Study Sample/Population
This cross-sectional survey of contraceptive use and knowledge was part of a larger cohort study called ‘Saving Children, Improving Lives’ (SCIL) project conducted by the Public Health Research Institute of India (PHRII). Data were collected from 2014- 2016 in Mysore District, Karnataka, India.(18) According to the last census, the total population of Mysore district is 2,994,744 persons. More than half (58.6%) of residents live in the district’s 1,332 rural villages. For rural residents, the annual per capita income is INR 16,086 [USD $322] and literacy rate is 63.3%, compared with an all-India annual per capita income of INR 38,005 [USD $760] and literacy rate of 74.0%. The majority of residents identify as Hindu (86%) and the remaining 14% as Muslim or belonging to other religions.(19, 20) Given the higher fertility rates(21), analyses were restricted to tribal groups (see Table 1).
Measures
Data were collected from all pregnant women attending the mobile medical clinics conducted by PHRII. During the study period, 303 pregnant tribal women were part of the cohort and consented to be interviewed. Sociodemographic data, medical history, obstetric history, knowledge, attitude and beliefs about HIV, and factors influencing institutional delivery were collected using an interviewer-administered survey in Kannada. Information on sampling methods is detailed in an article by Kojima et al (2017).(18) We assessed sociodemographic characteristics including participant age (years), education (none, primary, or >primary), employment (yes or no), caste (Jenu Kuruba, Betta Kuruba, Soliga, Yarava, or other tribal group), number of live children, monthly household income (<4,000, 4,001-10,000, or >10,000 rupees) and religion (Hindu or Muslim). We also assessed other factors commonly associated with contraceptive knowledge and use, including years of marriage, husband’s education (none, primary, or >primary), primary household decision-maker (participant only, participant’s husband, jointly with husband, others in the household, and jointly with others in the household). To assess awareness of birth control methods, we asked each participant if she had ever heard of (yes/no) or if she knew where to get (yes/no) four temporary forms of birth control (condoms, injectables, copper-T intrauterine device, and oral contraceptive pill) and two permanent methods (female sterilization and male sterilization). To assess birth control usage, we asked whether the participant or her partner had ever used (yes/no) each of the four contraceptive temporary methods or the two permanent contraceptive methods.
Analysis
All data were entered into Microsoft Access (Redmond, WA) and analyzed using StataSE V14.1 (College Station, TX). Descriptive statistics were performed using frequencies and proportions for categorical variables. Bivariate analyses were conducted to determine differences in sociodemographic characteristics between women who reported knowing at least one form of temporary contraception versus those who did not using univariate logistic regression. Multivariable logistic regression analyses were conducted to examine factors associated with contraceptive knowledge. Regression diagnostics to detect multicollinearity or redundancy were conducted and did not detect any important correlations that warrant omission of variables from the models. Results from the logistic regression are reported as adjusted odds ratios (AOR) and associated 95% confidence intervals (95%CI). Since a total of 99.7% of women were Hindu, we did not include religion of women in the regression models due to absence of variability.
Ethical considerations
The study was reviewed and approved by the Institutional Review Boards of the Public Health Research Institute of India and Florida International University, USA. All women provided written informed consent before participating in the program.