Social Network and Risk-Taking Behavior Most Associated with Rapid HIV Testing,Circumcision,and Preexposure Prophylaxis Acceptability Among High-Risk Indian Men

Participants and setting

This study was conducted from 2010 to 2011 at the intersection of two national highways in Hyderabad, Andhra Pradesh, India. Both truck drivers and young apprentice drivers (cleaners) between the ages of 16 and 65 were recruited in a proportional manner from 12 locations within 4 areas near entry points into the city. Trucks parked at these locations were approached in a systematic manner with information on eligible nonparticipants recorded. Truck drivers and cleaners who understood either Telugu or Hindi were eligible for participation. Of eligible truck drivers and cleaners, 7 truck drivers and 3 cleaners were randomly selected to be interviewed each day. A 60-min face-to-face survey was administered.

Survey

Blood samples

Dry blood spots (DBS) were collected using a validated finger-prick method with a safety lancet and prepared on Whatman No. 3 paper according to procedures utilized previously. ^{38 –40} Three sequential HIV antibody tests were conducted according to NACO guidelines ⁴¹ using standard methods and kits (Vironostika HIV Uni-Form Ag/Ab, bioMerieux, France; Vidas HIV Duo Ultra; bioMerieux, France; and HIV Tridot; J. Mitra, New Delhi, India). HSV-2 IgG (HerpeSelect 2 ELISA IgG, Focus Diagnostics, CA) and Treponemal pallidum IgG (Treponostika sandwich ELISA, bioMerieux, France) testing was also conducted as in previous work. ⁴²

Analysis

Descriptive statistics for the total sample and stratified by truck-driver/cleaner status are reported. Cronbach α was calculated for the five acceptability items within each of the three interventions. We examined the relationship of intervention acceptability and social network characteristics in separate bivariate analyses for each intervention. We then utilized a stepwise approach by creating three separate initial logistic regression models for rapid test, circumcision, and PrEP acceptability that included other sociodemographic, risk and HIV/STI serostatus variables significant at the p<0.05 level from bivariate analysis. Variables that remained statistically significant in the initial regression analyses were then combined into a final multiple logistic regression model for each of the three interventions separately. Truck driver status (truck driver or cleaner) was forced into all final models. Analyses were performed using STATA software version 11.0 (StataCorp, Austin, TX).

Ethical considerations

The protocol was approved by the Indian Council of Medical Research (ICMR), the National AIDS Control Organization (NACO), and institutional ethics panels in India and the United States. All study participants provided written informed consent and received a gift worth 120 Indian Rupees (∼$3 USD) for participation.

Respondent characteristics

Of the 1800 eligible truck-drivers and cleaners we achieved an 89% participation rate (Fig. 1) for a total of 1602 respondents. Table 1 depicts relevant sociodemographic, risk, and serostatus characteristics.

OPEN IN VIEWER

FIG. 1.

Enrollment and participation.

Social network characteristics and intervention acceptability

There was excellent internal consistency for acceptability items across each of the three interventions (Cronbach α=0.96–0.98). Figure 2 depicts acceptability of rapid testing, circumcision, and PrEP. High acceptance of rapid testing was reported by respondents: only 2.6% (42) were unlikely to accept such testing. Among uncircumcised participants, however, only 9.1% (113) were likely to be accepting of circumcision. For PrEP, 85.9% (1376) of study participants were likely to be accepting, however, a pericoital formulation was preferred to a once-daily PrEP formulation among 57.5% (800) of those likely to accept PrEP and 97.4% (222) of those unlikely to accept PrEP. Table 2 depicts the distribution of social network characteristics for respondents.

OPEN IN VIEWER

FIG. 2.

Novel prevention intervention acceptability among Indian truck drivers and cleaners (n=1602).

Discussion

Over the past few years there has been a substantial increase in the number and diversity of biomedical interventions available ranging from adult male circumcision to PrEP, earlier initiation of antiretroviral therapy; all of which require increased uptake of HIV testing to appropriately determine optimal strategies. ⁴³ The current study illuminates several important factors that are likely to influence novel prevention intervention acceptability among at risk Indian men. We found that acceptability of rapid testing was high, PrEP was accepted by most study participants—with a pericoital strategy preferred—and circumcision was generally unacceptable. Additionally, several social network factors were associated with prevention intervention acceptability. Prosocial factors were generally associated with prevention intervention acceptability for rapid testing and circumcision. Acceptability of rapid testing for example, was associated with having more close friends, having a very close friend, communicating daily with friends, feeling that friend advice is important and discussing sexual matters with friends. Finally, we found that the level of risk taking did not predict acceptability of most of the prevention interventions with the exception that individuals who were riskier had favorable attitudes of PrEP, but the only predictor of circumcision acceptability was prior history of sex with a man.

Although rapid testing was found to be highly acceptable which replicates previous work in this setting, ^44,45 overall rates of adults tested and who receive their results in India are some of the lowest in the world at 10.8%. ⁴⁶ These findings suggest that low rates of testing may reflect insufficient public health promotion of testing as part of comprehensive HIV prevention strategies. Strategies to decrease barriers to HIV testing in this highly mobile and at risk population are urgently needed given that only 19% of the sample overall and 8.8% of cleaners had ever received an HIV test previously.

In the case of PrEP, overall acceptability was much higher at 85.6% than in previous work with this population where initial willingness to consider PrEP was only 32.9%. ⁸ The reason for this may be due to differences in data collection methods utilized, analytic methods employed, changes in PrEP attitudes over time, or perhaps increasing use of antiretroviral medications for treatment over time in this context. For example, the study by Schneider et al. ⁸ utilized qualitative interviews for data collection with subsequent analysis of one acceptability item and a willingness to pay application that differed from the acceptability analysis utilized in this study. Prior studies have suggested that Indian patient satisfaction is enhanced when medications are offered by their physicians ^{47 –49} that would tend to corroborate oral prophylaxis as one of the more attractive prevention interventions in this study. However, considerable challenges remain requiring additional investigation such as the concept of taking pills long term, which often clashes with the symptomatic disease model in India in which patients often decide that they do not need medications when they feel well. ⁵⁰ The finding in this study of a general preference for pericoital dosing of PrEP may alleviate concerns for long-term use of medication in a daily format. Current work is underway to determine efficacy of other PrEP formulations such as fixed intermittent or pericoital compared to daily dosing. ⁵¹ Intermittent formulations, for example, may be highly relevant for this mobile population given that they have predictably episodic risk.

Finally, the low acceptability of circumcision in this setting is not surprising. The potential for negative response due to religious connotation (in a country where being circumcised is highly predictive of being Muslim) poses a major barrier to circumcision for HIV prevention in India. ^8,52 Several strata of uncircumcised men within our sample, however, were more likely to find circumcision acceptable: Christians and MSM. The finding of MSM as being more accepting of circumcision is of interest and worth further investigation. It is feasible that for some MSM circumcision may be considered as a form of cleanliness or even attractiveness, ⁵³ however, others have found that uncircumcised partners may be preferred. ⁵⁴ Studies of circumcision among MSM populations are currently underway ⁵⁵ and if found to be effective, may be a population worth considering for intervention in India ⁵⁶ , although use of circumcision outside of generalized epidemics would require careful examination. ⁵⁷

Our findings that higher risk practices are associated with intervention acceptability follow previous findings of perceptions of risk and willingness to use PrEP or circumcision. ⁸ Of concern from our data is that MSM were not more likely to find rapid testing or PrEP more acceptable than heterosexual truck drivers; two interventions with proven benefit in this population. ^58,59 The lack of a difference in acceptability between heterosexual truck drivers and those truck drivers who have sex with men could suggest that perception of risk for HIV is similar between the two groups. Truck driver targeted HIV behavior change communication may not include messages specific for MSM. Furthermore, we are unaware of empirical data that describes the behavior and risk practices of men who are both truck drivers and MSM.

Several of the relationships between social network characteristics and composition and acceptability of the prevention interventions are of interest. Notably, prosocial network characteristics were consistently associated with increased acceptability of rapid testing, potentially reflecting the increasingly normative nature of HIV testing in this population. Truck drivers are increasingly aware of their HIV infection risk, sometimes to the point of stigmatizing the profession. ⁶⁰ However, although prosocial factors were universally associated with acceptability of testing, this was not the case with PrEP where some prosocial features were associated with less acceptability of PrEP. For example there was discordance between having at least one close friend and speaking daily with friends and PrEP acceptability. Previous work suggests that daily communication, usually by phone, is likely with individuals who reside in the truck drivers' hometown. ³¹ Close relationships with such individuals, who may in turn be close to truck drivers' families might affect the willingness of the participants to acknowledge their HIV risk ⁶¹ and thereby limit their willingness to consider this prevention modality as acceptable. Close relationships with such friends whose norms may not include frequent contact with female sex workers at rest stops may be a source of cognitive dissonance ⁶² for the respondents given their described sexual behavior when away from home. Additional research is needed to elucidate how acceptability of different prevention modalities is established in relation to the different friend typologies that exist within distinct geographic spaces of this highly mobile population. ⁶³

Ability of social networks to influence attitudes may be a function of the invasiveness of the intervention. For things that are less intrusive such as oral testing and taking a pill, social networks may be playing a role, but for something requiring minor surgery at the individual level, social networks may not play as much of a role. In general, social network variables seemed to play a greater role in the likelihood of accepting HIV testing than in accepting PrEP, while none were significantly associated with circumcision likelihood. As the interventions become increasingly burdensome for the individual, the persuasiveness of social network effects in the individual's uptake of these interventions may diminish. Mobilizing communities to obtain HIV testing through social networks as is currently being tested (Project Accept ⁶⁴ ) and may be useful in the Indian setting if it is found to be efficacious.

Respondents with higher income were less likely to find PrEP acceptable. This is similar to findings from a PrEP acceptability study in the United States where high-risk men who were less educated were more interested in PrEP. ³⁴ Further research into the conceptual understanding of PrEP among diverse social and economic strata is required if effective adoption and adherence is to be obtained among high-risk men such as Indian truck drivers and younger apprentice drivers. Because younger apprentice drivers earn less and are less educated, they may require specific PrEP messaging and monitoring compared to their older and more experienced truck driver mentors.

Our study had several limitations. First, we focus on acceptability of HIV prevention interventions that are not yet available in India for HIV prevention in this population. However, as in other studies, initial acceptability has been shown to be linked to future behavior in multiple settings for multiple behaviors. ^11,65,66 Because the interventions examined in this study are either under investigation or currently unavailable for HIV prevention in India, actual uptake of the interventions could not be assessed. Additionally, we did not directly ask respondents about behavioral disinhibition and risk compensation, and these themes will be needed to be addressed in further studies of HIV prevention interventions that are found to be modestly effective in high-risk men. Furthermore, we focused on the close social networks, and particularly friends of these truck drivers and did not measure characteristics of expanded social networks. We also did not examine other potentially useful interventions such as an HIV vaccine or the use of microbicides.

Findings from this work can be further developed in order to inform NACP policy decisions. Obtaining an understanding of the social network characteristics of various types of prevention interventions could help target resources toward the individuals most likely to benefit from them. This study has demonstrated an often robust association between social interconnectedness and willingness to accept PrEP and HIV testing in Indian truck-drivers and cleaners. Further interventions could take advantage of high-risk Indian men's interconnectedness in order to increase the uptake of HIV prevention interventions within their social networks, recognizing that uptake may also be concomitantly driven by individual risk practice and risk perception.