HCV Communication Within Ego-centric Networks of Men and Women who Inject drugs

Abstract

Introduction:

Leveraging interpersonal communication among social networks of people who inject drugs (PWID) may be an innovative strategy to increase awareness and access to hepatitis C (HCV) care. However, little is known about HCV communication patterns among PWID and if these patterns vary by gender.

Methods:

Egocentric network data collected at baseline from HCV-infected PWID enrolled in a randomized HCV treatment intervention trial in Baltimore, Maryland was analyzed. Logistic generalized estimating models were conducted to identify predictors of HCV communication.

Results:

Among 227 PWID, the mean age was 43.8 (SD=10.3), 28.2% (n=64) were women and 71.8% (n=163) were men. Female participants reported 516 dyadic relationships and male participants 1139 dyadic relationships. While there were significant gender differences based on socio-demographics, risk behavior and network composition, there were few differences in HCV communication patterns. Both men and women had increased odds of HCV communication with alters who are currently enrolled in drug treatment (AOR 1.7, 95% CI: 1.3–2.4), alters with whom participants share drug preparation equipment (AOR 3.0, 95% CI: 1.9–4.6), alters who are sex partners compared to kin (AOR 3.0; 95% CI: 1.9–4.9) and alters with whom respondents have increased trust (AOR 1.1; 95% CI: 1.11.2) and daily/weekly interactions (AOR 1.7; 95% CI 1.3–2.1).

Conclusion:

PWID engaged with trusted alters about HCV disclosure and information, highlighting the important role network interventions could play in this vulnerable population.

Background

In 2017, the United States Department of Health and Human Services declared the opioid crisis a public health emergency. In addition to significant increases in overdose related morbidity and mortality (Hedegaard et al. 2017), there have been marked increases in incident hepatitis C (HCV) infection among people who inject drugs (PWID) (Holtzman et al., 2021; Zibbell et al., 2018). HCV infection is an important preventable public health problem. Over time, chronic infection can lead to substantial liver disease, including cirrhosis and hepatocellular carcinoma. Temporally, in PWID, HCV infection almost always precedes HIV infection. For instance, reports of HCV foreshadowed the recent HIV outbreaks in Indiana (Conrad et al., 2015) and, more recently, in West Virginia and Massachusetts (Centers for Disease Control and Prevention, 2011; Evans et al., 2018; Page et al., 2019; Zibbell et al., 2015).

Up until 2014, HCV treatment required the use of interferon, which was associated with severe side effects and low rates of HCV cure (Feld & Foster, 2016; Pawlotsky, 2016;Mushtaq et al., 2020). The recent availability of safe, effective direct-acting antiviral (DAA) agents, which are shown to cure over 95% of those who have access to these treatments, put the elimination of HCV as a public health challenge within reach (Falade-Nwulia et al., 2017). In 2016, the World Health Organization set laudable goals for hepatitis elimination by 2030 (WHO). However, awareness of HCV serostatus, testing, and cure remain low among PWID (Collier et al., 2015; Falade-Nwulia, Gicquelais, et al., 2020; Grebely et al, 2017; Kåberg et al., 2017), who comprise the core of the HCV epidemic in many high-income countries including the United States (Grebely & Dore, 2017; Hajarizadeh et al., 2013; Nelson et al., 2011; Shepard et al., 2005). HCV testing and treatment barriers for PWID have been extensively reported (Strauss et al., 2008; Swan et al., 2010; Treloar et al., 2013). Barriers include the perception that HCV is benign, concerns about lack of affordability, and stories of negative experiences with liver biopsy and HCV treatment-related side effects from the inferferon era that continue to be shared within peer networks of PWID (Christiani et al., 2008; Kurtz et al., 2005; Miller-Lloyd et al., 2020; Neale et al., 2008; Strauss et al., 2008; Swan et al., 2010; Treloar et al., 2013; Zeremski et al., 2013). PWID also face significant barriers to navigating and accessing healthcare in traditional settings due to competing priorities and stigmatization of PWID in healthcare interactions (Motavalli et al., 2020), further deterring HCV testing and treatment with oral DAA therapies. These barriers, as well as vulnerability to HCV, is heightened for women who inject drugs (WWID) due to biological factors and gender-based norms (Carr & Gramling, 2004; Falade-Nwulia et al., 2020; Malinowska-Sempruch, 2015). Compared to male counterparts, WWID report more frequent sharing of drug preparation equipment, which introduces greater opportunities for exposure to HCV (Meyers et al., 2020; Toro-Tobón et al., 2019; Bogart et al., 2005; Davey-Rothwell & Latkin, 2007; Des Jarlais et al., 2012; El-Bassel et al., 2014a, 2014b; Esmaeili et al., 2017). There is also greater stigma towards drug use for women than men due to women being stereotyped as promiscuous and unable to fulfill traditional gender roles as primary caregivers (El-Bassel et al., 2014a; Gibson & Hutton, 2021). These representations hinder healthcare access for WWID as they may hide their drug use for fear of being judged negatively and of having their children removed from their care (Meyers et al., 2020; Myers, Fakier, & Louw, 2009).

Given PWID’s barriers to navigating traditional healthcare systems, alternative strategies are needed to increase PWID’s awareness of and access to HCV testing and treatment. One successful strategy may be to leverage existing social networks to increase interpersonal communication about HCV testing and treatment. Interpersonal communication among network members is a critical source of information sharing, advice seeking and support (Goldsmith, 2004; Rogers Everett, 1995) and can contribute to uptake of harm reduction innovations (Rogers Everett, 1995). This strategy is especially important among marginalized and stigmatized populations who may have medical mistrust (Abadie et al., 2018). For example, interventions have successfully incentivized PWID to recruit network members into HCV care (Falade-Nwulia, Ward, et al., 2020) and act as peer mentors to provide support and advice as participants navigate HCV care (Crawford & Bath, 2013; Janda & Mergenhagen, 2017; Lafferty et al., 2018). However, little is known about how HCV is discussed within PWID networks outside the scope of an intervention. Understanding with whom PWID talk about HCV can provide insight into the types of interventions needed to increase awareness and uptake of HCV testing and treatment. Furthermore, despite WWID’s increased vulnerability to HCV, they are understudied within the larger drug-using population and there is a dearth of analyses that disaggregate data by gender (Malinowska-Sempruch, 2015). As a consequence, gender-specific needs of WWID are frequently overlooked in HCV prevention and treatment services (Iversen et al 2015; United Nations Office on Drugs Crime, 2014). Therefore the objective of this analysis is to identify how interpersonal characteristics impact HCV communication among PWID, with a focus on exploring differences between men and women.

Methods

Study Design and Sample

Data come from the baseline surveys of a randomized trial to enhance Hepatitis C and HIV prevention and care among people using substances in Baltimore, Maryland (Dayton et al., 2019). Inclusion criteria for enrollment included being age 18 years or older and having a positive HCV antibody test. Participants were also encouraged to recruit network members who were drug and/or sexual partners to join the study. The present analysis was restricted to participants who reported injection drug use in the prior 6 months.

Measures

Survey items assessed respondents’ own attributes and behaviors, as well as elicited characteristics of respondents’ personal social networks. Personal social networks were elicited by asking participants to define their social support network, which is the network of family, friends, neighbors, and community members that is available in times of need to give psychological, physical, and financial help (Lin et al., 1979). For the personal network inventory, participants enumerated the first names of network members who provided them advice, pitched in to help do things, loaned them money, whom they entrust with their money, and with whom they socialize, all within the prior 6 months. Once this list was established, respondents answered a series of questions about each network member (hereafter referred to as “alter”).

The outcome of interest was whether or not the respondent had discussed HCV with alters (a dyadic variable). This variable was a composite variable where communication included talking with an alter about the following HCV-related topics in the prior 6 months: respondent’s HCV status, alter’s HCV status, getting tested for HCV, the cure for HCV, or where to go for HCV medical care.

Predictor Variables

Respondent-level socio-demographics were obtained from participant baseline surveys and included measures of age (measured in years), race (Black, White or Other), education level (categorized as less than high school, high school/GED, and some college or more), relationship status (categorized as married/in a committed relationship vs separated/divorced/widowed/single), and whether or not the respondent had ever been diagnosed with HCV or HIV. Network compositional variables were aggregated from their social network survey and included network size (the total number of alters listed per respondent), average age of alters within networks, and the percent of their network that was comprised of alters who were Black, White or Other race/ethnicity, alters who were sex partners, kin, non-kin, and the percent of the network where respondents and alters were of the same race or gender (homophily).

Alter-level data were obtained from the social network survey and included measurement of alter’s race (Black, White, Other), gender (male or female), and whether the alter had ever had an STI or was enrolled in drug treatment at the time of the survey. Measures corresponding to the relationship between the respondent and alter (i.e. dyadic measures) included behaviors that occurred within the relationship, such as whether or not the respondent and alter drink alcohol together, share needles/cookers, and the frequency of their interactions (daily/weekly vs. monthly or less often). Relationship types were categorized based on previous work (Yang, Davey-Rothwell, & Latkin, 2013) and used to classify relationships as sexual partnerships, kin or non-kin. Sexual partnerships included spouse, boy/girlfriend, fiancé, and sex partner. Kin included any family member, such as parent, child, or sibling. Non-kin included friends, acquaintances, NA/AA program sponsors, and spiritual leaders. Other dyadic measures included trust (measured as a continuous variable with 1 indicating no trust and 10 indicated complete trust) and racial homophily, which is when both the respondent and alter report the same race (McPherson et al., 2001). Research suggests that people tend to associate with others who are similar, and that this shared identity facilitates communication (McPherson et al., 2001).

Analytic Approach

Differences in the distribution of socio-demographic and network compositions between men and women were assessed with chi-square and t-tests. The subsequent analytic aim was to examine variables at the level of the alter and dyad that predicted HCV communication within dyads. Because of our interest in comparing men and women, analyses were conducted separately by male and female gender. We subsequently conducted interaction analyses assessing for a differential impact of ego gender on HCV communication patterns by alter attribute within dyads. Logistic generalized estimating equations (GEE) models were fit using ‘xtgee’ in Stata version 14.2 (Rabe-Hesketh & Skrondal, 2008) to examine the bivariate relationship between each predictor variable and the primary outcome. The GEE approach is a general modeling strategy to adjust for the correlated (i.e. dependent) structure of social network data wherein alters were clustered around egos.

Factors were considered for inclusion in multivariable analysis if they demonstrated an association with the outcome at level of p<0.1 in bivariate analyses for either male or female respondents or if they were conceptually relevant. Correlation matrices among predictor variables were used to identify collinearity. Variables conceptually related to one another or with correlation threshold of r>0.30 were excluded from multivariable analyses. This led to the exclusion of the variable injects drugs with because it was highly correlated with sharing needles/cookers. We controlled for the following respondent-level variables: respondent HCV status, HIV status, age, race, and educational level.

Results

There were a total of 227 respondents, 71.8% of whom were men (n=163), and 28.2% were women (n=64) (Table 1). In comparing men and women, there was a higher proportion of women with less than a high school education compared to men (50.0% vs 33.7%; p=0.007). Fewer women were in committed relationships (62.5% vs 81.6%, p=0.002) compared to men. Women were also more likely to have ever received an HCV diagnosis (89.1% vs 74.9%; p=0.02). Related to network composition, women had larger networks than men (8.0 vs 6.9; p=0.08), and a higher proportion of their network comprised of kin, (0.3 vs 0.2; p=0.001) and a smaller proportion of non-kin (0.5 vs 0.6; p=0.001). On average, women and men both discussed HCV with four alters.

Table 1.

Differences in Socio-demographics, Risk Behaviors and Network Composition between Women and Men who Inject Drugs

Variable	Total (n=227)	Women who Inject Drugs (n=64)	Men who Inject Drugs (n=163)
	N(%)	N(%)	N(%)	p-value
Age (mean, SD)	43.8 (10.3)	44.0 (10.3)	43.7 (11.3)	0.8 0.4
Race
Black	104 (45.8)	31(48.4)	73 (44.8)
White	113 (49.8)	32 (50.0)	81 (49.7)
Other	10 (4.1)	1 (1.6)	9 (5.2)
Education level				0.007
Less than HS	87 (38.3)	32 (50.0)	55 (33.7)
HS/GED	92 (40.5)	20 (31.3)	72 (44.2)
College or more	48 (21.2)	12 (18.8)	36 (22.1)
Married/in committed relationship	173 (76.2)	40 (62.5)	133 (81.6)	0.002
Homeless	142 (62.6)	35 (52.7)	107 (65.6)	0.1
Total sex partnersa (mean, SD)	1.7 (0.2)	3.0 (0.6)	1.1 (0.1)	0.0
Transactional sexa	63 (39.4)	33 (62.3)	30 (28.0)	0.00
Shared cookers/cottonsa	144 (63.4)	37 (57.8)	107 (65.6)	0.27
Ever received HCV diagnosis	179 (78.9)	57 (89.1)	122 (74.9)	0.02
HIV positive	30 (13.2)	8 (12.5)	22 (13.5)	0.8
Personal Network Characteristics	7.3 (4.1)	8.0 (4.1)	6.9 (4.1)	0.08
Network size (mean, SD)
Percent Network Race
Black	0.6 (0.4)	0.6 (0.41)	0.5 (0.42)	0.7
White	0.4 (0.4)	0.3 (0.4)	0.5 (0.4)	0.3
Other	0.0 (0.1)	0.1 (0.09)	0.0 (0.1)	0.5
Percent network relationship type
Sex partners	0.1 (0.1)	0.1 (0.12)	0.1 (0.12)	0.14
Kin	0.3 (0.2)	0.3 (0.2)	0.2 (0.2)	0.001
Non-kin	0.6 (0.2)	0.5 (0.2)	0.6 (0.3)	0.001
Percent Network PWID	0.4 (0.2)	0.4 (0.23)	0.4 (0.24)	0.97
Percent Race homophily	0.8 (0.3)	0.8 (0.2)	0.8 (0.3)	0.16
Percent Gender homophily	0.6 (0.2)	0.5 (0.2)	0.6 (0.2)	0.09
Average age difference within personal networks (mean, SD)	11.3 (4.5)	10.8 (4.7)	12.3 (3.8)	0.00
Total number of alters with whom respondent discussed HCV (mean, SD)	3.9 (3.5)	4.2 (3.9)	3.8 (3.4)	0.44
number sex partners discussed HCV (mean, SD)	0.7 (0.7)	0.7 (0.7)	0.6 (0.8)	0.32
number drug partners discussed HCV (mean, SD)	1.4 (1.8)	1.5 (2.0)	1.4 (1.8)	0.3
number kin discussed HCV (mean, SD)	1.1 (1.4)	1.3 (1.3)	1.1 (1.3)	0.3
number non kin discussed HCV (mean, SD)	2.1 (2.7)	2.2 (3.0)	2.1 (2.6)	0.4

^aIn prior 90 days

Respondents’ personal networks comprised a total of 1,655 dyads: 516 dyads among the 64 female respondents and 1139 dyads among the 163 male respondents (Table 2). Both men and women discussed HCV within over half of all dyads elicited. Women disclosed their HCV status within 47.8% of dyads (n=246), discussed alters’ HCV status within 9.3% of dyads (n=47), getting tested for HCV within 13.4% of dyads (n=69), HCV cure within 19% of dyads (n=98) and where to go for HCV medical care within 13.6% of dyads (n=69). Men disclosed their HCV status within 47.8% of dyads (n=544), discussed alters’ HCV status within 6.5% of dyads (n=74), getting tested for HCV within 16.6% of dyads (n=189), HCV cure within 19.8% of dyads (n=222) and where to go for HCV medical care within 14.6% of dyads (n=166).

Table 2.

Distribution of Alter and Dyadic Attributes by Infectious Disease Care and Prevention Conversations Within Dyads of Women and Men who Inject Drugs

	Total PWID (n=1655)	Women (n=516 dyads)			Men (n=1139 dyads)			Total PWID (n=1655)
Characteristic		No discussion 247 (47.9) N%	Discussion 269 (52.1) N%	UOR (95%CI)	No discussion 519 (45.6) N%	Discussion 620 (54.4) N%	UOR (95%CI)	P value for interaction term with ego genderb
Alter Characteristics
Age (mean std)	45.2 (15.2)	44.4 (18.2)	47.9 (13.7)	1.0 (1.0–1.0)	44.2 (15.8)	45.5 (13.8)	1.0 (1.0–1.0)	0.3
Race
Black	915 (55.5)	162 (65.9)	151 (56.1)	REF	306 (59.2)	296 (48.1)	REF	REF
White	664 (40.3)	72 (29.3)	105 (39.0)	1.3 (0.8–2.3)	198 (38.3)	289 (46.9)	1.4 (1.1–1.8) *	0.8
Other	69 (4.2)	12 (4.9)	13 (4.8)	0.9 (0.4–2.0)	13 (2.5)	31 (5.0)	1.2 (0.7–19)	0.5
Gender
Male	870 (52.7)	105 (43.2)	127 (47.2)	REF	311 (60.0)	327 (52.7)	REF	REF
Female	780 (47.3)	138 (56.8)	142 (52.8)	0.8 (0.6–11)	207 (40.0)	293 (47.3)	1.6 (1.3–2.0) *	0.001
Sexual
Orientation
Homosexual	175 (10.9)	49 (21.8)	40 (14.9)	REF	31 (6.3)	55 (8.9)	REF	REF
Heterosexual	1426 (89.1)	176 (78.2)	228 (85.1)	1.3 (0.9–2.0)	462 (93.7)	560 (91.1)	0.8 (0.5–1.1)	0.1
Currently enrolled in drug treatment	343 (20.8)	40 (16.3)	72 (26.8)	2.0 (1.1–3.0) *	72 (13.9)	159 (25.7)	2.0 (1.4–2.7) *	0.5
Ever had an STI	111 (6.7)	11 (4.5)	24 (8.9)	2.2 (1.1–4.6) *	19 (3.7)	57 (9.2)	1.9 (1.3–3.0) *	0.8
Dyadic Characteristics
Race homophily	1355 (81.9)	208 (84.2)	48 (17.8)	1.2 (0.7–2.1)	447 (86.1)	479 (77.3)	0.8 (0.6–1.0)+	0.1
Drinks alcohol with alter	226 (13.7)	27 (11.0)	31 (11.5)	1.3 (0.9–2.1)	79 (15.3)	89 (14.4)	1.0 (0.7–1.5)	0.3
Injects drugs with alter	313 (18.9)	23 (9.4)	67 (24.9)	1.7 (1.0–2.9) *	59 (11.4)	164 (26.5)	2.2 (1.6–3.1) *	0.9
Has shared needles/cookers with altera	206 (12.5)	17 (6.9)	39 (14.5)	1.9 (1.0–3.5) *	33 (6.4)	117 (19.9)	3.3 (2.2–4.9) *	0.2
Relationship types
Sex Partner	185 (112)	14 (5.7)	47 (17.5)	4.0 (1.9–8.7) *	22 (4.3)	102 (16.5)	3.8 (2.3–6.6) *	0.9
Kin	516 (31.2)	97 (39.4)	80 (29.7)	REF	166 (32.2)	173 (27.9)	REF	REF
Non kin	949 (57.3)	135 (54.9)	142 (52.8)	1.0 (0.7–15)	327 (63.5)	345 (55.7)	0.9 (0.7–12)	0.7
Trust (mean, std)	7.1 (3.1)	6.8 (3.1)	7.3 (3.0)	1.1 (1.0–1.2) *			1.1 (1.0–1.1) *	0.4
Relationship length, months (mean std) Frequency of interaction	17.9 (16.9)	16.8 (16.6)	16.5 (15.6)	1.0 (1.0–10)	18.4 (17.1)	18.7 (17.3)	1.0 (1.0–10)	0.8
Daily/weekly	1128 (68.2)	152 (61.8)	203 (75.5)	1.9 (1.2–3.0)	320 (61.8)	453 (73.1)	1.7 (1.4–2.2) *	0.4
Monthly less	525 (31.8)	94 (38.2)	66 (24.5)	REF	198 (38.2)	167 (26.9)	REF	REF

^aIn prior 90 days;

^binteraction of ego gender and alter/dyadic characteristic;

^**significance at p<0.01;

^*significance at p<0.05;

⁺significance at p=0.10

Men and women had many significant predictors of HCV communication in common: Women and men both had had increased odds of HCV communication with alters currently enrolled in drug treatment (UOR 2.0, 95% CI: 1.1–3.0 and UOR 2.0; 95% CI: 1.4–2.7 respectively), alters who ever had an STI (UOR 2.2; 95% CI: 1.1–4.6 and UOR 1.9, 95% CI: 1.3–3.0 respectively), alters with whom they inject drugs (UOR 1.7, 95% CI: 1.0–2.9 and UOR 2.2, 95% CI: 1.6–3.1 respectively), alters with whom they share drug preparation equipment (UOR 1.9; 95% CI: 1.0–3.5; UOR 3.3, 95% CI: 2.2–4.9 respectively), alters who are sex partners compared to kin (UOR 4.0, 95% CI: 1.9–8.7 and UOR 3.8, 95% CI: 2.3–6.6 respectively), alters with increased levels of trust (UOR 1.1, 95% CI: 1.0–1.2 and UOR 1.1, 95% CI: 0.0–1.1 respectively and alters with whom they interact daily/weekly compared to monthly or less (UOR 1.9, 95% CI: 1.2–3.0 UOR 1.7, 95% CI: 1.4–2.2 respectively). Men and women also both had increased odds of communicating HCV related information with alters who were White, compared to Black. We found no evidence of gender specific differences in these communication patterns based on interaction analyses assessing for a differential effect of gender (Table 2).

Related to differences between men and women, men had significantly increased odds of discussing HCV with alters who were female (UOR 1.6; 95% CI: 1.3–2.0), an effect which was not found among female egos. The p-value for the interaction effect of ego gender on HCV communication by alter gender was statistically significant (p=0.001) consistent with differences in HCV communication with female alters depending on the egos gender.

Given the few differences in HCV communication between men and women, one multivariable model was conducted among all 1655 male and female dyads (table 3). In multivariable analysis, alters who are currently enrolled in drug treatment (AOR 1.7, 95% CI: 1.3–2.4), alters with whom participants share drug preparation equipment (AOR 3.0, 95% CI: 1.9–4.6), alters who are sex partners compared to kin (AOR 3.0; 95% CI: 1.9–4.9) and alters with whom respondents have increased trust (AOR 1.1; 95% CI: 1.11.2) and daily/weekly interactions (AOR 1.7; 95% CI 1.3–2.1) remained independently associated with having an HCV conversation.

Table 3.

Adjusted Odds Ratios (AOR) of Alter Attributes and Dyadic Attributes that Predict Health Communication Within Dyads

Characteristic	(n=1665 dyads)
	AOR (95% CI)
Alter Race
Black	REF
White	1.2 (0.9–1.7)
Other	1.1 (0.6–1.8)
Alter Female	1.2 (0.9–1.5)
Currently enrolled in drug treatment	1.7 (1.3–2.4) **
Racial Homophily	0.8 (0.5–1.1)
Ever had an STI	1.3 (0.8–2.1)
Has shared needles/cookers with altera	3.0 (1.9–4.6) **
Relationship type
Sex Partner	3.0 (1.9–4.9) **
Kin	REF
Non kin	0.9 (0.7–1.2)
Trust	1.1 (1.1–1.2) **
Frequency of interaction
Daily/weekly	1.7 (1.3–2.2) **
Monthly less	REF

^aIn prior 90 days;

controlling for HIV and HCV status, respondent age and education;

^**significance at p<0.01;

^*significance at p<0.05

Discussion

The current study assessed similarities and differences between men and women who inject drugs as it relates to socio-demographics, risk behaviors, network composition, and patterns of HCV communication within dyads. Despite differences in socio-demographics, risk behaviors and network composition, few differences were observed between men and women’s HCV communication patterns. Across both genders, many characteristics related to relational closeness, such as trust, frequency of interaction, sexual partnership and sharing of drug preparation equipment were key determinants of discussing HCV in relationships. One possible interpretation is that the social stigma of having these conversations may constrain HCV conversations to emotionally close relationships. Peer interventions that leverage social ties may therefore be successful when focused on relationally close relationships such as sexual and injection drug use partnerships of PWID. Couples-based interventions have been previously endorsed as an effective way of improving PWID engagement in HIV-related interventions, such as HIV testing and antiretroviral therapy adherence (Jiwatram-Negrón & El-Bassel, 2014; Reddon, Marshall, & Milloy, 2019), though remain less explored as it relates to HCV care (Jiwatram-Negrón & El-Bassel, 2014). Additionally, HCV conversations within relationships where drug preparation equipment is shared could also be an indication of serosorting, that is, to selectively share injection equipment with persons of like HCV status, which has been reported to occur within PWID populations (Burt et al., 2009; Smith et al., 2013). If prevalent, the practice of serosorting would provide further rationale to encourage and support HCV testing and treatment within networks of PWID.

In addition to HCV conversations within close relationships, conversations also were more likely to occur with alters currently enrolled in drug treatment. Alters enrolled in drug treatment may be able to act as “network bridges,” which are social network members that occupy a position between groups that are not otherwise connected (Kadushin, 2012). As such, PWID enrolled in drug treatment may be well positioned to accelerate the spread of new ideas and behaviors related to HCV across loosely or non-connected regions of PWID networks (Kadushin, 2012). For this reason, PWID enrolled in drug treatment may be particularly effective as peer mentors in targeted interventions to diffuse information about infectious disease innovations, such as HCV treatment or HIV Pre-exposure Prophylaxis (PrEP) for HIV prevention, to peers in their networks.

While interventions have successfully incentivized PWID to recruit network members into and provide support throughout HCV care (Falade-Nwulia, Ward, et al., 2020, Crawford & Bath, 2013; Janda & Mergenhagen, 2017; Lafferty et al., 2018), our finding that PWID discussed HCV within over half of social network relationships highlights the important role of social networks in disseminating HCV-related information outside the confines of a research study. Turning to trusted peers for HCV disclosure and information may be a way of compensating for PWID’s limited interactions with health care providers in traditional medical settings (Motavalli et al., 2020). It is therefore critical to ensure that information shared within these networks is accurate. Additionally, programs are needed to support PWID to enhance already existing HCV communication with tools to engage and support network members through all the stages of HCV care continuum from testing to HCV cure. Despite our study finding few gender differences in HCV communication, it is likely that there may be gender differences in how men and women who inject drugs access HCV care and treatment, which should be explored in future studies.

There are study limitations that should be considered. First, the cross-sectional design does not provide information on temporality, and causal claims cannot be made. Additionally, the sample consisted of participants in a randomized control trial of an intervention to enhance Hepatitis C and HIV prevention care. Therefore, findings from this analysis may not be generalizable to the broader population of men and women who inject drugs. These data were however collected prior to the study intervention and thus represent communication patterns prior to intervention. Additionally, based on the way survey items were asked, we cannot assess if respondents initiated HCV related conversations or if the alters did. Understanding who initiates these conversations would be valuable for understanding the nuance of information dissemination and should be explored in follow up studies. Despite these limitations, findings provide some of the first data to examine patterns of HCV related communication within social networks of men and women who inject drugs and have important implications to the design of peer and network based interventions seeking to increase PWID engagement in HCV care.

Highlights.

• Communication about HCV within social networks was similar for men and women

• PWID had higher odds of HCV communication with alters enrolled in drug treatment

• PWID had higher odds of HCV communication with sex partners

• PWID had higher odds of HCV communication with alters they trust

• PWID had higher odds of HCV communication with alters they interact with frequently