Influence of Peer-Based Needle Exchange Programs on Mental Health Status in People Who Inject Drugs: A Nationwide New Zealand Study

Alleviating the personal and social burden associated with substance use disorders requires the implementation of a comprehensive strategy, including outreach, education, community interventions, psychiatric treatment, and access to needle exchange programs (NEP), where peer support may be available. Given that substantial research underscores the potential benefits of peer support in psychiatric interventions, we aimed to conduct a national survey to examine key domains of mental health status in people who inject drugs (PWID) in New Zealand. PWID were recruited from 24 pharmacies and 16 dedicated peer-based needle exchanges (PBNEs) across the country. We focused on two mental health outcomes: (1) affective dysregulation, across the three emotional domains of the Depression Anxiety Stress Scale, due to its role in the maintenance of continued drug use, and (2) positive cognition and effective health- and drug-related information exchange with the provider, using the Satisfaction with Life Scale and an ad hoc questionnaire, respectively, in view of their association with improved mental health outcomes. We hypothesized that access to peer support would be associated with mental health benefits for PWIDs. Remarkably, the results of a multistep regression analysis revealed that irrespective of sex, age, ethnicity, main drug used, length of drug use, and frequency of visits to the NEP, the exclusive or preferential use of PBNEs predicted significantly lower depression and anxiety scores, greater satisfaction with life, and increased health-related information exchange with the service provider. These findings demonstrate for the first time an association between access to peer support at PBNEs and positive indices of mental health, lending strong support to the effective integration of such peer-delivered NEP services into the network of mental health services for PWID worldwide.

implementation of harm reduction strategies, currently defined as a comprehensive package aimed at the prevention, treatment, and care for people who inject drugs (PWID), is widely recognized as a satisfactory evidence-based approach to minimize the health risks associated with injecting drugs (3). Key elements of such intervention are the provision of safe equipment and materials, provision of up-to-date information about safe injecting practices, and access to health and counseling referrals.
Stimulated by changes of drug policy in the Netherlands that shifted the focus from detoxification to harm reduction, an environment permissive to self-organization of PWID led to the formation of the first underground exchanges (the "junkiebonden") in the early 1980s. Some years later, fueled by the positive impact these organizations were having on the community, funding initiatives from the Dutch Ministry of Health led to a spread of NEP schemes throughout the country (4). In New Zealand, the number of notified cases of AIDS had been growing through the 1980s (5), and in 1987, the government decriminalized the sale of needle and syringes, implementing in 1988 the first nationwide NEP in the world (6,7). Initially conceived as a pharmacy and general practitioner service, the New Zealand scheme quickly expanded with the addition of drug user groups contracted to provide educational support, later constituting charitable trusts and independent NEP operating a peer-to-peer system. At present, there are some 180 pharmacy outlets and 21 peer-based needle exchanges (PBNEs) participating in the New Zealand NEP scheme (www.needle.co.nz).
The peer-based strategy, which in New Zealand has evolved from the peer user groups of the 1980s, is an organic process whereby clients, rather than staff, provide direct service to their peers, particularly in the distribution of needles, syringes, and associated injection equipment. Building on existing community networks, trained peers assist PWID to access exchangers, distribute information about safer drug use and safer sex, and facilitate referrals to other user-friendly health services. Although not exempt of limitations and shortcomings (8), the peer service model, in addition to disseminating safer practices and making a variety of health services more accessible, emphasizes the importance of reconstructing relationships with family, friends, and the community at large (9,10), which may ultimately contribute to promote greater psychological stability and enhance the benefits of psychiatric interventions. In the area of mental health, there is considerable theoretical and experiential support for the notion that peer support may exert a positive influence in multiple relevant domains, including empowerment (11), symptom distress (12), self-esteem (13), and social integration (14). However, no previous studies have investigated the potential psychiatric benefits of peer support in the context of NEP on a large scale. We conducted a nationwide survey aimed at exploring the psychological well-being of PWID using pharmacies and PBNEs across New Zealand.
In the present study, we drew on the well-established triad of affect, cognition, and behavior, as key psychological mechanisms influencing health outcomes (15). By using this framework, we assessed two psychological health outcomes: (i) affect and (ii) cognitions and behaviors. First, we examined measures of general affect prompted by the close association between positive and negative affect states as possible inductors or outcomes for sustained drug use (16). Second, we assessed the extent to which individuals demonstrated positive cognitions about life and felt able to both share personal health-related information and access information on safe drug use, as these cognitions and behaviors are linked to wellness, enhanced risk perception, and self-efficacy (17).

MaTerials anD MeThODs sample
Peer-based needle exchanges and pharmacies (Levels 1 and 2 outlets, see www.needle.co.nz for a complete description of services and type) under the NEP scheme received the questionnaires by post. Included with the questionnaires there was a prepaid postal bag and an instructions sheet for staff. At each pharmacy or peerbased outlet, there was one main contact person (usually the manager of the pharmacy or NEP) who was briefed personally or over the phone and instructed to inform all staff to hand out the survey to PWID. PWID were asked to participate in the survey and those who expressed willingness were first given an information sheet with details of the general purpose of the study. The information sheet indicated that the information provided by participants would remain anonymous at all times and that through completion of the survey consent would be obtained to use the data for analysis and publication. For this purpose, each NEP was asked to allocate a private space within their premises. No interviewer was present to prevent bias. The survey was completed by the PWID with no time limit. The estimated time to complete it was 30 min approximately. To compensate for the time invested in completing the questionnaire, a NZ $10 supermarket voucher was given to the participants. Surveys were collected from 315 respondents (170 males, 141 females, with 4 who did not disclose their sex), aged from 19 to 66 years (mean age = 42.40, SD = 9.2), from regions throughout the north (Auckland, Palmerston, Midlands, Wellington, Northland, and Napier) and south (Christchurch/ Canterbury, Dunedin/Southland, and Nelson/Malborough) islands of New Zealand, including 24 pharmacies (16 in the north island and 8 in the south island) and 16 PBNEs (11 in the north island and 5 in the south island). The study was approved by the Human Ethics Committee of the University of Canterbury and by the New Zealand NEP, the Pharmaceutical Society, and the Pharmacy Guild of New Zealand. The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. Table 1 provides a summary of other demographic and drug use data relating to ethnicity, type and main drug used, length of drug use, preferred use of NEP, and frequency of visits to the NEP.

assessment of affective state
To assess the psychological status of the PWID in the affective domain, the Depression Anxiety Stress Scale 21 (DASS-21), a short form of Lovibond and Lovibond's 42-item self-report, was utilized (18). DASS-21 comprises three dimensions of seven items each: depression, anxiety, and stress. The scale aims to capture an    individual's distress by evaluating these three common negative emotional states. Each question is answered on a 4-point Likert scale: 0 (never), 1 (sometimes), 2 (often), and 3 (almost always). Evidence has shown the reliability and validity (i.e., internal consistency and test-retest) of the scale in both clinical and nonclinical settings (18)(19)(20).

assessment of Well-Being
The Diener Satisfaction with Life Scale (SWLS) was used as a reliable indicator of subjective well-being (21). This scale consists of five items designed to measure global cognitive judgments of one's life satisfaction. The participant is required to answer each question on a 7-point Likert scale: 1 (strongly disagree) to 7 (strongly agree). Past research supports the reliability and validity of this scale (22).

assessment of health-related information exchange
To examine the self-perceived ability of PWID to share experiential information in relation to their drug use, including health information, and to access information on drug use and safe practices, a 5-point Likert questionnaire consisting of eight ad hoc questions was devised. This health-related information exchange questionnaire (IEQ) included questions aimed to measure the degree to which PWID were willing to share personal experiences at the NEP, felt they had an adequate knowledge about safe practices, felt comfortable about asking questions regarding drug use and health referral services, and perceived the information available at the NEP as being useful and effective.

Missing Data
Across the 10 variables for which data were collected [of which three of the perceived variables (negative affect, satisfaction with life and information exchange) comprised 32 items], 90 cases were removed from the analysis due to missing data, leaving a cohort size of n = 225 (124 males, 101 females, mean age = 42.37, SD = 9.12). A series of independent group t tests and chi-square test [we used SPSS for Windows 22 (IBM Corp, 2013) for all statistical tests] were computed for those variables for whom individuals were removed due to having some missing data [ranging from n = 67 to 86 (as data remained for some of these individuals for other variables in the analysis)] and those individuals included in the analysis (n = 225). Table 2 shows that no statistical significant differences occurred for mean scores between both groups for the established variables: DASS, SWLS, age, length of use, frequency  of visits to the NEP, and preferred use of NEP. Furthermore, no significant association was found between inclusion and exclusion due to missing data for sex (χ 2 = 0.07, p = 0.797), ethnicity (χ 2 = 0.09, p = 0.765), and main drug of use (χ 2 = 3.03, p = 0.082).

clinical caseness
To demonstrate the clinical casesness of affective state among the respondents, the DASS provides severity ratings for each of the subscales of depression, anxiety, and stress. Table 3 shows a breakdown of the scores by each of the severity rating provided by the DASS manual scoring.

Factor analysis
With the measures we collected, we identified three possible domains: emotional evaluations of affect (comprising depression, anxiety, and stress of the DASS), cognitive evaluations of affect (comprising SWLS scores), and health-related information exchange (comprising the IEQ scores). To explore these as three separate outcome domains, we performed an exploratory factor analysis to explore the underlying structure of the items from the DASS, SWLS, and IEQ. The number of participants (225) to variables (34) ratio (6.6:1) exceeded the recommended minimum ratio needed for EFA of 5 to 1 (with a minimum number of participants of 150) (23, 24). All items were subjected to maximum likelihood analysis (Kaiser-Meyer-Olkin measure of sampling adequacy = 0.92; Bartlett's test of sphericity, ( 2 = 4,749.93, df = 561, p < 0.001). To determine the number of factors to extract, we performed a parallel analysis of Monte Carlo simulations (25) that allowed us to determine the number of factors by comparing the eigenvalues of a higher value with those that might be expected from purely random data. The fourth eigenvalue (11.40, 4.79, 2.62, and 1.26) failed to exceed the fourth mean eigenvalue (1.82, 1.71, 1.63, and 1.56) calculated from 1,000 generated data sets with 225 cases and 34 variables, suggesting that a 3-factor solution was appropriate. Therefore, a three-factor solution was explored using a promax rotation, as we expected the factors to be correlated, with delta set to 0. The results of this analysis are presented in Table 4.
Meaningful loadings were assessed using the criteria of 0.32 ("poor"), 0.45 ("fair"), 0.55 ("good"), 0.63 ("very good"), and 0.71 ("excellent") (26). From this solution, all the items from the DASS loaded on the first factor, with one item loading with a value of 0.33 and the rest of loadings ranging from 0.48 to 0.86. All the IEQ items loaded on the second factor with one item loading with a value of 0.34, and the rest of the loading values ranging from 0.52 to 0.89. Finally, all the SWLS items loaded on the third factor, with loading values ranging from 0.54 to 0.91. Moreover, the correlations between the factors ranged from −0.10 to −0.50, suggesting the factors had no more than 25% of shared variance. Therefore, the current findings suggest that the current items used fall under three different general domain: emotional evaluations of affect (in which higher scores represent higher levels of negative emotional affect), cognitive evaluations of affect (in which higher scores represent higher levels of positive cognitive affect), and self-perceived ability to share and access relevant health information (in which higher scores represent higher levels of health-related information exchange).
However, the finding that the items of the DASS load on one factor in this context questions whether the DASS scores is best represented, at least in this sample, by one overall score, or is best considered in terms of three factors, given theoretical and empirical evidence that the DASS commonly forms three factors of depression, anxiety, and stress (18). Therefore, we subjected just the items of the DASS scale to a confirmatory factor analysis (CFA). A key focus of CFA is to demonstrate the incremental value of proposed models (27). We compared the goodness of fit, a unidimensional model, representing an underlying latent factor structure of negative affect with a three-factor model comprising depression, anxiety, and stress. To assess the goodness of fit of the data, we looked at the five statistics recommended by Hu and Bentner (28) and Kline (29): the relative χ 2 (CMIN/ DF), the comparative fit index (CFI), the non-normed fit index (NNFI), the root mean square error of approximation (RMSEA), and the standardized root mean square residual (SRMR). The following are the criteria to assess whether the model fit was "acceptable": (i) CMIN/DF should be less than three to be acceptable, (ii) the CFI, GFI, and NNFI should exceed 0.90 to be acceptable, (iii) the RMSEA should not exceed 0.08 and be below 0.06 to be a "good" fit, and (iv) the SRMR values less than 0.08 are "acceptable" and those less than 0.05 are "good" (28,29). Improvement on the value of the model was assessed by changes in CFI (ΔCFI) being > 0.01 (30). Comparing the two models, the three-factor model (CMIN/ DF = 2.02, CFI = 0.93, GFI = 0.87, NNFI = 0.92, RMSEA = 0.068, SRMR = 0.046) shows acceptable and improved (as indicated by ΔCFI being > 0.01) goodness-of-fit statistics over the one-factor model (CMIN/DF = 2.98, CFI = 0.87, GFI = 0.78, NNFI = 0.85, RMSEA = 0.094, SRMR = 0.059). Therefore, we computed the following scale scores to represent three domains, in terms of overall scores for SWLS (α = 0.85) to assess cognitive evaluation of affect, overall scores for IEQ to assess information sharing (α = 0.87), and the three DASS subscale scores for depression (α = 0.90), anxiety (α = 0.84), and stress (α = 0.90) to assess emotional evaluations of affect.

Multiple regression
We ran 5 three-step multiple regressions, with overall SWLS and IEQ scores and DASS subscales scores used in the series as dependent variables, with sex, age, and ethnicity (recoded for NZ European [1] versus non-NZ European [2]) as the predictor variables in step 1; then main drug used (recoded as opiate [1] versus stimulant [2]), length of use, and frequency of visits to the NEP as predictor variables in step 2; and, finally, we incorporated preferred use of NEP as a predictor variable in step 3. An a priori sample size calculator for a multiple regression study, given a desired probability level of p < 0.05, the number of predictors in the model being 7, the anticipated effect size being medium ( f 2 = 0.15), and the desired statistical power level of .8, calculated the minimum required sample size as n = 103, suggesting that our current sample size of n = 227 used for this analysis exceeded these criteria. The variance inflation factors (VIFs) and tolerance factors for each of the single predictor variables were no larger than 1.20 and no smaller than 0.83, respectively. Therefore, they did not contravene the threshold values for VIF of at least 5 and tolerance statistics of less than 0.2 that are used to suggest collinearity between independent variables (30). For each regression (see Table 5

DiscUssiOn
Although community development programs providing access to injecting equipment exist in other countries, including European countries, Australia, and USA, there is a considerable resistance from governments to support integrated NEP of the kind currently implemented in New Zealand (31,32), with some NEP in Europe having been recently dismantled due to political opposition (33). In many countries, there remains considerable counterproductive, not evidence based, sociopolitical debate over the ethical and financial implications of harm reduction strategies, despite its proven cost effectiveness (3,34). Such debate fails to address not only the consequences for public health in terms of transmission of blood-borne viruses but also the psychological and social needs of vulnerable and marginalized populations. NEP activities are enabled in New Zealand by the Ministry of Health through dedicated (peer based) exchanges, mobile units, and participating pharmacy outlets. Pharmacies operate at different levels depending on the type of injecting equipment provided and distribute educational materials to PWID on safe injecting. Dedicated exchanges and associated mobile units are run and staffed largely by paid workers who have had life experiences as injecting drug users themselves and therefore understand the lifestyle and challenges PWID face. Albeit training is sparse, such workers are well placed to offer advice on safe injecting and available referral health services as appropriate. The emphasis of these PBNEs is not only on harm reduction but also on a model of intentional psychosocial support and outreach, where the peers employed in the support role are considered to be further  along in their recovery journey. Therefore, they are capable and willing to provide leadership, advice, shared life experiences, and emotional support to those who enter their services (35,36). Sharing credible positive experiences, practical strategies, and coping mechanisms has long been regarded as a form of support that people with experience in dealing successfully with personal mental health problems can offer others facing similar challenges (12,37). Albeit the harm reduction strategy implemented by way of NEP is widely recognized to minimize the health risks associated with injecting drugs (1), scientific evidence that peer support can be of psychological benefit for PWID in this specific context is lacking.
The current study was the first to examine on large-scale key aspects linked to mental health status and welfare of PWID using peer-based and non-peer-based NEP schemes. Participants were recruited from 24 pharmacies and 16 PBNEs across the country of New Zealand. We focused on three aspects of well-being, including emotional evaluations of negative emotional symptomatology (specifically depression, anxiety and stress) (38), cognitive evaluations of affect (satisfaction with life), and health-related information exchange, as both predictors and outcomes that might be linked to improved mental health status and recovery in PWID (16). First, the results indicated that the sample of PWID comprised a large number of subjects with psychiatric vulnerability, as revealed by the severity rating of the DASS (Table 3). Second, the findings clearly showed that the preference for PBNEs over pharmacy-based services (exchanges) shared unique variance with lower depression, lower anxiety, greater satisfaction with life, and increased health-related information exchange, while controlling for sex, age, ethnicity, main drug used, length of drug use, and frequency of visits to the NEP. In terms of the salience of these findings, considering the size of unstandardized and standardized beta coefficient, that effect was larger for the health-related information exchange, with comparatively smaller associations for depression, anxiety, and satisfaction with life. Notwithstanding, these data show for the first time that the use of PBNEs, in comparison with other similar services, is associated with positive mental health outcomes.
Where peers are employed as providers of services and support within traditional or specialized health agencies, peer support is generally framed within a model of wellness focused on effective functioning and recovery rather than the illness and its specific symptoms (8). A core feature defining the effectiveness of this model is the ability of the peer to engage with clients on the same level through increased empathy and deeper understanding of their challenges (39). Previous evidence suggests that such community-oriented, peer-delivered interactions have the potential to engage hard to reach or marginalized groups and positively impact on multiple layers of risk and behavioral change and particularly on people who regularly use drugs (40)(41)(42). Peer support has been consistently linked in the mental health literature with enhanced acceptance, self-esteem, self-efficacy, quality of life, community inclusion, empowerment, and willingness to work toward recovery through exposure to role modeling and alternative, more functional, worldviews (8,12,13,43). The current study was correlational in nature and therefore did not allow us to identify the specific mechanisms mediating the positive effects of peer services on the well-being and affective state of PWID. Notwithstanding, the findings for positive mental health outcomes and effective health-related information exchange variables revealed an association of the peer support service with positive enhancement processes, which may occur via the self or social factors (11,14). Given the nature of the data and the fact that observations were collected concurrently, threefold interactions may have given rise to such enhancement processes. First, both higher levels of effective health-related information exchange and reduced affective dysregulation could lead to a preference for PBNEs via heightened perceived well-being and self-efficacy, through a process akin to self-empowerment (11). Second, a preference for PBNEs might reflect elevated levels of social activation (14), which in turn could promote greater levels of positive mental health and more fluid health-related information exchange with the service provider. The third possibility is that the two processes combine together, be it within individuals, or across the sample.
In terms of recommendations for public and mental health policies, the current findings have far-reaching implications for the development of effective programs of harm reduction and especially for the psychological support and, in cases of psychiatric co-morbidity, rehabilitation of PWID. We showed here that patent psychological benefits were associated with access to peer support at PBNEs in the areas of well-being, affective regulation, and effective communication with the service providers. Such benefits are likely to impact positively at multiple levels, both psychological and social, and could go a long way in facilitating remission from continued drug use, recovery, and strengthening of social networks within the community. Taken together, these findings highlight the need to standardize peer support roles in terms of their values, skills, knowledge base, and remit and for further integration of such peer-supported programs within national policies regulating NEP and health care for PWID worldwide.

eThics sTaTeMenT
The study was approved by the Human Ethics Committee of the University of Canterbury and by the New Zealand NEP, the Pharmaceutical Society, and the Pharmacy Guild of New Zealand. PWID were asked to participate in the survey by staff at PBNEs and pharmacies and those who expressed willingness were first given an information sheet with details of the general purpose of the study. The information sheet indicated that the information provided by participants would remain anonymous at all times and that through completion of the survey consent would be obtained to use the data for analysis and publication. No additional considerations. aUThOr cOnTriBUTiOns BH, CH, and JC designed the study; BH collected the data; JM and JC analyzed the data; and JM and JC wrote the paper.