Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Jun 1.
Published in final edited form as: Drug Alcohol Depend. 2022 Apr 7;235:109441. doi: 10.1016/j.drugalcdep.2022.109441

INITIATION OF OPIOID AGONIST TREATMENT AND SUBSEQUENT SUBSTANCE USE AND OTHER PATTERNS AMONG ADOLESCENTS AND YOUNG ADULTS IN VANCOUVER, CANADA

Andreas Pilarinos a,b, Danya Fast a,c, Ekaterina Nosova a,c, Yandi Kwa d, Ronald Joe d, Jane A Buxton e, Kora DeBeck a,f
PMCID: PMC9872979  NIHMSID: NIHMS1866794  PMID: 35427979

Abstract

Background

Opioid agonist treatments (OAT) are effective interventions for reducing illicit opioid use; however, less is known about OAT among adolescents and young adults (AYA). This study sought to examine OAT retention and discontinuation among AYA.

Methods

Data were derived from the At-Risk Youth Study, a prospective cohort of street-involved AYA in Vancouver, Canada, between September 2005 and December 2018. Multivariable Cox regression analysis was employed to identify sociodemographic, substance use, and other factors associated with time to first OAT. Substance use, homelessness, service utilization, and criminal justice patterns among AYA who did and did not initiate OAT were examined using before and after analysis.

Results

Of 676 AYA who reported weekly illicit opioid use, 454 (67.2%) reported not being on OAT at some point over the study period and 217 (32.1%) initiated OAT over follow-up. In non-linear growth curve analysis, only participants retained in OAT demonstrated significant reductions in daily illicit opioid use when compared to ‘no OAT’ controls (p < 0.05). Nevertheless, reductions in homelessness (p = 0.070) and increases in difficulty accessing services (p = 0.078) were observed between participants retained in OAT vs. ‘no OAT’ controls, while reductions in homelessness (p = 0.085) and weekly non-medical prescription opioid use (NMPOU) (p = 0.061) were observed between ‘OAT discontinuers’ vs. ‘no OAT’ controls.

Conclusions

Despite finding that OAT retention was associated with significant reductions in daily illicit opioid use, no significant improvements in other key indicators were observed. This underscores the importance of providing supports alongside OAT to improve treatment outcomes among AYA.

Keywords: adolescents, young adults, opioid use disorder, opioid agonist treatment, longitudinal study, before and after analysis

1. INTRODUCTION

In North America, opioid-related overdose has become the leading cause of accidental death (Ahmad et al., 2020, Donroe et al., 2018, Lee and Mannix, 2018). To date, interventions have been expanded in response to the toxic drug crisis, including opioid agonist treatments (OAT) like methadone and buprenorphine-naloxone; however, evidence suggests AYA are less likely to access OAT versus adult populations (Alinsky et al., 2019, Chavez et al., 2020, Feder et al., 2017, Hadland et al., 2017). Similarly, there is a growing body of research demonstrating concerning sociodemographic-, substance use-, health system-, and criminal justice system-related barriers to OAT access that require further investigation in the Canadian context (Pilarinos et al., 2021).

Given the lack of high-quality evidence examining OAT among AYA (Chang et al., 2018, Viera et al., 2020), there have been calls for an improved understanding of OAT initiation, retention and outcomes among AYA (British Columbia Centre on Substance Use, 2018, Camenga et al., 2019). Therefore, this study sought to identify sociodemographic-, substance use-, health system-, and criminal justice-related factors associated with recent OAT among AYA who report recent weekly opioid use. Additionally, we sought to compare before and after substance use, difficulty accessing health services, incarceration, and homelessness patterns following OAT between AYA who: 1) initiated and were retained on OAT and 2) initiated and discontinued OAT, versus AYA who 3) reported opioid use but did not initiate OAT.

2. MATERIAL AND METHODS

2.1. Study procedures

Data were derived from the At-Risk Youth Study (ARYS), an open prospective cohort of street-involved AYA aged 14–26 years at recruitment (Wood et al., 2006). Participants were recruited through snowball sampling and self-referral, and study eligibility criteria included residing in the greater Vancouver area; reporting past-30-day illicit drug use (e.g., heroin, fentanyl, crystal methamphetamine); being “street-involved”, defined as being recently homeless or accessing services intended for homeless youth; and providing written informed consent at enrolment. An interviewer-administered questionnaire collecting information on socio-demographic characteristics, substance use, and health and social service utilization patterns was conducted at baseline, and participants were contacted by phone, mail or through street-based outreach to conduct follow-up interviews at six-month intervals. Participants received a $40 (CAD) honorarium upon each visit completion. Ethical approval for this study was provided by University of British Columbia/Providence Health Care Research Ethics Board.

2.2. Study sample

To compare before and after substance use, difficulty accessing services, incarceration, and homelessness patterns using non-linear growth curve analyses, the sample was restricted to participants who reported recent weekly illicit opioid use but who were not concurrently enrolled in OAT (last six months). This permitted the before and after comparison of participants who reported recent weekly illicit opioid use but who may or may not have initiated OAT. Additionally, participants who stopped or decreased their frequency of illicit opioid use in the absence of OAT were retained in the analyses from their first report of weekly illicit opioid use as they could still be eligible for OAT.

2.3. Study variables

Exposure to opioid agonist treatment was assessed based on the question: “in the last 6 months, have you received any medications for the treatment of your alcohol or drug use?”. Participants could report multiple types of OAT within the last 6-month period, as they may have initiated one type of OAT and switched to a more preferred OAT afterwards; however, participants who reported non-medical prescription opioid or OAT use (e.g., ‘diverted’ methadone) were not classified as having received OAT. Types of OAT included: Methadone/Methadose; buprenorphine-naloxone; slow-release oral morphine (SROM); injectable hydromorphone, diacetylmorphine, or other injectable opioids; as well as other substitution treatments (i.e., prescription opioids).

We considered self-reported socio-demographic, substance use-related, health and social service utilization, and behavioural factors in our analyses. These included: age (per year older); sex (male vs. female); ethnicity or race (self-reported Indigenous identity vs. white; other ethnic or racialized identities vs. white); relationship status (single vs. other); education level (>high school vs. other); self-reported diagnosis of mental illness (yes vs. no); any injection drug use (yes vs. no); and non-fatal overdose, defined as an overdose or acute reaction following drug use (yes vs. no) (Brugal et al., 2002). Past-six-month (hereafter referred to as ‘recent’) daily and weekly injection and non-injection substance use variables were derived by reading participants a list of substances and asking them to identify the types of substances used and their respective frequency of use over the previous six-month period. These included heroin or fentanyl use (yes vs. no), which were combined due to the increased presence of fentanyl within the heroin supply; non-medical prescription opioid use (NMPOU) (yes vs. no); cocaine use (yes vs. no); crack cocaine use (yes vs. no); and crystal methamphetamine use (yes vs. no). Due to the lower frequency of reported daily NMPOU or heroin or fentanyl use, we combined the NMPOU and heroin or fentanyl use (yes vs. no) into a single “daily illicit opioid use” variable in order to examine the relationship between high-intensity opioid use and recent OAT.

Additional factors were examined in the analyses including: incarceration, defined as being in detention, prison or jail overnight or longer (yes vs. no); homelessness, defined as sleeping on the street, having no fixed address, staying with friends or staying in a shelter or hostel (yes vs. no); employment (yes vs. no), defined as regular, temporary, or self-employment that is legal; difficulty accessing services, defined as needing but not being able to obtain health or social (yes vs. no); accessing a detoxification program (yes vs. no); accessing a recovery house (yes vs. no), defined as a communal living environment that supports participants’ recovery goals (Government of British Columbia, 2020); accessing a treatment centre (yes vs. no), defined as a short-term, intensive program that provides treatment and recovery supports (Government of British Columbia, 2020); accessing a counsellor (yes vs. no); and, accessing Narcotics Anonymous, Cocaine Anonymous, Alcoholics Anonymous, or Self-Management and Recovery Training (yes vs. no). All measures referred to exposures in the last six months, except for age, gender, and ethnicity or race.

2.4. Statistical analyses

2.4.1. Baseline comparisons

To compare the baseline characteristics between participants who initiated and did not initiate OAT, Mann-Whitney U-test was used for continuous variables and Pearson’s Chi-squared test was used for binary variables.

2.4.2. Cox regression analysis

To identify factors associated with the time to the first OAT initiation event, we conducted a preliminary Cox proportional hazard regression. To be eligible, participants were required to have reported recent weekly illicit opioid use but no recent OAT (in the last six months) at any point over the study period and subsequently attended a follow-up visit. Participants who reported continuous OAT enrolment from their baseline interview were excluded from analyses; however, if participants discontinued OAT, they were re-included in the analysis if they reported recent weekly opioid use and attended a subsequent study visit.

For this analysis, a bivariate Cox model examined the association between the explanatory variables of interest and time to OAT vs. no OAT (outcome), defined as the time between a participant’s first report of recent weekly illicit opioid use and a subsequent report of recent OAT (last six months). Given interviews were conducted bi-annually, time was calculated based on the number of days between a first and follow-up interview, whereas time to the event was calculated based on mid-time point between the first and follow-up interview. All substance use variables were lagged to the previous observation period to avoid reverse causality. We checked the variance inflation factor to test for collinearity. Two Cox regression models were employed in this analysis, including one that adjusted for daily illicit opioid use and another that adjusted for separate categories of weekly illicit opioid use.

To build a reduced multivariable Cox model, we retained all explanatory variables that met statistical significance at the bivariate level (p < 0.1) and used a backward selection process dropping variables with the largest p-value, except age, sex, and self-reported Indigenous, ethnic, or racialized identity, which were retained in the final models. We continued this process until the multivariable model with the lowest Akaike Information Criterion, a measure of model fit, was identified (Maldonado and Greenland, 1993, Rothman et al., 2008). This technique has been previously employed and described with the ARYS cohort (Feng et al., 2013, Hadland et al., 2012).

2.4.3. Non-linear growth curve analysis

To compare changes in substance use, difficulty accessing health services, incarceration, and homelessness before and after OAT, we established two case groups and one ‘no OAT’ control group. Each of these groups consisted of participants that provided three consecutive observations, or a ‘trio’ of observations (e.g., ‘before event’ – ‘event’ – ‘after event’), within a three-year period. This was done to account for any time participants were away because of treatment or for other reasons. Trios are representative of event-level data, meaning that participants could contribute multiple trios to both the case and control groups.

The two case groups consisted of AYA who (1) initiated and were retained in OAT, and (2) initiated and discontinued OAT, versus a ‘control’ group of AYA who (3) did not initiate OAT. Each case event was then matched to two control events and McNemar’s test and non-linear growth curve analysis was used to examine before and after substance use, difficulty accessing services, incarceration, and homelessness between case and control trios.

2.4.3.1. Case groups definition: ‘retained’ and ‘discontinued’ OAT

For the OAT retained group, participants were required provide consecutive reports of: recent weekly illicit opioid use but no OAT within that last six months at time 1 (‘before event’), OAT in the last six months at time 2 (‘event’), and OAT in the last six months at time 3 (‘after event’) within a three-year period. Participants could report different types of OAT due to transitioning from one to another. Similarly, contributing a trio of observations to the OAT discontinuation group required consecutive reports: of recent weekly illicit opioid use but no recent OAT within the last six months at time 1 (‘before event’), OAT in the last six months at time 2 (‘event’), and a no recent OAT at time 3 (‘after event’) within a three-year period.

2.4.3.2. Control group definition: ‘No OAT’

As with the case groups, contributing a trio of events to the control group required consecutive reports of: recent weekly illicit opioid use but no recent OAT at time 1 (‘before event’), no recent OAT at time 2 (‘event’), and no recent OAT and at time 3 (‘after event’) within a three-year period. This allows for a comparison between participants who initiated OAT and controls who qualified for but did not initiate OAT.

2.4.3.3. Matching case to control trios procedure

For this analysis, each case was matched to approximately two controls by randomization and based on participant characteristics at time 1 (‘before event’). The matching criteria were sex, ethnicity or race (Indigenous, ethnic, or racialized identity vs. white), age (within ± 5 years), recent injection drug use, years of injection drug use (within ± 3 years), and recent illicit opioid use (daily vs. weekly). We also matched for recent crack cocaine use (weekly vs. other) as we found a significant difference in recent weekly crack cocaine use between cases and controls.

To avoid bias, we applied a bootstrapping technique (Efron and Tibshirani, 1994) to increase the stability of the case and control estimates (DeBeck et al., 2009, Lake et al., 2016, Pilarinos et al., 2020, Vlahov et al., 2001). We repeated the matching procedure 50 times to create 50 samples and ran the McNemar and non-linear growth curve analyses 50 times and summarized the result as the mean of the results.

2.4.3.4. Within-group changes of case and control groups

McNemar’s test compared within-group differences in substance use, difficulty accessing health services, incarceration, and homelessness patterns before and after OAT between (1) ‘OAT continuers’ vs. matched ‘no OAT’ controls, and (2) ‘OAT discontinuers’ vs. matched ‘no OAT’ controls.

2.4.3.5. Between-group changes between case and control

We then proceeded with non-linear growth curve analyses to compare the between-group differences in the change in substance use, difficulty accessing health services, incarceration, and homelessness patterns from before to after OAT (Davidian and Giltinan, 2017, Grimm et al., 2011). The resulting slope of the line represents the degree and direction of change (p < 0.05) between each of the case groups and their matched controls, indicating whether the between-group differences were significant.

For these analyses, all significance tests were two-sided at a significance level of p < 0.05. R Foundation for Statistical Computing (Version 3.2.4; Vienna, Austria) was used to conduct all analyses (R Core Team, 2013). This technique to examine before and after patterns and behaviors has been employed in research among people who use drugs (DeBeck et al., 2009, Lake et al., 2016, Pilarinos et al., 2020, Vlahov et al., 2001).

3. RESULTS

Between September 2005 and December 2018, 676 participants reported recent weekly illicit opioid use; however, 101 (14.9%) participants were excluded from the analysis as they were continuously enrolled in OAT and their initiation period could not be derived, as were 121 (17.9%) participants who completed one study visit. A summary of the baseline characteristics of participants who reported recent weekly opioid use, stratified by recent OAT vs. no recent OAT, are presented in Table 1.

Table 1.

Baseline characteristics of adolescent and young adult participants who report recent weekly opioid use and no recent opioid agonist treatment (OAT) engagement at baseline and were included in the Cox regression, stratified by subsequent initiation of OAT (n = 217) at any point during the study period (September 2005 to December 2018) (n = 454).

Characteristic Total (%)
(n = 454)		 Opioid agonist treatment p - value
Yes (%)
(n = 217)		 No (%)
(n = 235)a
Age (<19 vs. ≥19)b,c 23 (21 – 25) 23 (21 – 24) 23 (21 – 25) 0.088
Sex (male vs. female) 307 (67.6) 145 (66.8) 162 (68.9) 0.630
Indigeneity (Indigenous vs. white) 117 (25.9) 72 (30.6) 45 (20.7) 0.021
Other ethnicities/races (other vs. white)d 41 (9.1) 19 (8.1) 22 (10.1) 0.752
Relationship status (single vs. other)e 313 (68.9) 146 (67.3) 166 (70.6) 0.441
High school educatione 177 (39.0) 84 (38.7) 91 (38.7) 0.998
Mental illnesse 283 (62.3) 133 (61.3) 148 (63.0) 0.712
Injection drug usee 295 (65.0) 161 (74.2) 132 (56.2) < 0.001
Non-fatal overdosee 121 (26.7) 68 (31.3) 52 (22.1) 0.025
Daily illicit opioid usee,f 252 (55.5) 139 (64.1) 111 (47.2) < 0.001
Weekly heroin or fentanyl usee 366 (80.6) 188 (86.6) 176 (74.9) 0.002
Weekly NMPO usee,g 147 (32.4) 59 (27.2) 87 (37.0) 0.026
Weekly cocaine usee 93 (20.5) 45 (20.7) 48 (20.4) 0.953
Weekly crack cocaine usee 128 (28.2) 63 (29.0) 64 (27.2) 0.578
Weekly CM usee,h 217 (47.8) 103 (47.5) 113 (48.1) 0.861
Incarceratione 99 (21.8) 56 (25.8) 42 (17.9) 0.038
Homelessnesse 297 (65.4) 142 (65.4) 153 (65.1) 0.960
Employmente 191 (42.1) 83 (38.2) 108 (46.0) 0.106
Difficult accessing servicese 409 (90.1) 193 (88.9) 214 (91.1) 0.449
Detoxificatione 75 (16.5) 40 (18.4) 35 (14.9) 0.312
Recovery housee 31 (6.8) 14 (6.5) 17 (7.2) 0.625
Treatment centree 23 (5.1) 13 (6.0) 10 (4.3) 0.481
Counsellore 49 (10.8) 23 (10.6) 26 (11.1) 0.874
NA/CA/AA/SMARTe,i 51 (11.2) 27 (12.4) 24 (10.2) 0.454
Open in a new tab
a.

Two participants were excluded from the baseline descriptive analysis due to missing outcome variable data

b.

Denotes median age

c.

Denotes the interquartile range

d.

Other ethnic or racialized groups include Black, Latinx, Middle Eastern, South Asian, and ‘other’ Asian

e.

Refers to activities in the last six months

f.

Refers to a combination of illicit opioids that includes heroin and non-medical prescription opioids

g.

Denotes non-medical prescription opioid use

h.

Denotes crystal methamphetamine

i.

Denotes narcotics anonymous, cocaine anonymous, alcoholics anonymous, and self-management and recovery training

3.1. Cox regression of time to recent OAT

A total of 454 participants reported recent weekly opioid use, contributing 2241 observations over a median follow-up of four study visits (IQR = 2–6). Of these, 217 (32.1%) participants reported initiating OAT at some point over follow-up, accounting for 1115.1 person-time risk years and an incidence rate of 23 (95% CI: 20–26) per 100 person-years. Among participants who did initiate OAT, 175 (38.5%) initiated OAT once, 38 (8.4%) initiated OAT twice, and 4 (0.9%) initiated OAT three times. Table 2 presents the adjusted and unadjusted Cox regression of factors associated with time to recent OAT.

Table 2.

Cox Proportional Hazard regression analysis of factors associated with time to first opioid agonist treatment among adolescents and young adults who report recent weekly opioid use in Vancouver, Canada at any point during the study period (between September 2005 and December 2018) (n = 454).

Unadjusted model Adjusted Model 1a Adjusted Model 2b
Characteristic Hazard Ratio (95% CI) Hazard Ratio (95% CI) Hazard Ratio (95% CI)
Follow-up period (per six-months later) 1.04 (1.02 – 1.06) 1.02 (1.00 – 1.05)c 1.02 (1.00 – 1.05)c
Age (<19 vs. ≥19) 0.72 (0.35 – 1.48) 1.01 (0.96 – 1.07) 1.03 (0.98 – 1.09)
Sex (male vs. female) 1.06 (0.80 – 1.41) 1.05 (0.76 – 1.47) 1.03 (0.74 – 1.44)
Indigeneity (Indigenous vs. white) 0.74 (0.53 – 1.04) 0.74 (0.50 – 1.11) 0.77 (0.51 – 1.15)
Other races/ethnicities (other vs. white)d 1.27 (0.82 – 1.98) 1.06 (0.58 – 1.94) 1.07 (0.59 – 1.94)
Relationship status (single vs. other) 1.17 (0.88 – 1.55)
Education (>high school vs. other) 1.23 (0.94 – 1.65)
Mental illness (yes vs. no) 0.92 (0.69 – 1.23)
Injection drug use (yes vs. no)e,f 2.76 (2.00 – 3.80)c 2.31 (1.52 – 3.52)c 2.59 (1.72 – 3.90)c
Non-fatal overdose (yes vs. no)e,f 1.53 (1.14 – 2.07)c 1.03 (0.70 – 1.52) 1.06 (0.72 – 1.55)
Daily illicit opioid use (yes vs. no)e,c,g 2.94 (2.14 – 4.04)c 2.74 (1.85 – 4.05)c –
Heroin/fentanyl use (≥weekly vs. <weekly)e,f 3.62 (2.49 – 5.27)c – 2.35 (1.53 – 3.62)c
NMPO use (≥weekly vs. <weekly)e,c,h 1.01 (0.71 – 1.42) –
Cocaine use (≥weekly vs. <weekly)e,f 1.19 (0.84 – 1.71)
Crack cocaine use (≥weekly vs. <weekly)e,f 0.87 (0.63 – 1.21)
CM use (≥weekly vs. <weekly 1.14 (0.86 – 1.50)
Incarceration (yes vs. no)e 1.00 (0.71 – 1.41)
Homelessness (yes vs. no)e 0.87 (0.66 – 1.15)
Employment (yes vs. no)e 0.71 (0.53 – 0.95) 0.80 (0.56 – 1.15) 0.81 (0.56 – 1.15)
Difficulty accessing services (yes vs. no)e 1.29 (0.89 – 1.87)
Detoxification (yes vs. no)e 2.45 (1.69 – 3.55)c 1.49 (0.92 – 2.42) 1.56 (0.94 – 2.59)
Recovery house (yes vs. no)e 2.03 (1.22 – 3.39)c 1.65 (0.91 – 3.00) 1.62 (0.91 – 2.91)
Treatment centre (yes vs. no)e 2.61 (1.63 – 4.16)c 2.01 (1.23 – 3.27)c 2.05 (1.28 – 3.27)c
Counsellor (yes vs. no)e 1.84 (1.21 – 2.81)c 1.54 (0.92 – 2.58) 1.56 (0.94 – 2.59)
NA/CA/AA/SMART (yes vs. no)e,j 1.76 (1.19 – 2.59) 1.19 (0.67 – 2.12) 1.11 (0.63 – 1.94)
Open in a new tab
a.

Adjusted model included a daily illicit opioid use co-variate

b.

Adjusted model included a weekly heroin or fentanyl use co-variate

c.

Refers to variables that were statistically significant at p < 0.05

d.

Other ethnic or racialized groups include Black, Latinx, Middle Eastern, South Asian, and ‘other’ Asian

e.

Refers to activities in the last six months

f.

Refers to activities lagged to the previous, available follow-up

g.

Refers to a combination of illicit opioids that includes heroin and non-medical prescription opioids

h.

Denotes non-medical prescription opioids

i.

Denotes crystal methamphetamine

j.

Denotes narcotics anonymous, cocaine anonymous, alcoholics anonymous, and self-management and recovery training

In the first adjusted Cox model, recent daily illicit opioid use [adjusted hazards ratio (AHR) = 2.74, 95% confidence interval (CI) = 1.85–4.05], injection drug use (AHR = 2.31, 95% CI = 1.52–3.52), accessing a treatment centre (AHR = 2.01, 95% CI = 1.23–3.27), and later follow-up period (AHR = 1.02, 95% CI = 1.00–1.05) were positively associated with recent OAT. In the second adjusted Cox model, recent weekly heroin or fentanyl use (AHR = 2.35, 95% CI = 1.53–3.62), injection drug use (AHR = 2.59, 95% CI = 1.72–3.90), accessing a treatment centre (AHR = 2.05, 95% CI = 1.28–3.27), and later follow-up period (AHR = 1.02, 95% CI = 1.00–1.05) were positively associated with recent OAT; however, there was no association between recent weekly NMPOU and OAT initiation.

3.2. Non-linear growth curve analysis

A total of 336 participants were eligible for inclusion in the before and after analysis, contributing 1528 observations. Of these, 100 participants contributed 108 trios to the ‘OAT continuers’ group, 48 participants contributed 52 trios to the ‘OAT discontinuers’ group, and 244 participants contribute 577 trios to the ‘no OAT control’ group. ‘OAT continuers’ (n = 100) reported receiving methadone (n = 91), buprenorphine-naloxone (n = 10), and SROM (n = 2), while ‘OAT discontinuers’ (n = 48) reported receiving methadone (n = 40), buprenorphine-naloxone (n = 9), and slow-release oral morphine (n = 1) at least once.

The within-group differences in substance use, difficulty accessing health services, incarceration, and homelessness patterns from the before and the after period are presented in Table 3, Table 4. In Table 3, participants retained in OAT demonstrated greater reductions than the control group in the prevalence of: any injection drug use (−12.1% vs. −11.7%), non-fatal overdose (−13.1% vs. −9.3%), daily illicit opioid use (−28.0% vs. −9.3%), weekly heroin or fentanyl use (−31.8% vs. −20.2%), weekly cocaine use (−6.5% vs. −4.5%), and homelessness (−22.4% vs. −8.6%); however, the opposite was observed for any weekly NMPOU (−20.6% vs. −21.6%), weekly crystal methamphetamine use (+0.9% vs. −3.0%), incarceration (+3.7% vs. −1.9%), and difficulty accessing health or social services (−2.8% vs. −6.9%).

Table 3.

Substance use, difficulty accessing health services, incarceration, and homelessness patterns in the period before and after opioid agonist treatment (OAT) among participants who were retained in OAT for two or more consecutive study periods (OAT continuers, n = 108 cases) vs. controls who did not access OAT (mean n over 50 runs = 207 controls) at any point during the study period (between September 2005 and December 2018).

Variables Opioid Agonist Treatment perioda p value
Before n (%) After n (%)
Injection drug use b
Continuers 88 (82.2) 75 (70.1) 0.012
Controls 171 (82.6) 147 (70.9) < 0.001
Non-fatal overdose b
Continuers 34 (31.8) 20 (18.7) 0.015
Controls 60 (28.9) 41 (19.6) 0.042
Daily illicit opioid use b,c
Continuers 84 (78.5) 54 (50.5) < 0.001
Controls 142 (68.7) 123 (59.4) < 0.001
Weekly heroin or fentanyl use b
Continuers 92 (86.0) 58 (54.2) < 0.001
Controls 172 (82.9) 130 (62.7) < 0.001
Weekly NMPO use b,d
Continuers 32 (29.9) 10 (9.3) < 0.001
Controls 68 (33.0) 24 (11.4) < 0.001
Weekly cocaine use b
Continuers 18 (16.8) 11 (10.3) 0.146
Controls 23 (11.0) 13 (6.5) 0.170
Weekly CM use b,e
Continuers 48 (44.9) 49 (45.8) 1.000
Controls 105 (50.8) 99 (47.8) 0.513
Incarceration b
Continuers 22 (20.6) 26 (24.3) 0.522
Controls 43 (21.0) 40 (19.1) 0.642
Homelessness b
Continuers 67 (62.6) 43 (40.2) < 0.001
Controls 123 (59.5) 105 (50.9) 0.052
Difficult accessing services b
Continuers 88 (82.2) 91 (85.0) 0.677
Controls 182 (88.0) 168 (81.1) 0.088
Open in a new tab
a.

Before and after values represent the mean number of cases and controls based on the mean numbers from 50 datasets. Each case matched to approximately two controls depending on randomization by sex (male vs. female), Ethnicity or race (Indigneous, ethnic or racialized identity vs. white), recent injection drug use (yes vs. no), age (±5 years), years injecting drugs (±3 years), and recent weekly crack use (yes vs. no)

b.

Refers to activities in the previous 6 months

c.

Refers to a combination of illicit opioids that includes heroin and non-medical prescription opioids

d.

Denotes non-medical prescription opioids

e.

Denotes crystal methamphetamine

Table 4.

Substance use, difficulty accessing health services, incarceration, and homelessness patterns in the period before and after opioid agonist treatment (OAT) among participants who access OAT for only one study follow-up (OAT discontinuers, n = 52 cases) vs. controls who did not access OAT (mean n over 50 runs = 95 controls) at any point during the study period (between September 2005 and December 2018).

Variables Opioid Agonist Treatment Perioda p value
Before n (%) After n (%)
Injection drug use b
Discontinuers 44 (86.3) 41 (80.4) 0.450
Controls 83 (87.4) 68 (71.2) 0.004
Non-fatal overdose b
Discontinuers 15 (29.4) 8 (15.7) 0.070
Controls 25 (25.9) 19 (19.6) 0.389
Daily illicit opioid use b,c
Discontinuers 43 (84.3) 34 (66.7) 0.052
Controls 64 (67.3) 52 (55.1) 0.088
Weekly heroin or fentanyl use b
Discontinuers 47 (92.2) 30 (58.8) <0.001
Controls 83 (86.9) 59 (62.0) 0.001
Weekly NMPO use b
Discontinuers 8 (15.7) 6 (11.8) 0.617
Controls 26 (27.7) 7 (7.5) 0.001
Weekly cocaine use b
Discontinuers 6 (11.8) 5 (9.8) 1.000
Controls 15 (15.9) 7 (7.7) 0.201
Weekly CM use b,d
Discontinuers 27 (52.9) 19 (37.3) 0.080
Controls 45 (47.5) 43 (45.7) 0.720
Incarceration b
Discontinuers 11 (21.6) 10 (19.6) 1.000
Controls 21 (22.2) 17 (18.2) 0.621
Homelessness b
Discontinuers 30 (58.8) 17 (33.3) 0.006
Controls 57 (60.1) 51 (53.6) 0.342
Difficulty accessing services b
Discontinuers 41 (80.4) 41 (80.4) 0.724
Controls 85 (89.1) 79 (83.2) 0.296
Open in a new tab
a.

Before and after values represent the mean number of cases and controls based on the mean numbers from 50 datasets. Each case matched to approximately two controls depending on randomization by sex (male vs. female), Ethnicity or race (Indigneous, ethnic or racialized identity vs. white), recent injection drug use (yes vs. no), age (±5 years), years injecting drugs (±3 years), and recent weekly crack use (yes vs. no)

b.

Refers to activities in the previous 6 months

c.

Refers to a combination of illicit opioids that includes heroin and non-medical prescription opioids

d.

Denotes crystal methamphetamine

Table 4 presents the results from the before and after comparison between case events that involved OAT discontinuation in comparison to control events that did not involve any OAT. In this comparison, events that involved OAT discontinuation demonstrated greater reductions than controls in the prevalence of any non-fatal overdose (−13.7% vs. −6.3%), daily illicit opioid use (−17.6% vs. −12.2%), weekly heroin or fentanyl use (−33.4% vs. −24.9%), weekly crystal methamphetamine use (−15.6% vs. −1.8%), and homelessness (−25.5% vs. −6.5%). Alternatively, our comparison of events that involved OAT discontinuation to controls demonstrated smaller reductions in the prevalence of any injection drug use (−5.9% vs. −16.2%), weekly NMPOU (−3.9% vs. −20.2%), weekly cocaine use (−2.0% vs. −8.2%), incarceration (−2.0% vs. −4.0%), and difficulty accessing any health or social services (0% vs. −5.9%).

The results of the non-linear growth curve analyses of between-group differences are reported in Table 5, Table 6. This analysis suggests that participants who were retained in OAT demonstrated greater reductions than ‘no OAT’ controls in the prevalence of any daily illicit opioid use (slope: −1.83 vs. −0.63, p = 0.015), although no other significant differences arose. Nevertheless, a non-statistically significant reduction in homelessness (p = 0.070) and an increase in difficulty accessing services (p = 0.078) was observed among AYA who were retained in OAT vs. ‘no OAT’ controls, while participants who discontinued OAT demonstrated non-statistically significant reductions in homelessness (p = 0.085) and weekly NMPOU (p = 0.061) when compared to ‘no OAT’ controls.

Table 5.

Non-linear growth curve analyses comparing substance use, difficulty accessing health services, incarceration, and homelessness patterns between cases who were retained on OAT for two or more consecutive study periods (OAT continuers, n = 108 cases) vs. controls who did not access OAT (mean n over 50 runs = 207controls) at any point during the study period (between September 2005 and December 2018).

Variables Slope (95% CIa) p valueb
Injection drug use c
Continuers −1.49 (−3.55, 0.58) 0.729
Controls −1.63 (−1.72, −1.54)
Non-fatal overdose c
Continuers −1.01 (−2.47, 0.46) 0.534
Controls −0.70 (−0.75, −0.65)
Daily illicit opioid use c,b
Continuers −1.83 (−3.23, −0.43) 0.015
Controls −0.63 (−0.68, −0.58)
Weekly heroin or fentanyl use c
Continuers −2.27 (−3.84, −0.70) 0.182
Controls −1.57 (−1.62, −1.52)
Weekly NMPO use c,d
Continuers −2.23 (−4.16, −0.30) 0.747
Controls −2.18 (−2.25, −2.11)
Weekly cocaine use c
Continuers −0.80 (−2.88, 1.27) 0.791
Controls −0.76 (−0.83, −0.69)
Weekly CM use c,e
Continuers 0.02 (−1.38, 1.42) 0.600
Controls −0.22 (−0.27, −0.17)
Incarceration c
Continuers 0.33 (−1.24, 1.90) 0.409
Controls −0.12 (−0.18, −0.06)
Homelessness c
Continuers −1.48 (−2.89, −0.07) 0.070
Controls −0.62 (−0.67, −0.57)
Difficulty accessing services c
Continuers 0.28 (−1.43, 1.99) 0.087
Controls −0.72 (−0.78, −0.66)
Open in a new tab
a.

Denotes a 95% Confidence Interval

b.

Refers to a combination of illicit opioids that includes heroin and non-medical prescription opioids

c.

Refers to activities in the previous 6 months

d.

Denotes non-medical prescription opioids

e.

Denotes crystal methamphetamine

Table 6.

Non-linear growth curve analyses comparing substance use, difficulty accessing health services, incarceration, and homelessness patterns between cases who accessed OAT for only one study follow-up (OAT discontinuers, n = 52 cases) vs. controls who did not access OAT (mean n over 50 runs = 95 controls) at any point during the study period (between September 2005 and December 2018).

Variables Slope (95% CIa) p valueb
Injection drug usec
Discontinuers −0.89 (−4.10, 2.32) 0.234
Controls −2.11 (−2.23, −1.99)
Non-fatal overdosec
Discontinuers −1.46 (−4.01, 1.08) 0.345
Controls −0.61 (−0.72, −0.50)
Daily illicit opioid usec,b
Discontinuers −1.41 (−3.53, 0.71) 0.409
Controls −0.80 (−0.87, −0.73)
Weekly heroin or fentanyl usec
Discontinuers −2.66 (−5.14, −0.19) 0.328
Controls −1.83 (−1.94, −1.72)
Weekly NMPO usec,d
Discontinuers −0.70 (−4.59, 3.20) 0.061
Controls −3.05 (−3.42, −2.68)
Weekly cocaine usec
Discontinuers −0.32 (−3.44, 2.80) 0.439
Controls −1.19 (−1.36, −1.02)
Weekly CM usec,e
Discontinuers −1.14 (−3.26, 0.97) 0.165
Controls −0.12 (−0.19, −0.05)
Incarcerationc
Discontinuers −0.14 (−2.26, 1.97) 0.697
Controls −0.29 (−0.39, −0.19)
Homelessnessc
Discontinuers −1.74 (−3.83, 0.35) 0.085
Controls −0.45 (−0.55, −0.35)
Difficulty accessing servicesc
Discontinuers 0.38 (−2.41, 3.17) 0.192
Controls −0.85 (−0.98, −0.72)
Open in a new tab
a.

Denotes a 95% Confidence Interval

b.

Refers to a combination of illicit opioids that includes heroin and non-medical prescription opioids

c.

Refers to activities in the previous 6 months

d.

Denotes non-medical prescription opioids

e.

Denotes crystal methamphetamine

4. DISCUSSION

The finding that 43% of ARYS participants who reported weekly opioid use did not report accessing OAT is concerning. Given evidence that OAT access among adults increased from 73.2% in 2006 to 78.9% in 2016 in Vancouver, Canada, findings indicate that many AYA are not accessing potentially life-saving OAT (Socías et al., 2018). Adding to this, our findings of frequent OAT re-initiation events complement existing research among adults demonstrating that the cessation and re-initiation of OAT is similarly common among AYA (Kerr et al., 2005). This may be explained by known barriers to treatment in the study setting, which include geographical barriers, being unable to secure treatment, being involuntarily discharged from treatment, high treatment costs, and long treatment wait times (Hadland et al., 2009, Phillips et al., 2014). However, this may also be attributed to a reluctance towards OAT among AYA, as reported in a recent qualitative investigation in the study setting, where OAT is not viewed as part of their long-term treatment and recovery goals (Giang et al., 2020).

Results also indicate that being retained on OAT was associated with significant reductions in recent daily illicit opioid use, and that recent daily illicit opioid use, recent weekly heroin use, and injection drug use were positively associated with a shorter time to OAT. Additionally, recent SUD treatment centre access was positively associated with a shorter time to OAT, while accessing detoxification, a counsellor, and a recovery house approached statistical significance, reinforcing calls for the incorporation of OAT into other health and treatment settings (Dunn et al., 2019, Fanucchi and Lofwall, 2016, Luli et al., 2020, Raheemullah and Lembke, 2019). Another notable finding was that being retained in OAT was associated with significant within-group reductions; however, the only between-group reduction observed was in daily illicit opioid use between participants retained in OAT vs. ‘no OAT’ controls. While reductions in daily illicit opioid use were observed among cases and controls, this may be attributed to the expansion of SUD treatments and programming over the duration of the study, such as the establishment of a low-barrier, youth-specific health service provider and outreach team (Mitchell et al., 2017). This may also explain the positive association between follow-up period and shorter time to OAT, given that the expansion of services may have increased AYAs’ access to OAT and other treatments within the study setting.

This study reinforces evidence that OAT is effective at reducing high frequency opioid use among AYA (Harcus et al., 1980, Marsch et al., 2005, Matson et al., 2014, Minozzi et al., 2014, Smyth et al., 2018) and supports recommendations for the provision of OAT to AYA (British Columbia Centre on Substance Use, 2018, Committee On Substance Use Prevention, 2016, The Society for Adolescent Health and Medicine, 2021). However, the absence of significant reductions between cases and controls with respect to weekly heroin or fentanyl use, weekly NMPOU use, weekly stimulant use, homelessness, incarceration, and difficulty accessing health services is concerning. Firstly, the continued use of opioids while on OAT aligns with a previous Cochrane review, which found that 46% of AYA reported continued opioid use while on methadone (Mattick et al., 2009). Secondly, results demonstrating continued crystal methamphetamine use in this sample complement qualitative literature from the study setting that found that some adults continue using stimulants to counter methadone’s sedating effects, among other reasons (McNeil et al., 2020). Hence, the absence of significant reductions in substance use, difficulty accessing health services, incarceration, and homelessness emphasize the need for treatment interventions for AYA who report polysubstance use and that meet AYAs’ recovery priority and goals.

Potential solutions to improving OAT retention among AYA have been proposed (Viera et al., 2020) and include implementing tailored approaches to OAT that acknowledge the differences between AYA and adults (Calvo et al., 2017, Corace et al., 2018, Unger et al., 1998). Examples include working collaboratively with AYA to determine an adequate OAT dose (Artenie et al., 2019, González-Saiz et al., 2008, González-Saiz et al., 2018, Viera et al., 2020), as a higher OAT dose is positively associated with reductions in opioid and stimulant use (Faggiano et al., 2003, Heikman et al., 2017); expanding access to take-home dosing (Dunn et al., 2021, Figgatt et al., 2021, Viera et al., 2020); supporting AYA with accessing safe and stable housing, employment and income assistance, and re-connecting with school and culture in a way that meet AYAs’ treatment goals (Giang et al., 2020, Simeone et al., 2017, Viera et al., 2020, Zhou et al., 2017); and, working with AYA to include family, friends, or peers in treatment programming (Kidorf et al., 2018, Lin et al., 2011).

Nonetheless, future research that seeks to understand the relationship between OAT receipt and health and social support access may be beneficial to improving the continuity of care for AYA, in addition to adolescent-specific research on ways to better engage adolescents in OAT. Adding to this, future research comparing AYA who were retained in OAT for longer durations to AYA retained in OAT for shorter durations may provide more insight into the long-term use of OAT and associated outcomes. Lastly, this study excluded naltrexone due to recommendations against its use in the study setting (British Columbia Centre on Substance Use, 2018); however, the American Academy of Pediatrics (Committee On Substance Use Prevention, 2016) recommends naltrexone for the treatment of OUD, warranting further investigation into its effectiveness on opioid use outcomes among AYA and in the context of the toxic drug crisis.

There are limitations to this study. Firstly, street-based outreach was used to recruit participants into the study, and therefore the ARYS cohort does not represent a random sample and findings may not be generalizable to other settings or populations. Additionally, the use of self-reported data increases recall and social desirability bias, although previous research has demonstrated self-reported data from people who use drugs is reliable (Brener et al., 2003). Unmeasured confounding may exist where a number of factors related to treatment delivery could not be accounted for (e.g., provision of other supports or treatments alongside OAT, dosage, quality of care received), which are known factors of treatment retention and substance use outcomes that may bias the results. Due to limitations in the study instrument, participants were asked about their OAT access over a six-month period and so treatment cessation over this period may not have been captured or participants may have transitioned between different OAT. This may have resulted in participants being included in the case groups despite having stopped and re-initiated treatment, leading to the inclusion of OAT discontinuers in the OAT retained group. Furthermore, a before and after comparison could not be conducted on variables that were used in matching (i.e. weekly crack cocaine use) and the use of event-level data does not allow for the examination of individual OAT trajectories. Additionally, a majority of participants reported polysubstance use, which may influence OAT retention and outcomes. Lastly, this study included participants who did not have a formal OUD diagnosis and who may therefore be less likely to be provided OAT.

5. CONCLUSION

Being retained on OAT was associated with significant reductions in daily illicit opioid use, which is promising and indicative that these treatments are beneficial among AYA in the study setting. Nevertheless, more than half of the participants did not initiate OAT over the study period and the absence of other significant reductions in substance use and other patterns among AYA who were retained in OAT or discontinued OAT vs. controls that did not access OAT is discouraging. Findings highlight the importance of providing AYA access to a continuum of health and social services, harm reduction interventions, and a more diverse array of therapeutic options that improve substance use and health outcomes.

HIGHLIGHTS.

  • This study examined opioid agonist treatment (OAT) outcomes among young people.

  • OAT retention was associated with reductions in daily illicit opioid use.

  • No other reductions, in substance use nor other outcomes, were observed.

  • Young people who use opioids require a continuum of health and social supports.

Acknowledgements

The study authors would like to thank ARYS participants for their involvement in this research, as well as all previous and current research staff supporting this study. We would also like to thank Lizzy Ambler for their support in drafting and reviewing this manuscript.

Role of funding source

The ARYS study is supported by the Canadian Institutes of Health Research [MOP-286532] and US National Institutes of Health [U01-DA038886]. Andreas Pilarinos is supported through a Four-Year Fellowship from the University of British Columbia. Dr. Kora DeBeck is supported by the Michael Smith Foundation for Health Research/MSt. Paul’s Hospital Foundation-Providence Health Care Career Scholar Award and a CIHR New Investigator Award. The funding organizations had no role in the design and conduct of this study, including data collection, management, analysis, and interpretation; in preparing, reviewing, or approving the final manuscript; nor in any decisions to submit the manuscript for publication.

Footnotes

Conflict of interest

No conflicts declared.

REFERENCES

  1. Ahmad F, Spencer LRM, Warner M, Sutton P, Statistics, N.C.f.H., 2020. Provisional drug overdose death counts.
  2. Alinsky R, Zima B, Bagley S, Rodean J, Matson P, Adger H, Hadland S, 2019. Receipt of addiction treatment following opioid-related overdose among medicaid-enrolled youth. J. Adolesc. Health 64 (2 Supplement), S17. 10.1016/j.jadohealth.2018.10.046. [DOI] [Google Scholar]
  3. Artenie AA, Minoyan N, Jacka B, Høj S, Jutras-Aswad D, Roy E, Gauvin L, Zang G, Bruneau J, 2019. Opioid agonist treatment dosage and patient-perceived dosage adequacy, and risk of hepatitis C infection among people who inject drugs. Cmaj 191 (17), E462–E468. 10.1503/cmaj.181506. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Brener ND, Billy JO, Grady WR, 2003. Assessment of factors affecting the validity of self-reported health-risk behavior among adolescents: evidence from the scientific literature. J. Adolesc. Health 33 (6), 436–457. 10.1016/s1054-139x(03)00052-1. [DOI] [PubMed] [Google Scholar]
  5. British Columbia Centre on Substance Use, 2018. A guideline for the clinical management of opioid use disorder: youth supplement, Provincial Government of British Columbia, June 31, 2018 ed.
  6. Brugal MT, Barrio G, Regidor E, Royuela L, Suelves JM, 2002. Factors associated with non-fatal heroin overdose: assessing the effect of frequency and route of heroin administration. Addiction 97 (3), 319–327. 10.1046/j.1360-0443.2002.00058.x. [DOI] [PubMed] [Google Scholar]
  7. Calvo M, MacFarlane J, Zaccaro H, Curtis M, Caban M, Favaro J, Passannante MR, Frost T, 2017. Young people who use drugs engaged in harm reduction programs in New York City: overdose and other risks. Drug Alcohol Depend. 178, 106–114. 10.1016/j.drugalcdep.2017.04.032. [DOI] [PubMed] [Google Scholar]
  8. Camenga DR, Colon-Rivera HA, Muvvala SB, 2019. Medications for maintenance treatment of opioid use disorder in adolescents: a narrative review and assessment of clinical benefits and potential risks. J. Stud. Alcohol Drugs 80 (4), 393–402. 10.15288/jsad.2019.80.393. [DOI] [PubMed] [Google Scholar]
  9. Chang DC, Klimas J, Wood E, Fairbairn N, 2018. Medication-assisted treatment for youth with opioid use disorder: current dilemmas and remaining questions. Am. J. Drug Alcohol Abus 44 (2), 143–146. 10.1080/00952990.2017.1399403. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Chavez LJ, Bonny AE, Bradley KA, Lapham GT, Cooper J, Miller W, Chisolm DJ, 2020. Medication treatment and health care use among adolescents with opioid use disorder in Ohio. J. Adolesc. Health 07, 07. 10.1016/j.jadohealth.2020.01.016. [DOI] [PubMed] [Google Scholar]
  11. Committee On Substance Use Prevention, 2016. Medication-assisted treatment of adolescents with opioid use disorders. Pediatrics 138 (3), 09. 10.1542/peds.2016-1893. [DOI] [PubMed] [Google Scholar]
  12. Corace K, Willows M, Schubert N, Overington L, Howell G, 2018. Youth require tailored treatment for opioid use and mental health problems: a comparison with adults. Can. J. Addict 9 (4), 15–24. 10.1097/CXA.0000000000000032. [DOI] [Google Scholar]
  13. Davidian M, Giltinan DM, 2017. Nonlinear Models for Repeated Measurement Data. Routledge. 10.1201/9780203745502. [DOI] [Google Scholar]
  14. DeBeck K, Kerr T, Li K, Milloy MJ, Montaner J, Wood E, 2009. Incarceration and drug use patterns among a cohort of injection drug users. Addiction 104 (1), 69–76. 10.1111/j.1360-0443.2008.02387.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Donroe JH, Socias ME, Marshall BD, 2018. The deepening opioid crisis in North America: historical context and current solutions. Curr. Addict. Rep 1–10. 10.1007/s40429-018-0228-5. [DOI] [Google Scholar]
  16. Dunn KE, Huhn AS, Strain EC, 2019. Differential adoption of opioid agonist treatments in detoxification and outpatient settings. J. Subst. Abus. Treat 107, 24–28. 10.1016/j.jsat.2019.10.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Dunn KE, Brooner RK, Stoller KB, 2021. Technology-assisted methadone take-home dosing for dispensing methadone to persons with opioid use disorder during the Covid-19 pandemic. J. Subst. Abus. Treat 121, 108197 10.1016/j.jsat.2020.108197. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Efron B, Tibshirani RJ, 1994. An Introduction to the Bootstrap. CRC Press. [Google Scholar]
  19. Faggiano F, Vigna-Taglianti F, Versino E, Lemma P, 2003. Methadone maintenance at different dosages for opioid dependence. Cochrane Database Syst. Rev 10.1002/14651858.CD002208. [DOI] [PubMed] [Google Scholar]
  20. Fanucchi L, Lofwall MR, 2016. Putting parity into practice—integrating opioid-use disorder treatment into the hospital setting. N. Engl. J. Med 375 (9), 811–813. 10.1056/NEJMp1606157. [DOI] [PubMed] [Google Scholar]
  21. Feder KA, Krawczyk N, Saloner B, 2017. Medication-assisted treatment for adolescents in specialty treatment for opioid use disorder. J. Adolesc. Health 60 (6), 747–750. 10.1016/j.jadohealth.2016.12.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Feng C, DeBeck K, Kerr T, Mathias S, Montaner J, Wood E, 2013. Homelessness independently predicts injection drug use initiation among street-involved youth in a Canadian setting. J. Adolesc. Health 52 (4), 499–501. 10.1016/j.jadohealth.2012.07.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Figgatt MC, Salazar Z, Day E, Vincent L, Dasgupta N, 2021. Take-home dosing experiences among persons receiving methadone maintenance treatment during COVID-19. J. Subst. Abus. Treat 123, 108276 10.1016/j.jsat.2021.108276. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Giang V, Thulien M, McNeil R, Sedgemore K, Anderson H, Fast D, 2020. Opioid agonist therapy trajectories among street entrenched youth in the context of a public health crisis. SSM - Popul. Health 11 (no Pagin-.-.). 10.1016/j.ssmph.2020.100609. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Gonzalez-Saiz F, Lozano Rojas O, Ballesta Gomez R, Bilbao Acedos I, Galiana Martinez J, Garcia Collantes MA, 2008. Evidence of reliability and validity of the Opiate Dosage Adequacy Scale (ODAS) in a sample of methadone maintenance patients. Heroin Addict. Relat. Clin. Probl 10 (1), 25–38. [Google Scholar]
  26. Gonzalez-Saiz F, Rojas OL, Trujols J, Alcaraz S, Sinol N, de Los Cobos JP, Group BNS, 2018. Evidence of validity and reliability of the Opiate Dosage Adequacy Scale (ODAS) in a sample of heroin addicted patients in buprenorphine/naloxone maintenance treatment. Drug Alcohol Depend. 183, 127–133. 10.1016/j.drugalcdep.2017.10.035. [DOI] [PubMed] [Google Scholar]
  27. Government of British Columbia, 2020. Recovery services & treatment support: substance use treatment options. 〈https://www2.gov.bc.ca/gov/content/overdose/what-you-need-to-know/recovery-services-treatment-support#services〉. (Accessed February 5, 2020).
  28. Grimm KJ, Ram N, Hamagami F, 2011. Nonlinear growth curves in developmental research. Child Dev. 82 (5), 1357–1371. 10.1111/j.1467-8624.2011.01630.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Hadland SE, Kerr T, Li K, Montaner JS, Wood E, 2009. Access to drug and alcohol treatment among a cohort of street-involved youth. Drug Alcohol Depend. 101 (1–2), 1–7. 10.1016/j.drugalcdep.2008.10.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Hadland SE, Werb D, Kerr T, Fu E, Wang H, Montaner JS, Wood E, 2012. Childhood sexual abuse and risk for initiating injection drug use: a prospective cohort study. Prev. Med 55 (5), 500–504. 10.1016/j.ypmed.2012.08.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Hadland SE, Wharam JF, Schuster MA, Zhang F, Samet JH, Larochelle MR, 2017. Trends in receipt of buprenorphine and naltrexone for opioid use disorder among adolescents and young adults, 2001–2014. JAMA Pediatr. 171 (8), 747–755. 10.1001/jamapediatrics.2017.0745. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Harcus A, Ward A, Smith D, 1980. The monitored release of buprenorphine: results in the young. J. Int. Med. Res 8 (2), 153–155. 10.1177/030006058000800210. [DOI] [PubMed] [Google Scholar]
  33. Heikman PK, Muhonen LH, Ojanperä IA, 2017. Polydrug abuse among opioid maintenance treatment patients is related to inadequate dose of maintenance treatment medicine. BMC Psychiatry 17 (1), 1–11. 10.1186/s12888-017-1415-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Kerr T, Marsh D, Li K, Montaner J, Wood E, 2005. Factors associated with methadone maintenance therapy use among a cohort of polysubstance using injection drug users in Vancouver. Drug Alcohol Depend. 80 (3), 329–335. 10.1016/j.drugalcdep.2005.05.002. [DOI] [PubMed] [Google Scholar]
  35. Kidorf M, Brooner RK, Peirce J, Gandotra J, Leoutsakos J-M, 2018. Mobilizing community support in people receiving opioid-agonist treatment: a group approach. J. Subst. Abus. Treat 93, 1–6. 10.1016/j.jsat.2018.07.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Lake S, Milloy M, Dong H, Hayashi K, Wood E, Kerr T, DeBeck K, 2016. Initiation into prescription opioid injection and associated trends in heroin use among people who use illicit drugs. Drug Alcohol Depend. 169, 73–79. 10.1016/j.drugalcdep.2016.10.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Lee LK, Mannix R, 2018. Increasing fatality rates from preventable deaths in teenagers and young adults. JAMA 320 (6), 543–544. 10.1001/jama.2018.6566. [DOI] [PubMed] [Google Scholar]
  38. Lin C, Wu Z, Detels R, 2011. Family support, quality of life and concurrent substance use among methadone maintenance therapy clients in China. Public Health 125 (5), 269–274. 10.1016/j.puhe.2011.01.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Luli AJ, Bounthavong M, Watanabe JH, 2020. Pharmacists on the front-line of solving a public health crisis: increasing access to medications for opioid use disorder. J. Am. Pharm. Assoc 60 (2), 297–303. 10.1016/j.japh.2019.10.014. [DOI] [PubMed] [Google Scholar]
  40. Maldonado G, Greenland S, 1993. Simulation study of confounder-selection strategies. Am. J. Epidemiol 138 (11), 923–936. 10.1093/oxfordjournals.aje.a116813. [DOI] [PubMed] [Google Scholar]
  41. Marsch LA, Bickel WK, Badger GJ, Stothart ME, Quesnel KJ, Stanger C, Brooklyn J, 2005. Comparison of pharmacological treatments for opioid-dependent adolescents - a randomized controlled trial. Arch. Gen. Psychiatry 62 (10), 1157–1164. 10.1001/archpsyc.62.10.1157. [DOI] [PubMed] [Google Scholar]
  42. Matson SC, Hobson G, Abdel-Rasoul M, Bonny AE, 2014. A retrospective study of retention of opioid-dependent adolescents and young adults in an outpatient buprenorphine/naloxone clinic. J. Addict. Med 8 (3), 176–182. 10.1097/ADM.0000000000000035. [DOI] [PubMed] [Google Scholar]
  43. Mattick RP, Breen C, Kimber J, Davoli M, 2009. Methadone maintenance therapy versus no opioid replacement therapy for opioid dependence. Cochrane Database Syst. Rev (3) 10.1002/14651858.CD002209.pub2. [DOI] [PubMed] [Google Scholar]
  44. McNeil R, Puri N, Boyd J, Mayer S, Hayashi K, Small W, 2020. Understanding concurrent stimulant use among people on methadone: a qualitative study. Drug Alcohol Rev. 39 (3), 209–215. 10.1111/dar.13049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Minozzi S, Amato L, Bellisario C, Davoli M, 2014. Maintenance treatments for opiate-dependent adolescents. Cochrane Database Syst. Rev (6) 10.1002/14651858.CD007210.pub3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Mitchell K, Durante SE, Pellatt K, Richardson CG, Mathias S, Buxton JA, 2017. Naloxone and the Inner City Youth Experience (NICYE): a community-based participatory research study examining young people’s perceptions of the BC take home naloxone program. Harm Reduct. J 14 (1), 1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Phillips M, DeBeck K, Desjarlais T, Morrison T, Feng C, Kerr T, Wood E, 2014. Inability to access addiction treatment among street-involved youth in a Canadian setting. Subst. Use Misuse 49 (10), 1233–1240. 10.3109/10826084.2014.891618. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Pilarinos A, Barker B, Nosova E, Milloy MJ, Hayashi K, Wood E, Kerr T, DeBeck K, 2020. Coercion into addiction treatment and subsequent substance use patterns among people who use illicit drugs in Vancouver, Canada. Addiction 115 (1), 97–106. 10.1111/add.14769. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Pilarinos A, Bromberg DJ, Karamouzian M, 2021. Access to medications for opioid use disorder and associated factors among adolescents and young adults: a systematic review. JAMA Pediatr. 10.1001/jamapediatrics.2021.4606. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. R Core Team, 2013. R: A language and environment for statistical computing. Vienna, Austria. [Google Scholar]
  51. Raheemullah A, Lembke A, 2019. Initiating opioid agonist treatment for opioid use disorder in the inpatient setting: a teachable moment. JAMA Intern. Med 179 (3), 427–428. 10.1001/jamainternmed.2018.6749. [DOI] [PubMed] [Google Scholar]
  52. Rothman KJ, Greenland S, Lash TL, 2008. Modern Epidemiology. Lippincott Williams & Wilkins. 10.1016/j.annemergmed.2008.06.461. [DOI] [Google Scholar]
  53. Simeone C, Shapiro B, Lum PJ, 2017. Integrated HIV care is associated with improved engagement in treatment in an urban methadone clinic. Addict. Sci. Clin. Pract 12 (1), 1–6. 10.1186/s13722-017-0084-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Smyth BP, Ducray K, Cullen W, 2018. Changes in psychological well-being among heroin-dependent adolescents during psychologically supported opiate substitution treatment. Early Inter. Psychiatry 12 (3), 417–425. 10.1111/eip.12318. [DOI] [PubMed] [Google Scholar]
  55. Socías ME, Wood E, Kerr T, Nolan S, Hayashi K, Nosova E, Montaner J, Milloy M-J, 2018. Trends in engagement in the cascade of care for opioid use disorder, Vancouver, Canada, 2006–2016. Drug Alcohol Depend. 189, 90–95. 10.1016/j.drugalcdep.2018.04.026. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. The Society for Adolescent Health and Medicine, 2021. Medication for adolescents and young adults with opioid use disorder. J. Adolesc. Health 68 (3), 632–636. 10.1016/j.jadohealth.2020.12.129. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Unger JB, Simon TR, Newman TL, Montgomery SB, Kipke MD, Albomoz M, 1998. Early adolescent street youth: an overlooked population with unique problems and service needs. J. Early Adolesc 18 (4), 325–348. 10.1177/0272431698018004001. [DOI] [Google Scholar]
  58. Viera A, Bromberg DJ, Whittaker S, Refsland BM, Stanojlovic M, Nyhan K, Altice FL, 2020. Adherence to and retention in medications for opioid use disorder among adolescents and young adults. Epidemiol. Rev 42 (1), 41–56. 10.1093/epirev/mxaa001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Vlahov D, Safaien M, Lai S, Strathdee SA, Johnson L, Sterling T, Celentano DD, 2001. Sexual and drug risk-related behaviours after initiating highly active antiretroviral therapy among injection drug users. Aids 15 (17), 2311–2316. 10.1097/00002030-200111230-00013. [DOI] [PubMed] [Google Scholar]
  60. Wood E, Stoltz J-A, Montaner JS, Kerr T, 2006. Evaluating methamphetamine use and risks of injection initiation among street youth: the ARYS study. Harm Reduct. J 3 (1), 1. 10.1186/1477-7517-3-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Zhou K, Li H, Wei X, Li X, Zhuang G, 2017. Relationships between perceived social support and retention among patients in methadone maintenance treatment in mainland China. Psychol. Health Med 22 (4), 493–500. 10.1080/13548506.2016.1164873. [DOI] [PubMed] [Google Scholar]

RESOURCES