|
nif:isString
|
-
Data for this study were drawn from a cross-sectional survey conducted in 2007–08 across 21 districts in four states of southern (Andhra Pradesh, Karnataka, Tamil Nadu) and western (Maharashtra) India, identified as high epidemic states by the Indian National AIDS Control Organisation prior to the year 2005 [19]. The data was collected by Population Council, Karnataka Health Promotion Trust and Annamalai University and can be downloaded freely from the Harvard dataverse network [20]. Study participants were recruited through a two-stage systematic sampling procedure. Geographical maps drawn for each district were used to list all migrant worker residential areas and worksites and to select cluster areas within these sites for recruitment. Clusters were created by combining smaller sites and dividing larger sites such that each cluster offered an area with approximately 5000 male migrant workers. Three clusters were then selected randomly from within each district, and migrant men within the chosen clusters were systematically sampled to obtain a minimum of 2500 participants per state. A total of 11,635 eligible male migrant workers satisfied the inclusion criteria- age 18 years or older, having migrated to at least two places in the past two years for work- were selected for the cross-sectional survey. Of these, 145 (1.2%) men refused to be interviewed, and 271 (2.3%) men did not complete their interview and were thus excluded from the analyses, providing our final sample size of 11,219.
Data were obtained through face-to-face interviews conducted in private locations close to the respondent's residence or workplace. Verbal consent was obtained from all respondents before being interviewed. Verbal consent was sought for two reasons: first, large proportions of male migrants were expected to be illiterates and cannot read and sign the consent form; second, no personal identifier or biological data were collected. The interviewers read the complete script of the consent form in to the respondent and explained if there was any doubt about any aspects of survey. For each respondent, a separate consent form was used and those who consented to be interviewed were asked to give their thumb impression on the consent form. Ethical approval for the study was obtained from the institutional review boards of the Population Council and the University of Manitoba, Canada.
Respondents who had viewed either adult movies with sex content or blue films (using compact discs or shown in video parlours) in the one month prior to the survey were categorised as having exposure to pornographic videos (coded as 1); else considered as no exposure to pornographic videos (coded as 0). In this study, this variable was first used as a dependant variable to understand the predictors of viewing pornographic videos and then as an independent variable to examine the effect of viewing pornographic videos on HIV-related sexual risk behaviours.
HIV-related sexual risk behaviours in the study were measured using the following indicators: paid sex (sex with a sex worker) and unpaid sex (sex with a female other than spouse who is not a sex worker), engagement in anal/oral sex, inconsistent condom use, experience of STI-related symptoms and alcohol consumption prior to sex. All respondents were asked a question with dichotomous response categories (no, yes) to examine whether they had sex with a sex worker in the 12 months prior to survey. A similar question was asked to examine if respondents had sex with a non-sex worker other than spouse (hereinafter referred to as non-sex worker) in the 12 months prior to survey to assess unpaid sex. Questions were also asked on the type of sex acts they engaged in the last time they had sex with these partners. Individuals who reported either engaging in anal or oral sex were considered to have engaged in anal/oral sex in the last sex. Inconsistent condom use was assessed separately for sex workers and non-sex workers. For each type of partners, respondents were asked about frequency of condom use (indicated by 1 = every time, 2 = almost every time, 3 = sometimes, 4 = never) during sex in the past 12 months. Individuals who had “always” used condoms in the last 12 months were coded as 0 (consistent condom users) and the rest were coded as 1 (inconsistent condom users). The survey also collected information on self-reported STI symptoms. Participants were defined as having STI-like symptoms if they reported any of the following in the past 12 months: genital ulcers; swelling in groin area; itching in genital area; or frequent painful urination. To examine alcohol consumption prior to sex, a question with dichotomous response categories was asked to those who ever had sex.
Data collected included the socio-demographic characteristics of migrants, including age, highest level of education completed, marital status including cohabitation status with spouse, income, ability to save money, living arrangement, and frequency of return to the native place (origin). These variables were used as independent variables while predicting the likelihood of exposure to pornographic videos and as covariates while predicting the risk associated with the exposure to pornographic videos for different HIV-related sexual risk behaviours.
Univariate, bivariate and multivariate analyses were performed. Univariate analysis was used to describe the profile of the study population. Bivariate analysis was used to present the prevalence of exposure to pornographic materials by socio-demographic characteristics. A series of multiple logistic regression models were generated; first to examine the predictors of viewing pornographic videos and second to examine the effect of viewing pornographic videos on HIV-related sexual risk behaviours. In the first logistic regression analysis, viewing pornographic videos was considered as the dependent variable and socio-demographic characteristics of migrants were considered as independent variables. In the rest of the logistic regressions, viewing pornographic videos was the key independent variable whereas socio-demographic factors were used as covariates and indicators of HIV-related sexual risk behaviours were considered as dependent variables. Results were presented in the form of percentages, adjusted odds ratios (AOR) and their corresponding 95% confidence interval (CI). All the analyses were carried out using STATA version 12.1 (StataCorp., College Station, TX, USA).
|
![[This page as RDF]](http://data.gesis.org/somesci/static/rdf-icon.gif){kind=icon}