Introduction

Globally, pregnancy termination, which includes stillbirths, miscarriages, and abortions-is a common and emotionally challenging experience for millions of women of reproductive age (15–49 years)1,2,3. Although regional estimates vary, research suggests that a significant proportion of pregnancies, both intended and unintended, end in termination, with rates ranging between 10 and 20% of all pregnancies4. Less developed countries (LDCs), however, report an exceptionally high pregnancy termination rate of 49%5. This alarming statistic is compounded by limited access to and utilization of sexual and reproductive health (SRH) care services in this region6,7.

In LDCs, a combination of factors disproportionately leads to pregnancy terminations3, with can have serious negative consequences. These include severe bleeding, infections, and death, as well as other complications like uterine perforation, cervical lacerations, and an increased risk of future reproductive health issues, such as infertility and menstrual disorders8,9. Additionally, socioeconomic, cultural, and behavioural factors, like early marriage, teenage childbearing, and limited access to family planning, influence pregnancy patterns and increase the likelihood of pregnancy loss among sexually active married women in LDCs10,11,12,13,14. While healthcare facilities may be theoretically accessible, many LDCs struggle to provide equitable maternal healthcare15,16,17. Limited access to quality sexual and SRH care, including prenatal care and diagnostic services, hinders the early detection and management of pregnancy complications18. Furthermore, inadequate infrastructure, a shortage of qualified healthcare professionals, and negative staff attitudes towards pregnant women-especially those seeking late prenatal care-contribute to poor maternal outcomes19,20. These factors can create a hostile environment that discourages women from utilising SRH services, ultimately leading to adverse pregnancy outcomes, including pregnancy loss.

Importantly, addressing pregnancy termination in LDCs is vital for achieving multiple Sustainable Development Goals (SDGs) as outlined by the United Nations21. Firstly, it contributes to SDG 3: Good Health and Well-being by reducing maternal and newborn deaths through improved healthcare. Secondly, the subject is also central to SDG 5: Gender Equality, as it disproportionately affects women, highlighting a lack of autonomy and the influence of patriarchal norms. By focusing on these factors, the research will also provide evidence that empowers women and improves their health. Lastly, the study aligns with SDG 10: Reduced Inequalities by investigating disparities in access to care, with the goal of informing targeted interventions to ensure equitable health outcomes for marginalized women. Imporntaly, the study is key in contributing to global efforts to achieve the SDGs, especially Target 3.1, which aims to reduce the maternal mortality ratio to less than 70 per 100,000 live births by 2030 of which LDCs account for majority of it.

Notwithstanding the challenges, significant efforts have been made to address pregnancy loss trends LDCs. Interventions typically focus on improving maternal health through antenatal and postnatal care, and providing skilled healthcare access22,23,24,25,26,27. Furthermore, promoting healthy behaviours before and during pregnancy-including ensuring essential nutrient intake, encouraging balanced diets, and discouraging harmful practices to reduce complications, creating a conducive environment for wanted pregnancies that do not end up in terminations6,28. Despite these concerted efforts, progress remains limited29,30. While some regional studies have explored contextual factors at sub-regional levels within LDCs31,32,33, a comprehensive understanding of individual, household, and community factors contributing to this issue in LDCs as a whole is lacking.

From the foregoing, it is clear that pregnancy terminations are a result of various interlinked factors. Therefore, this study uses the Socio-Ecological Model (SEM) as proposed by34,35,36 as its theoretical framework, which is built on four key premises. The first premise, multiple levels of influence, posits that an individual’s behaviour is shaped by factors at different levels: individual (e.g., knowledge), interpersonal (e.g., family), community (e.g., environment), organisational (e.g., procedures), and policy (e.g., laws). The second premise, interactions and interdependencies, acknowledges that these levels are interconnected, and a change at one level can affect others. The third premise, contextual factors, recognises the social, cultural, economic, and physical context in which people live and interact. Finally, the fourth premise, dynamic and reciprocal relationships, highlights that the relationships between individuals and their environments are constantly changing and mutually influencing.

Based on the SEM, pregnancy terminations in LDCs are influenced by a wide range of factors37. These include individual-level factors such as a lack of knowledge and limited access to reproductive healthcare, alongside demographic and socioeconomic characteristics like age, education level, and working status38. At the interpersonal or household level, social norms, partner support, and family expectations play a significant role39. The community level is defined by the availability and quality of reproductive health services and cultural factors like stigma, taboos, and cultural beliefs surrounding pregnancy termination40. Lastly, organisational and policy-level factors, encompassing the availability of trained providers, clear policies on the legality of terminations, and resources to ensure access for vulnerable and marginalised women, plays an important role41.

In this study we focused on married women because a significant proportion of these women in LDCs rely on their husbands for decision-making, including healthcare decisions due to patriarchal societal norms and limited autonomy42,43. We believe that by focusing on these women, our study will inform interventions that can effectively address their specific needs and challenges, ultimately improving their health outcomes. Therefore, this study sought to address the identified knowledge gap by investigating the contextual factors associated with pregnancy termination among sexually active married women in LDCs based on the SEM framework. Such insights are crucial for informing targeted interventions and preventive measures to improve maternal health outcomes and reduce the prevalence of unsafe pregnancy termination.

Methodology

This study employed the most recent Demographic and Health Surveys (DHS) data (2015 −2022) from 61 developing countries. The DHS programme is a comprehensive resource for women’s health and well-being data, particularly focusing on women of reproductive age (15–49 years). The analysis utilised the women recode file (IR), which contains relevant variables including demographics (age, marital status), socio-economic factors (religion, education, wealth index, residence), sexual and reproductive health (fertility, pregnancy termination, family planning, prenatal/postnatal care, marital patterns, sexual activity, HIV/AIDS), and others.

That said, it is important to acknowledge potential data variations across countries with the region. These variations may stem from specific data points, geographical location, or even questionnaire versions used. Despite this, the DHS programme prioritises standardisation of key indicators to facilitate trend analysis and comparisons between and within countries. Detailed information regarding DHS methodologies and tools is freely accessible at the following links: https://dhsprogram.com/Data/ and https://dhsprogram.com/What-We-Do/Survey-Types/DHS-Questionnaires.cfm.

The DHS utilises a stratified two-stage cluster sampling design. In the initial stage, Primary Sampling Units (PSUs), generally enumeration areas (EAs), were chosen from urban and rural strata using probability proportional to size (PPS) sampling. This approach ensured that EAs with larger populations had a greater likelihood of being selected. Subsequently, a comprehensive list of households within the chosen EAs served as the basis for randomly selecting a predetermined number of households for interviews. All women aged 15–49 residing within these selected households were eligible to participate in the interviews.

A pooled, weighted sample of 715 281 sexually active, non-sterile, fecund married women aged 15 to 49 was included in the analysis. Complete information on all relevant study variables was available for each participant. Regional distribution the sample was as follows: Central Asia (Tajikistan) 6 976; Latin America and Caribbean (Colombia, Dominican Republic, Guyana, Honduras, Haiti and Peru) 17 576; North Africa, Western Asia and Europe (Albania, Armenia, Egypt, Jordan, Kyrgyz Republic, Turkey and Yemen) 47 044; Oceania (Papua New Guinea) 6 422; South and South East Asia (Afghanistan, Bangladesh, India, Indonesia, Cambodia, Myanmar, Maldives, Nepal, Philippines, Pakistan and Timor-Leste) 447, 636; and Sub-Saharan Africa (Angola, Burkina Faso, Benin, Burundi, Congo DRC, Congo Brazzaville, Cote D’Ivoire, Cameroon, Ethiopia, Gabon, Ghana, The Gambia, Kenya, Comoros, Liberia, Lesotho, Madagascar, Mali, Mauritania, Malawi, Mozambique, Nigeria, Niger, Namibia, Rwanda, Sierra. We selected countries with Demographic and Health Surveys (DHS) data from 2015 onwards to align with the SDGs timeline. This approach allowed us to capture the most recent prevalences and patterns in pregnancy termination and provide timely insights for policies related to maternal and sexual and reproductive health in LDCs.

Study variables

Outcome variable

The outcome variable of interest for this study is pregnancy termination among sexually active non-sterile fecund women of child bearing age. In this study, pregnancy termination is defined as ‘a loss of a pregnancy by a woman resulting from a stillbirth, miscarriage or abortion (unsafe or safe)’ as evidenced from other similar studies31,44. The DHS collects this data as pregnancy termination resulting from the three mentioned conditions in the definition, and is collected as a binary variable ‘No’ if a woman did not experience any of the three conditions, and ‘Yes’ if she did. For analysis, these responses were coded numerally, with ‘0’ representing ‘No’ and ‘1’ representing ‘Yes’.

Independent variables

Individual level

In this study, the individual-levels factors considered in the analysis included: women’s age (15—24; 25–34 and 35 years and above); age at first sex, age at first birth, and age at first marriage (< 18 years, ≥ 18 years); modern contraceptive use (no, yes); education level (none, primary, secondary and tertiary); employment status (no, yes); and decision-making (alone or jointly, someone else), all based on previous similar studies32,45

Household level

The household-level factors included in this study are: sex of household head (male, female); partner’s education (none, primary, secondary and tertiary); household size (< 5 members, 6 or more members) and wealth index (poor, middle and rich).

Community level

In order to come up with community-level factors, we aggregated some of the individual and household-level bio-demographic, socio-economic and access to and utilisation of SRH services as follows bases on other similar studies, such as those by12,32. The factors aggregated were: type of place of residence (urban, rural); community wealth index (low, medium and high); community media exposure (low, medium and high); community exposure to family planning messages (low, medium and high); and community contraception use (low, medium and high).

Statistical analysis

The analysis for this study was conducted using Stata version 17. Univariate analysis was employed to describe the characteristics of study participants, while bivariate analysis (chi-square test) assessed the association between the outcome variable and individual, household, and community-level factors. To identify predictors of pregnancy termination among married women in LDCs, binary multilevel regression analysis was performed using the ‘melogit’ command. Models were sequentially built (I to V), starting with the outcome variable alone to establish random intercepts, followed by the inclusion of individual, household, and community-level factors. The Akaike Information Criteria (AIC) was used to evaluate model fit, with lower AIC indicating better performance. Prior to analysis, a multicollinearity test based on the variance inflation factor (VIF) confirmed minimal collinearity (mean VIF = 1.24; min = 1.02 and max = 1.58). All bivariate and multilevel results, including 95% confidence intervals, are presented in the findings.

Results

Results in Fig. 1 below shows that 18.5% of married women in LDCs have terminated a pregnancy, with Central Asia and North Africa, Western Asia, and Europe exhibiting the highest rates at 26.3% and 27.4% respectively. Conversely, Sub-Saharan Africa and Oceania reported the lowest percentages at 16.8% and 7.0% respectively.

Fig. 1
figure 1

Percentage distribution of pregnancy termination by married women in less developed countries.

Full size image

Table 1 below shows that the majority of the study participants were aged 25–34 (41.0%), with regional disparities favouring older women married women (35 years and older) in Latin America and the Caribbean (49.4%), and North Africa, Western Asia, and Europe (43.2%). Nearly all married women (97.2%) had initiated sexual activity before 18, with the exception of those in North Africa, Western Asia, and Europe (57.4%).

Table 1 Married womens individual, household and community characteristics in less developed countries, 2015–2022.
Full size table

A significant proportion (16.0%) had become mothers before 18, with higher rates in Sub-Saharan Africa (25,1%) compared to Central Asia (2.4%). Early marriage was also common (37.5%), particularly in Sub-Saharan Africa (46.1%) compared to Central Asia (13.6%). The study further revealed that less than half (42.6%) of the married women were using modern contraceptives, with Latin America and the Caribbean having twice the percentage of Sub-Saharan Africa (58.0% vs. 28.8%).

The study also found that 50% of married women in the sample had at least attained secondary education (38.2% secondary and 13.2% tertiary). Despite this, only 30.0% of married women were employed. There were clear regional disparities in employment rates. Women in sub-Saharan Africa and Latin America and the Caribbean had higher employment rates (61.1% and 57.7%) compared to those in North Africa, Western Asia and Europe, as well as South and South-East Asia (20.4% and 15.6%, respectively). Notably, the majority of married women (70.2%) reported having decision-making autonomy, either independently or in conjunction with their partners. Central Asia demonstrated the highest percentage of women with this power (67.7%), while South and South-East Asia had much lower levels (15.4%).

An examination of married women’s household characteristics reveals that the majority (86.1%) of the married women live in households headed by males. The study further observed that, despite more than half of the married women having attained at least secondary level education, fewer husbands or partners demonstrated comparable educational attainment, with only 22.7% having completed secondary education and 8.2% having completed tertiary education. Besides, the typical household size from which married women came from was found to be five members or fewer (53.2%). However, regional disparities were evident, with married women in Central Asia (70.5%) residing in bigger households (6 or more members) compared to those in Latin America and the Caribbean (31.4%). Additionally, household wealth distribution revealed a polarized pattern, with a substantial proportion of married women (41.3%) residing in rich households and a similar proportion (39.3%) residing in poor households, a pattern consistent across the sub-regions within the LDCs.

Further, community-level characteristics reveal that less than two-thirds (63.8%) of married women in developing countries reside in rural areas, with Oceania exhibiting a proportion twice as high that of North Africa, Western Asia and Europe (88.8% vs. 42.7%). Overall, 53.1% of married women in LDCs lived in households with a high community wealth index, with Latin America and the Caribbean demonstrating a higher prevalence than Oceania (54.6% vs. 46.9%). Similarly, slightly more than half (50.9%) of the women exhibited high community media exposure, with South and South-East Asia exhibiting a notably higher percentage compared to Oceania (54.6% vs. 28.1%). In addition, 51.6% and 48.8% of married women resided in communities with high levels of family planning messages exposure and contraceptive use, with South and South-East Asia displaying significantly higher percentages (66.0% and 58.8%) compared to Oceania (28.5%) and Sub-Saharan Africa (32.3%).

Table 2 below shows the association between pregnancy termination among married women and various individual, household, and community factors. Our findings indicate that older married women (35 + years) are more likely to have terminated a pregnancy (23.1%) than younger women (24 years or younger, 11.3%). Regionally, married women in Central Asia have a higher rate of pregnancy termination (35.6%) compared to those in Oceania (7.5%). Furthermore, women who initiated sexual activity or had their first child at 18 years or older were significantly more likely to have terminated a pregnancy (24.9% and 19.5%, respectively) compared to those who did so before 18 (18.3% and 17.8%, p < 0.001). Finally, married women who married before 18 had a slightly higher rate of pregnancy termination (18.9%) than those who married at 18 or later (18.2%, p-value = 0.001).

Table 2 Percentage distribution of pregnancy termination among married women of reproductive age by individual, household and community characteristics in less developed countries, 2015–2022.
Full size table

Additionally, married women using modern contraceptives (19.6%) reported a higher incidence of pregnancy termination compared to those not using contraceptives (17.7%). Regional disparities were evident, with Central Asia and North Africa, Western Asia and Europe reporting higher rates of pregnancy termination among contraceptive users (33.3% and 30.2%) than Oceania (6.5%). Similarly, results reveal that, overall, married women with no formal education had a lower prevalence of pregnancy termination (17.8%) than those with primary or higher education (p < 0.001). Similarly, employed married women (19.9%) were more likely to report prior pregnancy termination than those not employed (17.8%, p < 0.001). Regional variations persisted, with Central Asia and Oceania showing higher rates of pregnancy termination among employed married women (31.9%) compared to sub-Saharan Africa (18.4%). No statistically significant association was found between decision-making autonomy and the likelihood of pregnancy termination.

With regard to the association between married women having had a pregnancy termination and household characteristics, married women from households headed by males exhibited a slightly higher prevalence of pregnancy termination (18.5%) compared to those headed by females (18.1%, p < 0.05). Women whose partners had higher educational attainment reported a notably higher proportion of pregnancy terminations. For instance, women whose partners had no formal education had a rate of 17.6%, while those whose partners had tertiary education or higher had a rate of 21.0% (p < 0.001). While the overall trend was consistent, regional disparities were observed. For example, in Oceania, South and South-East Asia, and sub-Saharan Africa, women whose partners had no formal education had lower rates of pregnancy termination. Conversely, in Central Asia, Latin America and the Caribbean, and North Africa, Western Asia & Europe, women with primary education exhibited lower rates. Similarly, married women residing in wealthier households reported a higher proportion of pregnancy terminations (19.6%) compared to those in poorer households (17.2%). This trend was consistent across most regions, with exceptions in Central Asia and North Africa, Western Asia & Europe.

Furthermore, married women who live in urban areas reported significantly higher rates of pregnancy termination (20.5%) than their rural counterparts (17.3%, p < 0.001). This trend was consistent across most regions, with exceptions in North Africa, Western Asia, and Europe. A higher community wealth index was associated with increased pregnancy termination among married women (19.6%). However, this relationship was region-specific, being evident only in Oceania and South and South-East Asia. Additionally, married women with medium levels of community media and family planning exposure were linked to higher rate of pregnancy termination (19.1% and 19.6%) compared to low (17.2% and 18.3%) or high (18.8% and 18.0%) exposure. At a regional level, Community media exposure had a significant impact on pregnancy termination in North Africa, Western Asia, Europe, South and South-East Asia, and sub-Saharan Africa (p < 0.01), whereas community family planning message exposure was associated with pregnancy termination in Central Asia, Latin America/Caribbean (p < 0.05) and South and South-East Asia (p < 0.001). Similarly, married women from communities with medium contraceptive use exhibited higher pregnancy termination rates (19.4%) than those with low use (17.8%, p < 0.005). This effect was statistically significant in North Africa, Western Asia, Europe, (p < 0.01), and South and South-East Asia (p < 0.001).

Interaction effects

We used stepwise logistic regression, with the Wald test as the criterion, to identify variables for our multilevel analysis. The procedure removed household size (p = 0.8875) and community contraceptive use (p = 0.4664). The remaining variables were then used to create a full model, which included interaction terms. We then compared this to a reduced model that excluded the interaction terms. Further, for model comparison and final selection, a likelihood ratio test comparing the full and reduced models revealed no significant interaction (p = 0.2605), leading us to accept the null hypothesis and drop the interaction terms. This decision was further supported by the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) values, which were lower for the reduced model (596,656.8 and 596,804.4, respectively) than for the full model. Consequently, we proceeded with the multilevel regression analysis using the reduced model.

Predictors of married womens pregnancy termination in less developed countries (2015–2022)

Results for the multilevel regression analysis are presented in Table 3 below, based on odds ratios and their corresponding confidence intervals (CIs) at all levels. Additionally, the fixed effects and random effects are equally presented according to the appropriate significance levels as highlighted in the methodology. The results for this study are presented in five models.

Table 3 Multilevel estimates and adjusted odds of pregnancy termination among married women of reproductive ages in less developed countries, DHS 2015–2022.
Full size table

Model 1 is a null model, which consists only results of the random intercept. This model excludes individual, household and community-related factors to demonstrate the level of heterogeneity of married women within clusters nested within the 61 countries that formed part of the analysis. Since the Akaike Information Criterion revealed that model V which included all the levels had a better fit, the interpretation of the results focused more on this model.

Model V in Table 3 below shows that 13.0% of the observed pregnancy termination that occur in less developed countries is as a result of the individual, household, and community level factors. Specifically, concerning the individual-level factors, older married women 25–34 year and 35 years and above had higher odds of having had a pregnancy terminated compared to those age 24 years or below (AOR = 1.83, 95% CI: 1.80–1.87; and AOR = 2.69, 95% CI: 2.50–2.56, respectively).

On the other hand, married women who initiated sex at 18 years or older exhibited higher odds of having a pregnancy terminated compared to those who initiated below 18 years (AOR = 1.33, 95% CI: 1.27–1.39). Likewise, compared to married women who had never given birth at the time of the survey, married women whose age at first birth was below 18 years or at 18 years or older exhibited higher odds of pregnancy termination (AOR = 1.09, 95% CI: 1.06–1.11; and AOR = 1.13, 95% CI: 1.11–1.16, respectively). Contrarily, married women who got married at or above 18 years exhibited lower odds of pregnancy termination than those who got married before they turned 18 years (AOR = 0.77, 95% CI: 0.76–0.79).

Furthermore, married women whose attained level of education is primary, secondary and tertiary have higher odds of terminating their pregnancy compared to those with no formal education (AOR = 1.15, 95% CI: 1.22–1.17; AOR = 1.20 95% CI: 1.17–1.22 and AOR = 1.10, 95% CI: 1.07–1.13). Similarly, the odds of currently married women who were not working having had terminated a pregnancy were higher (AOR = 1.09, 95% CI: 1.08–1.11) compared to those who were working. In contrast, the study observed that married women who were able to make independent and joint decisions with their husband or partner exhibited lower odds of pregnancy termination compared to those whose decisions were made by someone else (AOR = 0.98, 95% CI: 0.96–0.99).

With regard to household-level factors, the full model revealed that, just like the married women themselves, married women whose partners level of education was primary, secondary and tertiary had higher odds of having had a pregnancy terminated compared to those with higher education (AOR = 1.04, 95% CI: 1.02–1.06; AOR = 1.04% CI: 1.02–1.06 and AOR = 1.07 95% CI: 1.07–1.04–1.10). Similarly, married who resided in households with 5 or fewer household members exhibited higher odds of pregnancy termination compared to those whose household size is 6 or more members (AOR = 1.07(1.05–1.06). On the contrary, married women whose household wealth index is middle and rich exhibited higher odds of prior pregnancy termination compared to those whose household wealth index is poor (AOR = 1.04, 95% CI: 1.02–1.06).

With regard to community-level factors, married women residing in rural communities of LDCs were found to have lower odds of pregnancy termination compared to those from urban areas (AOR = 0.88, 95% CI: 0.86–0.90). Conversely, married women whose community wealth index was medium and high were found to have higher odds of pregnancy termination compared to those whose community wealth index was low (AOR = 1.07, 95% CI: 1.1.04–1.10; and AOR = 1.05, 95% CI: 1.03–1.08).

Similarly, married women whose community media exposure was medium and high had higher odds of pregnancy termination compared to those whose media exposure was low (AOR = 1.06, 95% CI: 1.03–1.09 and AOR = 1.05, 95% CI: 1.03–1.07). Likewise, married women from communities with medium family planning message exposure exhibited higher odds of pregnancy termination compared to those with low exposure (AOR = 1.04, 95% CI: 1.04–1.06). Besides women whose community contraception use was medium and high exhibited higher odds of pregnancy termination compared to those whose community contraception use is high (AOR = 1.05, 95% CI: 1.02–1.08 and AOR = 1.03, 95% CI: 1.01–1.05).

Discussion

A multilevel regression analysis was employed to examine the determinants of pregnancy termination among married women residing in LDCs in light of the socio-ecological model consisting the interaction of individual, interpersonal, community, and society factors.

The analysis reveals a significant prevalence of pregnancy termination in LDCs, with variations across regions. The highest rates were observed in North Africa, Western Asia, and Europe, while Oceania exhibited the lowest. These findings align with previous research conducted in LDCs46,47. These observed results can be as a result of a confluence of factors such as legal and cultural restrictions surrounding pregnancy terminations, besides limited access to modern contraceptives in North Africa, Western Asia and Europe. In contrast, the lower rates for Oceania may reflect a combination of stronger public health infrastructure and different societal norms regarding reproductive health. The analysis further identified a complex interplay of individual, household, and community factors contributing to pregnancy termination, suggesting the need for a multifaceted approach to address this issue.

At the individual level, our study reveals that older married women are more likely to terminate due to financial and family completion reasons, while younger women are less likely to terminate due to a lack of knowledge and cultural stigma. This finding contradicts previous research in Ghana48 but aligns with studies suggesting older women may have completed their desired family size and as such they may not see the need to have another child at an older age49,50,51. Besides, previous studies indicate that poverty and economic difficulties can influence women’s reproductive choices, particularly among older women who may find it challenging to financially support additional children, potentially leading to pregnancy termination52,53. Furthermore, the dearth of sexual and reproductive healthcare infrastructure in LDCs can disproportionately impact older women, leading to unintended pregnancies and unsafe terminations3,54. Moreover, lack of family planning education, and societal pressures on women to bear children, may contribute to unplanned pregnancies that are subsequently terminated55.

Conversely, women who initiated sexual activity at a younger age were less likely to have terminated a pregnancy compared to older women. This association may be attributed to several interconnected factors. Firstly, tradition beliefs, cultural norms, and stigma surrounding sexual activity can influence young women decision-making with regard to them family planning, potentially leading to unintended pregnancies which they decide to keep hence56,57. Secondly, socio-economic barriers and limited access to healthcare facilities, particularly in marginalized communities, can hinder young women’s ability to obtain reliable contraception as most of the facilities may not provide or have the method of choice58. Additionally, younger women may have limited access to comprehensive sexual education, leading to a lack of knowledge about SRH care including access to safe pregnancy termination59,60.

This study’s finding that women with greater autonomy have lower odds of pregnancy termination provides compelling evidence that empowering women at the individual level leads to more informed reproductive choices similar to other studies61,62. This autonomy maybe a direct result of factors like delayed marriage, higher education and employment, which enhances their financial stability and thereby making informed choices with greater maturity about their family planning63. Moreover, delayed marriage may align with a stronger desire for parenthood, leading to more intentional and planned pregnancies among women who marry later. In contrast, younger women may have limited resources, or lack of personal autonomy, which can increase the likelihood of unintended pregnancies and subsequent terminations.

Similarly, though findings on the relationship between educational attainment and employment status and pregnancy termination among married women in LDCs are mixed64,65. However, this study reveals that higher levels of education and employment among married women in LDCs are associated with increased rates of pregnancy termination. This observed finding may be attributed to several factors. Educated women are more likely to be access and utilise SRH services due to increased financial independence and reduced reliance on their partners for economic support by utilising their health insurance when need be or paying out of pocket when not covered through insurance66. Furthermore, increased education and employment often correlate with greater knowledge of contraception and family planning, leading to more effective family planning practices and potentially reducing unintended and unwanted pregnancies. Therefore, empowering women through education, economic opportunities, and comprehensive SRH services is crucial for promoting gender equality and improving women’s health outcomes in LDCs.

A strong correlation between autonomy for women and lower odds terminations in this study provides compelling evidence that empowered women are more likely to assert their reproductive rights and make informed decisions about their bodies. Studies suggest that women with greater autonomy have lower odds of pregnancy termination due to increased control over their reproductive choices67,68. This autonomy often manifests in factors like education, employment, and access to contraception. When women have the power to make informed decisions about their bodies and families, they are more likely to plan pregnancies, access necessary resources, and prevent unwanted pregnancies, thereby reducing the need for termination. Additionally, higher levels of education and economic empowerment can lead to delayed marriage and childbearing, giving women more time to establish themselves and make informed family planning choices69.

Interpersonal and household-level factors also play a role in pregnancy termination among sexually active married women in LDCs. Specifically, the finding that married women whose husbands had higher education exhibited a greater likelihood of pregnancy termination highlights the significant influence of the interpersonal level of the socio-ecological model which are consistent with the practice observed in similar studies within the region68. This observed pattern can be attributed to several factors. A partner’s education often correlates with more progressive views on gender equality and reproductive rights, fostering a supportive environment that enables a woman’s autonomy and decision-making regarding pregnancy termination54,70. Besides, economic factors may also play a role, as higher-educated partners may have greater income or career prospects, influencing family planning decisions and the ability to support a larger family71.

Interestingly, our study findings reveal that women in households with fewer household members exhibited higher odds of pregnancy termination compared to those with more household members. A potential explanation for the observed association between smaller household size and higher rates of pregnancy termination lies in the complex interplay of socio-economic factors experienced by households in LDCs. For instance, in households with fewer members, the financial burden of child-rearing may be more pronounced, leading to increased stress and a perceived inability to adequately provide for a child. This finding suggests that smaller households may lack the social support system often found in larger, more extended families, which are more prevalent in some LDCs communities, thereby exacerbate feelings of isolation and contribute to decisions regarding pregnancy termination.

Furthermore, the study’s finding that married women from poor households were more likely to choose abortion underscores the complex interplay of financial constraints at the interpersonal or household level. Limited financial means can impede woman’s ability to continue with her pregnancy, potentially due to inadequate healthcare access and economic burden of raising a child52,72,73,74, standing out as one of the key factors within this level of the SEM.

At the community and societal level, our study found that place of residence, community wealth index, media exposure, family planning message exposure, and contraceptive use were all associated with pregnancy termination. The disparity between rural and urban areas maybe attributed to limited access to healthcare services and information in rural communities, which can make it more difficult for women to obtain termination services in relation to their counterparts from more affluent groups75,76,77.

Furthermore, the positive correlation between media and family planning message exposure and higher termination rates highlight the powerful role of information dissemination at the community level78,79. Media, through television, radio, and online platforms, can provide married women with knowledge about contraception, safe termination practices, and their reproductive rights80,81. This increased knowledge and awareness empower women to make informed decisions about their reproductive health, even if those decisions include pregnancy termination.

The counterintuitive finding that married women from communities with medium and high contraceptive use also had higher termination rates is subtle but crucial point. This can be can be attributed to a more open and informed approach to sexual and reproductive health within these communities, leading to increased overall awareness and use of contraceptive methods82,83. Moreover, this increased awareness might also coincide with a greater understanding of the options for managing unintended pregnancies, including termination84. This finding demonstrates how community-level awareness and openness can influence individual-level choices.

Study strengths and limitations

The primary strength of this study lies in the robust DHS data, which offer a comprehensive framework for in-depth analysis of pregnancy termination, particularly among married women in LDCs. These data provide large, representative samples, standardized data collection, and extensive coverage of reproductive health indicators, ensuring comparability across countries and regions. However, the study has some limitations, Firstly, pregnancy terminations might be underreported due to social stigma, legal restrictions, or cultural taboos. As such, women may not disclose their experiences with pregnancy termination, leading to biased estimates. Secondly, the study’s cross-sectional design limits the ability to establish causality or temporal relationships between variables. Thirdly, there is a lack of detailed explanatory variables. The DHS might not capture detailed and contextual information about pregnancy terminations such as: reasons for terminating, gestational age at termination, method of termination, complications, or outcomes. Additionally, abortion policies, access to and utilization of SRH, and economic and environmental factors, which could have provided additional context to the observed prevalence, were not included.

Conclusion

The findings of this multilevel regression analysis, framed within a socio-ecological model, underscore the interwoven nature of pregnancy termination among married women in LDCs. Our study provides compelling evidence that a woman’s decision to terminate a pregnancy is not a simple, isolated choice but rather a complex outcome influenced by a dynamic interaction of individual, interpersonal, household, community, and societal factors. At the individual level, the analysis reveals nuanced patterns, such as the increased likelihood of termination among older women, driven by financial considerations and family completion, and the counter-intuitive association with higher education and employment, which we posit reflects greater autonomy and more informed reproductive decisions. These individual-level dynamics are further shaped by household and interpersonal influences, where a partner’s educational attainment can foster a supportive environment for reproductive choices and financial constraints within smaller or poorer households can heighten the perceived need for termination. At the macro level, the study demonstrates the powerful role of information dissemination through media and family planning programmes, which appears to increase overall reproductive health literacy and awareness of termination as a viable option for managing unintended pregnancies.

Policy implications

The findings from this analysis carry critical policy implications for improving women’s reproductive health in LDCs. Firstly, policy should prioritise empowering women through access to comprehensive sexual and reproductive health (SRH) education, alongside increased opportunities for education and employment. This will empower women to assert their autonomy and make more informed decisions about family planning. Secondly, interventions must move beyond the individual and target the household and community levels. Public health programmes should actively engage men, particularly those with higher education, to foster supportive partnerships regarding reproductive rights. Lastly, and most crucially, there is a need to improve the accessibility and quality of family planning and safe pregnancy termination services, particularly in rural and marginalised communities so that no woman is left behind, while simultaneously leveraging media to disseminate accurate and destigmatising information about pregnancy terminations and SRH in general. Such policies will enhance the overall well-being of women, thereby contributing to the achievement of Sustainable Development Goals (SDG) 3, 5 and 10, ensuring universal good health and well-being for all, gender equality and economic empowerment.