Drug and Alcohol Dependence 62 (2001) 9 – 17 www.elsevier.com/locate/drugalcdep

Predictors of outpatient treatment retention: patient versus substance use characteristics Mary E. McCaul a,b,*, Dace S. Svikis a, Richard D. Moore b a

Department of Psychiatry and Beha6ioral Sciences, Johns Hopkins Uni6ersity School of Medicine, Comprehensi6e Women’s Center, 911 N. Broadway, Baltimore, MD 21205, USA b Department of Medicine, Johns Hopkins Uni6ersity School of Medicine, Baltimore, MD 21205, USA Received 12 October 1999; received in revised form 3 May 2000; accepted 15 May 2000

Abstract The present study examined predictors of participation and retention for patients treated at an urban, hospital-based outpatient substance abuse treatment clinic. All patients were interviewed using the Addiction Severity Index (ASI) at the time of admission. Based on lifetime diagnostic history of psychoactive substance abuse/dependence, patients (N = 268) were classified as: alcoholonly, drug(s)-only, and alcohol +drug(s). Alcohol-only patients were significantly older, more likely to be Caucasian, married, have less than a high school education, and be employed than drug-only or alcohol/drug patients. Using multiple regression analysis, substance use status did not predict treatment participation and retention, whereas race, gender and employment composite score were significant predictors. Specifically, patients attended more sessions and remained in treatment longer if they were Caucasian, male and had a high employment composite score. These findings suggest that type of substance abuse may be overemphasized as a predictor of outpatient drug-free treatment retention, and that greater emphasis should be placed on tailoring treatment to patients’ cultural, gender and vocational needs. © 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Outpatient treatment; Substance use diagnosis; Gender; Ethnicity; Employment status; Attendance

1. Introduction Drug-free outpatient treatment settings provide services to individuals with a variety of different substance use disorders, often including alcohol, cocaine, or heroin dependence or combinations of these disorders. Recent studies have begun to examine differences in patient characteristics and the relative effectiveness of outpatient drug-free services as a function of patients’ alcohol and/or drug use status. For example, in a study of admissions to a 28-day traditional rehabilitation program (Brown et al., 1993), patients diagnosed with only an alcohol use disorder were significantly older and more likely to be married or living common-law than patients diagnosed with both alcohol and cocaine use disorders. Alcohol-only patients also reported significantly longer periods of alcohol use and more severe * Corresponding author. Tel.: +1-410-9555439; fax: + 1-4109554769. E-mail address: [email protected] (M.E. McCaul).

alcohol-related problems at the time of treatment admission. Additionally, Brower et al. (1994) found that alcohol-only rehabilitation patients had higher psychiatric severity and were more likely to have had prior treatment for their substance use disorder than either patients who abused cocaine only or alcohol and cocaine. Several studies have suggested diminished outpatient drug-free treatment efficacy for patients with drug use disorders in combination with an alcohol use disorder compared to patients with an alcohol use disorder only (Mammo and Weinbaum, 1993; Wickizer et al., 1994). For example, alcohol and cocaine patients were more likely to report alcohol or drug relapse, fewer abstinent days, and to have a positive urine toxicology result than alcohol-only patients during the 6 months following treatment discharge (Brown et al., 1993). Similar findings of decreased treatment efficacy for patients with combined alcohol and drug use disorders have been reported by other investigators (Alterman et al., 1992; Carroll et al., 1993; Brower et al., 1994), and generally

0376-8716/01/$ - see front matter © 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 3 7 6 - 8 7 1 6 ( 0 0 ) 0 0 1 5 5 - 1

10

M.E. McCaul et al. / Drug and Alcohol Dependence 62 (2001) 9–17

suggest the need for more tailored and intensive treatment interventions to address the increased relapse risks. The present study had two goals. The primary goal was to examine the predictive utility of patient characteristics and drug use status on retention and participation in an outpatient, drug-free treatment program. A secondary goal was to characterize differences in demographic and psychosocial characteristics of patients admitted to drug-free outpatient treatment as a function of alcohol and or drug use status. It was hypothesized that patients with a drug use disorder would have poorer treatment participation and retention than patients with only an alcohol use disorder. The study was focused on treatment attendance and retention because: (1) high attrition is recognized as a key factor limiting substance abuse treatment effectiveness (Gainey et al., 1993; Carroll, 1997; Onken et al., 1997); and (2) treatment participation and retention have been consistently found to be positively related to post-discharge outcomes (Hubbard et al., 1989; Hoffman et al., 1996; McKay et al., 1998). 2. Methods

2.1. Subjects Subjects (N = 268) were consecutive admissions to an abstinence-oriented, urban, hospital-based, publicly funded outpatient treatment program. Patients reported their primary and, as applicable, secondary and tertiary drug problems as part of their intake interview using the Addiction Severity Index (ASI). The overall prevalence of different substance use problems (primary, secondary or tertiary) was: 81% alcohol; 48% cocaine; 36% marijuana; 30% heroin; 6% sedatives; 6% hallucinogens; and 2% other. On the basis of the ASI results, patients were categorized into one of three mutually exclusive groups based on lifetime substance use problems: alcohol-only (N =72), drug-only (N= 51), and alcohol+drug (N =145). Alcohol and drug ASI composite scores and lifetime number of prior treatment episodes supported the accuracy of group assignment. That is, alcohol composite scores were elevated in the alcohol-only (0.52) and alcohol+drug (0.57) groups compared to the drug-only group (0.07) (F = 81.05, P B 0.0001), whereas drug composite scores were elevated in the drug-only (0.57) and alcohol+ drug (0.54) groups compared to the alcohol-only group (0.04) (F= 115.21, PB 0.0001). Similarly, prior alcohol treatment episodes were reported only by subjects in the alcohol only (2.1) and alcohol+drug (2.6) groups (drug only= 0; F =16.06, P B 0.0001), whereas prior drug treatment episodes were reported only by subjects in the drug only (1.3) and alcohol+drug (2.4) groups (alcohol only =0; F =35.43, P B 0.0001).

2.2. Procedures Patient characteristics were assessed using the intake ASI (McLellan et al., 1980, 1985). The ASI is a semistructured interview that examines seven domains, including medical, employment, alcohol, drug, legal, family/social and psychological. Two summary measures are available for each ASI domain: interviewer severity rating, which represents the interviewer’s impression of the patient’s need for additional services (low: 0–3; moderate: 4–6; severe 7–9); and a composite score, which is mathematically calculated from data obtained during the interview (range 0–1). Interviews were conducted by trained assessment staff who participated in regular quality control procedures to enhance the accuracy and reliability of data collection. Based on Maryland state management information guidelines, patients were categorized as outpatient clinic admissions if they had at least two face-to-face therapeutic services within the 30 days following either their initial intake appointment for direct outpatient admission or transfer from residential or intensive outpatient services. Patients were discharged when they successfully completed their treatment goals, transferred to another treatment facility, exhibited violent or other extreme behaviors, or failed to attend treatment for 30 days.

2.3. Treatment ser6ices The present study focused exclusively on outcomes in the standard outpatient program; however, three substance abuse treatment modalities were available within the program. Residential treatment was delivered on a 16-bed dormitory unit. Intensive outpatient program (IOP) patients attended the clinic from 08:00 to 17:30 h daily. In both modalities, patients received 6 h of group therapeutic intervention daily as well as thrice-weekly individual counseling sessions during a 14-day length of stay. Standard outpatient treatment consisted of three therapeutic components: abstinence maintenance monitoring (breathalyzer and urinalysis testing); individual counseling; and group education/counseling sessions. Patients were expected to participate in three abstinence monitoring visits per week, one group counseling session per week, and two individual counseling sessions per month for a minimum of a 6-month duration. Patients could be directly admitted into any one of the three treatment modalities based on bed availability, medical and psychiatric acuity, patient preference, homelessness and substance abuse severity. Over half of the patients in the study (55%) were admitted directly into the residential program, 9% into the intensive outpatient program, and 36% into standard outpatient services. Patients were transferred into the outpatient program following completion of the residential or intensive outpatient modalities.

M.E. McCaul et al. / Drug and Alcohol Dependence 62 (2001) 9–17

The present paper examines attendance and retention only for patients transferred or directly admitted to the standard outpatient program. Completion rates for the residential program exceeded 87%, and, as a result, there was insufficient variability in outcomes to permit meaningful analyses. An advantage of these high completion rates is that very few subjects were lost from the outpatient treatment analyses. While completion rates for the IOP were somewhat lower, the small number of subjects admitted directly to this level of care (N= 27) precluded separate analyses of outcome predictors for this subsample. Level of care at the time of program admission (residential, IOP or direct outpatient) is included as a predictor variable in the analyses of outpatient program participation and retention to account for differences in treatment exposure immediately preceding outpatient program admission.

2.4. Statistical analysis Dependent measures included outpatient length of stay (number of calendar days from first to last face-toface program service), total number of days attended, number of individual counseling sessions, and number of group counseling sessions. Data were collected using monthly counselor-generated attendance and billing logs, with confirmation by retrospective chart reviews when discrepancies were identified. Data analyses were conducted using the SAS statistical program. Categorical data were analyzed using  2 analyses. Continuous variables were analyzed using analysis of variance as the primary statistical technique. Duncan multiple range test was used for post hoc comparisons of group means. Finally, a series of fully adjusted, step-wise multiple regression analyses were conducted on outpatient treatment retention and participation measures. Table 1 Demographic characteristics as a function of lifetime substance use status

Age (years) Female (%) African-Amer ican (%) Never married (%) BHigh school (%) a

Alc only (N= 72)

Drug only (N =51)

Alc+drug (N =145)

F/ 2 value*

42 25 28

28 43 67

32 26 50

48.14d 6.39a 19.52c

29

75

49

25.75d

61

41

44

25.35d

PB0.05; b PB0.01; PB0.001; d PB0.0001 are levels of statistical significance. * F values are presented for continuous measures;  2 values are presented for categorical measures. c

11

3. Results

3.1. Demographic, substance use and psychosocial characteristics As summarized in Table 1, alcohol-only patients were oldest and most likely to be Caucasian, married or separated/divorced, and to have dropped out of school prior to receiving their high school degree. Drug-only patients were youngest and most likely to be AfricanAmerican, never married, and to have a high school diploma or post-secondary education. Patients with alcohol and drug problems were intermediate between the other two groups. The three groups differed on a number of patient characteristics (Table 2). Alcohol-only patients were significantly less likely to have been recently hospitalized for detoxification than patients in the two drug-use groups, who had comparable detoxification rates. Alcohol + drug patients were most likely to have experienced a substance-related overdose in their lifetime as compared with alcohol-only or drug-only patients. For recent substance use problems (number of days in the past 30), alcohol+drug and drug-only patients reported comparable frequencies of drug problems, while alcohol+ drug patients reported nearly twice as many days with alcohol problems than alcohol-only patients. The three groups had comparable employment/support composite scores; however, analyses of specific employment and income items identified differences across the three groups (Table 2). Drug-only patients were most likely to be unemployed and to have never held the same full-time job for more than 1 year. Relatedly, source of income varied as a function of patients’ drug involvement. Alcohol-only patients were more likely to report recent employment income, with 2–3 times more alcohol-only patients reporting earned income greater than $500 per month than drug-only or alcohol+ drug patients. Drug-only and alcohol+drug patients were more likely to report illegal income. In line with these findings, there were significant differences in interviewer severity ratings reflecting need for employment counseling, with more drug-only and alcohol +drug patients receiving moderate to high ratings as compared with alcohol-only patients. As observed for employment composite scores, composite scores in the ASI legal domain also did not differ across the three groups, yet there were a number of group differences in legal status (Table 2). Alcohol-only patients were most likely to enter treatment at the prompting of the criminal justice system. Drug-related charges were more likely to be reported by patients in the drug-only and alcohol+ drug groups, whereas alcohol-only patients were more likely to report disorderly conduct or DWI charges. Alcohol+drug patients were most likely to report having been arrested and charged with robbery compared to the other two groups.

M.E. McCaul et al. / Drug and Alcohol Dependence 62 (2001) 9–17

12

Table 2 Alcohol and drug use, legal and employment characteristics from the intake Addiction Severity Index (ASI) as a function of lifetime substance use status F/ 2 value**

Alc only

Drug only

Alc+drug

(N =72)

(N =51)

(N =145)

Alcohol and drug characteristics Recent detox (%) Ever overdosed (%) No. days recent alcohol problems* No. days recent drug problems*

26 6 5.1 0

59 27 0 12.0

60 44 9.7 11.7

23.39d 33.38d 19.16d 31.32d

Employment characteristics Longest FT jobB1 year (%) Currently unemployed (%) Employment income \500/month (%) Moderate-high interviewer severity ratings (%)

17 29 33 47

48 39 10 73

18 34 14 77

44.22d 31.26b 23.62b 21.01d

Legal characteristics Treatment prompted by CJS (%) Recent illegal activity (%) Recent illegal income (%) Drug-related charges (%) DWI (%) Drunk/disorderly (%) Robbery (%)

45 4 3 6 43 54 4

18 24 17 47 4 4 4

18 23 19 48 25 45 14

20.11d 12.39b 15.46a 41.78d 36.31d 37.57d 8.54a

* Patient characteristic was measured over the 30 days prior to intake. ** F values are presented for continuous measures;  2 values are presented for categorical measures. a PB0.05; b PB0.01; c PB0.001; d PB0.0001 are levels of statistical significance.

Groups did not differ on ASI medical, psychiatric or social/family composite scores. Across the three groups, approximately one-third of the patients reported chronic medical problems. Also, comparable rates of psychiatric treatment were observed, with approximately 20% of patients reporting a history of inpatient and 30% reporting a history of outpatient psychiatric treatment. However, alcohol-only patients (27%) were twice as likely to report trouble concentrating compared to alcohol+drug (15%) or drug-only patients (12%) ( 2 =6.28, P = 0.04). There also was a tendency for more alcohol-only patients (36%) to report recent clinical anxiety as compared with alcohol+drug (24%) or drug-only patients (20%) ( 2 =5.30, P =0.07). Finally, there was a trend for alcohol+ drug patients (18%) to report recent serious thoughts of suicide compared with alcohol-only (7%) or drug-only (14%) patients ( 2 =4.96, P=0.08).

3.2. Treatment attendance and participation As summarized in Table 3, mean length of treatment stay was significantly different across the three substance use groups. Post hoc comparisons revealed that the alcohol-only patients had a significantly longer length of stay than drug-only patients, while length of stay was not significantly different for alcohol-only and

alcohol+ drug groups and for drug-only and alcohol+ drug groups. When length of stay was analyzed categorically, alcohol-only patients were almost twice as likely to remain active in treatment a minimum of 6 months (42%) compared to drug-only (20%) or alcohol + drug patients (26%) ( 2 = 17.34, P= 0.008). The substance use groups also differed on measures of treatment attendance, including number of individual and group counseling sessions and total number of clinic visits (Table 3). Post-hoc analyses indicated that alcohol-only patients participated in significantly more Table 3 Treatment participation and retention as a function of lifetime substance use status

Length of stay (days) No. of visits No. of individual sessions No. of group sessions a b

Alc only (N =72)

Drug only (N =51)

Alc+drug (N= 145)

F value

163.8

105.4

129.9

3.54a

38.5 14.8

19.7 7.5

30.3 10.9

4.64b 4.48b

10.1

4.7

7.6

4.31b

PB0.05; PB0.01 are levels of statistical significance.

M.E. McCaul et al. / Drug and Alcohol Dependence 62 (2001) 9–17 Table 4 Predictors of outpatient treatment retention Variable

i Coefficient

P

Intercept (days)

141

Substance use status Alcohol only (vs. alc+drug) Drug only (vs. alc+drug)

19 −7

0.29 0.73

Race African-American (vs. Caucasian) Other (vs. Caucasian)

−38 −19

0.05 0.72

Female (vs. male) Employment comp score

−32 42

0.05 0.02

Fig. 1. Race and gender as predictors of treatment retention. Length of stay represents the number of days from initial program contact to last face-to-face visit. The combined race/gender variable significantly predicted length of stay (P B0.001).

treatment activities than drug-only patients; attendance of alcohol + drug patients did not differ significantly from the other two groups. Treatment compliance was evaluated by examining the proportion of kept versus scheduled clinic appointments; compliance was defined as attendance of at least 50% of scheduled sessions. There were significant differences in treatment compliance across the three study groups ( 2 = 11.66, P = 0.003). Alcohol-only (83%) and alcohol+drugs patients (77%) were significantly more likely to be compliant with treatment appointments than drug-only patients (57%).

3.3. Predictors of outpatient treatment participation and retention Predictors of treatment participation and retention were examined using multiple regression analyses. In the initial analysis, predictor variables included substance use status (alcohol-only, drug-only, alcohol+

13

drug), initial treatment modality (residential, IOP, outpatient), ASI composite scores as well as the demographic characteristics that were significantly different across substance use groups. There was no evidence of a predictive relationship between substance use status or the ASI alcohol or drug composite scores and treatment participation or retention. The ASI composite scores for the medical, legal, family/social and psychiatric domains also did not predict treatment participation. Next, a regression analysis was conducted entering those variables that were robust predictors of treatment retention in the overall regression model. Substance use status was retained in this analysis as it was the primary variable of interest in this study. The intercept in Table 4 is the grand mean of length of treatment retention for the entire sample; beta weights show the change in length of stay from the intercept as a function of each variable. As shown in Table 4, race, gender, and ASI employment composite score were significant predictors of outpatient treatment retention. Specifically, length of stay was significantly shorter for African American compared to Caucasian patients. Women had a shorter treatment duration than men. Greater need for employment counseling was associated with longer lengths of stay. These same relationships were observed for measures of treatment participation including number of individual and group counseling session and total clinic visits. Because race and gender individually predicted treatment retention, the two variables were combined in a final regression analysis. The combined race/gender variable significantly predicted length of stay (PB 0.001). As shown in Fig. 1, length of stay for AfricanAmerican men was significantly shorter than that of White men (P= 0.0005). Similarly, African-American women evidenced approximately half the treatment duration of White men (P=0.003).

4. Discussion This study examined psychosocial characteristics and outpatient treatment participation as a function of patients’ lifetime substance use status. Alcohol-only patients were more likely to be Caucasian, older, married or separated/divorced, less educated, employed with higher legal incomes, and more likely to enter treatment at the prompting of the criminal justice system than drug-only patients. Group differences also were observed in treatment retention and participation; however, in a multivariate analysis, patients’ substance use status did not predict treatment retention. Rather significant treatment outcome predictors were patient demographic characteristics including gender, race and employment status. Patients who were Caucasian, male

14

M.E. McCaul et al. / Drug and Alcohol Dependence 62 (2001) 9–17

or had high need for employment counseling remained in treatment longer and participated in more treatment services than patients who were African-American, female or had low need for employment counseling. Interestingly, a recent study on treatment retention of perinatal substance abusers also found demographic compared to drug use characteristics to be more predictive of length of stay in substance abuse treatment (Ingersoll et al., 1995). The present study found that women had poorer treatment participation and retention than men in this mixed-gender treatment setting. To date, there have been very few efficacy studies systematically examining substance abuse treatment outcomes as a function of gender (Institute of Medicine, 1990; Moras, 1998). Analyses by gender also are infrequently reported in effectiveness, evaluation or health services research. Because women represent only approximately one-quarter of the patients treated in mixed-gender settings (Office of Applied Studies, 1997), it is often difficult to accumulate an adequate sample of women for gender analyses. When treatment outcomes have been reported as a function of gender, findings have been mixed, with some reporting no difference (Vannicelli, 1984; McCance-Katz et al., 1999) and others finding poorer treatment outcomes for women (Hoffman et al., 1996). Factors hypothesized to contribute to poorer retention rates for women include: the imbalance of male and female patients in outpatient treatment (Office of Applied Studies, 1997); the perceived irrelevance of much of the traditional, male-oriented program content (Vannicelli, 1984; Institute of Medicine, 1990); and the lack of opportunity to address male relationship issues, including frequent histories of physical and sexual abuse (Miller et al., 1989, 1993). In one of the only randomized efficacy studies examining women’s-specific treatment (Dahlgren and Willander, 1989), women who were treated in a women’s-specific program remained in treatment longer, had higher rates of program completion and reported improved outcomes on a variety of alcohol, psychosocial and morbidity measures compared with women treated in a mixed-gender setting. The current study also found that African American patients discontinued treatment earlier than Caucasian patients. This finding concurs both with reports of higher rates of premature drop-out and treatment termination among African-American compared to White substance abusers (Kleinman et al., 1992; Agosti et al., 1996) as well as overall decreased access and use of health care services for African Americans compared to Whites (Blendon et al., 1989; Himmelstein and Woodhandler, 1995). Etiologic, language and cultural differences suggest that substance abuse treatment programs serving racial and ethnic minorities may increase their effectiveness by tailoring treatment procedures and curricula to meet the special needs of these patients (Law-

son and Lawson, 1989; Rouse et al., 1995). One strategy has been to match therapists and patients on the basis of race/ethnicity. While this appears to increase satisfaction ratings and initial engagement of African-American patients (Atkinson, 1983), effects on treatment retention and outcomes are not consistently evident (Sue, 1988; Maddux and Desmond, 1996; Rosenheck and Seibyl, 1998). Another approach for treatment programs is to modify the nature and extent of services available for minority patients. For example, Westermeyer (1984) found ethnically-sensitive treatment programs were more successful in recruiting ethnic minorities into treatment and achieved enhanced treatment outcomes for more individuals. Clearly, additional research is needed to more fully characterize the treatment needs of substance abusing ethnic minorities and to develop specialized treatment services to more effectively target these needs. In the current study, treatment retention and participation were lowest for African-American women compared to the three other ethnic/gender groups. Research has shown that substance abusing, African-American women experience a variety of health, social and environmental disadvantages that may interfere with treatment access and increase their risk for early treatment drop-out and poor outcomes (Mammo and Weinbaum, 1993; Carter and Rogers, 1996; Sanders-Phillips, 1998). For example, African-American women are at elevated risk for HIV-infection associated with their own or a sexual partner’s drug use. African-American women entering substance abuse treatment have been found to have a high prevalence of medical problems associated with their addiction lifestyle (e.g. sexually transmitted diseases, anemia, and dental diseases) as well as significant untreated medical disorders including heart disease and breast masses (Curtiss et al., 1993). Rates of alcohol-related deaths are over two times higher among African-American compared with Caucasian women (Rouse et al., 1995). African-American women are more likely to be single heads of households and to have these households victimized by serious violent crime as compared with Caucasian women (Rouse et al., 1995). These considerable personal, social and environmental issues that African-American women face has led to development of specialized treatment interventions to better address their needs (Saulnier, 1996; Jackson et al., 1997); however, the effectiveness of these programs has not yet been investigated in controlled treatment outcome research. It was observed that patients with high need for employment counseling remained in treatment longer than patients with low employment counseling needs. This finding is counter to earlier reports that employment and higher socioeconomic status were positive predictors of alcoholism treatment outcome (Polich et al., 1981; McLellan et al., 1983; Jackson et al., 1997). It

M.E. McCaul et al. / Drug and Alcohol Dependence 62 (2001) 9–17

was hypothesized that need for employment counseling may operate in at least three ways as a determinant of outpatient treatment retention. First, none of the patients in this study were referred to treatment as a condition for continued employment; thus, the likelihood that employment could act as ‘leverage’ in the current sample to improve treatment is reduced. Second, employed patients may experience greater difficulty attending treatment sessions because of work conflicts that unemployed patients do not experience. Third, those patients who reported a high need for employment counseling received supportive services and referrals to vocational rehabilitation as appropriate through their treatment participation. In Maryland, employment training and placement services often will not accept referrals of substance abuse patients until they have achieved a minimum of 3 months of abstinence. Thus, unemployed patients may have had a strong incentive for continued treatment involvement. In the present study, the demographic heterogeneity of the treatment sample is a strength, with sufficient representation of a variety of demographic and substance use variables to allow concurrent examination of these potentially important factors as predictors of treatment outcome. Another strength was the independence of the findings from the impact of managed care on treatment participation and retention. Treatment services were publicly-funded and, as a result, there were no external controls on number and frequency of visits or length of stay. There are several weaknesses in the study design as well. First, drug status classification of patients was based on self-report without independent corroboration by urine toxicology screens. The validity of this group assignment strategy, however, is supported by the consistency between patient self-report and ASI interviewer severity ratings. This consistency suggests that, after reviewing the objective ASI data across all seven domains, the trained interview staff concurred with patient self-report of substance abuse problems. Also, it is important to note that there was no reason for patients to misrepresent their alcohol and drug use since they were told that this information did not impact their admission to the program, level of care or overall treatment experience. Nonetheless, the absence of urine toxicology results to confirm substance use status group assignment is a weakness in the present study. A second potential methodological weakness is that the data are limited to within-treatment variables, with no post-discharge outcome measures available on study participants. Earlier research has demonstrated a positive relationship between extent of participation in outpatient treatment settings and post-treatment alcohol and drug use and psychosocial outcomes (Hubbard et al., 1989; Hu et al., 1997). Examining factors that influence treatment retention as a surrogate for post-treatment follow-up outcomes re-

15

duces the high costs associated with community-based follow-up and broadens the opportunity for conducting outcome-oriented research in community-based treatment settings, which have not traditionally been amenable or accessible to research participation. A third limitation of the study is the exclusive reliance on the ASI for characterizing patient status. While the ASI has widespread acceptability, good psychometric characteristics and measures some very salient aspects of clinical presentation, findings are limited by being based on a single instrument. Fourth, although the study sample was heterogeneous for race and gender, it was relatively homogeneous for socioeconomic status, thus limiting generalization of the conclusions to low income substance abuse treatment populations. Findings also may not readily generalize to other cultural groups or treatment settings. Finally, 60% of the subjects presented to treatment with a primary diagnosis of alcohol dependence compared with smaller percentages of subjects with a primary dependence on other drugs. Thus, patients with primary alcohol dependence were somewhat over represented in the sample and generalization of the findings to other clinical populations must be done cautiously. Nonetheless, it is encouraging that several of the findings are consistent with earlier reports based on primary cocaine dependent patients (Brown et al., 1993; Brower et al., 1994). This study has important implications for future development of substance abuse treatment services. Increasingly, research suggests that psychosocial treatments should be tailored for such patient characteristics as gender, age and cultural background rather than type of drug use disorder. For example, women with diverse substance use disorders may be more likely to have similar gender-specific treatment needs than men and women with the same drug use disorder. Creating programs that are sensitive and specific to patient needs should increase treatment effectiveness beyond the current substance-based system.

Acknowledgements The authors want to thank Anne Gupman, M.A. and Ramana Goplan, M.D. for their assistance with data collection and analyses. This research was conducted using internal sources of support.

References Agosti, V., Nunes, E., Ocepeck-Welikson, K., 1996. Patient factors related to early attrition from an outpatient cocaine research clinic. Am. J. Drug Alcohol Abuse 22, 29 – 39. Alterman, A.I., Droba, M., McLellan, A.T., 1992. Response to day hospital treatment by patients with cocaine and alcohol dependence. Hosp. Community Psychiatry 43, 930– 932.

16

M.E. McCaul et al. / Drug and Alcohol Dependence 62 (2001) 9–17

Atkinson, D.R., 1983. Ethnic similarity in counseling psychology: a review of research. Couns. Psychol. 11, 79–92. Blendon, R.J., Aiken, L.H., Freeman, J.E., Corey, C.R., 1989. Access to medical care for Black and White Americans: a matter of continuing concern. J. Am. Med. Assoc. 261, 278–281. Brower, K.J., Blow, F.C., Hill, E.M., Mudd, S.A., 1994. Treatment outcome of alcoholics with and without cocaine disorders. Alcohol. Clin. Exp. Res. 18, 734–739. Brown, T.G., Seraganian, P., Tremblay, J., 1993. Alcohol and cocaine abusers 6 months after traditional treatment: do they fare as well as problem drinkers? J. Subst. Abuse Treat. 10, 545– 552. Carroll, K.M., 1997. Enhancing retention in clinical trials of psychosocial treatments: practical strategies. In: Onken, L.S., Blaine, J.D., Boren, J.J. (Eds.), Beyond the Therapeutic Alliance: Keeping the Drug-Dependent Individual in Treatment, vol. 165. National Institute on Drug Abuse, Rockville, MD, pp. 4–24. Carroll, K.M., Rounsaville, B.J., Bryant, K.J., 1993. Alcoholism in treatment-seeking cocaine abusers: clinical and prognostic significance. J. Stud. Alcohol 54, 199–208. Carter, J.H., Rogers, C., 1996. Alcoholism and African-American Women: a medical sociocultural perspective. J. Natl. Med. Assoc. 88, 81 – 86. Curtiss, A.M., Lenz, K.M., Frei, N.R., 1993. Medical evaluation of African– American women entering treatment. J. Addict. Dis. 12, 29 – 44. Dahlgren, L., Willander, A., 1989. Are special treatment facilities for female alcoholics needed? A controlled 2-year follow-up study from a specialized female unit (EWA) versus a mixed male/female treatment facility. Alcohol. Clin. Exp. Res. 13, 499– 504. Gainey, R.R., Wells, E.A., Hawkins, J.D., Catalano, R.F., 1993. Predicting treatment retention among cocaine users. Int. J. Addict. 28 (6), 487– 505. Himmelstein, D.U., Woodhandler, S., 1995. Care denied: U.S. residents who are unable to obtain needed medical services. Am. J. Public Health 85, 341–344. Hoffman, J.A., Caudill, B.D., Koman, J.J., Luckey, J.W., Flynn, P.M., Mayo, D.W., 1996. Psychosocial treatments for cocaine abuse: 12-month treatment outcomes. J. Subst. Abuse Treat. 13, 3–11. Hu, T., Hunkeler, E.M., Weisner, C., Li, E., Grayson, D.K., Westphal, J., McLellan, A.T., 1997. Treatment particiation and outcome among problem drinkers in a managed care alcohol outpatient treatment program. J. Ment. Health Adm. 24, 23 – 34. Hubbard, R.L., Arsden, M.E., Rachal, J.V., Harwood, H.J., Cavanaugh, E.R., Ginzburg, H.M., 1989. Drug Abuse Treatment: A National Study of Effectiveness. University of North Caroline Press, Chapel Hill, NC. Ingersoll, K.S., Lu, I.-L., Haller, D.L., 1995. Predictors of in-treatment relapse in perinatal substance abusers and impact on treatment retention: a prospective study. J. Psychoactive Drugs 27 (4), 375– 387. Institute of Medicine, 1990. Broadening the Base of Treatment for Alcohol Problems, National Academy of Sciences, Washington, DC. Jackson, M.S., Stephens, R.C., Smith, R.L., 1997. Afrocentric treatment in residential substance abuse care: the Iwo San. J. Subst. Abuse Treat. 14, 87 –92. Kleinman, P., Kang, S., Lipton, D., et al., 1992. Retention of cocaine abusers in outpatient psychotherapy. Am. J. Drug Alcohol Abuse 18, 29 – 43. Lawson, G.W., Lawson, A.W., 1989. Alcoholism and Substance Abuse in Special Populations. Aspen, Rockville, MD. Maddux, J.F., Desmond, D.P., 1996. Ethnic matching of caseworker and patient in methadone maintenance. J. Subst. Abuse Treat. 13, 233– 239.

Mammo, A., Weinbaum, D.F., 1993. Some factors that influence dropping out from outpatient alcoholism treatment facilities. J. Stud. Alcohol 54, 91 – 101. McCance-Katz, E.F., Carroll, K.M., Rounsaville, B.J., 1999. Gender differences in treatment-seeking cocaine abusers — implications for treatment and prognosis. Am. J. Addict. 8, 300–311. McKay, J.R., McLellan, A.T., Alterman, A.I., Cacciola, J.S., Rutherford, M.J., O’Brien, C.P., 1998. Predictors of participation in aftercare sessions and self-help groups following completion of intensive outpatient treatment for substance abuse. J. Stud. Alcohol 59, 152– 162. McLellan, A.T., Luborsky, L., Woody, G.E., O’Brien, C.P., 1980. An improved diagnostic evaluation instrument for substance abuse patients: the Addiction Severity Index. J. Nerv. Ment. Disord. 168, 26 – 33. McLellan, A.T., Luborski, L., Woody, G.E., O’Brien, C.P., Druley, K.A., 1983. Predicting response to alcohol and drug abuse treatments: role of psychiatric severity. Arch. Gen. Psychiatry 40, 620– 625. McLellan, A.T., Luborsky, L., Cacciola, J., Griffith, J., Evans, F., Barr, H.L., O’Brien, C.P., 1985. New data from the Addiction Severity Index: reliability and validity in three centers. J. Nerv. Ment. Disord. 173, 412– 423. Miller, B., Downs, W., Gondoli, D., 1989. Spousal violence among alcholic women as compared to a random household sample of women. J. Stud. Alcohol 50, 533– 540. Miller, B., Downs, W., Testa, M., 1993. Interrelationships between victimization experiences and women’s alcohol use. J. Stud. Alcohol Suppl. No. 11, 109– 117. Moras, K., 1998. Behavioral therapies for female drug users: an efficacy-focused review. In: Wetherington, C.L., Roman, A.B. (Eds.), Drug Addiction Research and the Health of Women, NIH Pub. No. 98-4290. National Institute on Drug Abuse, Rockville, MD, pp. 197– 222. Office of Applied Studies, 1997. Uniform Facility Data Set (UFDS): Data from 1996 and 1980– 1996, Rockville, MD: Department of Health and Human Services. Onken, L.S., Blaine, J.D., Boren, J.J., 1997. Treatment for drug addiction: it won’t work if they don’t receive it. In: Onken, L.S., Blaine, J.D., Boren, J.J. (Eds.), Beyond the Therapeutic Alliance: Keeping the Drug-Dependent Individual in Treatment, vol. 165. National Institute on Drug Abuse, Rockville, MD, pp. 1 – 3. Polich, J.M., Armor, D., Braiker, H.B., 1981. The Course of Alcoholism: Four Years After Treatment. Wiley, New York, NY. Rosenheck, R., Seibyl, C.L., 1998. Participation and outcome in a resiential treatment and work therapy program for addictive disorders: the effects of race. Am. J. Psychiatry 155, 1029–1034. Rouse, B.A., Carter, J.H., Rodriguez-Andrew, S., 1995. Race/ethnicity and other sociocultural influences on alcoholism treatment for women. In: Galanter, M. (Ed.), Recent Developments in Alcoholism: Women and Alcoholism, vol. 12. Plenum Press, New York, pp. 343– 367. Sanders-Phillips, K., 1998. Factors influencing health behaviors and drug abuse among low-income Black and Latino women. In: Wetherington, C.L., Roman, A.B. (Eds.), Drug Addiction Research and the Health of Women, NIH Pub. No. 98-4290. National Institute on Drug Abuse, Rockville, MD, pp. 197–222. Saulnier, C.F., 1996. African-american women in an alcohol intervention group: addressing personal and political problems. Subst. Use Misuse 31, 1259– 1278. Sue, S., 1988. Psychotherapeutic services for ethnic minorities: two decades of research findings. Am. Psychol. 43, 301– 309. Vannicelli, M., 1984. Treatment outcome of alcoholic women: the state of the art in relation to sex bias and expectancy effects. In: Wilsnack, S.C., Beckman, L.J. (Eds.), Alcohol Problems in Women: Antecedents, Consequences and Intervention. Guilford, New York, pp. 369– 412.

M.E. McCaul et al. / Drug and Alcohol Dependence 62 (2001) 9–17 Westermeyer, J., 1984. The role of ethnicity in substance abuse. In: Stimmel, B. (Ed.), Cultural and Sociological Aspects of Alcoholism and Substance Abuse. Haworth Press, New York, NY, pp. 9–18.

.

17

Wickizer, T., Maynard, C., Atherly, A., Frederick, M., Koepsell, T., Drupski, A., Stark, K., 1994. Completion rates of clients discharged from drug and alcohol treatment programs in Washington State. Am. J. Public Health 84, 215– 221.