Of the 124 individuals screened, 20 participants either failed to meet study eligibility or failed to show up for randomization, and 10 male partners were eliminated from analyses, leaving a total of 94 study participants included in these analyses. The number of randomized participants included 28 in OTP, 33 in PCS, and 33 in MMM.
The final sample included more men (58%) than women (42%), and most were white (58%) or Hispanic (28%). shows baseline demographic characteristics and related variables by treatment site. At baseline, participants in the three settings had a significantly different distribution for race (chi-square = 12.83; p = 0.01), with 24% Black (non-Hispanic) participants in the PCS site, whereas only 7% and 6% of participants were Black (non-Hispanic) in the OTP and MMM sites respectively. Similarly, 50% of the participants self-reported as “other” race/ethnicity (American Indian, Alaskan Native, Asian or Hispanics) in the OTP site but only 18% and 24% belonged to that category in the PCS and MMM sites, respectively.
Demographic characteristics and associated variables by treatment site.
shows baseline drug use characteristics. There were no significant baseline differences found among the participants groups across treatment sites.
shows that at the end of 9 weeks there was no significant difference across treatment sites in the number of opioid-negative urine tests as measured by the TES (F=1.96; p = 0.15). Similarly, there was no difference among the treatment sites in TES at the end of 20 weeks (F = 2.64; p = 0.08).
Main Study Findings by Treatment Site
Significant differences were found for mean dose prescribed to participants by treatment site (F = 5.91;p = 0.00). shows that participants at the PCS site were prescribed significantly lower doses than those at both OTP and MMM sites.
A significant association was found between TES and prescribed dose. The correlation between the TES at 9 weeks and mean dose over the entire study duration was −0.40 (p = 0.00), and the correlation between the TES at 20 weeks and mean dose over the entire study duration was −0.41 (p = 0.00), indicating that higher dosage is associated with a lower percentage of opioid-negative urine test results.
Retention was analyzed in two ways. One method compared the proportion of participants at each treatment site who stayed until the end of the two target time periods of 9 and 20 weeks. The second method computed the number of weeks in treatment at each site before drop out.
provides the results of analyses comparing the number of participants who stayed in the study through weeks 9 and 20. No difference was found in retention through Week 9 by treatment site (chi-square = 1.86; p = 0.39), however the proportion of participants who stayed in the study through Week 20 was significantly associated with treatment site (chi-square = 6.12; p = 0.05) with the MMM site associated with the highest percentage of participants retained through week 20 (51.5%). also shows the mean number of weeks that participants in each treatment condition remained in treatment, differences that were not statistically significant. Comparing the number of weeks in the treatment-by-treatment condition by using proportional hazards model shows that for participants who remained in the study past 9 weeks, OTP participants had a 4 times higher drop-out rate compared to MMM participants (p = 0.01), and a 6 times higher drop-out rate compared to PCS participants (p = 0.01). There was no difference in the percentage of participants at the PCS and MMM sites who remained in treatment for more than 9 weeks.
The number of weeks a participant remained in the study was significantly associated with opioid use as measured by the TES at both 9 weeks (r = 0.48; p < 0.00) and 20 weeks (r = 0.58;p < 0.00). That is, a higher percentage of opioid-negative urine test results was associated with longer treatment retention at both 9 and 20 weeks. After controlling for treatment site, a significant association remained between the TES at 9 weeks and number of weeks retained in the study (F = 10.17; p < 0.00). The results were similar for the TES at 20 weeks and the number of weeks retained in the study (F = 17.48; p < 0.00).
Addressing treatment site, the TES at 9 weeks was significantly associated with number of weeks in the study for the PCS (r = 0.43;p = 0.01) and MMM sites (r = 0.59;p = 0.00), but not the OTP site (r = 0.29; p = 0.12). The TES at 20 weeks, however, was associated with number of weeks retained in the study for all three treatment sites, OTP (r = 0.42; p = 0.03), PCS (r = 0.50; p = 0.00), and MMM (r = 0.72; p < 0.00).
A total of 24 (25%) of the randomized participants completed 52 weeks of treatment. Of the completers, 12 were in MMM, 10 were in PCS, and 2 were in OTP. Follow-up assessments were collected from 26 participants regarding reasons for not completing 52 weeks of treatment. Of the 26, 12 participants reported that they were no longer interested in buprenorphine treatment, with 5 of the 12 reporting that they lost interest on the first day of treatment. Other reasons for termination included: four administrative discharges for failure to keep clinic appointments, two participants developed medical problems unrelated to buprenorphine, two individuals were incarcerated, three participants requested detoxification, one went to the hospital with concurrent psychiatric problems, and two participants tapered off buprenorphine after 98 and 182 days of treatment.
A total of 64 participants received some form of psychosocial counseling during their participation in the study, with individual counseling occurring most often. A small number of participants attended other types of counseling such as group sessions, NA or AA meetings, or AIDS counseling. The importance of psychosocial counseling was examined by analyzing possible differences in the number of sessions attended and the number of minutes attended by participants at each treatment site.
shows that the mean number of counseling sessions attended was not significantly different across the three sites, but the mean number of minutes spent in each individual counseling session was significantly different across treatment site (F = 33.65; p < 0.00). There was a significant correlation between the number of weeks the participant stayed in the study and the mean number of individual counseling sessions attended (r = 0.31; p = 0.02), such that longer retention was associated with a greater number of counseling sessions attended. There was also a significant correlation between the mean number of individual counseling sessions attended and the TES at 20 weeks (r = 0.26; p = 0.05), such that a greater number of individual counseling sessions was associated with a higher percentage of opioid-negative urine test results.
Qualitative Information: Site Experiences and Study Feasibility
At the primary care office setting (PCS), patients were allowed to reschedule missed appointments due to problems with transportation, work schedule and childcare. Trying to accommodate the patients' schedules led to physician and staff frustration. Fitting late patients into the schedule led to a pattern of patients being chronically late. The physician’s office staff was satisfied with conducting the on-site analysis of urine test using the urine cups. The cups allowed the monitoring of the patients’ progress without requiring access to a laboratory.
At the CBT clinic (MMM), which had been a medication-free setting, there was a concern about introducing opioid-dependent patients taking a maintenance medication into the clinic environment. Because not all clinic policies were determined in advance of treatment, staff were troubled that the physician provided prescriptions for buprenorphine even when participants did not attend all the group therapy sessions scheduled. The study patients were not integrated into groups with other substance users but congregated among themselves, causing less of a problem than anticipated, although one prescription opioid user became an injection user in the context of befriending heroin users in the group.
Physicians at the OTP site benefited from established site procedures in treating and monitoring the population. Staff was familiar with administering and interpreting urine tests, as well as performing random callbacks of medication.