Declarations

1. The efficacy of psychosocial interventions to reduce sexual and drug blood borne virus risk behaviours among people who inject drugs: A systematic review and meta-analysis Gail Gilchrist,Davina Swan, KideshiniWidyaratna, Julia Elena Marquez, Liz Hughes, Noreen DadiraiMdege, Martyn-St James M, JuditTirado Muñoz

2. Declarations • No conflict of interest

3. Introduction • Among people who inject drugs (PWID), the prevalence of Hepatitis C virus (HCV) is estimated to be 5-90% (Hahne et al., 2013) and <1-50% (Mathers et al., 2008)for HIV • Sharing of needles/syringes pose the greatest risk of BBV transmission among PWID. Public Health England’s, “Shooting Up” report reported that while needle and syringe sharing is lower than a decade ago, around one in seven PWID continue to share needles and syringes and that there is an increased risk of infection for those who inject amphetamines and amphetamine-type drugs, such as, mephedrone. • Transmission probability of HCV through sharing injecting paraphernalia is over 8 times lower than that through needles/syringes (approximately 0.19-0.30% and 2.5%, respectively) (Corson et al., 2013) • In Scotland, it is estimated that 38% of HCV infections are a result of paraphernalia sharing and 62% are a result of needle/syringe sharing among PWID(Corson et al., 2013)

4. Rationale • Opiate substitution therapy and needle exchanges (MacArthur et al., 2014; Turner et al., 2011; Degenhardt et al., 2010) are effective in reducing HIV and HCV among PWID. • Increasing the coverage of these interventions can reduce HCV prevalence among PWID, these reductions are modest and psychosocial interventions are required to further decrease HCV prevalence (Degenhardt et al., 2010) by educating PWID about transmission risks and motivating them to reduce sexual and drug taking risk behaviours.

5. Need to update the evidence… • Several systematic reviews and meta-analysis of psychosocial interventions to reduce HIV and HCV risk behaviours among drug users have reported modest effects (Meader et al., 2013; Sacks-Davis et al., 2012; Hagan et al., 2011) • However some reviews included non RCTs • Most recent review includes studies up to 2011 • Some reviews do not discriminate between PWID and drug users who do not inject drugs • There has been no recent systematic review focusing on the efficacy of psychosocial interventions to reduce injecting and sexual risk behaviours among samples of PWID, rather than people who use drugs, regardless of their injecting status

6. Aims and definitions • to determine the efficacy of psychosocial interventions in reducing injecting and sexual risk behaviours compared to a control group, who received either an intervention of lesser time or intensity, or usual care  (Sacks-Davis et al., 2012). • Psychosocial interventions were defined as interventions that aimed “to change behaviour through the exchange of information, typically delivered by a clinician or educator” including “brief interventions, motivational interviewing, cognitive behavioural therapy, contingency management…” (Walsh et al., 2014).

7. Methods: Systematic review with meta-analysis • Studies were eligible for inclusion if: • participants were PWID or results presented for PWID; • they were RCTs; • outcome/s included: • (i) any injecting risk behaviour including sharing of needle/syringes or other injecting paraphernalia, and injecting frequency, reported separately or as an aggregated outcome, • (ii) any sexual risk behaviour including unprotected sex or number of sexual partners, reported separately or as an aggregated outcome; • psychosocial interventions were compared to a control group • Relevant RCTs published during 1 January 2000 – 26 May 2015: MEDLINE, PsycINFO, CINAHL and Cochrane Collaboration. Clinical Trials databases were searched to identify trials in this area and related publications, with update search in MEDLINE to 9 December 2016.

8. Sub group analyses • sub-group analyses were conducted to compare psychosocial interventions with • (1) treatment as usual; • (2) education or information; • (3) HIV testing and counselling; • (4) control interventions of lesser time or intensity with and (5) without OST. • length of time in months from the end of the intervention to the final follow-up of included trials were compared • where there was more than one intervention group, data from the most relevant psychosocial intervention to the aims of this review were compared to the control intervention.

9. Results: PRISMA diagram

10. Quality of trials • Cochrane risk of bias used to assess quality • Number of sessions ranged from 1-22 • For interventions with >1 session, retention or adherence to the intervention ranged from 50% to 95%

11. Study characteristics • Most studies (n=18) conducted in the US, delivered in drug treating settings (n=14) and delivered by staff • In total, 12,840 participants (35% female; range 0–100%) were enrolled; the majority were PWID (84.5%) (range 16–100%). • Psychosocial interventions were compared with usual care (n = 4), education or information (n = 9), HIV testing and counselling (n = 5), interventions of lesser time or intensity with (n = 12) and without OST (n = 2), 8 interventions incorporated peer mentoring from an index participant to change the behaviours of other PWID • Most interventions (n = 14) were delivered to individual participants; 8 were delivered to groups, 2 to couples and remaining 8 trials delivered interventions in a combination of ways

12. Efficacy of psychosocial interventions versus control interventions in reducing ANY injecting risk behaviours among people who inject drugs Each line represents effect estimate and weight for each study

13. Efficacy of psychosocial interventions versus control interventions in reducing ANY injecting risk behaviours among people who inject drugs Does cross 0 – so favours neither control nor intervention Doesn’t cross 0 – so, favours intervention

14. Efficacy of psychosocial interventions versus control interventions in reducing ANY injecting risk behaviours among people who inject drugs Overall effect estimate (all studies combined)

15. Efficacy of psychosocial interventions versus control interventions in reducing NEEDLE and SYRINGE SHARING among people who inject drugs

16. Efficacy of psychosocial interventions versus control interventions in reducing the SHARING of OTHER INJECTING PARAPHERNALIA among people who inject drugs

17. Efficacy of psychosocial interventions versus control interventions in reducing ANY SEXUAL RISK BEHAVIOURS among people who inject drugs

18. Limitations • Limitations include the low number of studies for inclusion in some of the sub-group analyses of behavioural outcomes and intervention delivery modes • Heterogeneityin terms of the interventions studied and their duration, as well as differences in sample sizes and characteristics, length of follow-up, and assessment methods used to determine risk behaviours. • Lack of consistency across studies may have contributed to the moderate levels of heterogeneity (I2=58%-61%) noted in the meta-analyses.

19. Conclusions and recommendations • Psychosocial interventions effectively reduce injecting and sexual risk behaviours among PWID compared to control conditions, with 16 of the 32 trials reporting greater reductions among PWID in the intervention conditions • Psychosocial interventions should be delivered to needle exchange attenders and those engaged with drug treatment to reduce BBV transmission and other injecting and sexually acquired infections

20. Acknowledgements • This open access review was funded by the National Institute for Health Research Health Technology Assessment (project number 13/17/04). • http://www.kcl.ac.uk/ioppn/depts/addictions/research/drugs/bloodborneviruses.aspx • Gail.Gilchrist@kcl.ac.uk

21. Analysis • The principal summary measure was the standardized mean difference (SMD) • As outcome data were presented as dichotomous or continuous data across included trials, odds ratios were recalculated as SMD to allow pooling of data   • I2of 25% was considered low heterogeneity, 50% moderate and 75% high heterogeneity [28]. In the inverse variance method, individual effect sizes were weighted according to the reciprocal of their variance calculated as the square of the standard error.