Prevalence and Determinants of Contraceptive use among Employed and Unemployed Women in Bangladesh

2.1. Sampling

This study utilized a representative set of cross-sectional data extracted from the Bangladesh Demographic and Health Survey (BDHS) 2011.[15] The survey was conducted under the authority of the National Institute of Population Research and Training (NIPORT) under the Ministry of Health and Family Welfare, Bangladesh. A nationally representative household based sample was created through a stratified, multistage cluster sampling strategy of which 600 primary sampling units was constructed (207 in urban and 393 in rural areas). The primary sampling units were derived from a sampling frame created for the Population and Housing Census 2011, provided by Bangladesh Bureau of Statistics (BBS). All ever-married women aged 12-49 years who were usual members of the selected households and those who spent the night before the survey in the selected households were eligible to be interviewed in the survey. A total of 18,222 ever- married women aged 12-49 years were identified in these households, and 17,842 were interviewed, yielding a response rate of 98%. Finally, a total of 16,166 currently married women (employed 1950; unemployed 14,666) aged 15-49 years were selected for this study after dropping divorced/widowed/separated women and women aged below 15 years. Detailed information on survey design and sampling procedures has been reported elsewhere.[15] The study is approved by the Department of Population Science and Human Resource Development, Faculty of Science, University of Rajshahi; Bangladesh.

2.2. Measures

The dependent variable, women’s current contraceptive use, was a dichotomous variable indicating respondents’ use of any method (modern or traditional) of contraception at the time of the survey. We included several theoretically pertinent sociodemographic variables. We classified women’s age into empirically important groups (younger [15-24 years], middle age [25-34 years], and older [35-49 years]). Women’s education level was defined in terms of the formal education system of Bangladesh: illiterate (0 year), primary (1-5 years), secondary and higher (6 years or more). Tertiles were used in classifying number of living children (none, 1-2 and 3 and above). Place of residence was categorized as rural versus urban. Religion was categorized as Muslims versus non-Muslims. We used the BDHS wealth index as a proxy indicator of socioeconomic position. The BDHS wealth index was constructed from data on household assets, including ownership of durable goods (such as televisions [TV] and bicycles) and dwelling characteristics (such as source of drinking water, sanitation facilities, and construction materials). We used principal component analyses to assign individual household wealth scores. These weighted values were then summed and rescaled to range from 0 to1, and each household was assigned to the five quintiles. Tertiles were used to classify age at first cohabitation (<15, 15-17 and 18 and above years) and hearing about family planning (FP) on TV during last few months of the survey was classified as yes versus no.

2.3. Statistical analyses

Prevalence of current contraceptive use was calculated for employed and unemployed women. Differences in current contraceptive use by sociodemographic characteristics were assessed by χ² –test, with significance for all analyses set at p <0.05. Firstly, a binary logistic regression model was fitted to assess the net effects of selected sociodemographic variables on the use of contraception among currently married women irrespective of their employment status. Secondly, we categorized the women according to their employment status (employed or unemployed) and another two binary logistic regression models were constructed as whether the employed women used contraceptive (yes or no) and whether the unemployed women used contraceptive (yes or no) to determine which factors were most strongly associated with current contraceptive use. We entered all the covariates simultaneously into the multiple binary logistic regression models. There is an important assumption in multiple regression analyses, either linear or logistic, that there is no multicollinearity problem (dependent each to other) among the independent variables. However, there is no exact method to detect the multicollinearity problem in multiple logistic regression analysis. In this study, the magnitude of the standard error (SE) was used to detect the multicollinearity problem, if the magnitude of the SE lies between 0.001 and 0.5; it is judged that there is no evidence of multicollinearity.[20] In multivariate analysis, binary logistic regression analysis (enter method) was performed to determine which factors might affect the probability of using contraceptives. The analyses were adjusted for age, education, husband’s education, religion, region, residence, number of living children, fertility preference, and heard about FP on TV. We estimated the odds ratios (ORs) to assess the strength of the associations for the 95% confidence intervals (CIs) for significance testing. Statistical Package for Social Sciences version 18.0 (SPSS Inc., Chicago, IL, USA) was used for statistical analysis.

3.1. Descriptive statistics

Table 1 showed contraceptive use status of the study participants. The prevalence of current use of contraception among employed and unemployed women were 67.2% and 60.9% respectively, among them 56.5% employed women and 51.7% unemployed women used modern contraceptive method. The most commonly used contraceptive method by employed women was pill (27.7%), followed by injection (11.4%), periodic abstinence (8.3%) and condom (7.2%) whereas these proportions were 26.8%, 11.2%, 7% and 5.9% respectively among unemployed women. Female sterilization was found higher than male sterilization.

The χ² test was used to assess the association between current use of contraception and socio-demographic indicators (Table 2). The highest prevalence of contraceptive use was recorded among employed women aged 25-34 years, those who were educated, residing in urban areas or in Rangpur division, got married in the early ages (15-17 years), heard about FP on TV during few months preceding the survey, having more than 2 children and who did not want children anymore. On the other hand, contraceptive use was highly pervasive among unemployed women aged 25-34 years, had higher education, residing in urban areas, married before 15 years of age, did not want children any more, and among non-Muslims.

Women in the age group 25-34years used contraceptives considerably more than that of younger and older counterparts. We found that 74% employed and 67% unemployed women of this age group used contraceptives and it was noticeably higher in comparison with other groups of women. There was a significant effect of women’s educational attainment (p <0.001) and husband’s education (p <0.001) on contraceptive use among unemployed women. This study elucidated that contraceptive use was higher among employed women of every level of education (primary 69%, secondary 68% and higher 68%) than their unemployed counterpart (primary 62%, secondary 59% and higher 65%) (Table 2). The proportion of women using contraception in accordance with husband’s education was almost same as those of the wife’s education (Table 2). Use of contraception was found higher in Rangpur division, followed by Rajshahi, Khulna, Barisal, Dhaka, Chittagong and Sylhet division. Women in Sylhet division were in vulnerable condition as only 52% employed and 47% unemployed women used contraceptives there.

The present study described that 69% and 66% employed women used contraceptives in urban and rural areas respectively, whereas, these proportions became 64% and 59% respectively among unemployed women. Tendency of using contraceptives was found lower among Muslim unemployed women (60%) than that of non-Muslim counterparts (67%). The maximum use of contraceptive was observed among employed women (74%) who had more than 2 children whereas the corresponding figure was 65% among unemployed women (p <0.001). Nearly 25% employed women and 31% unemployed women were in vulnerable condition because they no longer wanted any children but did not use any contraceptive (Table 2).

This study observed that employment status of women had a significant effect on contraceptive use (p <0.001). Probability of using contraceptive increased among employed women than their unemployed counterpart [unadjusted OR (95% CI): 1.319 (1.193-1.458)]. We found that employed women were 1.21 times more likely to use contraceptive than their unemployed counterparts [OR (95% CI): 1.211 (1.083-1.353)] (Table 3). It was observed among employed women that with the increase of educational level probability of contraceptive use increases. Completing higher education increased the probability of contraceptive use compared with illiterate [OR (95% CI): 1.646 (0.966-2.802)]. Probability of using contraceptive decreased among women aged 35-49 years in comparison with their youngest counterparts [OR (95% CI): 0.302 (0.207-0.441)]. Number of living children was likely to play a significant role in contraceptive use, since women who had 1-2 children and who had more than two children were more likely to use contraceptives compared with women who had no children [OR (95% CI): 6.782 (4.489-10.246), p =0.000 and OR (95% CI): 10.277 (6.246-16.909), p =0.000, respectively]. Women who wanted no more children were more likely to use contraceptive than their counterparts who desired for more children [OR (95% CI): 2.504 (1.871-3.351)] (Table 4).

Among unemployed women aged 35-49 years were less likely to use contraceptives than that of youngest counterpart [OR (95% CI): 0.648 (0.572-0.733)]. Women with higher level of education were found 1.469 (95% CI: 1.201-1.797, p =0.000) times more willing to use contraception compared with their illiterate counterparts. Women in Rangpur division were found 1.250 (95% CI 1.078-1.448; Table 4) times more likely and in Sylhet division were found 0.427 (95% CI: 0.368-0.494) times less likely to use contraceptives compared with women in Barisal division. Residing in the rural areas decreased the probability of using contraceptives compared with their urban counterparts [OR (95% CI): 0.784 (0.722-0.850), p =0.000]. Probability of using contraceptive was higher among non-Muslim than their Muslim counterparts [OR (95% CI): 1.439 (1.272-1.628)]. Women who had 1-2 children and who had more than two children were 5.335 (95% CI: 4.629-6.150) times and 7.176 (95% CI: 6.028-8.543) times respectively more likely to use contraceptive compared with women having no child. Hearing about FP on television during few months preceding the survey increased the probability of using contraceptive [OR (95% CI): 1.146 (1.051-1.250)]. Compared with women who wanted more children, the probabilities of contraceptive use increased for women who wanted no more children [OR (95% CI): 1.487 (1.341-1.649)] and decreased for women who were undecided about having children [OR (95% CI): 0.537 (0.399-0.721)] (Table 4).