This is the first report of pre-treatment and early treatment characteristics associated with treatment attrition in youth treated with Bup/Nal. In youth receiving 12 weeks of Bup/Nal, early adherence to prescribed Bup/Nal was associated with retention and an early non-abstinence indicator, an opioid positive urine, was associated with attrition. A few pre-treatment characteristics, use of prescribed and over the counter medications of any kind in the month prior to treatment, and lifetime non-heroin opioid abuse were associated with retention while prior 30 day hallucinogen use was associated with attrition. In youth receiving 2 weeks of Bup/Nal, only medications for sleep were associated with retention, although this was not an independent predictor. What is striking, however, is that characteristics such as gender, injection drug use, the presence of hepatitis C, current use of heroin compared with non-heroin opioids, concurrent use of other drugs and alcohol, and comorbid psychiatric symptoms were not associated with attrition in either group. Extended use of Bup/Nal appears to help improve retention in youth with a wide range of characteristics, at least in this sample. Medication dose and withdrawal symptoms in the first 2 weeks were also not related to attrition.
The timing of increased dropout in the DETOX group corresponded with the end of the Bup/Nal taper at day 14, with 8% leaving before week 2 and 36% leaving between weeks 2 and 4. In the BUP group dropout was only 8% in the first 4 weeks. These newly reported data about the timing of attrition in the first few weeks in DETOX provides additional evidence that continuing Bup/Nal for a longer time is a key determinant of treatment retention (Woody et al., 2008
), as in adults (Sees et al., 2000
, Katz et al., 2009
An early opioid positive urine test was associated with attrition in opioid addicted adults (Stein et al., 2005
), and an early cocaine positive urine with attrition in cocaine addicted adults (Kampman et al., 2001
), consistent with findings in this sample. This marker may reflect low motivation for treatment or inadequate dosing (Stein et al., 2005
), although mean maximum dosing in BUP was within the usual dosing range and consistent with study guidelines at 14.6 mg/day in the first 2 weeks. An early opioid positive urine was also associated with worse treatment outcomes (i.e., opioid positive urines at 12 weeks) in this sample (Subramaniam et al., 2011
), further highlighting the importance of this marker in identifying youth at risk for both attrition and poorer drug use outcomes.
Lower adherence to counseling visits was a significant, although not independent predictor of attrition in our sample, similar to findings in adults (Stein et al., 2005
). Adherence to counseling or medication use may directly impact retention or adherence to either may be an indicator of future adherence as well.
Receipt of at least one prescribed or over the counter medication prior to treatment entry was associated with retention in the BUP group, although no specific category of medications, such as those for sleep or pain, was identified as related. Possible explanations are that entry into treatment with generally reduced or resolved symptoms may impact retention by limiting symptom discomfort, or that prior medication use reflects a history of adherence and suggests continued adherence.
Youth in the BUP group who reported lifetime abuse of non-heroin opioids (with or without lifetime heroin use) were less likely to drop out compared with those with lifetime use of heroin and/or any other drug. This is an intriguing finding given the alarming increase in the non-medical use of non-heroin opioids and related emergency care in youth (Drug Abuse Warning Network, 2011
; Johnston et al., 2011
). Lifetime non-heroin opioid abusers may represent a different population that samples these drugs but adheres to treatment. However, there were no differences in dropout between (1) past 30-day abusers of non-heroin opioids (with or without past 30-day heroin use) compared to past 30-day abusers of heroin, or (2) past 30 day or lifetime abusers of only non-heroin opioids (as their choice of opioids) compared with heroin only abusers, consistent with a previous report that found no difference in attrition between adolescents using non-heroin or heroin opioids (Motamed et al., 2008
). Further assessment of the characteristics and outcomes of young users of non-heroin opioids will be important.
The finding that hallucinogen use in the 30 days prior to treatment entry was associated with dropout in the BUP group may be an artifact of the small number of hallucinogen users (n=7).
Based on the bivariate analyses, only medications for sleep assisted with retention during treatment in DETOX, suggesting that offering relief from sleep disturbance, possibly as a lingering symptom of withdrawal, helps youth remain in psychosocial treatment when only very short term Bup/Nal is available.
We found that co-occurring psychiatric symptoms,as measured by the YSR and YASR, were unrelated to attrition, although higher rates of depression were associated with retention (Gerra et al., 2004
) and more severe depression and psychopathology were associated with attrition (Pani et al., 2000
) in adults with opioid dependence. However, the YSR and YASR provide only general measures of internalizing or externalizing symptoms, and do not indicate the presence of a specific psychiatric disorder, which was not assessed in this study. Early withdrawal symptoms were also not associated with attrition in this analysis, consistent with some (Scherbaum, Heppkausen, & Rist, 2004
), but not all studies in adults (Soyka et al., 2008
). Maximum dose of Bup/Nal in the first 2 weeks or number of days at this dose were not associated with attrition in either group, similar to findings in adults treated with comparable mean daily doses of 9–12 mg (Soyka et al., 2008
; Gerra et al., 2004
; Vigezzi et al., 2006
). In one observational study with adults, initial induction doses of 16 mg were associated with better retention (Leonardi et al., 2008
). However, this study may not have had enough power to identify such an effect in youth. It may also be that dosing was adequate enough to reduce the discomfort of withdrawal and therefore not impact attrition.
Attrition is a very serious problem in the treatment of opioid dependent adolescents. The limited number of pre-treatment characteristics identified as associated with attrition highlights the importance of comprehensively measuring other factors that could have a meaningful impact on retention such as those related to family support, housing or transportation, in the search for areas in which clinicians and researchers can intervene to improve retention.
The main study (Woody et al., 2008
) had sufficient power to detect differences in the primary outcome of opioid positive urines at 4, 8, and 12 weeks. If the sample sizes had been larger we would have had more power to detect factors associated with attrition in both groups. Given the number of dropouts and non-dropouts in each group, we were limited in the number of variables that could be included in analyses of independent predictors of attrition and could have missed other potentially significant predictors. While findings in the multivariate models are consistent with bivariate analyses, actual odds ratios may have been sensitive to small sample sizes. Other factors, such as motivation to receive treatment or intent to remain in treatment may be associated with attrition; however, data about these factors were not collected. Finally, all findings must be confirmed in other youth samples.
In summary, early evidence of medication adherence and opioid abstinence was strongly predictive of treatment retention, reinforcing the need to aggressively target those with early non-adherence to medication or an early opioid positive urine drug screen to further improve treatment retention and outcome. It is salient that there were few notable findings among a large group of possible pre-treatment characteristics that could be related to attrition, other than assignment to brief or extended administration of Bup/Nal. In this sample extended Bup/Nal treatment was effective in improving treatment retention for youth with opioid dependence across a wide range of demographic characteristics, pre-treatment clinical severity, current use of heroin or non-heroin opioids, concurrent abuse of a variety of other drugs and alcohol, the presence or absence of comorbid psychiatric symptoms or hepatitis C infection, or concurrent medication use. If replicated this finding could be of considerable importance to clinicians. However, other factors that may be barriers to retention, such as housing, transportation, and family support of treatment should be measured and further investigated as well.