Drug and Alcohol Dependence 137 (2014) 48–54

Contents lists available at ScienceDirect

Drug and Alcohol Dependence journal homepage: www.elsevier.com/locate/drugalcdep

Interrupting the social processes linked with initiation of injection drug use: Results from a pilot study C. Strike a,b,∗ , M. Rotondi c , G. Kolla a , É. Roy d , N.K. Rotondi b , K. Rudzinski a,b , R. Balian e , T. Guimond a , R. Penn a , R.B. Silver e , M. Millson a , K. Sirois e , J. Altenberg e , N. Hunt f a

Dalla Lana School of Public Health, University of Toronto, 155 College Street, Toronto, ON M5T 3M7, Canada Social and Epidemiological Research Department, Centre for Addiction and Mental Health (CAMH), 33 Russell Street, Toronto, ON M5S 2S1, Canada c School of Kinesiology and Health Science, Faculty of Health, York University, 4700 Keele Street, Toronto, ON M3J 1P3, Canada d Faculté de médecine et des sciences de la santé, Université de Sherbrooke, Campus Longueuil 1111, rue St-Charles Ouest, Longueuil, QC J4K 5G4, Canada e South Riverdale Community Health Centre, 955 Queen Street East, Toronto, ON M4M 3P3, Canada f Centre for Research on Drugs and Health Behaviour, Faculty of Public Health and Policy, London School of Hygiene & Tropical Medicine, 15-17 Tavistock Place, London WC1H 9SH, United Kingdom b

a r t i c l e

i n f o

Article history: Received 21 August 2013 Received in revised form 17 December 2013 Accepted 5 January 2014 Available online 18 January 2014 Keywords: Injection drug use Initiation Peer intervention HIV prevention

a b s t r a c t Background: Injection drug use is a skill learned in social settings. Change the Cycle (CTC), a peerdelivered, one-session intervention, is designed to reduce among people who inject drugs (PIDs) injection initiation-related behaviours (i.e., speaking positively about injecting to non-injectors, injecting in front of non-injectors, explaining or showing a non-injector how to inject) and initiation of non-injectors. We hypothesized that participation in CTC would lead to reductions in initiation-related behaviours six months later. Methods: Using respondent driven sampling (RDS), 98 PIDs were recruited in Toronto, Canada to participate in pilot testing of CTC. The baseline session consisted of a structured interview, the peer-delivered CTC intervention, instructions regarding RDS coupon distribution, and an invitation to return in six months for a follow-up interview. For the 84 PIDs completing the six-month interview, we compared initiation-related behaviours at baseline with six-month follow-up. Results: The proportion of PIDs offering to initiate a non-injector was reduced from 8.4% (95% CI: 2.5, 15.9) at baseline to 1.59% (95% CI: 0.4, 3.7) at 6-month follow-up. The prevalence of speaking positively about injection to non-injectors also decreased significantly. The proportion of PIDs who helped a non-injector with a first injection at baseline was 6.2% (95% CI: 2.1, 11.3) and at follow-up was 3.5% (95% CI: 0.8, 7.1). Paired analyses of initiator baseline versus follow-up data showed a 72.7% reduction in initiation (95%CI: 47.7, 83.1). Conclusions: While further refinements remain to be tested, pilot study results suggest that CTC holds promise as a prevention intervention. © 2014 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Evidence suggests that human immunodeficiency virus (HIV) prevention programmes may have reached a plateau, and are not as effective to reduce transmission of hepatitis C (HCV; Kwon et al., 2009; Palmateer et al., 2010; Vlahov et al., 2004). Further, estimates of the median time to HCV infection point to a narrow window of opportunity to prevent new infections (Hagan et al., 2004; Roy et al., 2009). Scale-up of existing programmes may address these

∗ Corresponding author at: Dalla Lana School of Public Health, University of Toronto, 155 College Street, Toronto, ON M5T 3M7, Canada. Tel.: +1 416 978 6292; fax: +1 416 978 2087. E-mail address: [email protected] (C. Strike). 0376-8716/$ – see front matter © 2014 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.drugalcdep.2014.01.004

issues; however, new prevention strategies are also necessary (Degenhardt et al., 2010; Strathdee et al., 2010). Preventing the initiation of injection drug use holds promise to reduce transmission of both HIV and HCV (Day et al., 2005; Hunt et al., 1999). 1.1. Social learning theory and the initiation of injection drug use Social learning theory (Bandura, 1977, 1986) offers insight into the initiation of injection drug use and the potential to interrupt this process. This theory hypothesizes that people learn and modify their behaviours through interaction, observation, behavioural experimentation, and reinforcement with others in their environments (Bandura, 1977, 1986). Repeated exposure, either through verbal or visual modelling of a marginal or even feared behaviour can make the behaviour seem normal, acceptable,

C. Strike et al. / Drug and Alcohol Dependence 137 (2014) 48–54

and even desirable by desensitizing the observer to the possible risks of the behaviour (Bandura, 1977, 1986). Existing research supports this hypothesis with evidence showing that the vast majority of injectors report that prior to their first injection, they were exposed to injecting within in their social groups (i.e., family and friends), had observed others injecting and also heard about the positive effects of consuming drugs by injection (Abelson et al., 2006; Atlani et al., 2000; Bauman and Ennett, 1996; Crofts et al., 1996; Doherty et al., 2000; Durrant and Thakker, 2003; Frajzyngier et al., 2007; Harocopos et al., 2009; Khobzi et al., 2008; Kolla et al., 2009; McElrath and Harris, 2013; Neaigus et al., 2006; Roy et al., 2006; Sherman et al., 2002; Small et al., 2009; Stillwell et al., 2006; Strike et al., 2009; Witteveen et al., 2006). Moreover, most current injectors report that the decision to inject for the first time was their own and they actively sought out their first injection (Bryant and Treloar, 2007; Crofts et al., 1996; McElrath and Harris, 2013; Witteveen et al., 2006). The most common motives to begin injecting include: desire to consume drugs using a more efficient and economical method, to get a more intense ‘high’ to experiment, and/or to emulate injection drug using friends (Crofts et al., 1996; Small et al., 2009; Stillwell et al., 2006; Witteveen et al., 2006). Non-injecting drug users typically require the help of an experienced injector with their first injection (Bryant and Treloar, 2007; Frajzyngier et al., 2007; Harocopos et al., 2009; McElrath and Harris, 2013). The percentages of current injectors who have ever helped someone with a first injection range from 17% to 47% (Bryant and Treloar, 2008; Crofts et al., 1996; Hunt et al., 1998; Strike et al., 2009). Among those who report having helped someone with a first injection, many are ambivalent and/or regretful about helping non-injectors with their first injection (McElrath and Harris, 2013; Shelley et al., 1993; Sherman et al., 2002; Small et al., 2009). Small et al. (2009) note that the initiation of non-injectors is perceived as a moral boundary that is not to be, but is routinely, crossed by current injectors. While not all current injectors will cross this boundary, the majority of current injectors, including those who have not initiated someone, report engaging in initiation-related behaviours such as speaking positively about injecting to noninjectors, injecting in front of non-injectors, and explaining or showing a non-injector how to inject (Hunt et al., 1999). 1.2. Social learning theory and Change the Cycle Intervention Based on social learning theory, Change the Cycle (CTC) is an adaption of an intervention developed in the United Kingdom (Hunt et al., 1998) to reduce the occurrence of initiation-related behaviours. CTC integrates social learning theory by considering that if non-injectors are exposed to less injection-related talk (e.g., comments about the efficiency and intensity of the high to be gained from injecting versus other methods of consumption) and modelling of injection behaviours, the risk that they will develop an interest and motivation to inject drugs will be reduced. CTC operationalizes the idea that reducing initiation will require interrupting some of the social behaviours that influence initiation. Using preliminary research, we modified the UK intervention and instead of professional staff members we hired peer workers (i.e., people who currently injected drugs, lived in the community, with no training in social work or case management) trained in active listening methods to deliver the intervention. Active listening is a technique used within counselling, conflict resolution and training that requires the listener to verbally reflect back what a speaker has said (Helgesen and Brown, 1995). Built around a guided conversation, CTC incorporated this technique to help peer workers avoid introducing their own views or solutions. Active listening helps the peer workers assist intervention participants to think through and discuss ways in which they engage (or not) in initiation-related behaviours and to consider if and how they might avoid these

49

behaviours in the future. Using an intervention manual, the peer workers deliver CTC in one session encompassing seven modules to guide the conversation: the participant’s injection initiation event; their experiences, if any, of initiating others; the health, legal, and social risks of initiation for themselves and the noninjector; identification of aspects of their own behaviour that may inadvertently promote injecting to non-injectors (e.g., speaking positively about or modelling injection); and the generation and rehearsal of responses to a series of vignettes describing common initiation scenarios (see Table 1). CTC added a seventh new module about safer injection education to acknowledge that there are situations where initiation might happen anyway and also to target injection risks among current injectors. As Hunt et al. (1999) note, many current injectors engage in injection initiation-related behaviours. As such, we designed CTC for all current injectors in the hopes of reducing these behaviours, as well as initiation among those who have or might help someone with a first injection in the future. The objective of this paper is to assess among current PIDs changes to initiation-related behaviours (i.e., speaking positively about injecting, injecting in front of non-injectors, explaining or showing a non-injector how to inject) and initiation of noninjectors, following a peer-based intervention. 2. Methods 2.1. Recruitment and eligibility criteria We pilot tested CTC using a longitudinal study design and report data collected at baseline and six-month follow-up. Eligibility criteria included: aged 16 years and over; injected drugs in the past 30 days; lived in Toronto, Canada; spoke English; and able to provide informed consent. Since a sampling frame is not available for this population, we used respondent driven sampling (RDS) to recruit participants (Heckathorn et al., 2002; Heckathorn, 2002). RDS is similar to snowball sampling in its use of chain referral and peer recruitment. However, using Markov chain theory, Heckathorn (1997, 2002) showed that proportions (i.e., prevalence of a specific trait) in the sample will reach an equilibrium whereby they are no longer influenced by the choice of initial participants. For the initial participants (commonly referred to as “seeds”), we selected 10 people who: met the recruitment criteria; were connected with the study locale, a peer-based harm reduction programme at a community health centre; and, who were well known among other PIDs. After completing the baseline interview, participants were provided with three uniquely numbered RDS coupons, instructions about who and how to recruit, and invited to come back in six months for a follow-up interview. Participants received $25 CAD for the baseline interview and $5 CAD for each eligible participant that they recruited into the study (up to a maximum of six). Participants who completed the six-month follow-up interview were paid $25.

2.2. Measures All study procedures and data were collected by an interviewer, with research and frontline service delivery experience. The interviewer verified eligibility, administered pen-and-paper questionnaires and reminded participants of follow-up interview dates. Peer intervention workers delivered the CTC intervention, but did not collect any data. At baseline, recruitment seeds and potential participants presenting with a valid RDS coupon were asked questions to confirm eligibility. After providing written consent, each participant completed the baseline intervieweradministered questionnaire asking in reference to the past six months questions about: demographic characteristics, drug use, injection risk behaviours, Severity of Dependence Scale (Ferri et al., 2000; Gossop et al., 1995, 1997), and initiationrelated behaviours. For RDS weighting procedures each participant was also asked, ‘How many people do you know who inject drugs, who also know you, and who you’ve spoken to in the past 6 months?’ After the baseline interview, each participant completed the peer-delivered intervention session (see Table 1) and then a short interviewer-administered, post-intervention questionnaire. At the six-month follow-up interview, participants reconfirmed consent and completed a shortened version of the baseline interview with questions about changes to employment, income, housing, drug use, initiation-related behaviours and attitudes in the six months since the baseline interview. All study procedures took place at the study office. Baseline data were collected from June to September 2011 and six-month follow-up data from November, 2011 to February, 2012. Data from questionnaires were entered into Microsoft Access (Microsoft Office, 2010) by a research assistant and verified by another. All electronic data were stored on a secure server using a password-protected system.

50

C. Strike et al. / Drug and Alcohol Dependence 137 (2014) 48–54

Table 1 Summary of peer-delivered CTC guided conversation sessions modules. # 1 2 3 4 5

6

7

Description

Source

Introductions and summary of CTC and its goals Discuss participants own first injection (i.e., when, motivation, who present, feelings about it); summarize discussion Discussion about experiences of and/or perceptions about helping a non-injector inject for the first time and/or any refusals to help with a first injection; summarize discussion Discuss participant’s perceptions of the risks (health, social and legal) for the person who injects for the first time; summarize discussion Discuss participant’s perceptions of the risks (health, social and legal) for the person who helps someone inject for the first time; summarize discussion Ask about and discuss participant’s perceptions about the drug use behaviours linked with initiation (i.e., things people do that influence others to want to try injecting): • They see other people inject • When they feel like they are the odd one out because everyone else is injecting • When they hear about the good things like the rush and other benefits • When they hear that it is more economical to inject than use drugs another way • When they hear and see someone explain how to prepare a hit and its effects Link back to participants’ initiation story; summarize discussion Discussion and rehearsal of responses to six initiation scenarios, including if and how to avoid future requests from someone to help with a first injection. 1. Imagine you have a close friend who uses drugs but has never injected. He/she is there when you are preparing an injection and asks you to give her/him one. 2. Imagine that your friend says to you “I’m going to do it anyway, whether you help me or not, and if you don’t help me I’m going to make a mess of it”. 3. Imagine that your friend then says to you “If you don’t do it for me, I’m going to go to that scumbag to do it for me, and you know he won’t do a good job or watch out for me like you would” 4. Imagine that your friend then says to you “Look, if you do this for me, I’ll go and buy the drugs for both of us and I’ll give you a shot if you help me”. 5. Imagine that you are sick (withdrawal) and they offer you some drugs in exchange for help with a first hit. 6. Imagine that the person is drunk or stoned and is asking you for help with a first hit. Presentation of safer injection education video followed by peer led review of video content and discussion about sources of injection equipment in the community

Developed by CTC team Break the Cycle (BTC; Hunt et al., 1998)

2.3. Analyses Estimates of population proportions and 95% confidence intervals (CI) were obtained using the RDS analysis software, RDSAT (Volz et al., 2012). Statistical methods to appropriately adjust for individual weights and homophily were based on methods described by Heckathorn (1997, 2002, 2007); and Heckathorn et al. (2002). For reference, 95% CI derived from RDSAT were based on a bootstrap resampling procedure with 10,000 iterations. Our maximum recruitment chain was seven waves, which was sufficient for convergence of RDS estimates. RDS methods provide unbiased estimates of population proportions, thus our primary analysis focuses on the proportion of the population who modified their behaviour between baseline and six month follow-up interviews. This paired analysis improves statistical power as it accounts for the individual correlation of a participant’s responses across both time points. Thus, the primary analyses centre on three proportions of the population: (1) who did not change their behaviours between baseline and follow-up; (2) who did not report a behaviour at baseline but did so at follow-up; and (3) who exhibited the behaviour at baseline but not at follow-up. In this analysis, a 95% CI for a change in behaviour that includes zero suggests no evidence of change 6 months after the CTC intervention; while a 95% CI that does not include zero provides statistical evidence of a change in behaviour at the 5% level. Given the longitudinal nature of the RDS data, appropriate methods for missing data were required for the analysis of study results. Some individuals who were lost to follow-up were initiators at baseline and a simple presentation of prevalence at each time point would suggest a large reduction in the initiation behaviours due to the rarity of this behaviour and the small sample size. We used a multiple imputation strategy to incorporate variation in the follow-up data and measure individual-level changes over time. In this way, primary analyses focus on the proportion of individuals in the sample who improved their behaviours, did not improve their behaviours and did not exhibit any changes in their behaviours after participating in the CTC intervention. A multiple imputation algorithm was used in conjunction with the RDS library (Schonlau and Liebau, 2012) in Stata (StataCorp, 2011). For each variable of interest, we examined the baseline and six-month outcomes to estimate the probability of the behaviour at the six-month follow-up. For each of 10 imputations, we randomly generated the missing response value (yes or no) conditional on that individual’s baseline value for the outcome of interest. After generation of these responses, we created a complete dataset with the observed and generated values which were analyzed using the traditional RDS approach (creating a point estimate and bootstrap confident interval). We stored this point estimate and 95% CI for this imputation, and repeated for the remaining nine imputations. Next, we computed

BTC (Hunt et al., 1998) Adapted to the Canadian context from BTC (Hunt et al., 1998) Adapted to the Canadian context from BTC (Hunt et al., 1998) BTC (Hunt et al., 1998)

Adapted from BTC (Hunt et al., 1998) by CTC peer workers and research team members with lived experience to the Canadian context

Canadian AIDS Treatment Information Exchange, 2008. Hep C: Peer Voices and Safer Choices DVD. Toronto: Ontario. CATIE. Available at CATIE www.hepcinfo.ca/en/resources/safer-injection-demo. Accessed November 25, 2013

the average point estimates and average of the lower/upper CIs to obtain the final imputed prevalence and 95% CI. This approach increases the variation in the estimates and CIs compared to simple imputation techniques (Donder et al., 2006) and allows analysis of study results. Without the multiple imputation approach, potential analyses would be limited due to missing data and within group recruitment in RDS analyses.

3. Results 3.1. Population characteristics Of the 108 people who presented with valid coupons, 10 had not injected in the past 30 days and were deemed ineligible. We collected baseline data from 98 participants and six-month followup from 84 participants. At baseline, the majority of the population were male (66%), 40 years of age or older (67.2%), Caucasian (90.3%) and had completed less than a high school diploma (52.8%; see Table 2). Most were recipients of an income support programme (i.e., 37.6% social assistance welfare or 51.4% disability pension) and lived at ‘their own place’ (59.1%); others reported less stable living arrangements (i.e., 5.6% shelter; 10.4% hotel/rooming house/halfway house; or 3.3% street/outdoor location). Within the past six months, the majority had injected opiates other than heroin (61.4%; 95% CI: 47.0, 77.4) and/or heroin (63.1%; 95% CI: 48.2, 76.5) and many had also injected powder cocaine (41.9%; 95% CI: 31.5, 53.5) and/or crack cocaine (38.5%; 95% CI: 25.0, 52.3). 3.2. Baseline initiation-related behaviours In the six months prior to the baseline interview, 20.3% (95% CI: 10.4, 29.8) of the population reported that they had spoken positively about injection to a non-injector and 17.6% (95% CI: 9.3, 27.7)

C. Strike et al. / Drug and Alcohol Dependence 137 (2014) 48–54

51

Table 2 Demographic and drug use characteristics of participants. Sample (n = 98)

RDS weighted prevalence*

95% CI*

Gender Male Female

68 30

65.9% 34.1%

50.2–79.5 20.5–49.8

Age 20–29 years old 30–39 years old 40–49 years old 50 years and older

9 19 34 36

8.0% 24.8% 36.1% 31.1%

2.1–14.4 14.4–36.8 25.1–48.1 19.8–45.7

Race/ethnicity/cultural background (not mutually exclusive) White Aboriginal Black Other

89 17 9 6

88.4% 16.7% 8.3% 5.1%

73.0–98.0 7.8–27.4 2.6–16.4 0.7–12.0

Income support recipient Received social assistance Received disability pension

44 49

37.6% 51.4%

25.5–53.3 32.4–65.9

Education Less than high school diploma High school diploma Some post-secondary Completed post-secondary diploma/degree

60 14 19 5

52.8% 23.7% 19.2% 4.3%

36.4–68.3 12.4–37.9 7.4–32.5 0.9–9.7

Drugs injected in last 6 months (not mutually exclusive) Heroin Other non-prescription opioids Crack Cocaine Speedballs Methamphetamine

66 67 45 44 13 5

63.1% 61.4% 38.5% 41.9% 9.6% 2.9%

48.2–76.5 47.0–77.4 25.0–52.3 31.5–53.5 3.9–16.9 0.8–5.9

Non-injection drugs used in last 6 months – (not mutually exclusive) Cannabis Alcohol Crack Other opioids not Rx Cocaine Benzodiazepines (non-prescribed) Heroin Methadone (not Rx)

79 77 77 42 32 27 22 16

83.5% 77.9% 76.5% 39.9% 30.2% 28.6% 17.2% 16.0%

73.5–92.3 66.1–88.9 59.0–92.4 28.3–53.1 2.0–41.5 17.9–40.1 2.0–41.5 7.9–40.1

Housing status Living in their own place/supportive housing Living at someone else’s place Living in a shelter Living in a hotel/rooming house/halfway house Living on the street/in the bush

48 23 15 7 3

59.1% 21.6% 5.6% 10.4% 3.3%

44.1–73.4 9.0–35.0 0.6–24.3 2.9–17.8 0.3–8.2

*

Adjusted using RDSAT 6.0 (Volz et al., 2010) following procedures described in Heckathorn (1997).

had injected in front of a non-injector (see Table 3). Few reported that they had explained or shown how to inject to a non-injector (11.1%; 95% CI: 4.7, 18.9) and only 8.4% had offered to give a noninjector a first injection (95% CI: 2.5, 15.9) in six months prior to the baseline interview. A minority (25.7%; 95% CI: 9.8, 41.6) reported that they had ever initiated someone (i.e., given the first injection) and 6.2% (95% CI: 2.1, 11.3) had done so in the past six months.

3.3. Initiation-related behaviours at six-month follow-up At the follow-up interview, 11.9% (95% CI: 4.3, 21.55) reported that they had spoken positively about injection to a non-injector and 21.3% (95% CI: 11.6, 32.2) had injected in front of a non-injector in the previous six months (see Table 2). Few reported that they had explained or shown how to inject to a non-injector (11.4%; 95% CI:

Table 3 Population percentages of initiation related-behaviours at baseline and six-month follow-up adjusted using MI algorithm and Stata RDS approach. Baseline

Six month follow-up

Behaviour change scores No change in behaviour between baseline and six-month follow-up

Spoke positively about injection Injected in front of a non-injector Explained or showed how to inject Offered to give first injection Gave someone first injection

(%, 95% CI)

(%, 95% CI)

(%, 95% CI)

Increased behaviour (reported behaviour at six-month follow-up but not at baseline) (%, 95% CI)

20.3% (10.4, 29.8) 17.6% (9.3, 27.7) 11.1% (4.7, 18.9) 8.4% (2.5, 15.9) 6.2% (2.1, 11.3)

11.9% (4.3, 21.55) 21.3% (11.6, 32.2) 11.4% (4.0, 21.3) 1.5% (0.4, 3.7) 3.5% (0.8, 7.1)

85.2% (76.0, 92.2) 73.4% (56.8, 85.0) 90.3 (85.5, 95.2) 91.8% (84.1, 97.6) 94.8% (89.9, 98.5)

3.3% (0.0, 7.5) 15.6% (5.7, 29.9) 4.81% (0.8, 10.5) Cannot compute 0.7% (0.0, 3.6)

Decreased behaviour (reported behaviour at baseline but not at six-month follow-up) (%, 95% CI) 11.6% (5.4, 19.4) 12.0% (4.5, 22.1) 4.9% (1.1, 11.7) 8.2% (2.4, 15.9) 4.5% (1.0, 9.4)

52

C. Strike et al. / Drug and Alcohol Dependence 137 (2014) 48–54

4.0, 21.3) and very few participants offered to give a non-injector a first injection (1.5%; 95% CI: 0.4, 3.7) in the six months prior to the follow-up interview. At follow-up, 3.5% (95% CI: 0.8, 7.1) had given someone their first injection since the baseline interview. 3.4. Increase in initiation-related behaviours at six-month follow-up Paired analyses (i.e., baseline versus six-month follow-up) showed a small and non-significant percentage (3.3%; 95% CI: 0.0, 7.5) reported uptake of speaking positively about injecting to noninjectors in the six months after the intervention but not at baseline. As well, paired analyses showed a significant proportion 15.6% (95% CI: 5.7, 29.9) of the population who had not injected in front of non-injector at baseline reported doing so at six-month follow-up. Similarly, analyses showed a potential increase in the proportion of the population who had explained or shown a non-injector how to inject at follow-up but had not done so at baseline (Table 3). 3.5. No change or a decrease in initiation-related behaviours at six month follow-up Paired analyses showed no evidence of any non-initiators at baseline initiating non-injectors at the six-month interview. Paired analyses also showed that among those who reported at baseline to have spoken positively about injecting to non-injectors, a statistically significant 11.6% (95% CI: 5.4, 19.4) had not done so in the six months after the intervention. Analyses of injecting in front of non-injectors showed that 12.0% (95% CI: 4.5, 22.1) of the population who had injected in front of a non-injector at baseline reported not having done so at six-month follow-up after the intervention. Subsequent analyses showed that of the population who had explained or showed a non-injector how to inject at baseline, 4.9% (95% CI: 1.1, 11.7) did not report having done so at six-month follow-up after the intervention. For our primary outcome, paired analyses showed that a statistically significant 4.5% (95% CI: 1.0, 9.4) of the population who reported at baseline having given a first injection did not report doing so at six-month follow-up. 4. Discussion Most injection initiation prevention interventions have targeted non-injectors, and used educational models to prevent the transition to injection; however, these studies show little effect (Dolan et al., 2004). CTC and its predecessor Break the Cycle target the behaviours documented in the literature (Abelson et al., 2006; Crofts et al., 1996; Doherty et al., 2000; Frajzyngier et al., 2007; Harocopos et al., 2009; Khobzi et al., 2008; Neaigus et al., 2006; Roy et al., 2006; Sherman et al., 2002; Small et al., 2009; Stillwell et al., 2006; Witteveen et al., 2006) and anchored by social learning theory (Bandura, 1977, 1986) to be involved in the transition to injection drug use: speaking positively about injecting, injecting in front of non-injectors, explaining or showing a non-injector how to inject. Paired analyses showed support for the study hypothesis that after receiving the CTC intervention, there was a reduction in the number of participants who: (1) gave a non-injector their first injection; (2) spoke positively about injection in front of non-injectors; and (3) who offered to help with a first injection. In relation to social learning theory, this latter finding is important in terms of partially altering the social milieu that supports acquisition of knowledge about new behaviours. One of the goals of CTC is to reduce the modelling of injection drug use in front of non-injectors. This is important because as Bandura (1977, 1986) points out, visual modelling of a marginal or even feared behaviour can make the behaviour seem normal,

acceptable and even desirable. Six months after receiving the CTC intervention, the reports of injecting in front of non-injectors were not reduced. This result led to a re-examination of expected intervention impacts in light of what we know about the social nature of drug use. Injectors and non-injectors often buy and use drugs within the same social spaces, and this provides opportunities for people to learn about injecting by watching others inject (Abelson et al., 2006; Crofts et al., 1996; Doherty et al., 2000; Frajzyngier et al., 2007; Harocopos et al., 2009; Khobzi et al., 2008; Kolla et al., 2009; Neaigus et al., 2006; Roy et al., 2006; Sherman et al., 2002; Small et al., 2009; Stillwell et al., 2006; Strike et al., 2009; Witteveen et al., 2006). Given the social nature of these environments, it may be difficult for PIDs, due to barriers or lack of opportunities, to avoid injecting in front of others in these contexts. Given a high level of social disapproval towards encouraging or helping someone to inject for the first time (Small et al., 2009), we endeavoured to create a study environment conducive to reporting of this and other related initiation behaviours. We hired interviewers and peer intervention workers with many years of frontline experience and who had prior training regarding active listening and creating non-judgmental social spaces. When offered the RDS coupon by a study participant, potential participants were informed about the nature of the study. The interviewer also outlined the study topic prior to asking for informed consent to ensure that all participants were aware of the types of questions to be asked. We selected an interviewer-administered questionnaire after completing a preliminary assessment for this intervention on the same topic and with the same study population; when asked over 60% of participants requested an interviewer-administered questionnaire over self-complete format. Also, we embedded these and other initiation-related questions in the middle of the questionnaire to allow time for the participant to develop comfort with the interviewer. Participants disclosed information about many sensitive topics during the interview (e.g., source of income, types of legal and illegal employment, incarceration history) leading us to believe that the study environment was conducive to disclosure. At both baseline and follow-up, only one person reported encouraging non-injectors to inject for the first time (results omitted). This finding concurs with existing studies that report current injectors disapprove of encouraging non-injectors to inject for the first time (McElrath and Harris, 2013; Rhodes et al., 2011; Small et al., 2009). Two of the participants who disclosed having helped someone with a first injection did not return for a follow-up interview. Given that these participants had already disclosed initiation to the interviewer and other sensitive behaviours, we doubt that the reason for loss to follow-up was to avoid disclosing these behaviours. Nevertheless, we do not know if this or other reasons (e.g., moved away, incarcerated) contributed to the loss to follow-up. In the future, use of audio-computer assisted self-interviews (e.g., Macalino et al., 2002) may help to increase comfort to disclose sensitive behaviours such as encouraging or helping someone to inject drugs for the first time. This may allow us to better quantify the impact, if any, of responding in socially desirable ways to these and other sensitive questions. Designed as a pilot, this study had a small sample size which influenced its statistical power. Nevertheless, important differences over time were noted. The use of the RDS recruitment method is a notable strength as is the low rate of attrition over time. The use of RDS in a longitudinal framework required application of multiple imputation procedures to account for the loss of participants over the course of the study. Although imputed values relied on an individual’s baseline scores and the overall likelihood of changing behaviours in the sample, this ad hoc technique may require further methodological validation. Nonetheless, these results present a more appropriate estimate of the changes over time than other analytical options such as treating the missing data

C. Strike et al. / Drug and Alcohol Dependence 137 (2014) 48–54

as an outcome category or simple imputation techniques. These analytic approaches were required because the current literature lacks specification of methods to address RDS analysis of longitudinal studies. Moreover, there is the potential for participant loss to follow-up to have biased results, as two of seven baseline initiators (29%) required imputation techniques, compared to 12 of 91 (13%) non-initiators who were lost to follow-up. Note however that missing data mechanisms were not examined due to our small sample size and rarity of initiation behaviours (e.g., an increase of one baseline initiator translates to a 14% decrease in the attrition rate of baseline initiators). For this pilot study, we did not recruit a comparison sample, which, along with randomization, will be essential to future assessments of the CTC intervention. This is the first study to report follow-up data evaluating a peer-delivered initiation intervention aimed at reducing initiation to injection drug use. Analyses showed a reduction in initiation and potential initiation-related behaviours among current injectors. We measured outcomes six months after the intervention and we do not know if and for how long any behavioural changes were sustained beyond this timeframe. Future evaluations will need to determine the durability of the effect and if a ‘booster’ session – where the concepts presented in the initial intervention are reviewed with participants – might be worthwhile. Also, CTC is based on a one-on-one, peer-delivered intervention session. This model is costly and a group model version of the CTC intervention is being considered by the team. Finally, CTC could be further adapted into a network based intervention (see Latkin et al., 2013). While further refinements remain to be tested, results from this pilot study suggest that CTC holds promise as a prevention intervention. Role of the funding source Gillian Kolla receives a PhD. fellowship from the Canadian Institutes of Health Research. Tim Guimond receives a career salary award from the Ontario HIV Treatment Network. Rebecca Penn receives a PhD. fellowship from the Canadian Institutes of Health Research. Contributors CS, GK, ÉR, TG, RB, RBS, KS, MM, JA, NH conceived of the design of the CTC intervention and study. GK and RP recruited participants and collected baseline data. RP and GK trained and supervised the peer workers. GK and RBS were responsible for attrition prevention procedures and GK collected the six month follow-up interviews. RBS and three other peer workers delivered the CTC intervention. CS, GK and MR performed the data analysis and interpretation with input from GK, TG, ÉR, NR. CS and MR drafted the paper and all other authors provided critical revisions. All authors approved the final version of the paper for submission. Conflicts of interest The authors have no conflicts or other disclosures to report. References Abelson, J., Treloar, C., Crawford, J., Kippax, S., Van Beek, I., Howardet, J., 2006. Some characteristics of early-onset injection drug users prior to and at the time of their first injection. Addiction 101, 548–555. Atlani, L., Caraël, M., Brunet, J.-B., Frasca, T., Chaika, N., 2000. Social change and HIV in the former USSR: the making of a new epidemic. Soc. Sci. Med. 50, 1547–1556. Bandura, A., 1977. Social Learning Theory. Prentice Hall Inc., Englewood Cliffs, NJ. Bandura, A., 1986. Social Foundations of Thought and Action: A Social Cognitive Theory. Prentice Hall Inc., Englewood Cliffs, NJ. Bauman, K.E., Ennett, S.T., 1996. On the importance of peer influence for adolescent drug use: commonly neglected considerations. Addiction 91, 185–198.

53

Bryant, J., Treloar, C., 2007. The gendered context of initiation to injecting drug use: evidence for women as active initiates. Drug Alcohol Rev. 26, 287–293. Bryant, J., Treloar, C., 2008. Initiators: an examination of young injecting drug users who initiate others to injecting. AIDS Behav. 12, 885–890. Crofts, N., Louie, R., Rosenthal, D., Jolley, D., 1996. The first hit: circumstances surrounding initiation into injecting. Addiction 91, 1187–1196. Day, C.A., Ross, J., Dietze, P., Dolan, K., 2005. Initiation to heroin injecting among heroin users in Sydney, Australia: cross sectional survey. Harm Reduct. J. 15, 2–8. Degenhardt, L., Mathers, B., Vickerman, P., Rhodes, T., Latkin, C., Hickman, M., 2010. Prevention of HIV infection for people who inject drugs: why individual, structural, and combination approaches are needed. Lancet 376, 285–301. Doherty, M.C., Garfein, R.S., Monterroso, E., Latkin, C., Vlahov, D., 2000. Gender differences in the initiation of injection drug use among young adults. J. Urban Health 77, 396–414. Dolan, K., Clement, N., Rouen, D., Rees, V., Shearer, J., Wodak, A., 2004. Can drug injectors be encouraged to adopt non-injecting routes of administration (NIROA) for drugs? Drug Alcohol Rev. 23, 281–286. Donder, A.R., van der Heijden, G.J., Stijnen, T., Moons, K.G., 2006. Review: a gentle introduction to imputation of missing values. J. Clin. Epidemiol. 59, 1087–1091. Durrant, R., Thakker, J., 2003. Substance Use and Abuse: Cultural and Historical Perspectives. Sage Publications Inc., London, UK. Ferri, C.P., Marsden, J., de Araujo, M., Laranjeira, R.R., Gossop, M., 2000. Validity and reliability of the Severity of Dependence Scale (SDS) in a Brazilian sample of drug users. Drug Alcohol Rev. 19, 451–455. Frajzyngier, V.M., Neaigus, A., Gyarmathy, V.A., Miller, M., Friedman, S.R., 2007. Gender differences in injection risk behaviors at the first injection episode. Drug Alcohol Depend. 89, 145–152. Gossop, M., Best, D., Marsden, J., Strang, J., 1997. Test-retest reliability of the Severity of Dependence Scale. Addiction 92, 353. Gossop, M., Darke, S., Griffiths, P., Hando, J., Powis, B., Hall, W., Strang, J., 1995. The Severity of Dependence Scale (SDS): psychometric properties of the SDS in English and Australian samples of heroin, cocaine and amphetamine users. Addiction 90, 607–614. Hagan, H., Thiede, H., Des Jarlais, D.C., 2004. Hepatitis C virus infection among injection drug users: survival analysis of time to seroconversion. Epidemiology 15, 543–549. Harocopos, A., Goldsamt, L.A., Kobrak, P., Jost, J.J., Clatts, M.C., 2009. New injectors and the social context of injection initiation. Int. J. Drug Policy 20, 317–323. Heckathorn, D.D., 1997. Respondent-driven sampling: a new approach to the study of hidden populations. Soc. Probl. 44, 174–199. Heckathorn, D.D., 2002. Respondent-driven sampling II: deriving valid population estimates from chain-referral samples of hidden populations. Soc. Probl. 49, 11–34. Heckathorn, D.D., 2007. Extensions of respondent-driven sampling: analyzing continuous variables and controlling for differential recruitment. Sociol. Methodol. 37, 151–207. Heckathorn, D.D., Semaan, S., Broadhead, R.S., Hughes, J.J., 2002. Extensions of respondent-driven sampling: a new approach to the study of injection drug users aged 18–25. AIDS Behav. 6, 55–67. Helgesen, M., Brown, S., 1995. Active Listening: Introducing Skills for Understanding. Cambridge University Press, Cambridge, NY. Hunt, N., Griffiths, P., Southwell, M., Stillwell, G., Strang, J., 1999. Preventing and curtailing injecting drug use: opportunities for developing and delivering ‘route transition interventions’. Drug Alcohol Rev. 18, 441–451. Hunt, N., Stillwell, G., Taylor, C., Griffiths, P., 1998. Evaluation of a brief intervention to reduce initiation into injecting. Drugs (Abingdon Engl) 5, 185–193. Khobzi, N., Strike, C., Cavalieri, W., Bright, R., Myers, T., Calzavara, L., Millson, M., 2008. Initiation into injection: necessary and background processes. Addict. Res. Theory 17, 1–14. Kolla, G., Balian, R., Altenberg, J., Silver, R., Roy, E., Hunt, N., Millson, P., Strike, C., 2009. Exploring the context of injecting for the first time – a qualitative study. In: The Ontario HIV Treatment Network’s 2009 Annual Research Conference, Toronto. Kwon, J.A., Iversen, J., Maher, L., Law, M.G., Wilson, D.P., 2009. The impact of needle and syringe programs on HIV and HCV transmissions in injecting drug users in Australia: a model-based analysis. J. Acquir. Immune Defic. Syndr. 51, 462–469. Latkin, C., Donnell, D., Liu, T.Y., Davey-Rothwell, M., Celentano, D., Metzger, D., 2013. The dynamic relationship between social norms and behaviors: the results of an HIV prevention network intervention for injection drug users. Addiction 108, 934–943. Macalino, G.E., Celentano, D.D., Latkin, C., Strathdee, S.A., Vlahov, D., 2002. Risk behaviors by audio computer-assisted self-interviews among HIV-seropositive and HIV-seronegative injection drug users. AIDS Edu. Prev. 14 (5), 367–378. McElrath, K., Harris, J., 2013. Peer injecting: implications for injecting order and bloodborne viruses among men and women who inject heroin. J. Subst. Use 18, 31–45. Microsoft Office, 2010. Access. Microsoft, Redmond, WA. Neaigus, A., Gyarmathy, V.A., Miller, M., Frajzyngier, V.M., Friedman, S.R., Des Jarlais, D.C., 2006. Transitions to injecting drug use among non-injecting heroin users: social network influence and individual susceptibility. J. Acquir. Immune Defic. Syndr. 41, 493–503. Palmateer, N., Kimber, J., Hickman, M., Hutchinson, S., Rhodes, T., Goldberg, D., 2010. Evidence for the effectiveness of sterile injecting equipment provision in preventing hepatitis C and human immunodeficiency virus transmission among injecting drug users: a review of reviews. Addiction 105, 844–859.

54

C. Strike et al. / Drug and Alcohol Dependence 137 (2014) 48–54

Rhodes, T., Bivol, S., Scutelniciuc, O., Hunt, N., Bernays, S., Busza, J., 2011. Narrating the social relations of initiating injecting drug use: transitions in self and society. Int. J. Drug Policy 22, 445–454. Roy, É., Boudreau, J.F., Boivin, J.F., 2009. Hepatitis C virus incidence among young street-involved IDUs in relation to injection experience. Drug Alcohol Depend. 102, 158–161. Roy, É., Morissette, C., Haley, N., Gutiérrez, N., Rousseau, L., Denis, V., 2006. Pourquoi commencer? L’initiation à l’injection de drogues selon les jeunes de la rue. Drogues, Santé Société 2, 45–75. Schonlau, M., Liebau, E., 2012. Respondent driven sampling. Stata J. 12, 72–93. Shelley, Y., Jones, K., Hunt, N., 1993. Amphetamine Injecting: Risk, Sharing and Helpseeking: An Ethnographic Study. Maidstone Health Authority, Maidstone. Sherman, S.G., Strathdee, S., Smith, L., Laney, G., 2002. Spheres of influence in transitioning to injection drug use: a qualitative study of young injectors. Int. J. Drug Policy 13, 113–120. Small, W., Fast, D., Krusi, A., Wood, E., Kerr, T., 2009. Social influences upon injection initiation among street-involved youth in Vancouver. Canada: a qualitative study. Subst. Abuse Treat. Prev. Policy 4, 1–8. StataCorp, 2011. Stata Statistical Software: Release 12. StataCorp LP, College Station, TX.

Stillwell, G., Hunt, N., Taylor, C., Griffiths, P., 2006. The modelling of injecting behaviour and initiation into injecting. Addict. Res. 7, 447–459. Strathdee, S.A., Hallett, T.B., Bobrova, N., Rhodes, T., Booth, R., Abdool, R., Hankins, C.A., 2010. HIV and risk environment for injecting drug users: the past, present, and future. Lancet 376, 268–284. Strike, C., Kolla, G., Balian, R., Altenberg, J., Silver, R., Roy, E., Hunt, N., Millson, P., 2009. Giving someone their first hit – implications for HIV prevention programming. In: The Ontario HIV Treatment Network’s 2009 Annual Research Conference, Toronto. Vlahov, D., Wang, C.L., Galai, N., Bareta, J., Mehta, S.H., Strathdee, S.A., Nelson, K.E., 2004. Mortality risk among new onset injection drug users. Addiction 99, 946–954. Volz, E., Wejnert, C., Cameron, C., Spiller, M., Barash, V., Degani, I., 2012. Respondentdriven Sampling Analysis Tool (RDSAT) Version 7.1. Cornell University, Ithaca, NY. Volz, E., Wejnert, C., Degani, I., Heckathorn, D., 2010. Respondent-Driven Sampling Analysis Tool (RDSAT) Version 6.0 (beta). Cornell University, Ithaca, NY. Witteveen, E., Van Ameijden, E.J., Schippers, G.M., 2006. Motives for and against injecting drug use among young adults in Amsterdam: qualitative findings and considerations for disease prevention. Subst. Use Misuse 41, 1001–1016.