The study employed a concurrent mixed methods approach where the convergence model was used for triangulation of results. Two quantitative and one qualitative data collection procedures were utilized―a) Collection of existing programme data―Within the National AIDS Control Programme, anyone, irrespective of age and gender, can be tested for HIV in the ICTCs, whereas all

consenting pregnant women are tested for HIV in the PPTCTs (Prevention of Parent to Child Transmission centres). Basic socio-demographic information as well as risk categories and HIV test results are captured anonymously in the registers of these centres. We extracted three years’ (2012-’13 to 2014-’15) retrospective from ICTCs/PPTCTs in the three study districts; b) Collection of structured questionnaire-based quantitative data―through face-to-face interviews from the consenting ICTC/PPTCT married women attendees (who attended during the previous six month period) as well as their spouses to explore sexual and other risk behaviours and practices in greater details; in addition, migration-related history of the spouses of the married women attendees were also obtained. Here, we undertook a case-control approach. Based on the register information, for each HIV(+) married woman we randomly selected two HIV(−) married women belonging to the same district and within ±5 years (minimum 18 years) age range; spouses of these women were also asked to participate and provide information through questionnaire-based interview; and c) Collection of qualitative data―through Focus Group Discussions (FGD) among selected population groups including female sex workers, long-distance truckers and male migrants. Broadly, the topics included issues on socio-economic and demographics, sexual and other risk behaviors, sexual partnerships, condom use, knowledge of HIV/AIDS and STIs, migration-related information as well as social and programme support. The FGDs were conducted in local language and were audio recorded. The recordings, along with the field notes, were transcribed within a few hours and translated into English later. All participants were aged 18 years or older.

This research project was approved by the Institutional Ethics Committee of ICMR-National Institute of Cholera and Enteric Diseases, Kolkata―700,010 vide approval letter no. A-1/2014-IEC dated 09.02.2015. In addition, for extracting existing data from the ICTC/PPTCT registers, approval was taken from the Data Sharing Committee of the Assam State AIDS Control Society as per the NACO Data Sharing Guidelines, 2015. Informed consent was taken from each client while collecting quantitative or qualitative data. Strict Confidentiality of all data was maintained and no personal identifiers were used in the analysis.

The extracted program data as well as quantitative data collected through structured interviews were entered into a pre-designed computerized database using MS-Excel. Following data cleaning and editing, descriptive analyses were carried out using STATA statistical software (STATA SE ver. 8.2). Quantitative interview data were analyzed using a case-control approach (HIV+ vs. HIV−), to identify risk/preventive factors for HIV infection. We used a step-by-step participatory procedure to analyze qualitative data thematically. For this, we undertook an iterative process of data collection and analysis through debriefing notes and periodic team meetings to identify emerging themes and follow-up questions. After all FGDs were completed, a data analysis workshop was organized and a codebook was developed based on both a priori and emerging themes. All transcripts were systematically coded using the software NVivo 10.

Analysis of retrospective HIV testing data was done for the general clients and pregnant women in the ICTCs and PPTCTs in the three study districts for three years (2012-’13 to 2014-’15). A total of 19640 clients were tested during this period in these three districts. As indicated in Table 1, people from other districts

of Assam and even from other States also visited these ICTCs and they often had higher prevalence of HIV infection. Among those who visited from different districts of Assam, attendees belonging to Karimganj district showed the highest HIV positivity (7.41%).

Table 2 presents the gender-disaggregated HIV prevalence among the attendees from the three study districts according to their marital, educational and occupational status. It was found that persons in the middle age groups (20 - 40 years) and those who were divorced or widowed or separated in both genders were most at risk for HIV infection. However, education had no influence on being HIV positive. Truckers among males had the highest rate of HIV positivity.

Limited data were available for prevalence of HIV among the clients according to their reported risk behaviours (Table 3). It was very disturbing to note that among both males and females, those who reported unsafe heterosexual practices had unacceptably high HIV positivity, even more than some of the so-called high risk groups.

A case-control approach was undertaken for the questionnaire-based interview.

*Heterosexual risk behavior: Male or Female clients reportedly involved in unsafe heterosexual practices.

All HIV(+) married women who attended the ICTCs/PPTCTs of these three districts during the previous six-month period were asked to participate as “cases”. Two randomly-chosen HIV(−) married women who attended from the same district and were within ±5 years of the age of the “case” women were recruited as “controls”. Interview-based data were also collected from their spouses wherever possible. Written informed consent was obtained from each participant before collection of any data.

A total of 25 HIV(+) women (“cases”) and 41 HIV(−) (“control”) women could be recruited in the study. Spouses of 21 “case” women and 37 “control” women could also be interviewed (Table 4).

Socio-economic and demographic characteristics of study participants

Table 5 shows the major socio-demographic characteristics of the “case” and “control” women. The two groups of women were comparable in terms of their age, education status, religious beliefs, castes, marital status as well as their age at (first) marriage. Overall, almost all the women declared no income of their own; only two women had a monthly income up to Rs. 5000.

The general characteristics of the spouses of the women are shown in Table 6. The table highlights that the spouses of the two groups of women were also comparable for these characteristics, except for education status, which showed that a higher percentage of spouses of “control” women had Grade 11 or higher education.

*SC = Scheduled castes; OBC = Other backward classes.

Sexual and other risk behaviours/practices of the study participants

Our analyses indicated that the “case” and “control” women were comparable in terms of their age (in years) at first sex (21.1 ± 3.9 vs. 21.5 ± 3.7; p = 0.64). Only one woman in each group had history of having pre-marital sexual experience while none of the women reported any extra-marital sexual relationships. Again, only one woman in the “control” group gave history of having anal sex ever. Eight percent (2 of 25) of “case” women and 12.2% (5 of 41) “control” women ever experienced any signs/symptoms of sexually transmitted infections (p = 0.70). Regarding substance use, 20% (5 of 25) “case” women and 15% (6 of 40) “control” women shared history of smoking cigarettes; however, none of them gave history of consuming alcohol or any other addictive substance.

Table 7 demonstrates the patterns of risk behaviours among the spouses of these women. It shows that compared to the spouses of the “control” women, a significantly higher percentage of spouses of the “case” women had pre-marital sexual experience, had sex workers as partners of pre-marital sex, always used condoms during sex with their spouse (may have reflected their own risk perception) and had extra-marital sexual relationships. The patterns of condom use with spouse were also corroborated by the respective groups of women, who also narrated similar distributions. The partners of extramarital sex of the male spouses were female sex workers only and condoms were not always used with them. None of the male spouses also always used condoms during sex with their pre-marital partners. Moreover, spouses of the “case” women had significantly earlier initiation of sex compared to the spouses of the “control” women. Regarding substance use, 47.6% (10 of 21) and 72.9% (27 of 37) spouses of “case” and “control” women respectively, had history of cigarette smoking; alcohol intake was reported by 38.1% (8 of 21) and 32.4% (12 of 37) spouses, respectively. No other substance use, including injecting drugs, was reported.

Thus, the results of the quantitative data suggested that it was mostly the behaviours and practices of the male spouses of the women, especially their indulgence

in pre- and extra-marital sex, having sex workers as their sex partners and negligible use of condoms during high risk sex that might put the women at risk of contracting HIV in these populations. Thus, more programmatic attention may be needed for this section of people, apart from high risk groups and the antenatal clinic attendees.

The purpose of collecting and analyzing qualitative data was to supplement the findings from the above-mentioned quantitative data, in order to enhance our understanding about the possible reasons of increasing HIV prevalence among general population (as indicated by prevalence among ANC attendees) in the three study districts. A total of 12 focus group discussions (FGDs) were conducted to explore relevant issues among different population groups [FSWs (home/street/brothel-based) = 6, Truckers = 3, Male migrant labourers = 3]. Each FGD had 6 - 9 attendees of various age groups.

Analyses of these qualitatively collected data provided clues to explaining some of the findings that emerged from quantitative data analyses. For example, analysis of existing ICTC/PPTCT data revealed a high prevalence of HIV among the housewives, especially in the Nagaon and Karimganj districts (7.71% and 6.5%, respectively); during FGDs with the sex workers in these two districts, it was found that many of them were actually married and were doing the sex work on a part-time basis at a distant place on the pretext of being engaged in some kind of job and without the knowledge of their family members. Many of these women do not consider them as sex workers and told that they indulged in such activities only to earn some extra money for sustaining themselves or their family, as quoted below for example:

FSW (home-based)/45y/Cachar: If anybody does sex work, she (usually) has some bindings … She is doing this work because of her children … because of her livelihood.

It was also noted through the FGDs with FSWs that despite having fair to good knowledge about HIV, its transmission and the consequences, many of them would often agree for unprotected sex if they were offered higher amounts of money.

FSW (home-based)/45y/Cachar: When sex work is my business, then so many types of people come … Someone may offer only Rs.10 and someone may offer Rs.1000 … for the sake of money then, we are forced to do sex without condoms FSW (street-based)/38y/Karimganj: Forget about customers … even they (the FSWs) also do not use condoms … condoms are supplied to them, but still they do not use. Because they need money… without condom, they earn more money

However, sometimes lack of proper knowledge about the utility of using condoms during sex might have also played a role for non-use of condoms, as demonstrated below:

FSW (street-based)/22y/Karimganj: HIV can spread if condom is not used. (However), some sex workers do not use condoms, they use Mala tablet (govt. supplied oral contraceptive pills) instead.

This could be one explanation for rising HIV prevalence among the female sex workers that drove this project to explore its possible reasons.

Through the FDGs conducted with the male migrant (out-migrant) labourers, it was understood that many young people, mostly in the range of 18 to 35 years, worked and stayed away from their families for a considerable period at a distant place (other districts or even in other states, like in Mumbai). Many of them would visit their families during harvesting seasons, but would go back again for work. The main motivation was better earning opportunities. The following accounts will give some perspective:

Male Migrant Labourer/(Age/Educ: 26y/Grade 5; 25y/Grade 7; 24y/Grade 9)/Nagaon About 60% - 70% of adult male population of this area seek jobs outside. Male Migrant Labourer/(24y/Grade 9)/Nagaon Payment is too low here. So, they go outside. Male Migrant Labourer/(26y, Grade 7)/Nagaon …. Also, there are lack of facilities [so, they go outside for work] Male Migrant Labourer/(26y/Grade 5; 22y/Grade 6)/Nagaon There are no … almost no company [industry] in Nagaon. Male Migrant Labourer/(24y/Grade 9)/Nagaon I can earn almost double there … it has changed my lifestyle, food habit, and made my health better [other participants also echoed the same]. We can also meet new people, even foreigners, which is not possible if we stay here in the village.

Many of these young males often indulge in high-risk behaviours like visiting sex workers.

Male Migrant Labourer/(24y/Grade 9)/Nagaon When 10 - 12 persons stay together, someone of them may have information about a sex worker’s home or workplace in the vicinity … if one of them visits her, then others also start to see her―it becomes like a habit then.

Asked about whether they would agree to pay more money to the sex workers to have sex without condoms, almost all of them responded that they would. Like the FSWs as mentioned earlier, these people also possessed misconceptions about condoms and its use, as exemplified below:

Male Migrant Labourer/(24y/Grade 9)/Nagaon There is a medication. Whoever doesn’t use condom, after eating that [the medication], there will be no need to use condoms. … I have heard like this, but have not seen yet. There is a medical drug―whatever you do in five minutes, nothing is needed [if that drug is taken].

Similarly, many of the truckers (especially the long distance truckers) also indulge in different risk behaviours, particularly unprotected sexual contacts with female sex workers. These people also have lower levels of awareness about HIV transmission and its implications.

It was worrisome that during the focus group discussions, especially with the street-based female sex workers, it was stated that there was a growing number of female students being involved in this profession―again, mostly to earn money to meet their personal and/or educational needs. Since many students now often access higher education at a place distant to their home―thus, remain in a comparatively more free environment and with different kinds of entertainment / addiction opportunities―such risk taking behaviours seem not very unusual, specially with peer-pressures.

Putting together, it appears that despite having high level of awareness about HIV-related matters, felt need for additional income is playing a crucial role for the females to indulge in high risk behaviours, although many of them do not identify them as sex worker. On the other hand, lack of adequate facilities and opportunities push the young males to seek job away from their residence, where circumstances facilitate risk-taking behaviours among them. Thus, along with the females themselves to some extent, their spouses or partners who might be a migrant labourer, trucker and/or a client of sex workers, may be playing important roles behind the rising trend of HIV infection among the general population in select districts of Assam.