study might not have been a good test of the predictive power of the TRT scale.

On the other hand, high TRT scores were related to poorer treatment adherence and outcome in 7 of 10 measures among 107 alcohol and polydrug abusers in a 28-day, intensive, inpatient drug abuse program. Patients with more elevated TRT scores were more likely to drop out of treatment earlier, were more likely not to return to treatment after an initial screening interview, and were rated as having lower treatment motivation, poorer program participation, and poorer comprehension of program materials. Thus, elevated TRT scores were associated with treatment dropout and poorer outcomes while in treatment (16). In summary, the research literature is inconclusive as to the utility of the TRT scale as an outcome measure with substance abuse clients.

The Addiction Severity Index (ASI) (17) has also been used to predict the outcome of substance abuse treatment. The ASI, now in its fifth iteration, was commissioned by the National Institute on Drug Abuse, as a measure for the assessment of severity levels in the domains known to be affected by both alcohol and drug abuse. These domains are medical, employment, family/social, legal, and psychiatric/psychological. The instrument also provides separate severity levels for both alcohol and drug abuse. Factor analysis of the ASI revealed that the test has four dimensions: labeled chemical dependence, criminality, psychological distress, and health-related problems (18).

Although the Department of Veterans Affairs (DVA) has mandated that its substance abuse treatment programs use the ASI at intake and again at 6-month intervals to measure treatment outcome, it is currently reconsidering this requirement and is presently seeking other possible alternative measures for this purpose (R. Suschinsky, personal communication, November, 2002). Any such decision should be based on the scientific status of the instrument chosen for this purpose. Accordingly, we hope to add to the accretion of knowledge regarding the use of the ASI as an outcome measure for substance abuse treatment.

The purpose of this study was to determine if there were predictors of treatment outcome for a large sample of methadone maintenance patients after 1 year of treatment. We used two of the MMPI-2 Content Scales (TRT and CYN) and the Addiction Severity Index to predict relevant outcome variables associated with methadone treatment success. We hypothesized that these measures are significantly related to the treatment outcome measures used in this study.

METHOD

Participants

The patient sample consisted of 108 African American men, consecutively admitted to a methadone maintenance program from January 1, 2001 to December 31, 2001. All patients met SAMHSA (19) and DSM-IV (20) criteria for Opiate Dependence and signed a voluntary Consent to Treatment form.

If a patient was deemed eligible for methadone maintenance services at intake, then the patient was admitted and became part of this study. Operationally this meant that all patients had at least a 1-year opiate addiction history and had a toxicology urine sample positive for illicit opiates. A physician, Board Certified in Internal Medicine, made the diagnosis of opiate dependence. The average age of the sample was 48.84 years (SD 6.58). The average years of education was 12 (high school completion). The patient program demographics are 98% men and 2% women. Race is 95% African American, 3% Caucasian, and 2% Hispanic.

The patients are primary opiate addicted but many have associated cocaine abuse (50%), alcohol abuse (20%), and/or marijuana abuse (10%). Common medical co-morbidities are Hypertension (30%), Hepatitis C+ (29%), and Diabetes (11%). Common psychiatric co-morbidities include Depression (all diagnoses) (30%) and Posttraumatic Stress Disorder (15%). HIV + constitutes 5% of the population and PPD + (i.e., exposure to the TB virus) is 3%. Unemployment rates are 65%, and the majority of patients have personality disorders. The sample listed above was not significantly different from these program base rates in these variables and, hence, was not an atypical sample. However, because of the low number of women and Hispanics admitted during the period of this study, they were excluded from final data analysis.

Treatment Program

The program is a large VA hospital-based methadone maintenance outpatient program in an urban area and consists of outpatient methadone maintenance and associated services. Patients were required to ingest methadone daily for the first 90 days of treatment, and then they may become eligible for reduced medication pickup schedules per SAMHSA (19) regulations. Because the program is closed on Sunday, all patients are given a take-home dose of methadone for Sunday.

Methadone dose continues to be contentiously debated in the literature (5,21-25). This program does not assume a "high-dose" or "low-dose" philosophy. Rather, it strives to attain an individualized dose that will forestall the opiate withdrawal syndrome and reach a steady state where the patient no longer abuses heroin. A board-certified specialist in internal medicine saw each newly admitted patient daily and monitored the dose level until the opiate withdrawal syndrome was controlled. During the period of study the methadone doses ranged from 5 to 110 mg with the average dose of 53 mg. Again, this is consistent with program base rates for this variable.

In addition to daily ingestion of methadone, all patients are expected to have individual and/or group counseling at least twice a month. The program has a multidisciplinary staff assigned as case managers/counselors, including addiction therapists, social workers, and clinical psychologists. Because of this distinction, it was also possible to see if outcome differences varied by assigned staff.

Measures

We used two measures from the Content Scales from the Minnesota Multiphasic Personality Inventory-2 (13). The Negative Treatment Indicators (TRT) is a 26-item scale. In the MMPI-2 restandardization sample, internal consistency for men was 0.78 and 0.79 for women; test-retest reliability was 0.80 for men and 0.88 for women. The scale has two Content Component Subscales labeled "Low Motivation" and "Inability to Disclose" (26). Elevated scales on the TRT suggests negative attitudes toward physicians and mental health treatment, problematic beliefs, such as a feeling that change is not possible and that there is nothing in their life that needs change, and feeling uncomfortable discussing personal matters. They believe that no one understands them. The scale correlates with low self-esteem, apathy, and limited coping skills (27).

The Cynicism scale is a 23-item scale. In the MMPI-2 restandardization sample, internal consistency was 0.86 for men and 0.85 for women; test-retest reliability was 0.80 for men and 0.89 for women. The scale has two Content Component Subscales, labeled "Misanthropic Beliefs" and "Interpersonal Suspiciousness" (26). Elevated scores indicate a person who maintains negative attitudes toward people, even those close to them. They look for hidden motives behind people's behavior, believe that people use others and cannot be trusted, and believe that people are only friendly for selfish reasons. They view others as uncaring, selfish, and dishonest. High scorers are viewed as maladjusted, whiney, demanding, and quick-tempered (27).

We hypothesized that these two scales might logically be related to the process variable of attendance at counseling, which should be instrumentally related to subsequent outcome variables used here.

We also used the Addiction Severity Index (ASI) (17), which is a mandated measure by the Veterans' Administration for all patients with a substance abuse diagnosis. The ASI consists of approximately 200 items and yields computer-generated, quantitative composite scores (CS) of problem severity ratings in each of the seven domains during the 30-day period prior to the interview for each of the domains. The ASI takes approximately 45 minutes to administer via the computer. The use of the ASI with African American methadone maintenance patients as a measure of treatment outcome has previously been established (28).

Although the ASI provides for severity ratings by the examiner, this was thought to be too subjective for measurement purposes by the developers of the ASI. Therefore, they developed a composite measure that was mathematically derived and calculated by using the intercorrelations of items within each domain. Each composite score (CS), then, is the sum of answers to several questions within an ASI domain and is computer generated. They were designed to be measures of problems at treatment entry. These CS severity ratings are reliable across a range of patient populations and also across different types of examiners (17,29). Although there is limited validity data on the use of CS scores, recent research found that six of the seven CS scores significantly predicted its validity criterion among 310 methadone maintenance patients (30).

Dependent Measures

Length of Stay

We calculated the number of days in treatment up to 1 year after admission.

Number of Missed Medication Days

We calculated the number of days a patient missed methadone medication, expressed as a percentage of missed medication days. Because patients who dropped out before 1 year would obviously have fewer medication days scheduled than patients remaining in treatment for a longer period of time, we did not use the total number of medicated days. Rather, we selected the percentage of missed medication days as an outcome variable. The numerator was the number of missed days and the denominator was the total days for which methadone was prescribed (up to 1 year). This is a very objective and extremely accurate measure. All outpatient appointments are placed in the computer, which then records whether the patient both kept the appointment and received the medication for that day. One merely has to count the number of days for an appointment "no-show" to attain an accurate count of the number of days the patient was not medicated with methadone.

Toxicology, Urine Screens

This is another objective and accurate measure, available through computer medical records. The program mandates two random urine screens each month and is very stringent about ensuring that patients adhere to this requirement. The failure to provide a sufficient amount of urine samples for toxicology analysis results in patients receiving an administrative discharge. Accordingly, the program does not have a problem with patients submitting the required number of urine samples: Furthermore, these urines are monitored under direct observation. To maintain privacy, the patient enters a secure bathroom area to urinate. In an adjacent room, a health technician watches the patient urinate into a specimen cup through a two-way window. A hanging mirror is also affixed above the urinal to allow for total observation on all sides.

Each sample is analyzed for the presence of methadone, opiates, cocaine, marijuana, and benzodiazepines. Because the patients in the program do not have current histories of using or abusing amphetamines and barbiturates, the program does not routinely test for these drugs. However, periodically urines are analyzed for amphetamines and barbiturates. In addition, other toxicology screens (i.e., phencyclidine, etc.) are performed on an individualized basis, determined by the patient's current abuse pattern.

The raw scores for this variable were based on the number of urine samples free from illicit drugs (numerator) compared to the total number of urine samples provided over the total length of stay (denominator) (up to 1 year), expressed in percent "clean."

Ratings of Patient Progress

The remaining dependent variables are more subjective but have been shown to correlate with outcome criteria with this population in other studies (21). For the past two decades, program staff have had to submit monthly reports on the progress of patients assigned to their caseloads. Each month the counselor indicates the degree to which the patient is adhering to the counseling schedule stipulated in the patient's treatment plan (a process variable), whether the patient is employed full-time during the previous month, whether the patient's urine is clean for the past month, and then provides an overall rating of patient progress on a 5-point scale, according to the following stipulations: 1) excellent, 2) good, 3) fair, 4) poor, and 5) insufficient knowledge to assign a rating. Again, this simple rating scheme has been found to correlate with meaningful indices of progress in outpatient methadone maintenance programs (21). It should be noted that staff were unaware that this current study was in progress and that monthly staff ratings would be used as a measure. This meant that the monthly patient progress ratings were an unobtrusive measure and, therefore, unaffected by the measurement itself.

Although reduction and/or elimination of criminal behavior is a frequently used outcome measure in this field, our patients tend to come to us with little criminality. Although many have extensive criminal records, their age (ranging from 40 to 70 at admission) seems to preclude active criminal behavior (other than using drugs). Furthermore, base rates of arrest data among our population have consistently reflected that 1% of our patients are arrested within a year. Hence, this low base rate precluded our using arrest data as an outcome measure.

Procedure

Patients were given the TRT and CYN in paper-and-pencil version on the same day of admission but after they had been given their initial dose of methadone. This ensured the removal of any drug withdrawal stales and possible associated anxiety that might have interfered with test-taking attitudes. Per program and VA policy, staff were given up to 30 days after admission to complete the ASI. However, 85% of these were completed in 14 days or less. Again, drug withdrawal states had been eliminated prior to administering the ASI.

Each month thereafter, the computer appointment management system and the computerized medical record (toxicology results) were inspected and the relevant data abstracted onto a research worksheet. The number of days, listed as a "no-show" in the computer, was recorded, as was the urine toxicology results for that month. The monthly patient progress reports for each patient were then consulted and the relevant data abstracted and placed on the research worksheet for that month's tally. This procedure continued until the patient had dropped out of treatment or had a length of stay of 365 days in methadone maintenance treatment. The data in all categories were then aggregated and averaged over the length of stay for the final figures for analysis.

RESULTS

Table 1 presents the means and standard deviations for all variables. We tested for differences between counselors because prior research from this clinic had found substantial differences between assigned counselors on similar outcome variables used in this research (21). A univariate analysis of variance by counselor was conducted on both the independent and dependent variables. The only significant differences were in the ASI domain score of psychological (F = 3.53: df = 1,11: p < 0.0005; effect size = 0.29) and the outcome variable "percent of counseling sessions attended" (F = 4.12; df = 1,11; p < 0.0005; effect size = 0.32). A student Neuman-Keuls post hoc test indicated that counselor 8 had significantly less severe ratings in the domain of psychological severity than counselor 11; counselor 11 had significantly fewer patients remain in treatment than counselors 4, 9, and 12. No other comparisons were significant.

A Pearson Product Moment correlation was conducted between the TRT and CYN scales with the ASI domain ratings and the outcome variables. The results of these correlations appear in Table 2. TRT scores were significantly correlated with ASI severity ratings in the domains of Medical and Psychological; CYN scores were significantly correlated with ASI domain severity scores in psychological. No other correlations were significant.

Pearson Product Moment correlations were also conducted with the ASI domain severity scores and the dependent measures. Results appear in Table 3. High scores in medical severity and in employment severity were significantly correlated with low percentage of patients employed full time; high scores on alcohol severity were significantly correlated with low percent of patients who were clean; high scores on drug severity were significantly correlated with low percent of patients who were clean and high scores (e.g., poor progress) on ratings of patient progress; high scores on psychological severity significantly correlated with low percent of full-time employment and lower attendance at counseling sessions. [Higher legal severity scores significantly and negatively correlated with age (-0.20) (data not shown in Table 3).] No other correlations were significant.

Next regression equations were established to predict the major outcome variables used in this study. We used a stepwise linear regression, because we had continuous data (Logistic regression is used for discrete data.) Results of this analysis appear in Table 4. Absence of illicit drugs was predicted by low ASI drug severity scores and missing fewer days of medication; full-time employment was predicted by low scores on ASI-employment severity, longer lengths of stay, and a younger age; better attendance at counseling sessions was predicted by low scores on ASI psychological and longer lengths of stay; better ratings of patient progress were predicted by longer lengths of stay and lower ASI severity scores in the drug domain.

Finally, a stepwise discriminant function analysis was conducted to predict dropouts versus active patients. The variables that emerged were percent of missed medication, progress ratings, and percent clean. The eigenvalue was 0.834 and the canonical correlation was 0.674. This function correctly classified 85% of the participants. Discriminant functions are known to demonstrate lower values upon cross-validation, but we were unable to cross-validate this function because of the relatively small sample size.

DISCUSSION

This study assessed 108 methadone maintenance patients with the MMPI-2 Negative Treatment Indicatiors (TRT) and Cynical Attitudes (CYN) scales and with the Addiction Severity Index (ASI) and then followed these patients for 1 year. We found that although the MMPI-2 TRT and CYN content scales were significantly related to some ASI domains, they were unrelated to the outcome criteria used in this study. On the other hand, high ASI severity scores in the drug domain were significantly correlated with low percentage of clean urines at follow-up and poorer ratings of patient progress by staff. The Drug CS severity rating, along with the lower number of missed medication days, entered into the stepwise logistic regression equation as predictive of percent of toxicology urine screens free from illicit drugs. In more common parlance, these results indicate that having a low rating in drug severity at intake and ingesting methadone regularly was associated with better outcomes (i.e., "clean" urine) after 1 year of treatment.

The next important finding pertains to the role of psychological/ psychiatric severity in patient outcomes. Specifically, higher severity CS scores in the "Psych" domain were significantly correlated with lower rates of employment and less attendance at counseling sessions. This domain also entered into the regression equation predicting counseling attendance (or lack thereof). These results are consistent with previous studies that demonstrated that patients rated high in psychological/psychiatric severity levels show poorer treatment outcomes than patients rated lower on this dimension (31,32).

The CS in the area of employment was correlated with percent of patients employed during treatment and can be construed as a predictive validity for this ASI domain measure. It corroborates earlier findings that higher scores on this measure were associated with unemployment at treatment follow-up among methadone maintenance programs (30). In addition, these results add to the predictive validity of the ASI domains of alcohol and drug, which were related to continued use of illicit drugs and poor ratings of progress, and of psychological, which was related to lower rates of employment and poorer attendance at counseling. They are also consistent with recent data on the validity and, therefore, the use of the ASI domains (30).

We did not find any significant differences on most outcome variables according to assigned staff. One of 12 staff members had higher dropouts than 3 other staff, but the low number of assigned cases, averaging 8 per staff, precludes any extended discussion on this finding.

There are some limitations to our data. First, the sample was African American males, mostly in the age range from 40 to 60 and may not be generalized to other populations with different patient demographics. Second, these results need replication and may or may not be related to outcomes after discharge. They only address certain aspects of outcome up to 1 year of treatment. This is a matter for further research. Third, because methadone maintenance patients often remain in treatment for several years, researchers need to look for predictors of treatment outcome at data points further along in treatment than the 1-year limitation imposed in this study. Fourth, because correlation data are not causal, we are unable to determine what aspects of the treatment services actually accounted for these results. This also needs additional study. Fifth, other important factors not addressed here (e.g., methadone dose (a) levels) may also predict treatment outcome of methadone maintenance. Our study only looked at a selected number of predictors and outcomes, albeit frequently identified outcome measures. Sixth, we were not able to cross-validate these results, so that the results may be unique to this data set. All of these facts illustrate the difficulty and challenges of conducting "real-world" research in actual treatment clinics where researchers may not be able to influence or control for variables that might influence treatment outcome.

However, the results of this study and those of prior studies suggest that the ASI domain scores of both drug and alcohol appear to be reliable predictors of treatment outcome among methadone maintenance programs from 1 to 2 years of treatment, with higher severity scores predictive of poorer outcome. Similar results were found for employment severity scores, with higher CS employment scores associated with future unemployment. Thus, treatment personnel in these programs may want to target patients with higher severity scores in the alcohol, drug, and employment domains with special interventions designed to improve outcomes.

Finally, these results, as well as those from other studies, may bear on DVA decision on whether to retain the ASI as an outcome measure for its substance abuse programs. Our results, obtained in a DVA methadone maintenance program, suggest that the ASI seems to have predictive utility for the treatment of opiate dependence.

(a) We did not study dose because prior research with patients from this same treatment program found that methadone dose did not influence treatment outcome on these same variables. [From Ref. (21).]

REFERENCES

(1.) American Psychiatric Association. Position statement on methadone maintenance treatment. Am J Psychiatry 1994; 151:792-794.

(2.) Cooper JR, Altman F, Brown BS, Czechowicz D. Research on the Treatment of Narcotic Addiction: State of the Art. Rockville. MD: U.S. Department of Health and Human Service, National Institute on Drug Abuse, 1983.

(3.) Parrino MW. (Chair). State Methadone Treatment Guidelines. Center for Substance Abuse Treatment: Treatment Improvement Protocol (TIP) Series. Rockville, MD: Substance Abuse and Mental Health Services Administration, 1993.

(4.) Federal Regulation of Methadone Treatment. Rettig RA, Yarmolinsky A, eds. Washington, DC: Institute of Medicine. National Academy Press, 1995.

(5.) Ball JC, Ross A. The Effectiveness of Methadone Maintenance. New York: Springer-Verlag, 1991.

(6.) Hubbard RL. Drug Abuse Treatment: A National Study of Effectiveness. Chapel Hill, NC: University of North Carolina Press, 1989.

(7.) Maddux JF, Desmond DP. Methadone maintenance and recovery from opioid dependence. Am J Drug Alcohol Abuse 1992; 18:63-74.

(8.) Sells SB, Demaree RG, Hornick CW. Comparative Effectiveness of Drug Abuse Treatment Modalities. NIDA: Washington, DC, 1979.

(9.) Craig RJ. Personality diinensions related to premature termination from an inpatient drug abuse treatment program. J Clin Psychol 1984: 40:350-355.

(10.) Craig RJ. Can personality tests predict treatment dropouts? Int J Addict 1984:19:665-674.

(11.) Aaronson LA, Dent OB, Kline CD. Cross-validation of MMPI and MMPI-2 predictor scales. J Clin Psychol 1996:52:311-315.

(12.) Craig RJ. Assessing contemporary substance abusers with the MMPI. II. The MacAndrew Alcoholism Scale: a review. Subst Use Misuse. In press.

(13.) Butcher JN, Dahlstrom WG, Graham JR, Tellegen A, Kaemmer B. Minnesota Multiphasic Personality Inventory-2: Manual for Administration and Scoring. Minneapolis, MN: National Computer Systems, 1989.

(14.) Munley PH, Busby R. MMPI-2 negative treatment indicators scale and irregular discharge. Psychol Rep 1994:74:903-906.

(15.) Haller DL, Knisely JS, Elswick RK, Dawson KS, Schnoll SH. Perinatal substance abusers: factors influencing treatment retention. J Subst Abuse Treat 1997:14:513-519.

(16.) Gilmore JD, Lash SJ, Foster MA, Blosser SL. Adherence to substance abuse treatment: clinical utility of two MMPI-2 scales. J Pers Assm 2001; 77:524-540.

(17.) McLellan AT, Kushner H, Metzger DS, 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.

(18.) Rogalski CR. Factor structure of the Addiction Severity Index in an inpatient detoxification sample. Int J Addict 1987:22:981-992.

(19.) Substance Abuse and Mental Health Services Administration. Opioid Drugs in Maintenance and Detoxification Treatment of Opiate Addiction: Final Rule. January 17, 2001, Federal Register. 21 CFR Part 291:42 CFR Part 8. Wednesday. US Govt. Printing Press.

(20.) American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: APA, 1994.

(21.) Blaney T, Craig RJ. Methadone maintenance: does dose determine differences in outcome? J Subst Abuse Treai 1999:16:221-228.

(22.) Craig RJ. Effectiveness of low-dose methadone maintenance for the treatment of inner city heroin addicts. Int J Addict 1980; 15:701-710.

(23.) General Accounting Office. Methadone Maintenance: Some Treatment Programs are Not Effective--Greater Federal Oversight Needed. Washington, DC: GAO, Human Resources Division. 1990.

(24.) Strain EC, Stitzer ML, Liebson IA, Bigelow GE. Dose-response effects of methadone in the treatment of opioid dependence. Ann Intern Med 1993; 119:23-27.

(25.) Strain EC, Stitzer ML, Liebson IA, Bigelow GE. Methadone dose and treatmentoulcome. Drug Alcohol Depend 1993:33:105-117.

(26.) Ben-Porath YS, Sherwood NE. The MMPI-2 Content Component Scales: Development, Psychometric Characteristics and Clinical Application. Minneapolis, MN: University of Minnesota Press, 1993.

(27.) Craig RJ. Interpreting Personality Tests: A Clinical Manual for the MMPI-2, MCMI-III, CPI-R, and 16PF. New York: Wiley, 1999.

(28.) Mulvaney FD, Brown LS, Alterman AI, Sage RE, Cnaan A, Cacciola J, Rutherford M. Methadone-maintenance outcomes for Hispanic and African-American men and women. Drug Alcohol Depend 1989: 54:11-18.

(29.) Rogers R. Handbook of Diagnostic and Structured Interviewing. New York: Guilford, 2001.

(30.) Bovasso GB, Alterman AI, Cacciola JS, Cook TG. Predictive validity of the Addiction Severity Index's composite scores in the assessment of 2-year outcomes in a methadone maintenance population. Psychol Addict Behav 2001; 15:171-176.

(31.) McLellan AT, Childress AR, Griffith J, Woody GE. The psychiatrically severe drug abuse patient: methadone maintenance or therapeutic community? Am J Drug Alcohol Abuse 1984; 10:77-95.

(32.) 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.

Robert J. Craig, Ph.D., A.B.P.P. (1) * and Ronald E. Olson, Ph.D. (2)

* Correspondence: Robert J. Craig, Ph.D., A.B.P.P., Drug Abuse Treatment Center, VA Chicago Health Care System, 820 South Damen Ave., Chicago, IL 60612-3740, USA. Director; E-mail: Robert.Craig2@med.va.gov.

(1) Drug Abuse Treatment Center, VA Chicago Health Care System, Chicago, Illinois, USA

(2) Oakland University, Rochester, Michigan, USA

COPYRIGHT 2004 Taylor & Francis Ltd.
COPYRIGHT 2004 Gale Group