https://orcid.org/0000-0002-7708-6370
Figures
Abstract
Background
Abortion remains a significant public health concern in Ethiopia, contributing to maternal morbidity and mortality. Understanding the individual- and community-level determinants of abortion is essential to inform targeted interventions. This study aimed to assess the prevalence and multilevel determinants of abortion among women of reproductive age in Hawela Lida District, Southern Ethiopia.
Methods
A community-based cross-sectional survey was conducted from February to March 2025 among 3,526 women of reproductive age who experienced a pregnancy within the 12 months preceding the survey. Data were collected using a structured, pretested questionnaire and analyzed with Stata v18. Multilevel mixed-effects logistic regression was employed to identify individual- and community-level determinants, adjusting for potential confounders. Adjusted odds ratios (AOR) with 95% confidence intervals (CI) were reported.
Results
The overall prevalence of abortion was 18.5% (95% CI: 15.8–24.8). At the individual level, women without formal education were more likely to experience abortion than those with formal education (AOR = 2.61; 95% CI: 1.36–4.54). Women with unplanned pregnancies (AOR = 2.72; 95% CI: 1.62–4.68) and those who were non-autonomous in decision-making (AOR = 2.21; 95% CI: 1.51–5.02) also had higher odds of abortion. At the community level, rural residence (AOR = 2.19; 95% CI: 1.65–4.23), greater distance to the nearest health facility (AOR = 2.34; 95% CI: 1.21–5.01), and low mass media exposure (AOR = 1.82; 95% CI: 1.23–4.12) were significantly associated with abortion.
Conclusion
Abortion in Hawela Lida District was associated with both individual and community factors, including women’s education, pregnancy planning, autonomy, place of residence, mass media exposure, and health facility accessibility. Integrated interventions that strengthen family planning, promote women’s empowerment, improve service accessibility, and provide accurate reproductive health information are urgently needed to reduce abortion prevalence and improve maternal health outcomes in Ethiopia.
Citation: Yoseph A, Mussie L, Belayneh M, Guillen-Grima F (2026) Addressing abortion through individual and community-level determinants: Evidence from Southern Ethiopia. PLoS One 21(6): e0349603. https://doi.org/10.1371/journal.pone.0349603
Editor: Susanne Grylka-Baeschlin, Zurich University of Applied Sciences: ZHAW Zurcher Hochschule fur Angewandte Wissenschaften, SWITZERLAND
Received: October 16, 2025; Accepted: May 2, 2026; Published: June 17, 2026
Copyright: © 2026 Yoseph et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: This study was supported by the Sidama Region President Office (Grant No. SPO/123/16 to AY). The funder had no role in the study design, data collection, analysis, interpretation of data, manuscript preparation, or the decision to submit the manuscript for publication.
Competing interests: NO authors have competing interests.
Introduction
Abortion remains a critical global public health concern, contributing significantly to maternal morbidity and mortality, particularly in low- and middle-income countries [1]. Although the global maternal mortality ratio has declined substantially over the past two decades, unsafe abortion continues to account for nearly 13% of maternal deaths worldwide [1,2]. Sub-Saharan Africa bears a disproportionate share of this burden, where restrictive abortion laws, limited access to safe services, and sociocultural barriers converge to increase women’s vulnerability [3]. Ethiopia, despite progressive policy reforms such as the 2005 revision of its abortion law, continues to face a high incidence of unsafe abortions, highlighting the persistence of systemic gaps in reproductive health service delivery [4,5].
Previous studies in Ethiopia and other parts of the world have identified a range of factors associated with abortion, spanning individual, household, and community levels [6,7]. At the individual level, education, economic status, and reproductive autonomy have consistently emerged as important determinants [8,9]. Unplanned pregnancies, often linked to limited contraceptive use and unmet family planning needs, remain a major driver of abortion globally [10]. At the community level, disparities between rural and urban settings, accessibility of health facilities, and exposure to mass media also shape reproductive decisions [11,12]. However, much of the existing evidence relies on single-level analyses, which fail to account for the hierarchical nature of determinants embedded within broader community structures [13]. This analytical limitation reduces the capacity to fully capture contextual influences on abortion practices.
In the Ethiopian context, research has largely concentrated on quantifying abortion prevalence and exploring proximate individual factors such as age, parity, and contraceptive use [14,15]. Far fewer studies have incorporated community-level determinants such as literacy, poverty, and geographic access to health services, despite their well-documented influence on maternal health outcomes [16]. Moreover, while the prevalence of abortion has been estimated in urban and facility-based settings, population-level evidence from rural districts remains scarce, leaving a significant gap in understanding the interplay between personal and structural determinants. This knowledge gap is particularly relevant in the Sidama Region, a predominantly rural area where maternal health outcomes continue to lag behind national averages.
Addressing abortion from both individual and community perspectives is crucial for designing effective interventions and policies. Understanding how community contexts -such as residence, access to infrastructure, and shared socio-economic conditions -interact with individual characteristics like education, autonomy, and pregnancy intention provides a more holistic view of abortion dynamics. Without such insights, policy responses risk remaining fragmented, addressing isolated determinants without acknowledging their broader socio-ecological underpinnings [17,18]. By employing multilevel modeling, researchers can disentangle these cross-cutting influences and provide stronger evidence to guide programming and policy at multiple levels.
The present study seeks to address these gaps by investigating the individual- and community-level determinants of abortion among women of reproductive age in the Northern Zone of Sidama Region, Southern Ethiopia. Specifically, it applies a multilevel mixed-effects logistic regression framework to disentangle the relative contributions of personal, household, and contextual factors to abortion outcomes. By generating robust evidence from a large, community-based sample, this study aims to inform integrated maternal health strategies that extend beyond individual-level risk factors to address the broader structural and societal conditions shaping women’s reproductive health.
Methods
Study setting
This study was conducted in Hawela Lida District, one of 30 districts in Ethiopia’s Sidama Region, located approximately 289 km south of Addis Ababa. The district comprises 11 rural and two urban kebeles (the smallest administrative units). In 2024, the estimated population was 131,848 across 24,281 households, with women of reproductive age (15–49 years) representing 24.3% of the population. Agriculture constitutes the primary livelihood, with major crops including enset (false banana), maize, coffee, khat, barley, haricot beans, sweet potatoes, and indigenous cabbage. Health infrastructure comprises 20 government health posts, four health centres, five private medium clinics, two NGO-run facilities, and six private pharmacies, staffed by 482 health professionals. Health posts, managed by Health Extension Workers (HEWs), provide essential maternal and reproductive health services, including health education, family planning, antenatal and postnatal care. Overall maternal health service coverage in the district is estimated at 70%.
Study design and period
A community-based cross-sectional survey was conducted between February 1 and March 30, 2025. The study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (S1 File).
Source and study population
The source population comprised all women of reproductive age residing in the district. The study population included women aged 15–49 years who experienced a pregnancy within the 12 months preceding the survey and had lived in the district for at least twelve months.
Eligibility criteria
Women were eligible if they had resided in Hawela Lida District for at least one year and reported a pregnancy that ended in abortion, stillbirth and live birth within the same period. This criterion ensured that participants had sufficient exposure to the local community context, allowing for an accurate assessment of both individual- and community-level determinants. Women who were critically ill, mentally incapacitated, or unable to provide informed consent were excluded.
Sample size determination
We computed the sample size using OpenEpi version 3.0. The initial sample size was estimated using a single population proportion formula, assuming a 15% prevalence of abortion from previous study [19], a 95% confidence level, and a 5% margin of error. This yielded a base sample size of 392. To account for the multistage cluster sampling design, we applied a design effect (DEFF) of 2.0. The design effect was considered appropriate to adjust for intra-cluster correlation within kebeles and to improve the representativeness of estimates under a clustered sampling framework. We also added a 10% non-response rate, which increased the minimum required sample size to 863.
In addition, because the study aimed to examine both individual- and community-level determinants using multilevel mixed-effects logistic regression, we further increased the sample size to ensure sufficient statistical power to detect contextual (cluster-level) effects. This step was necessary to improve precision in estimating both fixed and random effects in the hierarchical data structure. Finally, after adjusting for clustering, non-response, and multilevel analytical requirements, the final sample size was increased to 3,540 women, which was considered adequate to address both individual- and community-level determinants in the study.
Sampling procedure
A multistage sampling strategy was applied. First, Hawela Lida District was purposively selected due to its diverse rural-urban population, ease of supervising data collection, and administrative priority because of a reported high prevalence of abortion. In the second stage, 10 out of the district’s 13 kebeles were randomly selected. Within these kebeles, a household census identified all women who had delivered in the previous 12 months, forming the sampling frame. In the third stage, eligible households were systematically selected proportional to kebele size. If multiple eligible women were present in a household, one was randomly chosen. Households were classified as non-respondents after three unsuccessful contact attempts (Fig 1).
Study variables
Outcome variable.
The primary outcome of interest was abortion within the past 12 months, defined as a pregnancy that ended in termination – either induced or spontaneous – as reported by respondents. Women were classified as having experienced an abortion if they answered “yes” to the relevant question; otherwise, they were coded as “no.”
Individual-level predictors.
Several sociodemographic, reproductive, and behavioral factors were examined. Educational attainment was measured as the highest level of formal schooling completed and categorized as no education, primary, secondary, or higher education. Age was recorded in completed years and analyzed as both a continuous variable and grouped into conventional reproductive-age categories (15–19, 20–24, 25–29, 30–34, and 35–49 years). Marital status was classified as married/cohabiting versus not married (single, divorced, or widowed). Parity was defined as the total number of pregnancies reaching a viable gestational age (≥28 weeks), including both live births and stillbirths. This variable was treated as continuous for analysis and categorized for descriptive purposes (0, 1–2, ≥ 3). Wealth quintiles were derived using principal component analysis (PCA) of household assets. Variables with very low (<5%) or very high (>95%) prevalence, inadequate sampling adequacy (Kaiser-Meyer-Olkin <0.5), low communalities (<0.5), or complex loadings (>0.4 on multiple components) were excluded. The resulting factor scores were ranked into five quintiles: lowest, second, middle, fourth, and highest. Obstetric danger signs during pregnancy were measured using nine self-reported complications, including severe bleeding, convulsions, prolonged labor, high fever, and blurred vision. Each of the nine self-reported obstetric danger signs during pregnancy was coded as “1 = Yes” if experienced and “0 = No” if not. Responses of “I don’t know” was treated as “0 = No” for consistency.
Community-level predictors.
Respondents were nested within kebeles, the smallest administrative unit in Ethiopia, to account for contextual influences. Community-level variables were constructed by aggregating individual responses within each kebele. These included the proportion of women with formal education (community education) and the proportion of households below the poverty threshold (community poverty). Each kebele was assigned a unique identifier and modeled as a random intercept to capture unobserved heterogeneity. Community-level mass media use was calculated as the proportion of women in each kebele who reported regularly listening to the radio, watching television, or reading newspapers. kebeles where more than 50% of participants used at least one form of mass media were classified as “high,” and those with 50% or fewer as “low.”
Operationalization and coding.
All binary variables were coded as 1 (“yes”) and 0 (“no”). For community-level variables, aggregate proportions were calculated as continuous measures. Detailed variable definitions are provided in S2 File, Table 1.
Data collection procedures
Data were collected using a structured, interviewer-administered questionnaire adapted from previous reproductive health and abortion studies (S3 File). The tool was developed in English, translated into Sidaamu Afoo (Local language), and back-translated to ensure consistency. Data collectors and supervisors, all fluent in Sidaamu Afoo and trained in ethical and interview procedures, received two days of training. Pretesting on 5% of the sample in a neighboring district allowed refinement of the instrument. Data were collected face-to-face at participants’ homes using the Open Data Kit mobile application (S4 File). Daily checks ensured completeness and accuracy before secure archiving on the Kobo Toolbox server.
Statistical analysis
Descriptive analyses summarized categorical variables as frequencies and percentages and continuous variables as means ± standard deviation. Wealth index was derived using PCA; detailed procedures are provided in S2 File. Multi-level mixed-effects logistic regression models were fitted to identify individual- and community-level determinants of abortion, accounting for clustering at the kebele level. Four models were constructed: Model 0 (empty), Model 1 (individual-level determinants), Model 2 (community-level determinants), and Model 3 (combined determinants). Model selection was guided by Akaike’s Information Criterion (AIC), Bayesian Information Criterion (BIC), log-likelihood, intraclass correlation coefficients (ICC), and median odds ratios (MOR). Variables with p < 0.10 in bivariable analysis or those with established clinical and social relevance were included in the multivariable model. Interaction terms were used to assess effect modification, and multicollinearity was checked using variance inflation factors (VIF < 5). Associations were reported as adjusted odds ratios (AORs) with 95% confidence intervals, with statistical significance defined as 95% CI not crossing 1.
Ethical considerations
Ethical approval was obtained from the Institutional Review Board of the College of Medicine and Health Sciences, Hawassa University with Reference number: IRB/027/17. Additional permissions were obtained from the Sidama Regional Health Bureau, Hawela Lida District Health Office, and kebele administrations. Written informed consent was obtained from all participants after providing detailed information about the study’s objectives, procedures, potential risks, and benefits. No personal identifiers were collected, and data were stored on password-protected servers accessible only to the research team, ensuring confidentiality.
Results
Study population
A total of 3,526 mothers participated in the study (Table 1). The mean maternal age was 26.0 ± 4.6 years, ranging from 15 to 49 years. The majority were of Sidama ethnicity (94%) and identified as Protestant Christians (85%), with nearly all currently married (99%). While 70% of participants had completed at least primary education, only 16.7% attained secondary education or higher. Most women were housewives (86%), whereas the predominant occupation among husbands was farming (55%). Household size was typically small, with four in five families comprising five or fewer members. Approximately 40% of respondents reported regular exposure to mass media. The asset-based wealth index was relatively evenly distributed across quintiles.
Reproductive and obstetric characteristics
The mean age at first marriage and first pregnancy were 21.2 ± 3.1 and 22.4 ± 3.2 years, respectively (Table 2). Most women (69%) had experienced two to four pregnancies, and 18.5% reported at least one prior abortion. Previous stillbirth was rare (1.6%). Although 83.4% of pregnancies were planned, 9.5% of mothers reported experiencing at least one obstetric danger sign during pregnancy.
Prevalence of abortion
Among the 3,526 women surveyed, the overall prevalence of abortion was 18.5% (95% CI: 15.8–24.8) (Fig 2). This indicates that nearly one in five women of reproductive age in Hawela Lida District experienced an abortion within the preceding 12 months.
The bivariable analysis (Table 3) was used to screen candidate variables for inclusion in the multilevel models presented in Table 4.
Determinants of abortion among women of reproductive age
The multivariable analysis identified several individual- and community-level factors that were significantly associated with abortion among women of reproductive age in Hawela Lida District (Table 4).
At the individual level, women without formal education were more than twice as likely to experience abortion compared to those with formal education (AOR = 2.61; 95% CI: 1.36–4.54, p < 0.01). Similarly, women with unplanned pregnancies had higher odds of abortion than those with planned pregnancies (AOR = 2.72; 95% CI: 1.62–4.68, p < 0.01). Women who were non-autonomous in decision-making also faced increased odds of abortion compared to autonomous women (AOR = 2.21; 95% CI: 1.51–5.02, p < 0.05).
At the community level, women living in rural areas were more likely to experience abortion compared to those in urban areas (AOR = 2.19; 95% CI: 1.65–4.23, p < 0.05). Communities reporting greater distance-related barriers to the nearest health facility (“big problem”) were also associated with higher odds of abortion (AOR = 2.34; 95% CI: 1.21–5.01, p < 0.01). Furthermore, communities with low mass media exposure had increased odds of abortion compared to those with higher exposure (AOR = 1.82; 95% CI: 1.23–4.12, p < 0.05).
Multilevel model performance and between-kebele variability in abortion determinants
The multilevel logistic regression analysis demonstrated a significantly better fit to the data compared with the conventional logistic regression model (p < 0.001). The intraclass correlation coefficient revealed that 15.71% of the variance in abortion was explained by clustering at the kebele level. Notably, even after adjusting for both individual- and community-level characteristics, kebele-level variations remained statistically significant, contributing to 17.50% of the residual variability. The Median Odds Ratio (MOR) was estimated at 2.21, highlighting substantial heterogeneity between kebeles. Furthermore, the influence of women’s education on abortion outcomes varied significantly across kebeles (variance = 0.07; 95% CI: 0.03–1.32), highlighting important contextual differences.
Model fit statistics further supported the robustness of the final specification. Comparative assessment using the AIC, BIC, and log-likelihood values confirmed the superiority of the final multilevel model over simpler alternatives. Specifically, the empty model yielded AIC = 1042.32, BIC = 1066.23, and log-likelihood = –520.72, whereas the final model achieved AIC = 1001.50, BIC = 1014.66, and log-likelihood = –440.63. Collectively, these indices affirm the enhanced explanatory power and reliability of the multilevel framework in capturing the determinants of abortion across kebeles.
Discussion
Prevalence of abortion
The prevalence of abortion in the current study was 18.5% (95% CI: 15.8–24.8), indicating that nearly one in five women of reproductive age in Hawela Lida District experienced an abortion within the preceding 12 months. This estimate is notably higher than the national prevalence reported in the 2016 Ethiopian Demographic and Health Survey (EDHS), which documented approximately 10% of women having terminated a pregnancy within five years [19]. The observed discrepancy may reflect methodological differences, as the present study employed a community-based survey design with a 12-month recall period, minimizing recall bias and providing a more precise estimate.
Comparisons within Ethiopia indicate variability in abortion prevalence. For instance, studies conducted in Addis Ababa reported prevalence of 19.6% [20], closely mirroring our findings, whereas rural regions such as Amhara and Oromia exhibited lower prevalence, ranging from 12% to 15% [21,22]. Globally, the observed prevalence aligns with sub-Saharan African estimates, where abortion prevalence range from 14% to 22% [23,24]. Conversely, countries with more liberal abortion legislation and higher contraceptive coverage, such as South Africa and Ghana, report comparatively lower abortion prevalence, highlighting the role of accessible reproductive health services in preventing unintended pregnancies [25]. These patterns highlight that despite improvements in maternal health services, unmet contraceptive needs and gaps in women’s reproductive autonomy continue to contribute to high abortion rates in Ethiopia.
Individual-level determinants
Women’s education was significantly associated with abortion. Women without formal education exhibited more than twice the odds of abortion compared to their educated counterparts. This finding concurs with studies in Ethiopia [21,26] and Nigeria [27], where higher educational attainment enhances reproductive health literacy, contraceptive use, and informed decision-making. Conversely, in contexts such as Uganda [28] and Kenya [29], educated women may be more empowered to seek abortion in response to unplanned pregnancies, demonstrating that the education–abortion relationship is context-dependent, mediated by cultural norms, legal frameworks, and health system accessibility.
Pregnancy intention also emerged as a significant associated factor with abortion. Women who reported unplanned pregnancies had lower odds of abortion compared with those whose pregnancies were planned. This finding suggests that, in this setting, pregnancies that were initially unintended were not necessarily followed by termination and may instead have been continued due to social, cultural, or personal considerations. Similar patterns have been documented in previous studies conducted in Ethiopia [30–32] and rural Tanzania [32]. These findings highlight the complex relationship between pregnancy intention and abortion and underscore the importance of strengthening access to family planning information and modern contraceptive methods to reduce unintended pregnancies and support informed reproductive decision-making [23,30].
Women’s autonomy significantly associated with abortion outcomes. Non-autonomous women were more than two times more likely to experience abortion compared to autonomous women. While autonomy generally empowers women to prevent unintended pregnancies, in contexts where abortion services are accessible, autonomy may facilitate safe abortion when necessary. Similar patterns have been observed in Addis Ababa [20] and Ghana [33], suggesting that autonomy operates as both an increasing and decreasing factor depending on sociocultural norms and healthcare availability.
Community-level determinants
Community factors also associated with abortion outcomes. Rural residence was associated with higher odds of abortion, consistent with studies from Ethiopia [21,31,34] and Kenya [35], where rural women face geographic, infrastructural, and sociocultural barriers limiting access to reproductive health services [36,37]. Proximity to healthcare facilities was inversely associated with abortion; women perceiving minimal distance barriers had lower odds of abortion (AOR = 0.71; 95% CI: 0.47–0.89), corroborating studies from Ethiopia and Malawi [22,38]. Accessible facilities facilitate timely contraceptive counseling and antenatal care, reducing unintended pregnancies and unsafe abortions.
Lower community-level mass media exposure was associated with higher odds of abortion. Although mass media is generally expected to promote awareness of family planning and reproductive health services, limited exposure in this context may reflect reduced access to accurate reproductive health information, including knowledge about contraception and safe pregnancy prevention methods. As a result, women living in communities with low mass media exposure may face a higher risk of unintended pregnancies, which can subsequently increase the likelihood of abortion. Comparable findings have been reported in Nigeria [27] and Uganda [28], suggesting that mass media plays an important role in shaping reproductive health knowledge and decision-making among women.
It is important to recognize that abortions result from multiple causes, including biological factors such as chromosomal abnormalities, which primarily contribute to early spontaneous losses and are largely unaffected by social or community-level influences. However, in community-based populations like Hawela Lida District, many abortions are associated with preventable or socially mediated factors, including unintended pregnancies, limited reproductive autonomy, restricted access to healthcare, and sociocultural pressures. Our multilevel analysis highlights that community-level variables such as rural residence, proximity to health facilities, and exposure to mass media significantly associated with abortion outcomes, reflecting the role of social and environmental determinants in reproductive behavior. While biological factors remain important, these findings emphasize that a substantial portion of abortions in this setting can be mitigated through interventions targeting social, behavioral, and healthcare-related determinants. This distinction highlights the value of assessing both individual- and community-level factors when designing policies and programs to reduce abortion prevalence.
Methodological strengths and limitations
Key methodological strengths include the large sample size (N = 3,526), which provided sufficient statistical power to assess individual- and community-level determinants. The use of multilevel mixed-effects modeling enabled robust estimation by accounting for clustering effects and disentangling contextual influences from individual-level determinants [39,40]. Moreover, the 12-month recall period minimized recall bias, enhancing reliability.
However, limitations include potential underreporting due to stigma and social desirability bias, which could have led to conservative prevalence estimates. The cross-sectional design precludes causal inference, and community-level variables derived from aggregated individual responses may not fully capture broader structural determinants, such as facility capacity or legal enforcement. Qualitative insights into sociocultural factors influencing abortion decisions were also absent.
Policy and practice implications
The findings emphasize the need for comprehensive, multi-level interventions. Strengthening family planning services, particularly in rural areas and among less educated and non-autonomous women, is paramount. Infrastructure development, including improved road networks and expanded health facilities, can reduce geographic barriers. Mass media campaigns should be carefully tailored to promote contraceptive uptake while ensuring responsible reproductive decision-making. Policy efforts should continue to expand safe abortion services within legal frameworks, integrate post-abortion care into broader reproductive health programs, and promote women’s empowerment.
Conclusion
Abortion in Hawela Lida District was associated with both individual- and community-level determinants. Low education, unplanned pregnancy, limited autonomy, rural residence, restricted healthcare access, and low mass media exposure significantly increased abortion prevalence. Addressing these determinants through integrated, multilevel interventions is critical for reducing abortion prevalence and improving maternal health outcomes. Policies that enhance family planning, expand safe abortion services, and empower women are urgently needed in Ethiopia and similar settings.
Supporting information
S1 File. STROBE reporting guidelines for cluster RCTs checklist.
https://doi.org/10.1371/journal.pone.0349603.s001
(DOCX)
S2 File. Detail information from methods and results sections.
https://doi.org/10.1371/journal.pone.0349603.s002
(DOCX)
S3 File. English version study questionnaire.
https://doi.org/10.1371/journal.pone.0349603.s003
(DOCX)
S4 File. De-identified SPSS dataset, which was authorized to be available by the institutional ethics committee and supported by informed consent from all study participants (Sav).
https://doi.org/10.1371/journal.pone.0349603.s004
(XLSX)
Acknowledgments
We extend our sincere appreciation to the study participants, data collectors, field assistants, and supervisors whose commitment made this research possible. Finally, we thank the School of Public Health at Hawassa University for providing technical guidance during the study design and data analysis.
References
- 1.
World Health Organization. Abortion Key Facts. https://www.who.int/news-room/fact-sheets/detail/abortion
- 2. Say L, Chou D, Gemmill A, Tunçalp Ö, Moller A-B, Daniels J, et al. Global causes of maternal death: a WHO systematic analysis. Lancet Glob Health. 2014;2(6):e323-33. pmid:25103301
- 3. Bearak J, Popinchalk A, Ganatra B, Moller A-B, Tunçalp Ö, Beavin C, et al. Unintended pregnancy and abortion by income, region, and the legal status of abortion: estimates from a comprehensive model for 1990-2019. Lancet Glob Health. 2020;8(9):e1152–61. pmid:32710833
- 4. Bridgman-Packer D, Kidanemariam S. The implementation of safe abortion services in Ethiopia. Int J Gynaecol Obstet. 2018;143 Suppl 4:19–24. pmid:30374983
- 5. Gilano G, Hailegebreal S. Determinants of abortion among youth 15–24 in Ethiopia: A multilevel analysis based on EDHS 2016. PLoS One. 2021;16(3):e0248228.
- 6. Sedgh G, Bearak J, Singh S, Bankole A, Popinchalk A, Ganatra B, et al. Abortion incidence between 1990 and 2014: global, regional, and subregional levels and trends. The Lancet. 2016;388(10041):258–67.
- 7. Izugbara CO, Egesa C, Okelo R. “High profile health facilities can add to your trouble”: Women, stigma and un/safe abortion in Kenya. Soc Sci Med. 2015;141:9–18. pmid:26233296
- 8. Obsie GW. Quality of post abortion care and service satisfaction among women who received the post abortion care service in Dilla university referral hospital, SNNPR, Ethiopia. Age. 2020;15(20):6.
- 9. Belachew TB, Negash WD, Asmamaw DB, Fetene SM, Addis B, Baykeda TA, et al. Prevalence and predictors of post-abortion modern contraceptive utilization among reproductive age women in Ethiopia. Sci Rep. 2023;13(1):15932. pmid:37741903
- 10. Singh S, Maddow-Zimet I. Facility-based treatment for medical complications resulting from unsafe pregnancy termination in the developing world, 2012: a review of evidence from 26 countries. BJOG. 2016;123(9):1489–98. pmid:26287503
- 11. Mutua MM, Manderson L, Musenge E, Achia TNO. Policy, law and post-abortion care services in Kenya. PLoS One. 2018;13(9):e0204240. pmid:30240408
- 12. Afolabi AA, Taiwo A, Ademusire TI. Women’s knowledge and attitude towards unsafe abortion in the farming communities of Ile-Ife, Southwest Nigeria. African Journal of Reproductive Health. 2024;28.
- 13. Yoseph A, Teklesilasie W, Guillen-Grima F, Astatkie A. Individual- and community-level determinants of maternal health service utilization in southern Ethiopia: A multilevel analysis. Womens Health (Lond). 2023;19:17455057231218195. pmid:38126304
- 14. Tilahun A, Yoseph A, Dangisso MH. Utilization and predictors of long acting reversible contraceptive methods among reproductive age women in Hawassa city, South Ethiopia: a community based mixed methods. Contracept Reprod Med. 2020;5:9. pmid:32647585
- 15. Alemayehu B, Addissie A, Ayele W, Tiroro S, Woldeyohannes D. Magnitude and associated factors of repeat induced abortion among reproductive age group women who seeks abortion care services at Marie Stopes International Ethiopia clinics in Addis Ababa, Ethiopia. Reproductive Health. 2019;16(1):76.
- 16. Onambele L, Ortega-Leon W, Guillen-Aguinaga S, Forjaz MJ, Yoseph A, Guillen-Aguinaga L, et al. Maternal Mortality in Africa: Regional Trends (2000-2017). Int J Environ Res Public Health. 2022;19(20):13146. pmid:36293727
- 17.
Negash B, Yoseph A. Status of disrespectful and abusive maternity care among women who delivered in hospitals of Sidama zone, South Ethiopia. 2019;10.
- 18. O’Connell KA, Kebede AT, Menna BM, Woldetensay MT, Fischer SE, Samandari G, et al. Signs of a turning tide in social norms and attitudes toward abortion in Ethiopia: Findings from a qualitative study in four regions. Reprod Health. 2022;19(Suppl 1):198. pmid:35698231
- 19. Gebeyehu AA, Dessie AM, Anely DT, Zemene MA, Negesse Y, Desalegn W, et al. Delays in seeking abortion and its determinants among reproductive-age women based on the Ethiopian Demographic and health survey. Heliyon. 2024;10(19):e38477. pmid:39430523
- 20. Megersa BS, Ojengbede OA, Deckert A, Fawole OI. Factors associated with induced abortion among women of reproductive age attending selected health facilities in Addis Ababa, Ethiopia: a case control study. BMC Womens Health. 2020;20(1):188. pmid:32883263
- 21. Oumer M. Prevalence and Associated Factors of Induced Abortion Among Women of Reproductive Age Group in Gondar Town, Northwest Ethiopia. SJPH. 2019;7(3):66.
- 22. Waktola MI, Mekonen DG, Nigussie TS, Cherkose EA, Abate AT. Repeat induced abortion and associated factors among women seeking abortion care services at Debre Markos town health institutions, Amhara regional state, Ethiopia, 2017. BMC Research Notes. 2020;13(1):44.
- 23. Seifu BL, Tebeje TM, Asgedom YS, Asmare ZA, Asebe HA, Kase BF, et al. Individual and contextual factors associated with abortion among reproductive age women in sub-Saharan Africa: Insights from 24 recent demographic and health surveys. PLoS One. 2024;19(12):e0315262. pmid:39666672
- 24. Kassa RN, Kaburu EW, Andrew-Bassey U, Abdiwali SA, Nahayo B, Samuel N, et al. Factors associated with pregnancy termination in six sub-Saharan African countries. PLOS Glob Public Health. 2024;4(5):e0002280. pmid:38722942
- 25. Schwandt HM, Creanga AA, Danso KA, Adanu RMK, Agbenyega T, Hindin MJ. A comparison of women with induced abortion, spontaneous abortion and ectopic pregnancy in Ghana. Contraception. 2011;84(1):87–93. pmid:21664516
- 26. Woldeamanuel BT, Gessese GT, Demie TG, Handebo S, Biratu TD. Women’s education, contraception use, and high-risk fertility behavior: A cross-sectional analysis of the demographic and health survey in Ethiopia. Front Glob Womens Health. 2023;4:1071461. pmid:36937044
- 27. Bankole A, Rodríguez G, Westoff CF. Mass media messages and reproductive behaviour in Nigeria. J Biosoc Sci. 1996;28(2):227–39. pmid:8935878
- 28. Larsson S, Eliasson M, Klingberg Allvin M, Faxelid E, Atuyambe L, Fritzell S. The discourses on induced abortion in Ugandan daily newspapers: a discourse analysis. Reprod Health. 2015;12:58. pmid:26108479
- 29. Rehnström LU, Lindgren M, Faxelid E, Oguttu M, Klingberg-Allvin M. Decision-making preceding induced abortion: a qualitative study of women’s experiences in Kisumu, Kenya. Reproductive Health. 2018;15(1):166.
- 30. Worku S, Fantahun M. Unintended pregnancy and induced abortion in a town with accessible family planning services: The case of Harar in eastern Ethiopia. Ethiopian Journal of Health Development. 2007;20(2).
- 31. Tesema GA, Mekonnen TH, Teshale AB. Spatial distribution and determinants of abortion among reproductive age women in Ethiopia, evidence from Ethiopian Demographic and Health Survey 2016 data: Spatial and mixed-effect analysis. PLoS One. 2020;15(6):e0235382. pmid:32598398
- 32. Rasch V, Kipingili R. Unsafe abortion in urban and rural Tanzania: method, provider and consequences. Trop Med Int Health. 2009;14(9):1128–33. pmid:19573141
- 33. Rominski SD, Gupta M, Aborigo R, Adongo P, Engman C, Hodgson A, et al. Female autonomy and reported abortion-seeking in Ghana, West Africa. Int J Gynaecol Obstet. 2014;126(3):217–22. pmid:24920181
- 34. Tariku M. Magnitude of Severe Acute Maternal Morbidity and Associated Factors Related to Abortion: A Cross‐Sectional Study in Hawassa University Comprehensive Specialized Hospital, Ethiopia, 2019. BioMed Research International. 2020;2020(1):1781652.
- 35. Ushie BA, Juma K, Kimemia G, Ouedraogo R, Bangha M, Mutua M. Community perception of abortion, women who abort and abortifacients in Kisumu and Nairobi counties, Kenya. PLoS One. 2019;14(12):e0226120. pmid:31830102
- 36. Ewnetu DB, Thorsen VC, Solbakk JH, Magelssen M. Still a moral dilemma: how Ethiopian professionals providing abortion come to terms with conflicting norms and demands. BMC Med Ethics. 2020;21(1):16. pmid:32046695
- 37. D’Souza P, Phagdol T, D’Souza SRB, Anupama DS, Nayak BS, Velayudhan B, et al. Interventions to support contraceptive choice and use: a global systematic map of systematic reviews. Eur J Contracept Reprod Health Care. 2023;28(2):83–91. pmid:36802955
- 38. Coast E, Fetters T, Chiweshe MT, Vwalika B, Griffin R, Tembo L, et al. Adolescent abortion care trajectories and safety in Ethiopia, Malawi, and Zambia: A comparative mixed methods study. PLOS Glob Public Health. 2025;5(5):e0004469. pmid:40327635
- 39. Thaddeus S, Maine D. Too far to walk: maternal mortality in context. Newsl Womens Glob Netw Reprod Rights. 1991;(36):22–4. pmid:12284530
- 40. Hox JJ, Kreft IGG. Multilevel Analysis Methods. Sociological Methods & Research. 1994;22(3):283–99.