WOMEN’S AUTONOMY AND REPRODUCTIVE AND PERINATAL HEALTH OUTCOMES IN AFGHANISTAN

Abstract

Studies show that women’s autonomy (WA)--as measured by household decision making, healthcare decision making, and permission to go out--facilitates access to resources, including education, employment, and healthcare resources. In return, these resources determine maternal and child health. Prior studies, conducted in countries other than Afghanistan, have resulted in inconsistent findings because the degree and direction of the association between WA and health outcomes differ among countries based on their laws and cultural context. There is no research examining the impact of WA on reproductive and perinatal health outcomes in Afghanistan. Conducting this research in Afghanistan is critical given the country’s unique political and socio-cultural environment that shape its population health. For instance, less than one in two of Afghan women make decisions about their healthcare, while 53% of them have experienced physical domestic violence. The 2021 ruling of the country by the Taliban has exacerbated the violation of women’s rights in Afghanistan, making the examination of the impact of WA on reproductive health outcomes even more pertinent. My dissertation addresses these gaps in the literature by examining the association between women’s autonomy and 1. experience of domestic violence types (physical, sextual, and emotional), 2. unintended pregnancy, and 3. pregnancy loss. I also assess the moderating effect of education on the association of WA and domestic violence, and the mediating effect of domestic violence on the association of WA and unintended pregnancy and pregnancy loss. I used data from 19,098 married women aged 15-49, who completed the 2015 Afghanistan Demographic and Health Survey- the first and only national survey administered in the country. WA was measured across 5 domains (healthcare, visiting family, household purchases, spending, and contraceptive use). Adjusted odds ratios and 95% confidence intervals for the association between WA and the outcomes of interest were estimated using multiple logistic regression and adjusted for relevant confounders including age, ethnicity, education, wealth, residency, and parity. Additionally, to assess moderation and mediation, interaction terms and casual mediation models were used for each respective analysis. I found that about one in two Afghan women did not have autonomy in making decisions and experienced domestic violence, and approximately one in ten women experienced unintended pregnancy and pregnancy loss. In Aim 1, in the adjusted Model 2 (adjusting for confounders), women’s autonomy in healthcare decisions (Adjusted odds ratio [AOR]=0.70, CI: 0.60-0.81), spending (AOR=0.58, CI: 0.51-0.66), visiting families (AOR=0.69, CI: 0.60-0.80), household purchases (AOR=0.59, CI: 0.52-0.68), and not using contraception (AOR=0.66, CI: 0.46-0.93) were significantly associated with decreased experience of physical violence. In addition, women’s autonomy in healthcare decisions (AOR = 0.51, 95% CI: 0.39-0.65), spending (AOR=0.62, CI: 0.48-0.80), and household purchases (AOR=0.56, CI: 0.43-0.72) were significantly associated with decreased experience of sexual violence. Lastly, women’s autonomy in healthcare (AOR=0.82, CI: 0.72-0.94), spending (AOR=0.61, CI: 0.53-0.71), visiting families (AOR=0.79, CI: 0.70-0.88), and not using contraception (AOR=0.58, CI: 0.42-0.80) were significantly associated with decreased experience of emotional violence. We also found a greater protective effect of WA in visiting family among women with some education (vs. no education) across each domestic violence outcome. In Aim 2, Model 2, women’s autonomy in healthcare decisions (AOR=0.87, CI: 0.77 - 0.97) and spending (AOR=0.86, CI: 0.76 - 0.97) were significantly associated with decreased experience of unintended pregnancy. However, women’s autonomy in visiting families (AOR=1.15, CI: 1.02 - 1.29) was associated with slightly increased experience of unintended pregnancy. This association was partially mediated by physical and sexual domestic violence (21% of total effect mediated by domestic violence) in that the reduction of violence with WA attenuated the effect of WA on unintended pregnancy. Lastly, in Aim 3, Model 2, women’s autonomy in healthcare decisions (AOR=0.86, CI: 0.78 - 0.95), household purchases (AOR=0.87, CI: 0.79 - 0.97), and visiting families (AOR=0.90, CI: 0.82 - 1.00), were significantly associated with decreased experience of pregnancy loss. Some of the protective effects of autonomy in healthcare (18%) and household purchase (15%) on pregnancy loss were mediated through reduced experience of domestic violence during pregnancy. Beyond filling the gap in the literature, this study provides evidence and brings awareness (about the impact of the lack of women’s autonomy on adverse health outcomes) that is needed to urge policymakers and program implementers in targeting and improving women’s autonomy and health outcomes in Afghanistan. The findings of this study can also be translated into Dari and Pashto and be shared with the general public in Afghanistan. Interventions that may improve women’s autonomy and health outcomes in Afghanistan, such as establishing laws about women’s rights, granting access to education, awareness campaigns, microfinancing, and community-based distribution programs, could foster gender equity and improve maternal and child health – moving to more sustainable development, consistent with Sustainable Development Goal (SDG) 3 and 5.