More than 25 years of clinical experience and numerous research studies have shown that methadone maintenance programs are effective in the treatment of opiate addiction and provide many additional benefits beyond a decrease in illicit opiate use R15601BABCAFGD. However, methadone maintenance programs across the country vary widely in the level of ancillary services provided to patients, and these differences have a significant impact on treatment outcome R15601BABBGEHJ. In the past decade, the spread of HIV among injection-opiate users R15601BABBAJHG, as well as the prevalence of cocaine abuse and psychiatric disorders in this population R15601BABCIHEA, has underscored the need for additional services in methadone maintenance programs. However, the questions of what type and intensity of services are optimal and what is the most cost-effective way to provide these services remain open.
One study in this area R15601BABBIFBB seems to suggest a clear "dose-response" relation in treatments provided to methadone-maintained patients. That study compared three intensities of outpatient psychosocial treatments in a Department of Veterans Affairs (VA)-based methadone maintenance program and found that patients receiving the highest-intensity treatment, which included access to on-site medical/psychiatric care, family therapy, and employment counseling, had the best outcomes. While these findings suggest that providing more intensive psychosocial services in methadone maintenance programs results in better outcomes—a conclusion that agrees with clinical intuition—a cost analysis of that study R15601BABEAHDD revealed that this level of treatment was not the most cost-effective, indicating a possible divergence of clinical efficacy and cost-effectiveness.
To date, no study has evaluated psychosocial interventions in methadone maintenance programs that are in the upper range of the treatment intensity spectrum (i.e., American Society of Addiction Medicine level II services R15601BABBEADA. Previous research (e.g., reference R15601BABBIFBB has focused on interventions in the low to middle range of the spectrum, from basic methadone services to enhanced services. Thus, the question of what level of outpatient psychosocial services in methadone maintenance programs is optimal, in terms of both clinical outcome and cost, cannot at present be answered. Given the centrality of methadone maintenance programs in treating opiate addiction and promoting psychosocial rehabilitation, this is an important issue to address.
The current study attempted to bridge this gap by comparing an intensive day treatment program with an enhanced standard methadone maintenance treatment program. The lower-level treatment, enhanced standard care, was similar in intensity to the highest-level intervention investigated previously by McLellan and colleagues R15601BABBIFBB. In addition to standard methadone maintenance, enhanced services included a weekly cognitive-behavioral coping skills training group R15601BABCJAAA that covered the same thematic components provided in the day treatment program, but at a lower intensity, together with referrals to on- and off-site services. The primary goals of the study were to determine whether an intensive day treatment program was more effective than enhanced methadone maintenance treatment, to calculate the relative costs of each level of services, and to determine whether it was possible to identify subgroups of patients in methadone maintenance programs who differentially benefited from the two treatment intensities.
Study participants were recruited from a community-based methadone maintenance program. Among 308 eligible individuals, 291 began treatment; 242 were newly entering methadone maintenance treatment (60 were "methadone naive" and 182 had a history of methadone maintenance), and 49 were currently receiving methadone and were referred in lieu of detoxification because of unremitting drug use. The mean age of the 291 participants was 36.8 years (SD=6.9); 205 (70%) were male, and 86 (30%) were female; 171 (59%) were white, 83 (29%) were African American, 36 (12%) were Hispanic, and one (0.3%) was Native American. All were unemployed; 183 (63%) had completed high school. The participants had been using opiates for a mean of 12.7 years (SD=8.3). Opiates were administered intravenously by 214 (74%) of the subjects and intranasally by 77 (26%). Ninety percent (N=263) also reported using cocaine, and they had been using it for an average of 8.9 years (SD=7.1). Forty-nine percent (N=130) of the cocaine users used it intravenously, 27% (N=72) used the free-base form, and 23% (N=61) used it intranasally. Of the entire study group, 46% (N=134) met the DSM-III-R criteria for cocaine dependence; 21% (N=62) had a history of cocaine dependence, and 5% (N=16) met the criteria for cocaine abuse. Twenty-two percent met the DSM-III-R criteria for antisocial personality disorder, 25% met the criteria for current major depression, and 32% had a history of major depression. After complete description of the study to the subjects, written informed consent was obtained.
Twice-weekly urine samples were analyzed for the presence of opiates (≥200 ng/ml) and the cocaine metabolite benzoylecgonine (≥300 ng/ml). The Treatment Services Review R15601BABBDIBI was administered monthly. The assessment battery, administered before treatment, at the end of 12 weeks of treatment, and 6 months after completion of treatment, included the Addiction Severity Index R15601BABCACCH and the HIV Risk Assessment Battery R15601BABDFAIG.
All patients received daily methadone, to a target dose of 85 mg/day, and case management. By means of urn randomization, patients were assigned to one of two 12-week treatment programs: an enhanced methadone program (N=146) or a high-intensity day treatment program (N=145). Enhanced standard care included 2 hours per week of a manual-guided, cognitive-behavioral group intervention, which used nine of the 27 sessions described in the manual developed by Monti et al. R15601BABCJAAA, and three additional groups on physical health, vocational skills, and community resources, together with referral to other services. The day treatment program was a 5-hours-per-day, 5-days-per-week, manual-guided program, which was facilitated by the same master’s-level therapists who provided the enhanced standard care and which provided groups in five general areas: 1) substance abuse treatment, 2) physical and emotional health, 3) community development, 4) development of alternative reinforcers, and 5) basic daily living skills. Receipt of the intended level of services was verified through attendance records and patients’ charts.
The Kaplan-Meier method and log rank test were used to compare the survival time to dropout. Data from urine tests were analyzed with the generalized estimating equation method and the Wald statistic, as described by Liang and Zeger R15601BABBFHFJ, with the specification of a logit link function, binomial error, and exchangeable working correlation structure. The generalized estimating equation method uses all available data without imputation of missing data. Using retention in treatment and urine screen data as outcomes, we entered eight patient characteristics into the model as covariates: sex, race, severity of psychiatric disorder, addiction severity, diagnosis of cocaine dependence, diagnosis of antisocial personality disorder, methadone history, and treatment intensity preference. Change in drug use status (i.e., from abstinence to relapse) from the end of treatment to 6-month follow-up was analyzed with the use of McNemar’s test R15601BABCGCDE. Changes in scores on the Addiction Severity Index and the HIV Risk Assessment Battery were analyzed in a 2 (treatment intensity) × 3 (time: pretreatment, posttreatment, follow-up) mixed analysis of variance (ANOVA). Significant main effects were followed by 2 (treatment intensity) × 2 (time: pretreatment to posttreatment, pretreatment to follow-up, posttreatment to follow-up) ANOVAs in which all available data for each pair of time points were used. An alpha level of 0.05 was chosen for determining significant differences in primary outcome between the two treatment intensities; Bonferroni corrections were made for all other comparisons.
The approach taken to estimate costs involved multiplying the number of units of service utilized by each patient during the 12-week program by the estimated unit costs, adjusted to 1995 dollars to represent the middle year of the 3-year treatment phase of the study. Only participants who completed treatment were included in the cost analysis because complete Treatment Services Review data, which constituted the basis for secondary cost estimates, were available for these patients but not for patients who dropped out.
Program costs comprised all costs directly related to providing the two intensities of treatment, including methadone, urine screens, medical care, counseling, group therapy, and other related expenses. The unit cost of these services included salaries, fringe benefits, and overhead (e.g., the unit cost of methadone includes the costs of methadone, dispensing, staff, and overhead). The overhead rate was calculated by determining the indirect/direct cost ratio in 1995 at the clinic where the study was conducted and is included in each unit cost. Clinicians’ hourly rate was estimated as detailed by Rosenheck et al. R15601BABCCBJI. Utilization of on-site services was derived from attendance records and patient charts.
A number of secondary costs potentially affected by participation in the study were derived from published literature: unit costs for off-site health care (including hospitalization, emergency room visits, and medical and mental health outpatient visits) R15601BABDCCEB, criminal justice R15601BABDHFBG, vocational and legal services, and the administrative cost of transfer payments R15601BABBAGIH. Productivity from earnings was based on the Addiction Severity Index data as a negative cost. Utilization of secondary services was based on self-report data provided in the Treatment Services Review.
Student’s t tests, with the alpha level set at <0.02 after Bonferroni correction, were conducted to compare the two levels of treatment intensity. Program costs and secondary costs were grouped separately because primary cost and utilization data were based on internal accounting documents and attendance records, whereas the secondary cost data were derived from national cost averages and patients’ self-reports.
Eighty-one percent (N=237) of the 291 patients completed the 12-week program; 82% (N=119 of 145) completed the day treatment program, and 81% (N=118 of 146) completed enhanced standard care, with no difference in survival time (day treatment program: mean=10.9 weeks, SD=2.7; enhanced standard care: mean=10.7 weeks, SD=2.9; log rank=0.17, p=0.68). Nine dropouts were transferred to inpatient drug treatment, six were incarcerated, four were hospitalized for medical reasons, two requested detoxification, and the rest were administratively discharged for missing three consecutive methadone doses. The patients who completed treatment did not differ from the treatment dropouts in demographic characteristics, psychiatric diagnosis, or drug use history with one exception: patients with a diagnosis of antisocial personality disorder were less likely to be retained, regardless of assigned treatment intensity (χ2=7.99, df=1, p<0.005).
There was no significant difference between the day treatment program and enhanced standard care in methadone dose (day treatment program: mean=81.7 mg/day, SD=17.0; enhanced standard care: mean=78.1 mg/day, SD=14.9; t=1.75, df=232, p=0.08); in the number of days methadone was received out of 84 possible days (day treatment program: mean=81.7 days, SD=3.2; enhanced standard care: mean=81.5 days, SD=3.4; t=0.46, df=234, p=0.64); or in hours of case management (day treatment program: mean=6.51 hours, SD=3.18; enhanced standard care: mean=6.02 hours, SD=2.70; t=1.28, df=230, p=0.20). Patients who completed the day treatment program attended a mean of 191.54 hours (SD=56.28) of day treatment, and patients completing enhanced standard care attended a mean of 17.31 hours (SD=4.86) of skills training groups (t=33.60, df=120, p<0.0001). A multivariate analysis of variance (MANOVA) conducted on the seven Treatment Services Review variables revealed a main effect for treatment intensity: patients assigned to the day treatment program received more services overall than patients assigned to enhanced standard care (F=72.1, df=7, 201, p<0.0001). Subsequent analyses of these seven variables with the use of a significance level of 0.007 after Bonferroni correction revealed that patients assigned to the day treatment program received significantly more medical services (t=8.66, df=172, p<0.001), employment services (t=11.78, df=145, p<0.001), psychological services (t=16.39, df=121, p<0.001), and drug treatment services (t=15.49, df=155, p<0.0001) in the methadone program. Outside the program, day treatment program patients received significantly more medical services (t=2.73, df=207, p<0.007).
As shown in T1, there were no differences between groups in rates of opiate- and cocaine-free urine screens during treatment or at follow-up. There was a significant increase across weeks in the percentage of patients who were abstinent from opiates, from cocaine, or from both drugs (p values <0.001). Follow-up urine data were available for 74% of the patients who completed the 12-week treatment program (75% in the day treatment program; 73% in enhanced standard care). At the follow-up interview, 71% (N=75) of the day treatment program patients and 67% (N=64) of the enhanced standard care patients continued to receive standard methadone maintenance, which included drug counseling and random urine testing. Abstinence status did not change significantly from the end of treatment to 6-month follow-up (McNemar’s test: for cocaine, χ2=2.27, df=1, p=0.13; for opiates, χ2=0.39, df=1, p=0.54); 65% of the 85 patients available for follow-up who were abstinent from opiates and cocaine at treatment completion were still abstinent at the 6-month follow-up.
A series of 2 × 3 mixed ANOVAs conducted on Addiction Severity Index and HIV Risk Assessment Battery composite scores revealed no main effects of treatment intensity and no significant treatment intensity-by-time interactions. However, main effects of time were found on two of the seven Addiction Severity Index scores at the 0.007 alpha level with the Bonferroni correction: severity of drug abuse (F=61.62, df=2, 398, p<0.0001) and severity of legal problems (F=26.92, df=2, 398, p<0.001). Both problem areas decreased from pretreatment to posttreatment, with no significant change from posttreatment to 6-month follow-up. Changes on the remaining five Addiction Severity Index scores failed to reach statistical significance (employment problems: F=3.46, p=0.03; medical problems: F=0.26, p=0.77; family/social problems: F=1.36, p=0.26; psychiatric problems: F=0.64, p=0.52; alcohol problems: F=0.84, p=0.43; all df=2, 398).
On the three Risk Assessment Battery composite scores, with a corrected alpha level of <0.02, there were main effects of time on the overall Risk Behaviors Index composite score (F=22.74, df=2, 370, p<0.001), the Drug Risk Index composite score (F=26.03, df=2, 370, p<0.001), and the Sexual Risk Index composite score (F=8.79, df=2, 370, p<0.001). Drug and sexual high-risk behaviors decreased significantly from pretreatment to posttreatment, with further reductions in the overall risk index and drug risk behaviors from posttreatment to 6-month follow-up (overall risk index: F=5.16, df=1, 185, p<0.02; drug risk index: F=11.59, df=1, 185, p<0.001).
At treatment completion, 10.1% (N=12) of the 119 patients who completed the day treatment program and 14.9% (N=17) of the 118 patients who completed enhanced standard care had gained employment, a nonsignificant difference (χ2=1.04, df=1, p=0.33). The difference in number of hours engaged in work or seeking work during the 12-week treatment phase also failed to reach statistical significance (enhanced standard care: mean=71.2 hours, SD=120.7; day treatment program: mean=43.1 hours, SD=81.8; t=1.99, df=211, p=0.05).
A MANOVA conducted on the eight treatment satisfaction ratings revealed a significant interaction between satisfaction and treatment intensity (F=2.56, df=7, 1568, p=0.01). Subsequent univariate analyses, with alpha at <0.006, revealed that patients in the day treatment program were significantly more satisfied with two of eight dimensions: employment counseling (t=2.87, df=219, p=0.005) and leisure time activities (t=2.46, df=221, p=0.01). Differences between the two treatment intensities in satisfaction with the remaining six dimensions failed to reach statistical significance (satisfaction with methadone: t=0.09, df=224, p=0.92; with quality of groups: t=0.44, df=224, p=0.66; with relapse prevention: t=0.52, df=224, p=0.61; with help for medical problems: t=0.30, df=157, p=0.76; with help for family problems: t=0.29, df=222, p=0.77; with community referrals: t=2.87, df=210, p=0.03).
Costs of Each Treatment Intensity
Because the two treatments compared in this study were equivalent in overall outcome, it was immediately clear, without formal cost-effectiveness analysis, that the lower-intensity treatment was the more cost-effective. However, in order to provide specific cost data for the two treatment intensities, we calculated the total costs per patient for each of the two 12-week interventions. As shown in T2, total program costs were 3.1 times greater for the day treatment program than for enhanced standard care. This represents a per-patient per-day program cost of $55 for the day treatment program and $18 for enhanced standard care. The inclusion of secondary costs in total costs reduced the cost ratio of the day treatment program to enhanced standard care to 2.7.
Analysis by Patient Subgroup
Eight patient characteristics were selected for purposes of the current analysis: sex, race, diagnosis of cocaine dependence, diagnosis of antisocial personality disorder, severity of psychiatric disorder, addiction severity, history of methadone maintenance, and treatment intensity preference. Before random assignment to treatment, the majority (78%, N=228) of the patients reported that if they were to be given the choice, they would prefer to receive enhanced standard care; however, treatment preference did not predict differential response to the intensity of the treatment actually received. One patient characteristic predicted response to treatment (corrected alpha=0.006): patients with no prior history of methadone maintenance treatment were more likely to be retained in treatment if assigned to the lower-intensity enhanced standard care (13% dropout rate) than if assigned to the day treatment program (40% dropout rate) (p=0.01, Breslow-Day test), and at the 6-month follow-up, they were significantly more likely to provide an opiate-free urine sample if they had received the lower-intensity intervention (enhanced standard care: 71% provided opiate-free urine; day treatment program: 33% opiate-free) (χ2=5.66, df=1, p=0.02). Generalized estimating equation analysis of the urine data during the 12 weeks of treatment found no significant interactions between treatment intensity and any of the selected patient characteristics.
In this controlled investigation of two randomly assigned intensities of psychosocial intervention provided at a community-based, inner-city methadone maintenance program, we found that an intensive day treatment program yielded no greater benefit with respect to drug abstinence initiation and maintenance, resolution of drug-related problems, or reduction of HIV risk behaviors than a lower-intensity intervention provided at approximately one-third the cost of the day treatment program. Relative to pretreatment measures, clinical improvement was found for patients receiving either treatment intensity. Substantial reductions in drug use, in severity of addiction-related problems, and in prevalence of HIV risk behaviors were found. These improvements were maintained by the majority of the patients 6 months after completion of the study. Although patient satisfaction ratings after treatment tended to be higher for the day treatment program than for enhanced standard care, there was no evidence from the post hoc matching analysis that assignment to treatment intensity on the basis of patient preference at entry into treatment would have changed the clinical outcome.
The finding of no overall difference in outcome between the two intensities of treatment is in general agreement with findings from several large-scale studies in the addictions R15601BABBDDHH–R15601BABDCEBJ, suggesting a surprising lack of agreement between conventional clinical wisdom that "more treatment is better" and results of clinical trials R15601BABDFIAA. However, there are several important issues to be considered when interpreting results of these studies. Commentators on the conclusion drawn from the Fort Bragg Demonstration Project data that "more is not necessarily better" R15601BABCDFJH have asked, "More of what?"R15601BABBBFBB and "Relative to what?" R15601BABDHBHB. These are also valid questions to consider when interpreting the present data.
Before acceptance of a study’s conclusion that more intensive services are unnecessary, it must be established that the intended level of services was actually provided to, and received by, the study’s participants. Previous research suggests that drug users underutilize available services R15601BABBIFBB, R15601BABECEDC, making it difficult to draw conclusions regarding the optimal level of services. In the current study, we ensured that patients in the high-intensity treatment actually received the intended intensity of services and more of specific types of services (e.g., more medical, employment, psychological, and drug treatment services). However, patients assigned to this intensity of treatment received these services in the absence of a formal assessment of need. The finding of no overall benefit for this level of services suggests that providing more of something that the patient may not necessarily need is neither therapeutic nor an efficient use of societal resources. Furthermore, attending a day treatment program in a community-based methadone maintenance program may provide not only more unnecessary and costly services but also more exposure to drug-related cues and increased opportunity for fraternization with other drug users, which may promote or sustain drug use. Our data provide some evidence for the interpretation that more treatment may even be countertherapeutic for certain subgroups of patients. Patients with no prior history of methadone maintenance were more likely to be retained in treatment and were more likely to be abstinent at follow-up if they had received the lower-intensity intervention (see also reference R15601BABBAHGG.
In the current study, patients assigned to the lower-intensity intervention received a level of services previously shown to be superior to standard care R15601BABBIFBB—a manual-guided intervention provided by trained professionals, case management, and referral to on- and off-site services. Thus, the current study compared two active, effective treatments. Interpreting null results in studies using "active treatment" comparisons is complex and potentially misleading R15601BABDHBHB, highlighting issues of internal and external validity. In the current study, internal validity was addressed in several ways. Patients were randomly assigned to treatments, there was differential receipt of services in the two treatment intensities, and the treatments were delivered as specified in their respective manuals. Our primary outcome, illicit drug use, was based on an objective test, urine toxicology screens administered twice weekly for patients in both treatment intensities, making it highly unlikely that any instance of illicit drug use would go undetected. Maintenance of treatment effects was assessed in a follow-up interview, which included a urine toxicology screen, conducted 9 months after intake. The study had ample statistical power (90%) to detect a clinically meaningful difference (20%) relative to our objective outcome measure of drug use. Issues of external validity were addressed by conducting this study in an inner-city, community-based methadone maintenance program, with a sexually and racially heterogeneous patient sample. Our results are consistent with previous findings of the clinical efficacy of enhanced methadone services in an all-male, VA-based methadone maintenance program R15601BABBIFBB, suggesting the generalizability of our clinical findings.
We recognize, however, that it may be problematic to compare results of the cost analyses conducted in the current study to those conducted with respect to the VA study, because of possible differences in cost estimation methods as well as potential differences between the cost structures of VA-based clinics and community-based clinics R15601BABDABCD. Although our findings suggest that enhanced methadone services are more cost-effective than a day treatment program, we cannot conclude that enhanced standard care is the most cost-effective level of service in a community-based clinic, because we did not include an even lower intensity of treatment as a comparison. Whether less treatment is even more cost-effective in a community-based clinic, as was found with a VA sample R15601BABEAHDD, awaits further investigation. It may also be important to include societal costs of different levels of service in methadone programs in future studies, as we attempted to do in the current study, because these may interact in complex ways with on-site methadone program costs, either to augment or to reduce total treatment costs (compare with reference R15601BABBCJGE. For example, we found that including some societal costs in total treatment costs reduced the cost ratio of the day treatment program in relation to enhanced standard care from 3.1 to 2.7. It is possible that including a broader range of societal costs that could be directly linked to treatment participation would significantly affect the cost ratios of different treatment intensities, which may, in turn, have important policy implications R15601BABCADJJ.
The limitations of the study—which include loss of 17 randomly assigned patients before treatment began, loss of urine data on 26% of the participants at the 6-month follow-up, and the potential for insufficient power to detect significant interactions between treatment intensity and certain of the selected patient characteristics—do not limit the interpretation of the primary finding of no significant difference between the two treatment intensities in either retention in treatment or drug use during treatment.
The current findings, in conjunction with findings from the study conducted by McLellan et al. R15601BABBIFBB, suggest the following. Methadone maintenance with adjunctive psychosocial services produces clinically and socially meaningful outcomes in a traditionally difficult-to-treat, inner-city, opioid-dependent patient population at relatively low cost. The optimal treatment approach for the majority of patients entering methadone maintenance treatment is a thorough assessment of individual patient need, provision of a weekly psychosocial intervention, and referral to on- and off-site services, with follow-up to ensure that the necessary services are received. Only those patients identified as not benefiting from this level of treatment need to be referred to more intensive services. Future research is needed to investigate which components (e.g., group or individual counseling, on- or off-site referral resources) contribute to the effectiveness of enhanced methadone services.
Received Jan. 7, 1998; revision received June 8, 1998; accepted Aug. 4, 1998. From the Yale University School of Medicine; the VA Medical Center, West Haven, Conn.; and the National Institute on Drug Abuse, Rockville, Md. Address reprint requests to Dr. Avants, Yale Substance Abuse Center, 34 Park St., New Haven, CT 06519; firstname.lastname@example.org (e-mail). Supported by National Institute on Drug Abuse grants DA-08754 and DA-00277 (to Dr. Avants), DA-00112 (to Dr. Kosten), and DA-00089 (to Dr. Rounsaville). The authors thank patients and staff of the APT Foundation, Inc., for their contributions to this project.