Injection drug use and overdose among young adults who use prescription opioids non-medically

Abstract

Introduction

Non-medical prescription opioid (NMPO) use is a critical public health problem in the United States, with 2.1 million new initiates annually. Young adult NMPO users are at high risk for initiating injection drug use. We assessed correlates of injection drug use among young adult NMPO users in Rhode Island, a state heavily impacted by opioid overdose.

Methods

We used data from the Rhode Island Young Adult Prescription Drug Study (RAPiDS), which recruited 199 residents aged 18–29 who reported past-30-day NMPO use (65.3% male). We compared individuals who reported ever having injected with individuals who reported never injecting, using logistic regression to identify independent correlates of injection.

Results

Among eligible participants, the mean age was 24.6 years and 61.3% were white. Over one-quarter (n=59, 29.6%) of the sample had ever injected drugs. The majority (n=46, 78.0%) of participants who had ever injected drugs reported injecting heroin as her/his first drug; the majority also reported previously snorting her/his first drug that was injected (n=46, 78.0%). In multivariable analyses, white race, older age, lifetime homelessness, and ever having overdosed or seen someone overdose were independently associated with an increased likelihood of ever injecting drugs.

Conclusions

These findings demonstrate a high prevalence of lifetime injection drug use among young adults who use prescription opioids non-medically. Given the observed associations between injection drug use and witnessing as well as experiencing overdose, interventions are urgently needed to improve overdose education and naloxone distribution to young adult NMPO users who inject drugs.

1. BACKGROUND

Non-medical prescription opioid (NMPO) use—defined as intentional use of opioids outside of prescribed parameters or without a prescription (Lessenger & Feinberg, 2008; Voon & Kerr, 2013)—is a public health problem marked by increasing morbidity and mortality in the United States (Kuehn, 2007). In 2015, there were 12.5 million individuals who used an NMPO in the past year (representing 4.7% of the population aged 12 or older) and 2.1 million new initiates (Hughes et al., 2016). As a result, opioid overdose has become an epidemic in the United States, with one and a half times more drug overdose deaths than deaths due to motor vehicle crashes in 2015 (Rudd, Seth, David, & Scholl, 2016; Traffic Safety Facts, 2016). There were 20,101 prescription opioid overdose deaths in 2015 compared to 18,893 in 2014, an increase of 6.4% (ASAM, 2016; CDC, 2014). Yearly, opioid overdose costs $20.4 billion due to lost productivity and medical expenses (Inocencio, Carroll, Read, & Holdford, 2013). Rhode Island has been heavily impacted by NMPO use and opioid overdose. In 2015, 290 individuals died of an overdose, the fifth highest age-adjusted rate in the nation ("Drug Overdose Death Data"; "Drug Overdose Deaths").

The prevalence of NMPO use is particularly high among young adults; the percentage of past-year users was highest among 18- to 25-year-olds, reaching 8.5% in 2015 (Hughes et al., 2016). These rates of NMPO use trail only marijuana as the most commonly reported drug of use among young adults (Hughes et al., 2016). Among young adults, NMPO use has been associated with an increased likelihood of injection initiation and transitioning to heroin use (Carlson, Nahhas, Martins, & Daniulaityte, 2016; Cerda, Santaella, Marshall, Kim, & Martins, 2015; DeBeck et al., 2016; Mars, Bourgois, Karandinos, Montero, & Ciccarone, 2014). A range of injecting behaviors, including sharing needles and syringes as well as other injecting equipment, can result in an increased risk of infectious disease acquisition, including HIV and hepatitis C virus (HCV) (Hagan et al., 2001; Peters et al., 2016; Thorpe et al., 2002; Zibbell, Hart-Malloy, Barry, Fan, & Flanigan, 2014). Injection drug use also places individuals at higher risk for overdose than other forms of drug consumption (Darke & Hall, 2003). A study of 596 young adults in New York City and Los Angeles, California, found that recent injection drug use was associated with increased odds of non-fatal overdose (Silva, Schrager, Kecojevic, & Lankenau, 2013). In a study of 560 street-involved youth and young adults in Vancouver, Canada, 41.1% reported ever having injected drugs (Kerr et al., 2009). In this sample, non-fatal overdose was independently associated with injection drug use (Kerr et al., 2009).

Analysis of recent injection drug use trends shows a sharp increase in injection drug use among whites compared to blacks and Hispanics (Wejnert et al., 2016). Heroin use in particular, which was previously most frequent among low-income, non-white individuals living in urban areas, has become more common among middle-class white men and women living in nonurban settings (Cicero, Ellis, Surratt, & Kurtz, 2014). The increasing prevalence of heroin use is largely explained by use among young adults, with initiation rates two to seven times higher than other age groups and a rate of past-year use three and a half times higher than older adults (Ihongbe & Masho, 2016). Additionally, white race, ever having been arrested, and past-year NMPO use have been linked to increased odds of recent and lifetime heroin use among young adults (Ihongbe & Masho, 2016). As the need for evidence-based interventions to combat opioid overdose and opioid overdose mortality increases (Haegerich, Paulozzi, Manns, & Jones, 2014), it is necessary to assess the evolving correlates of injection drug use among young adults who use prescription opioids non-medically, a rapidly growing population.

This study aimed to assess sociodemographic, structural, childhood, and drug-related correlates of lifetime injection drug use among young adults who use prescription opioids non-medically in Rhode Island. Our primary hypothesis was that certain sociodemographic characteristics, such as white race and rural residence, structural factors, such as ever having been arrested, childhood exposures, such as emotional, physical, and sexual abuse, and drug-related experiences, such as ever having overdosed by accident, are correlated with injection drug use among young adult NMPO users. We also sought to establish a profile of the injection-related experiences of young adult NMPO users in order to inform targeted overdose prevention and education interventions.

2. MATERIAL AND METHODS

2.1 Study Design and Sample Selection

This analysis uses data from 199 participants in the Rhode Island Young Adult Prescription Drug Study (RAPiDS). Participants were recruited between January 2015 and February 2016 through a combination of targeted canvassing and mixed internet-based recruitment. In brief, targeted canvassing relied primarily on bus advertisements and flyering in areas where drug-using young adults were known to congregate, and mixed internet-based recruitment included regular posts on online classified sites (e.g., Craigslist), social media (i.e., Facebook), and forums (e.g., Reddit). The RAPiDS recruitment methodology has been detailed previously (Evans, Hadland, Clark, Green, & Marshall, 2016; Liebling et al., 2016; Macmadu, Carroll, Hadland, Green, & Marshall, 2017).

Eligibility criteria included: age 18 to 29 at the time of the interview, Rhode Island resident, not currently enrolled in substance use treatment, and able to complete an interview in English. Written informed consent was obtained. Participants were also required to confirm non-medical use of prescription opioids in the past 30 days by indicating which opioids she/he recently used non-medically, based on a modified version of the Substance Abuse and Mental Health Service Administration’s “pill card A” (SAMHSA, 2009).

Eligible participants completed a computer-assisted interview of approximately 45 minutes with a trained interviewer and sensitive portions of the survey were self-administered using a computer. Study participants were provided $25. The Brown University Institutional Review Board approved the study protocol.

2.2 Measures

The primary outcome for this analysis was self-reported lifetime injection drug use. The operational definition of this outcome was ever having used a needle to chip, fix, muscle, or inject drugs even once (yes vs. no).

The RAPiDS survey instrument measured sociodemographic characteristics, structural factors, childhood experiences, and drug-related experiences. The selection of variables to include in the analysis was guided by the a priori hypotheses described above. Specifically, the analysis included measures of sociodemographic factors such as age, sex, ethnicity, race, sexual orientation, educational attainment, employment, and geographic residence. Age was defined as a continuous variable measured by year. Sex was defined as a binary measure of sex at birth (male vs. female). Ethnicity was defined as being of Hispanic or Latino descent (yes vs. no). The race variable included the possible responses: American Indian or Alaska Native, Asian, black (African, Haitian, or of Cape Verdean descent), Native Hawaiian or other Pacific Islander, white, mixed, bi-racial, or multi-racial, and something else. For the purpose of these analyses, the categories were collapsed into white and non-white races. Sexual orientation was also collapsed into two categories: straight vs. gay, lesbian, bisexual, queer, or something else. Educational attainment was measured by highest level of education received. Employment was defined as being currently employed full-time or part-time (yes vs. no). Geographic residence type used self-reported current ZIP code or town of residence and was categorized as urban, suburban, or rural according to standard US census definitions and the Rhode Island land use survey (Rhode Island Land Use Trends and Analysis (Including Land use Surveys for the Period 1970–1995), 2000).

Structural factors included homelessness, juvenile detention, arrest, and incarceration. Homelessness was defined as ever having been homeless. History of juvenile detention, arrest, and incarceration were assessed by asking individuals if she/he had ever been detained in a juvenile detention center or training school, ever been arrested, and had ever been detained or incarcerated in an adult jail or prison, respectively.

Childhood experiences of abuse included emotional abuse, physical abuse, and sexual abuse. Specifically, participants were asked if, before the age of 18, a parent, stepparent, or adult living in her/his home swore at her/him, insulted her/him, or put her/him down (emotional abuse), a parent, stepparent, or adult living in her/his home hit her/him so hard that she/he had markers or were injured (physical abuse), and if she/he was ever sexually assaulted or abused. Drug-related experiences included in the analysis were ever having seen someone overdose, ever having overdosed by accident, ever having used heroin, and ever having used cocaine.

Participants who reported ever injecting drugs were asked to report a wider range of injection-related experiences. Participants were asked to report her/his age the first time they injected any drug, what drug she/he injected for the first time, if she/he had been snorting this drug before first injecting it, and what other drugs she/he was using in the week before her/his first injection. These participants were also asked to report which drugs she/he has ever injected. Participants who reported ever injecting drugs were also asked to report how often she/he injected in the past six months. These participants were asked to report where she/he acquires her/his needles and works. Possible responses included a close friend, a casual friend or acquaintance, a sex partner, a parent, other family members, a dealer, a needle exchange or dropin center, and a pharmacy. Multiple responses were acceptable and participants could also state sources that were not listed.

2.3 Statistical Analyses

First, we used descriptive statistics to summarize the characteristics of the complete sample and the sample of participants who reported ever injecting drugs. Next, we conducted Pearson’s chi-squared tests for categorical variables and ANOVA for continuous variables to assess differences for each variable between those participants who reported ever injecting and those who did not. We used Fisher’s exact test when one or more of the cells included fewer than 5 observations.

Next, we constructed a multivariable model using logistic regression, comparing individuals who reported ever injecting drugs with those who reported never injecting drugs. Variables that were found to be significant at the p<0.05 level in Table 2 were included in a baseline multivariable model, in addition to sex and recruitment source (i.e., targeted canvassing vs. mixed internet-based recruitment). To obtain a more parsimonious model, we first calculated the variance inflation factor (VIF) for all sets of variables and removed collinear covariates. Ever using cocaine was not eligible for inclusion in the multivariable modeling due to cells including zero participants. Then, a backward selection process was applied until the effect estimates for each variable were significant at p<0.10. The final model was adjusted for recruitment source and all p-values are two-sided. Analyses were conducted using Stata SE 13.1.

Table 2.

Description of RAPiDS participants by injection drug use (n=198^†)

	Never injected drugs 139 (70.2%) n (%)	Ever injected drugs 59 (29.8%) n (%)	χ2 (df)	p - value
Mean Age (Standard deviation)	23.9 (3.31)	26.0 (2.57)	2.07 (11)*	0.024
Sex			1.35 (1)	0.246
Male	87 (62.6%)	42 (71.2%)
Female	52 (37.4%)	17 (28.8%)
Ethnicity			(1)**	0.073
Hispanic or Latino descent	24 (17.3%)	4 (6.8%)
Non-Hispanic	115 (82.7%)	55 (93.2%)
Race			32.71 (1)	<0.001
White	67 (48.2%)	54 (91.5%)
Non-white	72 (51.8%)	5 (8.5%)
Sexual orientation			3.13 (1)	0.077
Straight	123 (88.5%)	47 (79.7%)
LGBQ and other	15 (10.8%)	12 (20.3%)
Education			(2)**	0.272
Less than high school	19 (13.7%)	4 (6.8%)
High school/GED	54 (38.8%)	21 (35.6%)
Beyond high school	66 (47.5%)	34 (57.6%)
Employment Status			1.51 (1)	0.219
Full or part-time	65 (46.8%)	22 (37.3%)
Unemployed	74 (53.2%)	37 (62.7%)
Geographic residence type			2.64 (2)	0.267
Rural	13 (9.4%)	10 (16.9%)
Suburban	14 (10.1%)	4 (6.8%)
Urban	107 (77.0%)	43 (72.9%)
Ever detained in juvenile detention center			1.20 (1)	0.274
Yes	30 (21.6%)	42 (71.2%)
No	109 (78.4%)	17 (28.8%)
Ever arrested			10.04 (1)	0.002
Yes	89 (64.0%)	51 (86.4%)
No	50 (36.0%)	8 (15.6%)
Ever incarcerated in jail or prison			12.35 (1)	<0.001
Yes	54 (38.9%)	39 (66.1%)
No	85 (61.2%)	20 (33.9%)
Ever homeless			14.27 (1)	<0.001
Yes	63 (45.3%)	44 (74.6%)
No	76 (54.7%)	15 (25.4%)
Childhood emotional abuse			5.41 (1)	0.020
Yes	88 (63.3%)	48 (81.4%)
No	48 (34.5%)	11 (18.6%)
Childhood physical abuse			0.08 (1)	0.777
Yes	49 (35.3%)	23 (39.0%)
No	84 (60.4%)	36 (61.0%)
Childhood sexual abuse			2.10 (1)	0.147
Yes	34 (24.5%)	21 (35.6%)
No	100 (71.9%)	38 (64.4%)
Ever seen someone overdose			21.93 (1)	<0.001
Yes	55 (39.6%)	45 (76.3%)
No	83 (59.7%)	14 (23.7%)
Ever overdosed by accident			26.23 (1)	<0.001
Yes	22 (15.8%)	30 (50.9%)
No	117 (84.2%)	29 (49.2%)
Ever used heroin			(1)**	<0.001
Yes	29 (20.9%)	56 (94.9%)
No	110 (79.1%)	3 (5.1%)
Ever used cocaine			(1)**	<0.001
Yes	72 (51.8%)	59 (100.0%)
No	67 (48.2%)	0 (0.0%)
Recruitment method			15.87 (1)	<0.001
Field	42 (30.2%)	36 (61.0%)
Internet	95 (68.3%)	23 (39.0%)

Notes:

^†1 participant excluded due to missing data.

Not all columns sum to 100% due to missing data and/or rounding.

Significance ascertained using a chi-square test unless otherwise noted.

^*Significance tested using an ANOVA.

^**Significance ascertained using a Fisher’s exact test.

LGBQ: Lesbian, gay, bisexual, or queer.

GED: General Education Development.

3. RESULTS

3.1 Descriptive Statistics

Among 199 participants, the mean age was 24.6 (SD=3.23), 65.3% (n=130) of the sample was male, and 14.1% (n=28) of the participants were of Hispanic or Latino descent. The majority (n=122, 61.3%) was white. Among the participants, 85.9% (n=171) reported their sexual orientation as straight and 13.6% (n=27) as lesbian, gay, bisexual, or queer (LGBQ). Among participants, 11.6% (n=23) had less than a high school education and 54.3% (n=108) were unemployed.

The distribution of other variables of interest are provided in Table 1. Of note, the majority (n=108, 54.3%) of the sample had ever been homeless. Among all participants, the majority (n=101, 50.8%) reported ever having seen someone overdose and over one-quarter (n=53, 26.6%) reported ever having overdosed by accident.

Table 1.

Characteristics of RAPiDS participants (n=199)

	Participants 199 (100%) n (%)
Mean Age (Standard deviation)	24.6 (3.23)
Sex
Male	130 (65.3%)
Female	69 (34.7%)
Ethnicity
Hispanic or Latino descent	28 (14.1%)
Non-Hispanic	171 (85.9%)
Race
White	122 (61.3%)
Non-white	77 (38.7%)
Sexual orientation
Straight	171 (85.9%)
LGBQ and other	27 (13.6%)
Education
Less than high school	23 (11.6%)
High school/GED	76 (38.2%)
Beyond high school	100 (50.3%)
Employment status
Full or part-time	87 (43.7%)
Unemployed	108 (54.3%)
Geographic residence type
Rural	23 (11.6%)
Suburban	18 (9.0%)
Urban	151 (75.9%)
Ever detained in juvenile detention center
Yes	48 (24.1%)
No	151 (75.9%)
Ever arrested
Yes	141 (70.9%)
No	58 (29.2%)
Ever incarcerated in jail or prison
Yes	94 (47.2%)
No	105 (52.8%)
Ever homeless
Yes	108 (54.3%)
No	91 (45.7%)
Childhood emotional abuse
Yes	136 (68.3%)
No	59 (29.6%)
Childhood physical abuse
Yes	72 (36.2%)
No	120 (60.3%)
Childhood sexual abuse
Yes	55 (27.6%)
No	138 (69.3%)
Ever seen someone overdose
Yes	101 (50.8%)
No	97 (48.7%)
Ever overdosed by accident
Yes	53 (26.6%)
No	146 (73.4%)
Ever used heroin
Yes	85 (42.7%)
No	114 (57.3%)
Ever used cocaine
Yes	132 (66.3%)
No	67 (33.7%)
Ever used a needle to inject drugs
Yes	59 (29.6%)
No	139 (69.8%)
Recruitment method
Field	79 (39.7%)
Internet	118 (59.3%)

Notes: Not all columns sum to 100% due to missing data and/or rounding.

LGBQ: Lesbian, gay, bisexual, or queer.

GED: General Education Development.

Over one-quarter (n=59, 29.6%) of the sample had ever injected drugs. The drugs that were most commonly reported as ever being injected were heroin (n=55, 93.2%), cocaine or crack (n=37, 62.7%), oxycodone (n=35, 59.3%), speedballs (i.e., heroin and cocaine mixed) (n=23, 39.0%), and hydromorphone (n=23, 39.0%).

Among participants who reported ever injecting drugs, the age of first injection drug use ranged from 11 to 27, with a mean age of 21.4 (SD=3.4). Comparatively, the age of first NMPO use among this cohort ranged from 12 to 25, with a mean age of 17.8 (SD=2.8). The majority (n=46, 78.0%) of participants who had ever injected reported injecting heroin as her/his first drug; 8.5% (n=5) reported injecting oxycodone first. Additionally, the majority (n=46, 78.0%) of participants reported first snorting the drug that they subsequently went on to inject. In the week before first injection, the other drugs most commonly used were cocaine or crack (n=18, 30.5%), oxycodone/acetaminophen (n=16, 27.1%), and oxycodone (n=14, 23.7%). The reported frequencies of injection drug use in the past six months were never (n=8, 13.6%), once or a couple of times (n=12, 20.3%), about once a month (n=5, 8.5%), at least every week (n=13, 22.0%), and daily (n=20, 33.9%). The majority (n=41, 80.4%) of these participants reported acquiring needles and works from a pharmacy.

3.2 Correlates of Injection Drug Use

A comparison of individuals who reported ever injecting drugs and those who did not is shown in Table 2. We observed a significant racial difference between the two groups. The vast majority (n=54, 91.5%) of participants who had ever injected drugs were white, compared to fewer than half (n=67, 48.2%) of participants who reported never injecting drugs. Additionally, approximately three-quarters (n=44, 74.6%) of individuals who had ever injected drugs had ever been homeless, while fewer than half (n=63, 45.3%) of individuals who had never injected drugs had ever been homeless. Over three-quarters (n=45, 76.3%) of individuals who had ever injected drugs had ever seen someone overdose, compared to 39.6% (n=55) of individuals who had never injected drugs. Similarly, approximately half (n=30, 50.9) of individuals who had ever injected drugs had ever overdosed by accident, compared to 15.8% (n=22) of individuals who had never injected drugs.

The VIF between ever being arrested and ever being incarcerated was 1.58, demonstrating moderate collinearity. The mean VIF of the model without ever being arrested before stepwise removal was 1.16; the mean VIF of the model without ever being incarcerated before stepwise removal was 1.19. As such, incarceration was excluded from the final model.

In the final model, factors independently associated with ever injecting drugs compared to never injecting drugs are shown in Table 3. Older age was associated with increased odds of ever having injected drugs. Compared to non-white participants, white participants were significantly more likely to report ever injecting drugs. Additionally, ever having been homeless and ever having seen someone overdose were each associated with approximately two and a half times the adjusted odds of ever injecting drugs compared to never having been homeless and never having seen someone overdose, respectively. Finally, participants who reported ever having overdosed by accident had almost three times the adjusted odds of ever having injected drugs compared to participants who reported never having overdosed by accident.

Table 3.

Adjusted odds ratios of ever injecting drugs vs. never injecting drugs: RAPiDS (n=195^*)

	Ever injected drugs (n=59)
	Adjusted odds ratio	95% Confidence Interval (CI)	p - value
Age (per year)	1.18	(1.03, 1.36)	0.015
Race
White	10.58	(3.51, 31.85)	<0.001
Non-white	1.00	(Ref)
Ever homeless
Yes	2.51	(1.09, 5.81)	0.031
No	1.00	(Ref)
Ever seen someone overdose
Yes	2.58	(1.09, 6.12)	0.032
No	1.00	(Ref)
Ever overdosed by accident
Yes	2.90	(1.22, 6.87)	0.016
No	1.00	(Ref)

Notes: Model adjusted for recruitment source.

^*The final model excludes 4 participants due to missing data.

The final model uses logistic regression and has 6 degrees of freedom.

The log likelihood of the model before stepwise removal is −74.28.

The log likelihood of the model after stepwise removal is −76.64.

The Nagelkerke R-squared of the model before stepwise removal is 0.520.

The Nagelkerke R-squared of the model after stepwise removal is 0.504.

The mean variance inflation factor for the model before stepwise removal is 1.16.

The mean variance inflation factor for the model after stepwise removal is 1.16.

4. DISCUSSION

The results of this study demonstrate a high prevalence of lifetime injection drug use among young adults who use prescription opioids non-medically. Important differences were observed between individuals who had ever injected drugs and those who had not. Sociodemographic factors such as older age and white race, structural factors such as ever having been homeless, and drug-related factors such as ever having seen someone overdose and ever having overdosed by accident were correlated with lifetime injection drug use.

Our findings are generally consistent with the existing literature on risk factors for injection drug use among high-risk youth. For example, homelessness has been found to be associated with injection drug use in a sample of street-involved youth (Feng et al., 2013), which our results confirm among young adults in Rhode Island who use prescription opioids non-medically. Prior studies have found a correlation between injection drug use and non-fatal overdose among young adults (Kerr et al., 2009; Silva et al., 2013); our results extend this finding to young adults who use prescription opioids non-medically. Over nine in ten participants who reported ever injecting drugs were white, and white race was associated with injection drug use in the multivariable model. These results likely reflect a fundamental shift in recent decades in the population of individuals who use opioids and/or inject drugs (Martins, Sarvet, & Santaella-Tenorio, 2017; Wejnert et al., 2016). Whereas the distribution of heroin initiates was evenly split between whites and non-whites as recently as the 1970s, almost 90% of initiates in the past decade have been of white race (Cicero, Ellis, Surratt, & Kurtz, 2014).

Certain risk factors for injection drug use among young adults which have been established by previous research were not found in our results. For example, previous studies of young adults have found a correlation between childhood sexual abuse and injection drug use, as well as earlier injection initiation (Hadland et al., 2012; Ompad et al., 2005); this was not supported by our findings. However, we did observe a difference in the bivariate analysis in the prevalence of childhood emotional abuse between participants who ever injected and those who did not. Although a history of childhood emotional abuse, ever being arrested, and ever experiencing incarceration were associated with lifetime injectio n drug use, these variables were not significant in multivariable analyses. Future studies with larger sample sizes are needed to identify the independent effects of sociodemographic characteristics, drug-related experiences, and structural factors on the likelihood of injection drug use among young adults who use prescription opioid non-medically.

These results indicate the importance of overdose prevention strategies among young adults who report NMPO use and also inject drugs. Overdose education and naloxone distribution (OEND) programs are important interventions that can likely prevent overdose deaths (Clark, Wilder, & Winstanley, 2014; Dichtl, Stover, & Dettmer, 2016). A review of OEND programs demonstrated effectiveness in preventing overdose mortality among people who inject drugs, as well as the cost-effectiveness of distributing naloxone to young adults who use heroin (Mueller, Walley, Calcaterra, Glanz, & Binswanger, 2015). In Massachusetts, OEND programs with enrollments of one to 100 individuals reduced rates of opioid overdose mortality by over one-quarter, while programs with enrollments of greater than 100 individuals resulted in reductions of almost half (Walley et al., 2013). New OEND programs or extensions of existing programs should be designed for and implemented among young adults. Intervention design should consider that young adults who use prescription opioids non-medically are often not engaged by traditional community-based services and harm reduction strategies and may be better reached through their social networks (Marshall, Green, Yedinak, & Hadland, 2016; Yedinak et al., 2016). Future research should assess the impact and effectiveness of these programs, which remain understudied specifically among young adults. In this analysis, a large proportion of individuals who reported never injected drugs also reported ever having seen someone overdose. Importantly, these individuals may be able to intervene when overdoses occur and should be included in OEND programs.

Additionally, the majority of participants who reported having recently injected drugs reported acquiring needles and works from a pharmacy. As such, naloxone training and distribution should be routinely offered alongside syringes and prescription opioid medications at pharmacies. In Rhode Island and Massachusetts, pharmacists can initiate the prescription and distribution of naloxone (Green, Dauria, Bratberg, Davis, & Walley, 2015). As a result, pharmacy-based naloxone comprised one-quarter of all naloxone distributed in Rhode Island from January 2014 to May 2015 (Green et al., 2015). There are a range of pharmacy-based naloxone models, including collaborative pharmacy practice agreements, pharmacy standing orders, and pharmacist as prescriber, which can be adapted by states to improve naloxone training and distribution (Green et al., 2015). Given the discrimination experienced by young adults who use prescription opioids non-medically and/or inject drugs, especially in clinical settings (Ahern, Stuber, & Galea, 2007; Liebling et al., 2016), peer-led interventions may also be effective. A study of young adults who inject drugs showed that a peer education intervention focused on education and skills-building was effective in reducing overall injection risk (Garfein et al., 2007). This peer-based approach can likely be adapted to additionally reduce the risk of opioid overdose.

Although this analysis does not directly assess the pathways for injection initiation among participants, the results suggest that most participants progressed from NMPO use, to non-injection (snorting) heroin use, to heroin injection. This progression has been identified in previous studies of young adults who inject heroin, but represents a shift from previous trends of prescription opioid injection prior to heroin injection (Lankenau et al., 2012; Mars et al., 2014). However, the majority of participants in this sample also reported experience with injecting prescription opioids (primarily oxycodone), suggesting that heroin and prescription opioid injection may occur contemporaneously.

This study has limitations that should be noted. First, the study relied on self-reported data, which may be subject to under-reporting or socially desirable reporting biases. To minimize the potential for these biases, sensitive portions of the survey were self-administered using a computer. Second, this study relied on cross-sectional data. This analysis uses lifetime experiences instead of exposures with shorter recall periods (e.g., within the past month) where applicable in an effort to maintain consistency with the recall period for the outcome of interest. Nonetheless, given the cross-sectional nature of our data, precise temporality cannot be ascertained. Therefore, the results of this analysis are purely correlative. Third, the sample size resulted in some small cell sizes, which increased the type II error rate and could have resulted in an inability to detect meaningful differences between groups. Fourth, although diverse methods were used to recruit participants residing throughout the state of Rhode Island, this population was not randomly sampled, which may limit the generalizability to other populations and settings.

5. CONCLUSIONS

This study found a high prevalence of injection drug use among young adult non-medical opioid users in Rhode Island. Older age, white race, ever having been homeless, ever having seen someone overdose, and ever having overdosed by accident were correlated with ever injecting drugs. Public health interventions, including improved overdose education and increased naloxone training and distribution among young adult NMPO users who inject drugs and may experience and witness accidental overdoses, are needed to reduce overdose-related mortality.

Highlights.

• Over one-quarter of participants had ever injected drugs.

• Over three-quarters of participants who ever injected drugs first injected heroin.

• White race, older age, homelessness associated with likelihood of injecting drugs.

• Past overdose or seeing an overdose associated with likelihood of injecting drugs.

Abbreviations

NMPO

Non-medical prescription opioid

RAPiDS

Rhode Island Young Adult Prescription Drug Study

LGBQ

Lesbian, gay, bisexual, or queer

GED

General Education Development