Date of Award

Spring 2021

Degree Type

Open Access Dissertation

Degree Name

Public Health, DPH


School of Community and Global Health

Terms of Use & License Information

Terms of Use for work posted in Scholarship@Claremont.

Rights Information

© Copyright Daniel Woytowich, 2021. All rights reserved



Rates of unintended births (UIBs) are disproportionately high in low- and middle-income countries (LMICs) where the capacity to provide care to unexpected mothers and their offspring is often lagging. To decrease the prevalence of UIBs and their negative impacts on children, women, families, communities, and health systems of developing nations, global health stakeholders must understand the characteristics of a woman's life in these regions that increase her risk for UIBs. This project identified and analyzed predictors of UIBs in Sub-Saharan Africa (SSA) and South-East Asia among all women of reproductive age. It built on findings from previous studies while also testing novel determinants in predictive models. The overall goal was to add to the conceptual understanding of sociodemographic, interpersonal, and family dynamic situations that predispose a woman to UIBs while factoring out overly specific regional influences. This can guide future research and inform public health practice in regions where comprehensive and context-specific studies on UIBs have not yet been done.


Nationally representative Demographic and Health Survey datasets from 27 LMICs across Africa and South-East Asia were appended. Weighted prevalence and 95% confidence intervals (CIs) were calculated while a Rao-Scott design-adjusted Chi-square test with second-order

correction estimated bivariate associations between predictors and UIBs. Multivariate logistic regression models were used to predict odds of UIBs across three blocks of predictor variables. The first block produced unadjusted odds ratios by treating country of residence as the only predictor. The second block added sociodemographic and sexual and reproductive health (SRH) variables, while the third added variables about the woman's partner, family power dynamics, and intimate partner violence. The regression analyses produced adjusted odds ratios (AORs), accompanying 95% CIs, and p-values for each predictor.


The final sample (n=380,577) had an UIB prevalence of 19.4% (CI = 19.2 – 19.6). Model 3 showed the highest odds of UIBs among women from Lesotho (AOR = 11.13, CI = 8.54 – 14.51), as compared to all other countries; Africa (AOR = 2.62, CI = 2.09 – 3.29) as opposed to South-East Asia; and fragile regions (AOR = 1.44, CI = 1.30 – 1.59) compared to non-fragile regions. Also with the highest odds of UIBs were women aged 15-20 years (AOR = 1.65, CI = 1.40 – 1.94); women who were never married (AOR = 1.82, CI = 1.61 – 2.05), compared to those currently and formerly married; those with a primary education (AOR = 1.59, CI = 1.18 – 2.16); women with a parity of nine or more (AOR = 5.54, CI = 4.37 – 7.03), compared to women with parities of


Although not statistically significant, relatively low odds of UIBs were observed in women with low SES, no education, without knowledge of modern contraceptives, and whose partners had no education. These findings may indicate that decreased levels of empowerment lead to a lack of FP or women feeling unable to classify births as unintended. Governments and donors associated with Lesotho, Malawi, Namibia, and South Africa are encouraged to increase efforts towards FP outreach and the prevention of UIBs. Stakeholders must pay special attention to UIBs in fragile settings and SSA since these regions had significantly higher odds of UIBs compared to non- fragile regions and South-East Asia, respectively. Women 20 years of age and younger; women not currently married; women married at age nine or younger; women with high parity; women who have their healthcare choices made for them by a family member; and women who had sex forced on them are at significantly higher risk of UIBs. Therefore, SRH practitioners are urged to focus FP programming on these subgroups of women when comprehensive and context-specific studies from which they can inform their practice are not available. Lastly, since several of the sociodemographic and SRH associations with UIBs observed in Model 2 lost statistical significance after adding partner and interpersonal covariates in Model 3, it is important for researchers and survey implementers to take indicators reflective of family dynamics into account in subsequent analyses on UIBs.