Menu
Menu
Menu
Menu
The current study investigated the trends and factors associated with the unmet need for family planning (FP) for limiting and spacing births among married Tanzanian women between 1999 and 2016. The study used Tanzania Demographic and Health Survey (TDHS) data for the years 1999 (N = 2653), 2004–2005 (N = 2950), 2010 (N = 6412), and 2015–2016 (N = 8210). Trends in the unmet need for FP were estimated over the study period. Multivariable multinomial logistic regression models were used to investigate the association between community-level, predisposing, enabling, and need factors with the unmet need for FP in Tanzania. The results showed no significant change in percentage of married women with an unmet need for birth spacing between 1999 and 2016. The proportion of married women with an unmet need for limiting births decreased from 9.5% (95% confidence interval (CI): 7.9%, 10.6%) in 1999 to 6.6% (95% CI: 5.9%, 7.3%) in 2016. Residing in a rural area, parity between 1–4 and 5+, visiting a health facility for any health services within twelve months, and planning to have more children (after two years and/or undecided) were factors positively associated with the unmet need for FP-spacing. Women with parity of 5+ were more likely to experience an unmet need for FP-limiting. Women’s age between 25–34 and 35–49 years, women’s employment status, watching television, women’s autonomy of not being involved in household decisions, and planning to have more children were factors associated with lower odds of having an unmet need for FP-spacing. Women’s age between 25–34 years, watching television, autonomy, and planning to have more children were factors with lower odds of having an unmet need for FP-limiting. Improving FP uptake among married Tanzanian women can reduce the unmet need for FP. Therefore, reducing unmet needs for FP is attainable if government policies and interventions can target women residing in rural areas and other modifiable risk factors, such as parity, health facility visits, planning to having more children, employment, watching television, and women’s autonomy.
The study used TDHS data from 1999 to 2016, 1999 (N = 2653), 2004/05 (N = 6950), 2010 (N = 6412), and 2015–16 (N = 8210). The National Bureau of Statistics (NBS), Office of the Chief Government Statistician (OCGS) in Zanzibar, and Inner-City Funds collected the data. The project was funded by the Government of United Republic of Tanzania, Global Affairs Canada, and the United States Agency for International Development (USAID) [16]. Data for maternal health, including FP, child health, infant nutrition, and other health-related data, were collected based on a nationally representative population in Tanzania [17,18,19,20]. The data were collected from eligible women aged 15–49 years who were married or cohabiting and were residents in the household 24 h prior to the survey using a two-stage stratified cluster sampling technique. In stage one, enumeration areas (EAs) were selected proportional to each geographical zone of Tanzania. The EAs were based on the 1988, 2002, and 2012 Tanzania Population and Housing Censuses, respectively [21,22,23]. In stage two, a systematic random sampling technique was used to select households after the complete household listing was conducted in each EAs. The response rates in the surveys were 98% in 1999, 97% in 2005–2004, 96% in 2010, and 97% in 2015–2016. The full methodological approaches used in the surveys are provided in the respective TDHS reports [17,18,19,20]. The present study was conducted based on a total weighted sample of 24,225 married and/or cohabiting women who were not using FP on the day of the interview but reported a willingness to use modern FP methods [24,25,26,27]. The United Republic of Tanzania has 31 regions (28 in mainland Tanzania and 3 in Tanzania Island) with a total of 184 districts. Geographically, Tanzania is divided into regions, districts, divisions, wards, and villages. Tanzania is the largest country in East Africa, covering a total of 947,300 km2. Based on 2022 National Census results, Tanzania has a total of 61,741,120 people, of whom 30,053,130 (48.8%) are men and 31,687,990 (51.3%) are women [28]. The 2015–2016 TDHS indicated that the total fertility rate was 5.2 [19] The study outcome was the unmet need for FP, measured as the proportion of women who: (1) are not pregnant or postpartum amenorrhoeic and are considered fecund and want to postpone their next birth for 2 or more years or stop childbearing altogether, but are not using a contraceptive method; (2) have a mistimed or unwanted current pregnancy; or (3) are postpartum amenorrhoeic and their last birth in the last 2 years was mistimed or unwanted [10]. The unmet need for FP was divided into: (i) unmet need for spacing births; and (ii) unmet need for limiting births (that is, women who want to space or limit births but are not currently using FP methods) [10]. Dividing the unmet need for FP into the need for spacing and limiting births provides additional and detailed information about issues related to non-use of FP methods among currently married women in Tanzania. In this study, we assumed that currently married women were sexually active, which was consistent with past studies [25,29,30,31,32,33]. The concept of unmet need for FP reflects the gap between the desire for childbearing and the use of FP methods. Using the Bradley et al. model [34], the unmet need for FP was computed. We categorized the unmet need for FP as “no unmet need or met need for FP (currently using FP),” “unmet need for spacing,” and “unmet need for limiting.” For this study, the exposure variables were selected based on previous studies from LMICs [24,25,26,27,35] and availability in the DHS data. The study factors were broadly classified as community-level, predisposing, enabling, and need factors based on the Anderson conceptual model of health service utilization and health-seeking behaviors [36]. The adopted conceptual model was consistent with past studies that examined the relationship between study factors and the unmet need for FP among married women [24,25,26,35,37,38]. Community-level factors included the place of residence, which was categorized as ‘rural’ or ‘urban’. Predisposing included sociodemographic and health knowledge factors. Sociodemographic factors included women’s age (grouped as ‘15–24 years,’ ‘25–34 years,’ or ‘35–49 years’), parity (grouped as ‘zero,’ ‘one to four,’ or ‘five or more’), women’s and husband’s education (grouped as ‘no schooling’ or ‘primary education or higher’), women’s and husband’s employment (grouped as ‘no employment,’ ‘formal employment,’ or ‘informal employment’), household wealth index (grouped as ‘poor,’ ‘middle,’ or ‘rich’), and the number of partners a woman has (grouped as ‘one’ or ‘more than one partner’). Household wealth index was calculated by the National Bureau of Statistics and Inner-City Funds using principal component analysis (PCA), considering the ownership of household assets such as toilets, electricity, television, radio, fridge, and bicycle, as well as the availability of the source of drinking water and the floor material used for the main house [39]. Health knowledge factors included listening to the radio, watching television, and reading newspapers/magazines (which were grouped as ‘yes’ or ‘no’), visit by health care workers within the last 12 months prior to the survey (grouped as ‘yes’ or ‘no’), and health facility visit within the last twelve months (grouped as ‘yes’ or ‘no’). Enabling factors included the distance to a health facility (grouped as ‘big problem’ or ‘not a big problem’ based on Measure DHS classification), women’s autonomy (grouped as ‘involved in household decision making’ or ‘not involved in household decision making’), household head (grouped as ‘male headed’ or ‘female headed’). The TDHS also collected information on reasons for not using FP methods, such as infrequent sex, sub-fecund, amenorrhea, breastfeeding, no need for more children, mother opposition, partner opposition, religion, not knowing FP methods, not knowing source of FP methods, fear of side effects, lack of access, too much cost, inconvenient to use FP methods, and interferes with body’s natural process (grouped as ‘yes’ or ‘no’). Need factors included planning to have more children (grouped as ‘do not want more,’ ‘want within two years,’ ‘want after two years and above or undecided’). All the study factors were grouped based on previous evidence from past studies from LMICs [24,25,26,27,31,38,40,41,42,43] (Figure 1). Conceptual model for unmet need for FP adopted from Anderson’s health service utilization model [36]. Our analytical approach was stepwise, based on previous studies conducted elsewhere [24,25,26,38,40,41,42,43]. The first step involved the estimation of frequencies and percentages of each study factor. Second, the prevalence of and trends in the unmet need for spacing and limiting births were calculated across the study factors (community-level, predisposing, enabling, and need factors) for each year from 1999 to 2016. Third, univariable logistic regression was conducted to establish associations between the study factors and the outcome variable of the unmet need for FP for spacing and limiting births. Fourth, multivariable logistic regression models were used to examine the associations between community-level, predisposing, enabling, and need factors with the outcome variable of the unmet need for FP. The multinomial regression model used in the analysis was based on the Anderson model [35] and was also based on other previous studies conducted in LMICs [24,25,26,31,35,37,38,40,41,42,43]. Community-level factor was entered into the stage 1 model to measure its association with the outcome variable, while adjusting for predisposing, enabling, and need factors. The same strategy was used for the predisposing factors (sociodemographic and health knowledge factors) model to establish the relationship with outcome variable, while adjusting for community-level, enabling, and need factors (stage 2). The corresponding modeling procedure was used for enabling factors and need factors in the third and fourth models (stage 3 and 4), respectively. In addition, the multivariable regression model was conducted for combining the TDHS data from 1999 to 2016: (i) to examine the trends and factors associated with the unmet need for spacing and limiting; (ii) to provide the exceptional opportunity to compare the unmet need for FP over time; (iii) to improve the statistical power of the study. Odds ratios (ORs) with 95% confidence intervals (CIs) were provided as the measure of association between the study variables and outcome variable. Analysis was performed using STATA version 14 software (Stata Corp, College Station, TX, USA), with the ‘svy’ command used to adjust for sampling weights, clustering, and stratification in both year-specific and combined datasets. The ‘mlogit’ function was used in the multinomial logistic regression models [44]. Furthermore, multi-collinearity was checked to examine for any influential variables related to each other using the ‘regress’ command [45] and reasons for not using FP were eliminated from the model due to high multi-collinearity.
N/A
