Abstract
Until the mid-1990s, the prevalence and incidence of HIV infection was uniformly low in countries across the Central and Eastern European region. In the past decade, however, this has changed dramatically, with a rapid increase in HIV infections in the region, especially in Eastern Europe where 41% of new HIV infection cases were among injecting drug users (IDUs) and as much as 66% of IDUs are infected with HIV in certain regions. While Russia, the largest country in Eastern Europe, has the fastest growing HIV rates in the world, the situation is different in Central Europe. For example, Hungary has low levels of HIV infection – estimated less than 1% of IDUs. Understanding the role of network factors in the spread and prevention of HIV could not only enable us to keep the HIV rates low among IDUs in countries like Hungary, but also provide a means for the effective prevention of other blood-borne and sexually transmitted infections (STIs) that share similar routes of transmission as HIV. Rogers’ diffusion of innovations theory may help explain why HIV rates among IDUs are low in Hungary. Valente’s related exposure or contagion model postulates that the more individuals within a social network adopt an innovation or a practice, the greater the probability of an individual is to adopt this innovation or practice. Personal network exposure (PNE), measured both within egocentric and sociocentric networks quantifies the extent to which a person is exposed to risk through their social network. The aim of this analysis was to assess the association of PNE and other correlates with injecting equipment sharing among IDUs in Budapest, Hungary.
INTRODUCTION
Until the mid-1990s, the prevalence and incidence of HIV infection was uniformly low in countries across the Central and Eastern European region. In the past decade, however, this has changed dramatically, with a rapid increase in HIV infections in the region, especially in Eastern Europe (Hamers and Downs 2003; Aceijas et al. 2004; European Centre for the Epidemiological Monitoring of AIDS 2005). In Eastern Europe, 41% of new HIV infection cases were among injecting drug users (IDUs). Estimates suggest that in certain regions, as much as 66% of IDUs are infected with HIV (Hamers and Downs 2003; Kelly and Amirkhanian 2003; Rhodes et al. 1999). The largest country in Eastern Europe, Russia, has the fastest growing HIV rates in the world (Kalichman et al. 2000). However, the situation is different in Central Europe. In 2004 fourteen Central European countries reported a total of 217 new HIV infections contracted through injecting drug use (European Centre for the Epidemiological Monitoring of AIDS 2005). It is estimated that in Central European countries excluding Poland and the former Yugoslavia, under 2% of IDUs are infected with HIV (Hamers and Downs 2003; Dehne et al. 1999; Robinson 2000; Kilibarda 1993). Hungary, like many Central European countries, has low levels of HIV infection. At the end of 2004, Hungary (population 10 million) had 1,175 diagnosed cases of HIV infection, and the estimated prevalence of HIV among IDUs was less than 1% (European Centre for the Epidemiological Monitoring of AIDS 2005).
In addition to being at risk of HIV infection, IDUs are also at risk of acquiring Hepatitis C (HCV) and Hepatitis B (HBV). In contrast to the very low prevalence of HIV among IDUs in Hungary, the prevalence of HCV among IDUs in Hungary was about 30% in the late 1990s (European Monitoring Centre for Drugs and Drug Addiction 2002; Ujhelyi et al. 1998; Zacher 2003; Topolánszky 2002), and, while HBV infection rates among IDUs in Hungary were 13% for women and 8% for men, only about 7% reported having been vaccinated against HBV (Topolánszky 2002). The presence of HCV and HBV epidemics among IDUs in Hungary is a concern not only in its own right, but also because it suggests that injecting and sex risk behaviors and risk networks among IDUs are very common (Hamers and Downs 2003; Stark et al. 1996; Mikl et al. 1998). This implies that many of the conditions for an expansive and rapid HIV epidemic are present in the Central European region among drug injectors.
Social network theory, an approach that recognizes that individuals are embedded in social and risk networks (Trotter II, Bowen, and Potter Jr 1995; Neaigus et al. 1994), may help explain the discrepancy of HIV rates between Central and Eastern Europe. According to this theory, the potential for epidemic spread of HIV infection within a specific risk-group depends on many supra-individual factors, including the probability of transmission, the average number of contacts (Anderson and May 1988), risk network composition and structures (which includes both the centrality of high-risk individuals and mixing patterns within a network (Friedman et al. 1997; Neaigus et al. 1994), and secondary prevention. Understanding the role of network factors in the spread and prevention of HIV could not only enable us to keep the HIV rates low among IDUs in Central European countries like Hungary, but also provide a means for the effective prevention of other blood-borne and sexually transmitted infections (STIs) that share similar routes of transmission as HIV, such as HBV and HCV, as well as other STIs.
A theory which may help explain why HIV rates among IDUs are low in Hungary is the diffusion of innovations theory (Rogers 2000). This theory describes how an innovation, an idea, practice, or object, is spread among members of a community (Rogers 2003). The exposure or contagion model (Valente 2005), which is related to the diffusion theory, postulates that the more individuals within a social network adopt an innovation or a practice, the greater the probability of an individual is to adopt this innovation or practice. Personal network exposure (PNE) quantifies the extent to which a person within a network is exposed to such an innovation, idea, practice or object (Valente 1995). PNE can be measured both within egocentric and sociocentric networks. Within egocentric networks, PNE is the proportion of adopters among network members and thus measures the direct influence of network members (Valente 2005). Within sociocentric networks, PNE measures both direct and indirect network influence and is calculated by weighting exposure by the inverse of the shortest paths (geodesic distances) between any two members of the sociocentric network (Valente 1995). To incorporate the influence of opinion leaders, PNE can be weighted by network measures (e.g. centrality) (Valente 2005).
PNE to risk or protective behaviors through direct and indirect ties to other IDUs who engage in these risk behaviors and who are members of the sociocentric network thus may influence an individual IDU’s injecting behavior. We hypothesize that if PNE is positively associated with injecting equipment sharing, then, according to the contagion model, sharing may be a social epidemic – the more individuals are exposed to injecting equipment sharing, the more likely they are to share. The aim of this analysis was to assess the association of PNE and other correlates with injecting equipment sharing among IDUs in Budapest, Hungary.
METHODS
Between October 2005 and February 2006, 83 injecting drug users were recruited in Budapest, Hungary from non-treatment settings using a combination of street outreach and chain referral methods (Heckathorn 1997; Gyarmathy and Neaigus 2005; Gyarmathy et al. 2006) for a United States National Institute on Drug Abuse funded study on network risk for HIV, HBV and HCV among IDUs in Budapest (study ongoing). Participants were recruited from areas of the city where drug users often congregate, from the needle exchange, and through referral by other IDUs who were in the study. Candidate participants who self-reported injecting drugs in the past 30 days were eligible to participate. Eligibility of drug use was verified by urine tests (heroin, cocaine, amphetamine and methamphetamine) for those reporting drug use in the past 2-3 days. Injecting route of drug administration was verified by inspection of injecting marks for those who reported injecting in the past 30 days. Participants were paid HUF 2000 (about USD 10) for participation and HUF 500 (about USD 2.5) for bringing in their nominated network members or other IDUs who were eligible to participate in the study. The Institutional Review Boards at the National Development and Research Institutes, Inc. in New York, USA and the Hungarian Academy of Sciences in Budapest, Hungary approved all human subjects procedures.
After providing their informed consent, eligible participants were administered a structured survey interview that took about 2 hours to complete. Then participants received counseling about and were tested for HIV, HAV, HBV, HCV and syphilis infections.
Measures
The interview included questions about demographics, socio-economic status, drug use and injecting behaviors, sexual risk behaviors, self-reported HIV and HCV infections, knowledge about HIV and HCV transmission, heroin dependence and injecting equipment sharing. The dependent variables in the analysis were five types of injecting equipment sharing: sharing cookers, sharing filters, receptive syringe sharing, distributive syringe sharing and syringe-mediated drug sharing (backloading) in the past 30 days. Since having an IDU sex partner has been found to be associated with injecting equipment sharing (Evans et al. 2003; Lum, Sears, and Guydish 2005), and we had found strong infection disclosure norms among IDUs in Hungary (Gyarmathy et al. 2006), we created four-way interaction variables for self-report of HCV infection and having an IDU sex partner (1. no reported HCV infection and no IDU sex partner (reference category), 2. no reported HCV infection and IDU sex partner, 3. reported HCV infection and no IDU sex partner, 4. reported HCV infection and IDU sex partner).
Participants were asked using standard naming stimuli (Friedman et al. 1999) to provide us with the real or fictitious name of friends or family whom they would go to for advice, asked a favor from in the past 30 days, with whom they have had sex or used non-injected or injected drugs in the past 30 days. Then, for every nominated network member, more detailed data was collected (Friedman et al. 1999; Neaigus et al. 2006; Weeks et al. 2002) about the network member's demographics, drug use and injecting behaviors and sexual risk behaviors, and how close they are emotionally to the network member. We also assessed the relationships among nominated network members.
Ties among participants who were interviewed for the study were ascertained based on each participant’s nominations and on reports of relationships of other participants about their network members. We used four methods to verify links (Friedman et al. 1999): 1. “storefront link”, when participants brought their network members in and linkage data was recorded, 2. “field link”, when the participant identified the network member to the staff in the field, 3. “ethnographic link”, when staff observed links in the field, and 4. “data set link”, when identifying data was used to establish links. We used UCINET (Borgatti, Everett, and Freeman 2002) to create social network measures based on this relationship data. First, we symmetrized the data in order to incorporate into the assessment of network exposure “modeling” or indirect communication (Neaigus et al. 2006), i.e., ties between participants who share injecting equipment. Geodesic distances between each participant were then calculated, along with two measures of centrality. Closeness centrality (how may steps away an individual is from others within the entire sociocentric network that this individual is a member of) was calculated to indicate potential roles of opinion leadership, and degree centrality (the number of ties of each individual within their egocentric network) to indicate the size of their personal network. PNE measures for each of the injecting equipment sharing variables were calculated in SAS® by transposing individual equipment sharing data into an exposure matrix, weighting exposure by closeness centrality (Valente 2005), and multiplying weighted exposure by the inverse of the geodesic distances. PNE measures were then standardized to have a mean of 0 and a standard deviation of 1 for easier interpretation.
ANALYSIS
Univariate odds ratios and corresponding 90% confidence intervals were calculated to assess associations between candidate variables and each of the equipment sharing variables. Then, multivariate logistic regression models with stepwise selection (with model entry and retention significance of 0.10) identified significant (p<0.10) correlates of equipment sharing, with PNE forced in as the main exposure variable. In each of these models, equipment-specific PNE measures were used (e.g., for assessing correlates of sharing cookers, PNE to sharing cookers were used; for correlates of sharing filters, PNE to sharing filters was used, etc.)
RESULTS
Sample description
The mean age of the 83 participants was 28.5 years (median: 28; SD=5.8; range: 18-41), 21 (25%) were female, 9 (11%) were Roma (Gypsy) (Table 1). About a third (n=28; 34%) had a high-school degree and 9 (11%) reported that they were still in school. Most were single (n=63; 76%) and had worked at some point at a legal job (n=63; 76%). Altogether 47 (57%) reported having an average monthly income of HUF 100,000 (approximately USD 500), and 16 (19%) reported being homeless. Altogether 48 (58%) participants reported sharing cookers, 47 (57%) sharing filters, 15 (18%) distributive syringe sharing, 17 (21%) receptive syringe sharing, and 26 (31%) backloading in the past 30 days. Almost all agreed that syringes can transmit HIV (n=77; 94%) and HCV (n=78; 95%) infections (data not shown in table). About half agreed that cookers, cotton and dissolved drugs can transmit HIV (n=37; 45%) and HCV (n=44; 54%) (data not shown in table). Altogether 26 (31%) reported incorrect knowledge about HIV and 21 (25.3%) about HCV (data not shown in table). The mean degree centrality was 4.4 (SD=5.3). Non-standardized PNE values show that mean network exposure to cookers and filters is the highest, followed by backloading, receptive syringe sharing and distributive syringe sharing.
Table 1.
Description of the sample – injecting drug users in Hungary, 2005-2006
| N (%) | N (%) | ||
|---|---|---|---|
| Total | 83 (100.0) | Shared cookers in the past 30 days | 48 (57.8) |
| Age - mean (SD) | 28.5 (5.8) | Shared filters in past 30 days | 47 (56.6) |
| Gender | Distributive syringe sharing in past 30 days | 15 (18.1) | |
| male | 62 (74.7) | Receptive syringe sharing in past 30 days | 17 (20.5) |
| female | 21 (25.3) | Backloading in past 30 days | 26 (31.3) |
| Ethnicity | Degree centrality - mean (SD) | 4.4 (5.3) | |
| non-Roma | 74 (89.2) | ||
| Roma | 9 (10.8) | PNE for sharing cookers (non-standardized) - mean (SD) | 136 (78.6) |
| High school degree | 28 (33.7) | PNE for sharing filters (non-standardized) - mean (SD) | |
| Currently in school | 9 (10.8) | 147 (85.5) | |
| Marital status single | 63 (75.9) | PNE for receptive syringe sharing (non-standardized) - mean (SD) | 50.8 (29.4) |
| Ever had a legal job | 63 (75.9) | ||
| Homeless | 16 (19.3) | PNE for distributive syringe sharing (non-standardized) - mean (SD) | 36.0 (21.2) |
| Average monthly income above HUF 100 000 (USD 500) | 47 (56.6) | PNE for backloading (non-standardized) - mean (SD) | 84.8 (49.8) |
Univariate correlates of equipment sharing
Significant (p<0.10) correlates of equipment sharing are presented in Tables 2-6. We found that equipment-specific standardized PNE was positively associated with sharing filters and inversely associated with distributive syringe sharing, and we found no significant associations between PNE and sharing cookers, receptive syringe sharing and backloading. Degree centrality was positively associated with sharing filters and backloading. Heroin dependence showed a negative association with sharing cookers, and both incorrect knowledge of HIV transmission and incorrect knowledge of HCV transmission were positively associated with both sharing cookers and sharing filters. Having an IDU sex partner and self-report of being HCV negative was positively associated with sharing cookers, sharing filters and distributive syringe sharing, while having no IDU sex partner and self-report of being HCV infected was associated with distributive syringe sharing. Being emotionally close to an IDU was associated with all forms of equipment sharing, except with backloading. Of the demographic variables, female gender was associated with sharing cookers, receptive syringe sharing and backloading, while younger age was associated with sharing cookers, sharing filters and backloading.
Table 2.
Univariate correlates of sharing cookers in the past 30 days – injecting drug users in Hungary, 2005-2006
| Characteristic | did not share | shared | OR (90% CI) | sig. p-value |
|---|---|---|---|---|
| N (%) | N (%) | |||
| Total | 35 (42.2) | 48 (57.8) | (reference) | |
| Personal network exposure - mean (SD) | −.02 (0.95) | 0.02 (1.04) | 1.0 (0.7, 1.5) | 0.863 |
| Degree centrality - mean (SD) | 4.0 ( 5.7) | 4.7 ( 5.0) | 1.0 (1.0, 1.1) | 0.5766 |
| Heroin dependence - high | ||||
| no | 13 (32.5) | 27 (67.5) | (reference) | * 0.0876 |
| yes | 22 (51.2) | 21 (48.8) | 0.5 (0.2, 1.0) | |
| Incorrect knowledge about HIV transmission | ||||
| no | 30 (52.6) | 27 (47.4) | (reference) | * 0.0063 |
| yes | 5 (19.2) | 21 (80.8) | 4.7 (1.8, 12) | |
| Incorrect knowledge about HCV transmission | ||||
| no | 31 (50.0) | 31 (50.0) | (reference) | * 0.0179 |
| yes | 4 (19.0) | 17 (81.0) | 4.2 (1.6, 12) | |
| no HCV and no IDU sex partner | 25 (52.1) | 23 (47.9) | (reference) | * 0.0406 |
| no HCV and IDU sex partner | 2 (16.7) | 10 (83.3) | 5.4 (1.4, 21) | 0.3866 |
| HCV and no IDU sex partner | 5 (38.5) | 8 (61.5) | 1.7 (0.6, 5.0) | 0.2135 |
| HCV and IDU sex partner | 3 (30.0) | 7 (70.0) | 2.5 (0.7, 8.6) | |
| Emotionally close to an IDU | ||||
| no | 26 (56.5) | 20 (43.5) | (reference) | * 0.0040 |
| yes | 9 (24.3) | 28 (75.7) | 4.0 (1.8, 9.0) | |
| Gender | ||||
| male | 30 (48.4) | 32 (51.6) | (reference) | * 0.0547 |
| female | 5 (23.8) | 16 (76.2) | 3.0 (1.2, 7.7) | |
| Age - mean (SD) | 30.2 (5.3) | 27.3 (5.8) | 0.9 (0.9, 1.0) | * 0.0265 |
| Ethnicity | ||||
| non-Roma | 31 (41.9) | 43 (58.1) | (reference) | 0.8836 |
| Roma x | 4 (44.4) | 5 (55.6) | 0.9 (0.3, 2.9) |
p<0.10
Table 6.
Univariate correlates of syringe mediated drug sharing (backloading) in the past 30 days – injecting drug users in Hungary, 2005-2006
| Characteristic | did not share | shared | OR (90% CI) | sig. p-value |
|---|---|---|---|---|
| N (%) | N (%) | |||
| Total | 57 (68.7) | 26 (31.3) | (reference) | |
| Personal network exposure - mean (SD) | −.09 (1.00) | 0.20 (1.00) | 1.3 (0.9, 2.0) | 0.2308 |
| Degree centrality - mean (SD) | 3.7 ( 5.3) | 6.0 ( 5.1) | 1.1 (1.0, 1.2) | * 0.0831 |
| Heroin dependence - high | ||||
| no | 26 (65.0) | 14 (35.0) | (reference) | 0.4870 |
| yes | 31 (72.1) | 12 (27.9) | 0.7 (0.3, 1.6) | |
| Incorrect knowledge about HIV transmission | ||||
| no | 42 (73.7) | 15 (26.3) | (reference) | 0.1485 |
| yes | 15 (57.7) | 11 (42.3) | 2.1 (0.9, 4.7) | |
| Incorrect knowledge about HCV transmission | ||||
| no | 45 (72.6) | 17 (27.4) | (reference) | 0.1913 |
| yes | 12 (57.1) | 9 (42.9) | 2.0 (0.8, 4.7) | |
| no HCV and no IDU sex partner | 35 (72.9) | 13 (27.1) | (reference) | |
| no HCV and IDU sex partner | 6 (50.0) | 6 (50.0) | 2.7 (0.9, 8.0) | 0.1349 |
| HCV and no IDU sex partner | 8 (61.5) | 5 (38.5) | 1.7 (0.6, 5.0) | 0.4277 |
| HCV and IDU sex partner | 8 (80.0) | 2 (20.0) | 0.7 (0.2, 2.7) | 0.6432 |
| Emotionally close to an IDU | ||||
| no | 33 (71.7) | 13 (28.3) | (reference) | 0.5028 |
| yes | 24 (64.9) | 13 (35.1) | 1.4 (0.6, 3.0) | |
| Gender | ||||
| male | 46 (74.2) | 16 (25.8) | (reference) | * 0.0670 |
| female | 11 (52.4) | 10 (47.6) | 2.6 (1.1, 6.2) | |
| Age - mean (SD) | 30.0 ( 5.6) | 25.1 ( 4.7) | 0.8 (0.8, 0.9) | * 0.0008 |
| Ethnicity | ||||
| non-Roma | 52 (70.3) | 22 (29.7) | (reference) | 0.374 |
| Roma x | 5 (55.6) | 4 (44.4) | 1.9 (0.6, 6.2) |
p<0.10
Multivariate correlates of equipment sharing
PNE was significantly (p<0.10) and positively associated with sharing filters and backloading, and significantly and inversely associated with distributive syringe sharing (Table 7). Degree centrality showed a significant association with distributive syringe sharing, and incorrect knowledge about HIV transmission with sharing cookers and filters. Having no IDU sex partner and self-report of being HCV infected was associated with distributive syringe sharing, and being emotionally close to an IDU with sharing cookers and filters. Of the demographic variables, female gender showed an association with sharing cookers, distributive syringe sharing and receptive syringe sharing, while younger age with all forms of equipment sharing except for distributive syringe sharing.
Table 7.
Multivariate correlates of sharing cookers, sharing filters, distributive and receptive syringe sharing
| Sharing cookers | Sharing filters | Distributive syringe sharing |
Receptive syringe sharing |
Backloading | |
|---|---|---|---|---|---|
| aOR (90%CI) | aOR (90%CI) | aOR (90%CI) | aOR (90%CI) | aOR (90%CI) | |
| Personal network exposure | 1.0 (0.65, 1.5) | 1.8 (1.1, 2.8)b | 0.2 (0.09, 0.47)b | 0.78 (0.45, 1.4) | 1.6 (1.0, 2.5) a |
| Degree centrality | N/S | N/S | 1.3 (1.1, 1.5)b | N/S | N/S |
| Incorrect knowledge about HIV transmission | 3.7 (1.3, 10.5)b | 4.3 (1.5, 12.8)b | N/S | N/S | N/S |
| Self-reported HCV infection and no IDU sex partner | N/S | N/S | N/S | 13.6 (3.2, 58.2)b | N/S |
| Emotionally close to an IDU | 2.7 (1.1, 6.6)a | 4.0 (1.6, 10.1)b | N/S | N/S | N/S |
| Female gender | 3.2 (1.1, 9.8)a | N/S | 4.1 (1.3, 13.1)b | 7.2 (2.3, 22.5)b | N/S |
| Age | 0.91 (0.84, 0.98)b | 0.90 (0.84, 0.98)b | N/S | 0.88 (0.80, 0.97)b | 0.82 (0.74, 0.90)b |
Note:
p<0.10
p<0.05
DISCUSSION
Contrary to our hypothesis, we found that PNE was not positively associated with all forms of injecting equipment sharing, but only with sharing cookers and filter, and it was inversely associated with distributive syringe sharing. In addition, we found an interaction of having an IDU sex partner and self-report of being infected with HCV as they relate to receptive syringe sharing.
Levels of injecting equipment sharing among IDUs in this sample are similar to reports on injecting equipment sharing from studies conducted in the United States and in other Central European countries, and lower than among IDUs in Eastern Europe. Studies have found that about 20% of IDUs in the US, Estonia, Czech Republic and Poland shared syringes in the past month (Lundgren, Amodeo, and Chassler 2005; European Monitoring Centre for Drugs and Drug Addiction 2002). In contrast, around the same time, studies among IDUs in Saint Petersburg reported 41% sharing syringes, and in Moscow, between 40-75% (Somlai et al. 2002; Reilley et al. 2000; Gore-Felton et al. 2003). The relatively low rates of syringe sharing but high rates of other injecting equipment sharing may be one reason why levels of HIV are low but of the prevalence of HCV is relatively high among IDUs in Hungary.
The positive association between PNE and sharing filters may indicate that these two sharing behaviors may be linked to a social epidemic. In a previous study we found that filters are often reused by IDUs in Hungary as a backup drug supply (Gyarmathy et al. 2006). In addition, ethnographic observations of injecting events and discussions with participants in this study indicate that retaining and giving away filters may be an act of giving someone a favor so that, in need, this person can also be asked for a favor. In this context, favor usually refers to an alternative drug supply in the form of a second cooking of the filter. The inverse association between PNE and distributive syringe sharing suggests that rather than being a social epidemic, distributive syringe sharing may be situation-specific and reflect the demand and supply of syringes in particular injecting settings, e.g. if there are N new syringes and N+1 IDUs, and one IDU gives his or her used syringe to the IDU without a new syringe, there is no need for the other IDUs to give away theirs. Indeed, the association of degree centrality with distributive syringe sharing corroborates this, since it shows that the more IDUs there are in an IDU’s egocentric network, the more likely they are to share syringes. Furthermore, ethnographic findings from this study suggest that syringe sharing occurs either in sexual partnerships or if somebody “falls in”, i.e., shows up unexpectedly in a group about to inject drugs, and this unexpected person does not have their injecting equipment handy.
Lack of the association between PNE and injecting equipment sharing may suggest that sharing is not a function of exposure, but rather that all IDUs in a given population are influenced by common structural conditions, such as laws against the possession of syringes and other injecting equipment, that operate independently of exposure. Under this scenario, the model might represent a form of “structural equivalence” (Burt 1987) so that a given behavior occurs because of macro-conditions that affect all IDUs (even singletons) rather than being mediated through social networks. We believe that sharing cookers may be due to such structural conditions. IDUs in this study reported during ethnographic observations that packaging of drugs by dealers is such that the packaged quantity is too much for one person to use all at one time. Thus, several people buy one package. However, because they find it difficult and unfair to share it dry, they dissolve the drug in one cooker and share it wet instead. In addition, incorrect knowledge about HIV transmission seems to be another reason why IDUs in this study share filters and cookers, since many are not aware that infections can also be transmitted though injecting equipment other than syringes.
The association of receptive syringe sharing with self-report of HCV infection and not having an IDU sex partner raises concerns. IDUs in this sample believe that once they have HCV infection, “it’s all the same” for them, not realizing that they are still at risk for HIV or potentially other blood borne infections. In addition, studies have shown that having an IDU sex partner is associated with sharing both syringes and other injecting equipment (Evans et al. 2003; Neaigus et al. 1995; Strathdee et al. 1997). While in a sexual partnership either partner may be equally likely to give or take used syringes regardless of HCV serostatus, reasoning that “they have unprotected sex anyway”, they may not receive used syringes from a third person. However, an IDU who does not have an IDU sex partner may be more likely to receive a used syringe from another person than either one of a couple. Due to their “informed altruism” (Des Jarlais et al. 2004), HCV infected IDUs may put themselves at risk for HIV infection.
Limitations of the study include the small sample size and that we allowed a higher than customary Type I error when calculating confidence intervals. Hence, while on the one hand we may have missed associations due to the small sample size, we may also have identified associations that may not be strong enough for statistical significance in a larger sample. In addition, due to the recruitment methods, the sample may not be representative of all IDUs in Budapest, since we may not have had access to IDUs that are more socially integrated (Gyarmathy and Neaigus 2005) and thus more hidden. Missing links between participants is another limitation, which could lead to underestimating PNE. However, we recruited most nominated network members and we also used links reported by other participants in our analysis, which we believe minimized the underreporting of links.
Network interventions among IDUs in Hungary combined with the provision of legal, sterile syringes and other injecting equipment, and regular testing and counseling services can help to prevent the outbreak of HIV and to control the existing HCV epidemic due to parenteral routes of transmission (Latkin and Knowlton 2005; Gyarmathy et al. 2004; Gyarmathy and Neaigus 2005). Interventions need to focus on disseminating proper knowledge about HIV and Hepatitis infections, especially concerning the sharing of injecting equipment other than syringes. While the currently existing disclosure norms should be encouraged (Gyarmathy et al. 2006), those with HCV infection need to be counseled about the fact that they are still uninfected with HIV and thus are at risk for HIV infection if they take used syringes from other injectors. In addition to focusing on correcting misconceptions about the risks of sharing both syringes and other injecting equipment, infection prevention efforts among IDUs in Hungary need to focus on increasing the distribution, practice and legalization of carrying sterile syringes in order to minimize the situation-specific factors that may lead to syringe and other injecting equipment sharing. Moreover, as much as possible, IDUs should be provided access to appropriate drug treatment so that they can be helped to stop the practice of drug injecting.
Table 3.
Univariate correlates of sharing filters in the past 30 days – injecting drug users in Hungary, 2005-2006
| Characteristic | did not share | shared | OR (90% CI) | sig. p-value |
|---|---|---|---|---|
| N (%) | N (%) | |||
| Total | 35 (42.2) | 48 (57.8) | (reference) | |
| Personal network exposure - mean (SD) | −.02 (0.95) | 0.02 (1.04) | 1.0 (0.7, 1.5) | 0.863 |
| Degree centrality - mean (SD) | 4.0 ( 5.7) | 4.7 ( 5.0) | 1.0 (1.0, 1.1) | 0.5766 |
| Heroin dependence - high | ||||
| no | 13 (32.5) | 27 (67.5) | (reference) | * 0.0876 |
| yes | 22 (51.2) | 21 (48.8) | 0.5 (0.2, 1.0) | |
| Incorrect knowledge about HIV transmission | ||||
| no | 30 (52.6) | 27 (47.4) | (reference) | * 0.0063 |
| yes | 5 (19.2) | 21 (80.8) | 4.7 (1.8, 12) | |
| Incorrect knowledge about HCV transmission | ||||
| no | 31 (50.0) | 31 (50.0) | (reference) | * 0.0179 |
| yes | 4 (19.0) | 17 (81.0) | 4.2 (1.6, 12) | |
| no HCV and no IDU sex partner | 25 (52.1) | 23 (47.9) | (reference) | * 0.0406 |
| no HCV and IDU sex partner | 2 (16.7) | 10 (83.3) | 5.4 (1.4, 21) | 0.3866 |
| HCV and no IDU sex partner | 5 (38.5) | 8 (61.5) | 1.7 (0.6, 5.0) | 0.2135 |
| HCV and IDU sex partner | 3 (30.0) | 7 (70.0) | 2.5 (0.7, 8.6) | |
| Emotionally close to an IDU | ||||
| no | 26 (56.5) | 20 (43.5) | (reference) | * 0.0040 |
| yes | 9 (24.3) | 28 (75.7) | 4.0 (1.8, 9.0) | |
| Gender | ||||
| male | 30 (48.4) | 32 (51.6) | (reference) | * 0.0547 |
| female | 5 (23.8) | 16 (76.2) | 3.0 (1.2, 7.7) | |
| Age - mean (SD) | 30.2 (5.3) | 27.3 (5.8) | 0.9 (0.9, 1.0) | * 0.0265 |
| Ethnicity | ||||
| non-Roma | 31 (41.9) | 43 (58.1) | (reference) | 0.8836 |
| Roma x | 4 (44.4) | 5 (55.6) | 0.9 (0.3, 2.9) |
p<0.10
Table 4.
Univariate correlates of distributive syringe sharing in the past 30 days – injecting drug users in Hungary, 2005-2006
| Characteristic | did not share | shared | OR (90% CI) | sig. p-value |
|---|---|---|---|---|
| N (%) | N (%) | |||
| Total | 68 (81.9) | 15 (18.1) | (reference) | |
| Personal network exposure - mean (SD) | 0.10 (0.98) | −.46 (1.01) | 0.6 (0.4, 0.9) | * 0.0548 |
| Degree centrality - mean (SD) | 4.4 ( 5.3) | 4.7 ( 5.8) | 1.0 (0.9, 1.1) | 0.8353 |
| Heroin dependence - high | ||||
| no | 32 (80.0) | 8 (20.0) | (reference) | 0.6603 |
| yes | 36 (83.7) | 7 (16.3) | 0.8 (0.3, 2.0) | |
| Incorrect knowledge about HIV transmission | ||||
| no | 47 (82.5) | 10 (17.5) | (reference) | 0.8531 |
| yes | 21 (80.8) | 5 (19.2) | 1.1 (0.4, 3.0) | |
| Incorrect knowledge about HCV transmission | ||||
| no | 52 (83.9) | 10 (16.1) | (reference) | 0.4320 |
| yes | 16 (76.2) | 5 (23.8) | 1.6 (0.6, 4.5) | |
| no HCV and no IDU sex partner | 42 (87.5) | 6 (12.5) | (reference) | |
| no HCV and IDU sex partner | 7 (58.3) | 5 (41.7) | 5.0 (1.5, 16) | * 0.0275 |
| HCV and no IDU sex partner | 10 (76.9) | 3 (23.1) | 2.1 (0.6, 7.7) | 0.3475 |
| HCV and IDU sex partner | 9 (90.0) | 1 (10.0) | 0.8 (0.1, 5.1) | 0.8257 |
| Emotionally close to an IDU | ||||
| no | 41 (89.1) | 5 (10.9) | (reference) | * 0.0646 |
| yes | 27 (73.0) | 10 (27.0) | 3.0 (1.1, 8.2) | |
| Gender | ||||
| male | 53 (85.5) | 9 (14.5) | (reference) | 0.1552 |
| female | 15 (71.4) | 6 (28.6) | 2.4 (0.9, 6.3) | |
| Age - mean (SD) | 28.7 ( 5.8) | 27.7 ( 5.9) | 1.0 (0.9, 1.1) | 0.5376 |
| Ethnicity | ||||
| non-Roma | 60 (81.1) | 14 (18.9) | (reference) | 0.5709 |
| Roma x | 8 (88.9) | 1 (11.1) | 0.5 (0.1, 3.3) |
p<0.10
Table 5.
Univariate correlates of receptive syringe sharing in the past 30 days – injecting drug users in Hungary, 2005-2006
| Characteristic | did not share | shared | OR (90% CI) | sig. p-value |
|---|---|---|---|---|
| N (%) | N (%) | |||
| Total | 66 (79.5) | 17 (20.5) | (reference) | |
| Personal network exposure - mean (SD) | −.00 (1.02) | 0.01 (0.93) | 1.0 (0.6, 1.6) | 0.9577 |
| Degree centrality - mean (SD) | 4.3 ( 5.5) | 4.9 ( 4.7) | 1.0 (0.9, 1.1) | 0.6431 |
| Heroin dependence - high | ||||
| no | 32 (80.0) | 8 (20.0) | (reference) | 0.9164 |
| yes | 34 (79.1) | 9 (20.9) | 1.1 (0.4, 2.6) | |
| Incorrect knowledge about HIV transmission | ||||
| no | 49 (84.2) | 9 (15.8) | (reference) | 0.1226 |
| yes | 16 (76.2) | 5 (23.8) | 1.3 (0.5, 3.5) | |
| no HCV and no IDU sex partner | 41 (85.4) | 7 (14.6) | (reference) | |
| no HCV and IDU sex partner | 10 (83.3) | 2 (16.7) | 1.2 (0.3, 4.9) | 0.8567 |
| HCV and no IDU sex partner | 7 (53.8) | 6 (46.2) | 5.0 (1.6, 15 | * 0.0195 |
| HCV and IDU sex partner | 8 (80.0) | 2 (20.0) | 1.5 (0.3, 6.3) | 0.6683 |
| Emotionally close to an IDU | ||||
| no | 40 (87.0) | 6 (13.0) | (reference) | * 0.0672 |
| yes | 26 (70.3) | 11 (29.7) | 2.8 (1.1, 7.2) | |
| Gender | ||||
| male | 54 (87.1) | 8 (12.9) | (reference) | * 0.0053 |
| female | 12 (57.1) | 9 (42.9) | 5.1 (1.9, 13) | |
| Age - mean (SD) | 29.0 ( 5.5) | 26.5 ( 6.4) | 0.9 (0.8, 1.0) | 0.1185 |
| Ethnicity | ||||
| non-Roma | 57 (77.0) | 17 (23.0) | (reference) | 0.9687 |
| Roma x | 9 ( 100) | 0 ( 0.0) | zero cell |
p<0.10
Acknowledgments
The study was funded by the United States National Institute on Drug Abuse, international supplement grant 5R01DA014515-02S1 „Network Mixing and HIV Risk Among Young IDUs in Budapest” (Principal Investigator: Dr. Alan Neaigus; Co-Principal Investigator: Dr. V. Anna Gyarmathy). We would like to thank our local investigator, Dr. Eszter Ujhelyi and our local research staff, Ms. Nóra Zelei and Mr. Ferenc Meczenzof. Dr. Carl Latkin and Dr. Elaine Doherty provided valuable feedback on the analysis and interpretation of the results. We would also like to thank our participants for their involvement in the study.
Footnotes
National Development and Research Institutes, Inc., New York, NY, USA
Contributor Information
V. Anna Gyarmathy, NDRI2 and Johns Hopkins Bloomberg School of Public Health.
Alan Neaigus, NDRI and Columbia University, Mailman School of Public Health.
REFERENCES
- Aceijas C, Stimson GV, Hickman M, and Rhodes T 2004. Global overview of injecting drug use and HIV infection among injecting drug users. AIDS, 18(17): 2295–303. [DOI] [PubMed] [Google Scholar]
- Anderson RM, and May RM. 1988. Epidemiologic parameters of HIV transmission. Nature, 333: 514–18. [DOI] [PubMed] [Google Scholar]
- Borgatti SP, Everett MG, and Freeman LC. 2002. Ucinet for Windows: Software for Social Network Analysis. Harvard, MA: Analytic Technologies. [Google Scholar]
- Burt RS. 1987. Social Contagion and Innovation: Cohesion versus Structural Equivalence. AJS 92: 1287–335. [Google Scholar]
- Dehne KL, Khodakevich L, Hamers FF, and Schwartlander B. 1999. The HIV/AIDS epidemic in Eastern Europe: recent patterns and trends and their implications for policy-making. AIDS 13: 741–49. [DOI] [PubMed] [Google Scholar]
- Des Jarlais DC, Perlis T, Arasteh H, Hagan H, Milliken J, Braine N, Yancovitz S, Mildvan D, Perlman DC, Maslow H, and Friedman SR. 2004. "Informed altruism" and "partner restriction" in the reduction of HIV infection in injecting drug users entering detoxification treatment in New York City, 1990-2001. J Acquir Immune Defic Syndr 35(2): 158–66. [DOI] [PubMed] [Google Scholar]
- European Centre for the Epidemiological Monitoring of AIDS. 2005. HIV/AIDS surveillance in Europe, end year report 2004. 71 . Institute de Veille Sanitaire, Saint Maurice. [Google Scholar]
- European Monitoring Centre for Drugs and Drug Addiction. 2002. Annual report on the state of the drugs problem in the European Union and Norway 2002. Luxembourg: Office for Official Publications of the European Communities. [Google Scholar]
- Evans JL, Hahn JA, Page-Shafer K, Lum PJ, Stein ES, Davidson PJ, and Moss AR. 2003. Gender differences in sexual and injection risk behavior among active young injection drug users in San Francisco (the UFO Study). J Urban Health 80(1): 137–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Friedman SR, Curtis R, Neaigus A, Jose B, and Des Jarlais DC. 1999. Social networks, drug injectors' lives, and HIV. New York: Plenum. [Google Scholar]
- Friedman SR, Neaigus A, Jose B, Curtis R, Goldstein MF, Ildefonso G, Rothenberg RB, and Des Jarlais DC. 1997. Sociometric risk networks and risk for HIV infection. American Journal of Public Health 87(8): 1289–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gore-Felton C, Somlai AM, Benotsch EG, Kelly JA, Ostrovski D, and Kozlov A 2003. The influence of gender on factors associated with HIV transmission risk among young Russian injection drug users. Am J Drug Alcohol Abuse 29(4): 881–94. [DOI] [PubMed] [Google Scholar]
- Gyarmathy VA, and Neaigus A 2005. Marginalized and Socially Integrated Groups of IDUs in Hungary: Potential Bridges of HIV Infection. J Urban Health 82(3) Suppl 4: iv101–iv112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gyarmathy VA, Neaigus A, Ujhelyi E, Szabó T, and Rácz J 2006. Strong HIV and hepatitis disclosure norms and frequent risk behaviors among Hungarian drug injectors. Drug Alcohol Dependence (in Press) . [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gyarmathy VA, Racz J, Neaigus A, and Ujhelyi E 2004. The urgent need for HIV and hepatitis prevention in drug treatment programs in Hungary. AIDS Educ Prev 16(3): 276–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hamers FF, and Downs AM. 2003. HIV in central and eastern Europe. Lancet 361(9362): 1035–44. [DOI] [PubMed] [Google Scholar]
- Heckathorn DD. 1997. Respondent-driven sampling: A new approach to the study of hidden populations. Social Problems 44(2): 174–99. [Google Scholar]
- Kalichman SC, Kelly JA, Sikkema KJ, Koslov AP, Shaboltas A, and Granskaya J 2000. The emerging AIDS crisis in Russia: review of enabling factors and prevention needs. International Journal of STD and AIDS 11: 71–75. [DOI] [PubMed] [Google Scholar]
- Kelly JA, and Amirkhanian YA. 2003. The newest epidemic: a review of HIV/AIDS in Central and Eastern Europe. Int J STD AIDS 14(6): 361–71. [DOI] [PubMed] [Google Scholar]
- Kilibarda M 1993. HIV infection among drug abusers in the Belgrade area. Bull Narc 45: 135–46. [PubMed] [Google Scholar]
- Latkin CA, and Knowlton AR. 2005. Micro-social structural approaches to HIV prevention: a social ecological perspective. AIDS Care 17 Suppl 1: S102–13. [DOI] [PubMed] [Google Scholar]
- Lum PJ, Sears C, and Guydish J 2005. Injection risk behavior among women syringe exchangers in San Francisco. Subst Use Misuse 40, no. 11: 1681–96. [DOI] [PubMed] [Google Scholar]
- Lundgren LM, Amodeo M, and Chassler D 2005. Mental health status, drug treatment use, and needle sharing among injection drug users. AIDS Educ Prev 17, no. 6: 525–39. [DOI] [PubMed] [Google Scholar]
- Mikl J, Vyslouzilova S, Bruckova M, and Jedlicka J 1998. High rate of HIV risk behaviors among drug users in Prague, Czech Republic 1996-7. Int Conf AIDS abstract no. 23197. [Google Scholar]
- Neaigus A, Friedman SR, Curtis R, Des Jarlais DC, Furst RT, Jose B, Mota P, Stepherson B, Sufian M, Ward T, and Wright JW. 1994. The relevance of drug injectors' social networks and risk networks for understanding and preventing HIV infection. Social Science and Medicine 38(1): 67–78. [DOI] [PubMed] [Google Scholar]
- Neaigus A, Friedman SR, Goldstein MF, Ildefonso G, Curtis R, and Jose B 1995. Using dyadic data for a network analysis of HIV infection and risk behaviors among injecting drug users. NIDA Research Monograph 151: 20–37. [PubMed] [Google Scholar]
- Neaigus A, Gyarmathy VA, Miller MM, Friedman SR, and Des Jarlais DC. 2006. Transitions to injecting among non-injecting heroin users: social network influence and individual susceptibility. JAIDS (in Press) . [DOI] [PubMed] [Google Scholar]
- Reilley B, Burrows D, Melnikov V, Andreeva T, Bije M, and Veeken H 2000. Injecting drug use and HIV in Moscow: Results of a survey. J Drug Issues 30(2): 305–22. [Google Scholar]
- Rhodes T, Ball A, Stimson GV, Kobyshcha Y, Fitch C, Pokrovsky V, Bezruchenko-Novachuk M, Burrows D, Renton A, and Andrushchak L 1999. HIV infection associated with drug injecting in the newly independent states, eastern Europe: the social and economic context of epidemics. Addiction 94(9): 1323–36. [DOI] [PubMed] [Google Scholar]
- Robinson NJ. 2000. HIV infection in Poland (1987-1996). Rev Epidmiol Sante Publique 48 , no. Suppl 1: 17–31. [Google Scholar]
- Rogers EM. 2000. Diffusion theory: a theoretical approach to promote community-level change. Handbook of HIV prevention. editor Peterson JL, and DiClemente RJ, 57–65. New York: Kluwer Academic. [Google Scholar]
- Rogers EM. 2003. Diffusion of innovations. 5th ed. New York, NY: Free Press. [Google Scholar]
- Somlai AM, Kelly JA, Benotsch E, Gore-Felton C, Ostrovski D, McAuliffe T, and Kozlov AP. 2002. Characteristics and predictors of HIV risk behaviors among injection-drug-using men and women in St. Petersburg, Russia. AIDS Educ Prev 14(4): 295–305. [DOI] [PubMed] [Google Scholar]
- Stark K, Sieroslawski J, Muller R, Wirth D, Godwod-Sikorska C, and Bienzle U 1996. Determinants of current HIV risk behaviour among injecting drug users in Warsaw, Poland. European Journal of Epidemiology 12: 315–17. [DOI] [PubMed] [Google Scholar]
- Strathdee SA, Patrick DM, Archibald CP, Ofner M, Cornelisse PG, Rekart M, Schechter MT, and O'Shaughnessy MV. 1997. Social determinants predict needle-sharing behaviour among injection drug users in Vancouver, Canada. Addiction 92(10): 1339–47. [PubMed] [Google Scholar]
- Topolánszky Á, editor. 2002. Jelentés a magyarországi kábítószerhelyzetrõl. Budapest: GYISM. [Google Scholar]
- Trotter RT II, Bowen AM, and Potter JM Jr.. 1995. Network Models for HIV Outreach and Prevention Programs for Drug Users. NIDA Research Monograph 151: 144–80. [PubMed] [Google Scholar]
- Ujhelyi E, Tarjan V, Szomor K, Zacher G, Szabo J, and Lepes P 1998. Prevalence of HIV, HBsAg, and HCV infection among Hungarian drug users. Int Conf AIDS abstract no. 60099. [Google Scholar]
- Valente TW. 1995. Network models of the diffusion of innovations. Cresskill, NJ: Hampton Press. [Google Scholar]
- Valente TW. 2005. Network models and methods for studying the diffusion of innovations. Models and methods in social network analysis. editors Carrington PJ, Scott J, and Wasserman S. New York, NY: Cambridge University Press. [Google Scholar]
- Weeks MR, Clair S, Borgatti SP, Radda K, and Schensul JJ. 2002. Social networks of drug users in high-risk sites: fiding the connectinos. AIDS and Behavior 6(2): 193–206. [Google Scholar]
- Zacher G 2003. Drogtúladagolások - fõvárosi helyzet. Jelentés a magyarországi kábítószerhelyzetrõl. ed Ritter I, and Felvinczi KBudapest: Gyermek, Ifjúsági és Sportminisztérium. [Google Scholar]