Postpartum fertility desire in Uganda: does child sex composition matter? A cross-sectional analysis of the 2022 Uganda demographic and health survey

¹College of Business and Management Sciences (CoBAMS), School of Statistics and Planning, Department of Population Studies, Makerere University, Kampala, Uganda, ²Union for African Population Studies,University of Ghana, Legon, Accra, Ghana ³College of Business and Management Sciences (CoBAMS), School of Statistics and Planning, Department of Statistical Methods and Actuarial Sciences, Makerere University, Kampala, Uganda

^&Corresponding author
Patricia Ndugga, College of Business and Management Sciences (CoBAMS), School of Statistics and Planning, Department of Population Studies, Makerere University, Kampala, Uganda

Introduction    

Sub-Saharan Africa (SSA) continues to have the highest fertility levels globally, with women reporting an average desired family size of about five children compared with approximately three in most other regions [1]. High fertility in the region has been attributed to persistent preferences for large families, low contraceptive use, and, in some settings, preferences for children of a particular sex [2,3]. In Uganda, the total fertility rate is 4.5 children per woman, while women report an ideal family size of about five children, suggesting that preferences for relatively large families remain common [4]. Fertility preferences are influenced by several factors, including socioeconomic circumstances, reproductive experiences, partner preferences, and cultural norms [5]. For this study, postpartum women were defined as women who had a live birth within the two years preceding the survey. This extended postpartum period is a critical stage during which women reassess their fertility desires while balancing maternal recovery, breastfeeding, and infant care [6-8]. The World Health Organization recommends an interval of at least two years before attempting another pregnancy to reduce the risk of adverse maternal and child health outcomes [9]. Despite these recommendations, closely spaced pregnancies remain common in Uganda and other low- and middle-income countries, highlighting the need to understand fertility preferences during this period [10-12].

Among the factors that may influence fertility desire is the sex composition of existing children. Previous studies suggest that couples may continue childbearing when their existing children are all boys or all girls, in pursuit of a preferred combination of sons and daughters [5,13,14]. However, whether child sex composition independently influences fertility desire among postpartum women, after accounting for women's reproductive circumstances, remains unclear. This distinction is important because fertility desire during the postpartum period may reflect not only preferences regarding child sex but also current reproductive circumstances, including achieved family size, maternal age, previous pregnancy experiences, and time since childbirth. Determining whether child sex composition independently influences fertility desire is particularly relevant during this period, which provides an important opportunity to offer family planning services tailored to women's reproductive intentions.This study was guided by the Traits-Desires-Intentions-Behaviour (TDIB) framework [15,16], which proposes that fertility desires emanate from underlying childbearing motivations but are modified by women's current reproductive circumstances. In this study, these circumstances include the number of living children, sex composition of living children, pregnancy intention, and postpartum characteristics that may influence fertility desire during the extended postpartum period. Although factors associated with fertility desire have been widely studied among women of reproductive age, evidence focusing specifically on postpartum women remains limited, particularly in high-fertility settings such as Uganda. Moreover, little is known about whether the sex composition of living children independently influences fertility desire during the postpartum period after accounting for other reproductive characteristics. This study examined whether child sex composition was independently associated with fertility desire among postpartum women in Uganda after accounting for sociodemographic, reproductive, and postpartum characteristics.

Methods     

Data source and study design: this study employed a cross-sectional design using data from the 2022 Uganda Demographic and Health Survey (UDHS). The UDHS used a two-stage stratified cluster sampling design. In the first stage, clusters were selected using probability proportional to size, while in the second stage, households were systematically selected from the sampled clusters. The present analysis used data from the Women's Individual Recode (IR) file.

Study population: of the 18,251 women interviewed in the 2022 UDHS, the analysis included a weighted sample of 3,659 women aged 15-49 years who had a live birth within the two years preceding the survey. Women who were infecund or sterilized, had fewer than two living children, or were unsure or undecided about future childbearing were excluded. Women who were sterilized or infecund were excluded because they are biologically unable to have another child. Similarly, women who were unsure (2.8%) or undecided (9.3%) about future childbearing were excluded because their fertility intentions could not be clearly classified. Women with fewer than two living children were excluded because classification of the primary exposure, sex composition of living children, requires at least two children to distinguish between those with children of both sexes and those whose children were all of the same sex. Individual sample weights (v005) were applied to account for unequal probabilities of selection and nonresponse. The survey design was specified using the primary sampling units and sampling strata to account for the multistage, stratified cluster sampling design. All analyses incorporated sampling weights, clustering, and stratification using stata survey (svy) procedures.

Variables and measures

Outcome variable: the outcome variable was fertility desire. Fertility desire was measured using women's responses to the DHS question on whether they wanted another child and, among those who wanted another child, when they wanted the next child. In the DHS, fertility desire was described as a nominal variable with seven options: wants within 2 years, wants after 2 years, wants-unsure timing, undecided, wants no more, sterilized (respondent or partner), and declared infertile. Responses were categorized into three mutually exclusive groups: wanting another child within two years, wanting another child after two years, and wanting no more children (Figure 1). Women who were "undecidedʺ, "unsureʺ, "sterilized,ʺ or "infecundʺ were excluded from the analysis as they did not specify their preferred fertility desire or were biologically unable to have any more children.

Explanatory variables: variables were selected based on theoretical relevance and previous literature. Socioeconomic and demographic variables: age was categorized into five groups (15-24, 25-29, 30-34, 35-39, 40-49), educational attainment (highest level of education attainment: 0) None, i) Primary, ii) Secondary/higher), religious affiliation grouped into; i) Protestant, ii) Roman Catholic, iii) Muslim, iv) Others, marital status categorized as: i) Never married, ii) Currently married, iii) Formerly married), place of residence i) Urban and ii) Rural), and employment status (grouped into i) working, ii) Not working. Fertility and postpartum factors: number of living children, sex composition of children, time since birth, postpartum abstinence, and whether the last child was intended. Intention for the most recent pregnancy was categorized as wanted, wanted later (i.e., the pregnancy occurred earlier than desired), or unwanted (i.e., no more children were desired). Sex composition of living children was derived from the sex composition of all living children. The resulting variable classifies women into three categories based on the sex composition of their children: i) Has mixed sex composition of children, ii) Has only boys, and iii) Has only girls. Postpartum abstinence status was defined as whether a woman had resumed sexual intercourse following the birth of her most recent child. The variable was categorized as: i) Sexually active (had resumed sexual intercourse) and ii) Postpartum abstinent (had not resumed sexual intercourse).

Data management and analysis: descriptive statistics were performed, and the results were summarized using frequencies and percentages. A multinomial logistic regression model was fitted using Stata's svy: mlogit procedure to identify factors associated with wanting a child soon and wanting no more children. Relative risk ratios (RRRs) were computed to analyse the likelihood of wanting a child soon and wanting no more in comparison with wanting after two years, which was specified as the reference outcome. The 'wants after two years' category was used as the reference because it represented the largest outcome category and provided a meaningful comparison with women preferring either an earlier or no further birth. Statistical significance was determined using a p-value threshold of 0.05. Analysis was performed using STATA version 17. Multicollinearity was assessed using variance inflation factors (VIFs). All VIF values were below 5, indicating no evidence of problematic multicollinearity.

Ethics declaration: the study was conducted in accordance with the ethical standards of the Uganda Bureau of Statistics. Written informed consent was obtained from all participants before enrolment. The study was conducted in accordance with the ethical principles of the Declaration of Helsinki.

Results     

Characteristics of study participants: a total of 3,659 postpartum women with at least two living children were included in the analysis. Most women had primary education (60.2%), were married (89.4%), resided in rural areas (71.9%), and were employed (68.8%), as shown in Table 1. Table 2 shows that approximately three-quarters (76.4%) of the women had children of mixed sex, while 11.3% had only sons and 12.3% had only daughters. Nearly half (48.8%) were 12-23 months postpartum, and most (79.6%) had resumed sexual activity. Overall, 62.6% of women desired another child (5.3% wanted another child within 2 years, 57.3% wanted another child after 2 years).

Factors associated with fertility desire: multinomial logistic regression was used to analyse the factors associated with fertility desire among postpartum women, with wanting another child after two years as the reference category (Table 3). Overall, the model was statistically significant (F = 20.29, p < 0.001), indicating that the included covariates jointly explain variation in fertility desire. Variables considered potential confounders based on the literature and the conceptual framework were included in the multivariable model.

Child sex composition and fertility desire: at the bivariate level, child sex composition was significantly associated with fertility desire (χ2 = 30.42, p < 0.001) (Table 2). Women with mixed-sex children were more likely to report wanting no more children (42.9%) than those with only boys (23.3%) or only girls (16.0%). Conversely, women with only boys or only girls were more likely to want another child after two years than women with mixed-sex children (72.0% and 77.6% vs. 51.8%, respectively). However, after adjustment for sociodemographic, reproductive, and postpartum characteristics, child sex composition was not independently associated with fertility desire. Compared with women with mixed-sex children, those with only boys had a lower relative risk of wanting another child within two years rather than after two years (RRR = 0.69, 95% CI: 0.40-1.21), while those with only girls had a similar but non-significant pattern (RRR = 0.82, 95% CI: 0.47-1.42). For wanting no more children rather than another child after two years, the RRR was 1.23 (95% CI: 0.86-1.76) among women with only boys and 0.73 (95% CI: 0.51-1.05) among women with only girls. None of these associations reached statistical significance.

Other factors associated with fertility desire: the results showed that age significantly influenced "wanting no more childrenʺ, and not "wanting another child within two yearsʺ. Compared with women aged 15-24 years, the likelihood of wanting no more children was increased for women aged 30-34 years (RRR = 2.74), women aged 35-39 years (RRR = 5.13), and women aged 40-49 (RRR = 14.01). Compared with women with no formal education, those with Primary and secondary+ education were less likely to want another child within two years compared with the reference category (RRR= 0.50 and RRR=0.42, respectively). Married women were less likely than unmarried women to report wanting no more children rather than wanting another child after two years (RRR = 0.36). Compared with women who were 0-5 months postpartum, women who were 6-11 months and 12-23 postpartum were more likely to desire another child within two years rather than after two years (RRR = 2.10 and RRR=3.54, respectively). Table 3 further indicates that number of living children also influenced fertility desire. Compared with women with two living children, those with four living children were more likely to want no more children rather than another child after two years (RRR=3.55, 95% CI: 2.49-5.08), while those with five or more living children had even higher relative risk ratios (RRR=5.01, 95% CI: 3.37-7.44). In comparison with women whose last child was wanted at the time of pregnancy, women whose last pregnancy was unwanted were more likely to report not wanting any more children (RRR = 12.92).

Discussion     

The main finding of this study was that child sex composition was associated with fertility desire at the bivariate level but was not independently associated after adjustment for sociodemographic, reproductive, and postpartum characteristics. Women with only boys or only girls differed from those with mixed-sex children in their unadjusted fertility desires; however, these differences were no longer statistically significant after accounting for factors including maternal age, number of living children, education, marital status, time since childbirth, and previous pregnancy intention. These findings suggest that, among postpartum women with at least two living children, current reproductive circumstances may play a greater role in shaping fertility desire than the sex composition of existing children. One possible explanation of the absence of an independent association is that, in a context of declining fertility, changing gender norms, and increasing exposure to education and reproductive health information, achieved family size may increasingly outweigh child sex composition in fertility decision-making. Nevertheless, because the UDHS does not directly measure gender norms or explicit son preference, these explanations should be interpreted cautiously and warrant further qualitative investigation. Similar findings have been reported in multi-country analyses from sub-Saharan Africa and elsewhere, where parity consistently exerts a stronger influence on fertility preferences than child sex composition [17-20].

Although most women desired another child, the majority preferred to delay the next birth by at least two years, consistent with WHO recommendations on healthy birth spacing [21]. Similar findings have been reported in Uganda, Malawi, and Ethiopia, where postpartum women generally desire additional children but prefer to postpone pregnancy for at least two years because of breastfeeding, maternal recovery, and infant care responsibilities [22-24]. Older women, particularly those aged 40-49 years, were more likely to report wanting no more children. This finding is consistent with studies from Uganda, Tanzania, and Ethiopia showing that fertility desires shift towards limiting births as women attain their desired family size and approach the end of their reproductive years [17,19,25]. Older women may also perceive greater health risks associated with additional pregnancies, reinforcing the preference to discontinue childbearing. From a TDIB perspective, increasing age may reflect both accumulated reproductive experience and progression through the reproductive life course, making women more likely to transition from desiring additional children to limiting childbearing.

Education was associated with postponing rather than wanting another child soon. Women with primary or secondary+ education were less likely to desire another child within two years than after two years, consistent with evidence from Uganda, Kenya, and Ethiopia demonstrating that education enhances reproductive autonomy, knowledge of healthy birth spacing, and utilization of family planning services [18,20]. Education may therefore modify fertility traits by improving reproductive autonomy and knowledge, enabling women to express fertility desires that favour healthy birth spacing. Married women were less likely than unmarried women to report wanting no more children, which may reflect greater expectation for continued childbearing within unions. Marriage often provides a socially and economically supportive environment for continued childbearing, and married women may anticipate additional births to achieve shared family goals with their partners. Similar patterns have been documented in Uganda and Ethiopia, where stable unions are associated with stronger desire to continue childbearing [19, 26,27].

Women who were 6-23 months postpartum were more likely to desire another child within two years than women in the first six months postpartum. These findings illustrate how fertility desires evolve as women's postpartum circumstances change. As recovery progresses and childcare demands lessen, women become more willing to consider another pregnancy, demonstrating that fertility desires are dynamic rather than fixed. This reinforces the importance of providing continuous postpartum family planning counselling throughout the extended postpartum period rather than limiting counselling to the immediate postpartum period. Similar trends have been observed in Ethiopia and Kenya and support continuous postpartum family planning counselling throughout the extended postpartum period rather than focusing solely on the immediate post-delivery period [28,29]. As expected, number of living children was one of the strongest factors associated with fertility desire. Compared with women with two living children, those with four and five or more living children were more likely to report wanting no more children (RRR = 3.55, 95% CI: 2.49-5.08; and RRR = 5.01, 95% CI: 3.37-7.44, respectively). This finding may reflect the well-established inverse relationship between number of living children and the desire for additional children and suggests that women increasingly prefer to limit childbearing as they approach or exceed their desired family size. Similar findings have been reported in Uganda [30], Ethiopia [18] and elsewhere [26], where higher parity consistently predicts fertility limitation. This finding reinforces the TDIB proposition that current reproductive circumstances, particularly achieved family size, play a central role in shaping fertility desire.

Women whose most recent pregnancy was unwanted were more likely to report no desire for additional children, while those with pregnancies that were wanted later preferred to delay subsequent births. Women who experience an unwanted pregnancy may perceive that they have already achieved or exceeded their desired family size and therefore become more motivated to avoid future pregnancies. Comparable findings have been reported from Senegal [31], highlighting unintended pregnancy as an important predictor of future fertility limitation. Previous pregnancy experiences therefore appear to modify subsequent fertility desires, indicating that women continuously reassess their childbearing goals in response to changing reproductive circumstances. Overall, these findings can be interpreted within the TDIB framework by demonstrating that postpartum fertility desires are shaped by the interaction between underlying traits and women's current reproductive circumstances. While characteristics such as education influence reproductive motivations, immediate reproductive experiences including number of living children, pregnancy intention, age, marital status, and time since childbirth were the strongest factors associated with fertility desire. Notably, child sex composition, conceptualized as an underlying motivational trait, did not independently predict fertility desire once reproductive circumstances were considered.

Limitation: first, the cross-sectional design prevents causal inference between the identified factors and fertility desire. Second, fertility desire was self-reported and may have been influenced by recall or social desirability bias. Third, child sex composition was used as a proxy for potential child-sex preferences; however, the UDHS does not directly measure women's or couples' preferences for sons or daughters. Consequently, the absence of an association should not be interpreted as evidence that gender preferences are absent in Uganda. Fourth, the survey did not capture important contextual factors such as cultural norms and quality of postpartum family planning counselling, which may also influence women's fertility desire. Finally, the analysis was restricted to women with at least two living children because the primary exposure, sex composition of living children, could not be meaningfully assessed among women with fewer children. Consequently, the findings may not be generalizable to women with only one living child, among whom fertility desires and the influence of child sex may differ.

Conclusion     

Although most postpartum women desired additional children, the majority preferred to delay the next birth for at least two years. While child sex composition was associated with fertility desire at the bivariate level, this association disappeared after adjustment for reproductive and sociodemographic characteristics. Among postpartum women with at least two living children, fertility desire was more strongly associated with current reproductive circumstances-including age, parity, previous pregnancy intention, education, marital status, and time since childbirth-than with child sex composition. These findings have important implications for postpartum family planning programmes and policy in Uganda. Counselling and service delivery should prioritize women's reproductive intentions, parity, previous pregnancy experiences, and stage in the postpartum period rather than assuming that child sex preferences are a major determinant of fertility decisions. Strengthening continuous, woman-centred postpartum family planning services throughout the first two years after childbirth could better support women to achieve their desired birth spacing and family size. Future research should employ longitudinal and qualitative approaches to examine how fertility desires evolve over the postpartum period and to explore the influence of gender norms, partner preferences, and household decision-making on fertility intentions. Such evidence would provide a deeper understanding of the social and cultural factors that shape reproductive decision-making and inform interventions aimed at supporting informed and voluntary fertility choices.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

Patricia Ndugga: conceptualized the study, curated the data, conducted the formal analysis, led the investigation, and managed the project. Nyovani Janet Madise: contributed to the study conceptualization, supervised the formal analysis, and provided guidance throughout the investigation. Benard Odur and Betty Kwagala contributed to data curation, methodology development, and the formal analysis. All authors critically reviewed the manuscript, and agreed to be accountable for all aspects of the work. All the authors have read and agreed to the final manuscript.

Acknowledgments     

Our gratitude goes to Uganda Bureau of Statistics (UBOS) that provided us with the UDHS 2022 dataset that was used in this study. Further gratitude goes to Dr. Frank Namugera and Mr. Susan Habert Sendege of the School of Statistics and Planning, Makerere University, for their statistical support and for critically reviewing the statistical aspects of the manuscript.

Tables and figure     

Table 1: sociodemographic characteristics of postpartum women and their distribution by fertility desire, Uganda, UDHS 2022 (N=3,659)

Table 2: postpartum characteristics of postpartum women and their distribution by fertility desire, Uganda, UDHS 2022 (N=3,659)

Table 3: a multinomial regression of factors associated with fertility desire among postpartum women, Uganda, UDHS 2022

Figure 1: conceptual framework adapted from the traits-desires-intentions-behaviour (TDIB) model

References