Time to first birth and its determinants among female youths in Ethiopia using accelerated failure time shared frailty models

Abstract

Early motherhood is associated with increased public health risks, which can significantly influence a woman’s life trajectory and opportunities. Despite ongoing efforts to reduce youth pregnancies, the prevalence of early childbirth remains a major challenge, affecting not only potential health outcomes but also broader social and economic consequences. Although several studies have examined fertility trends, most have focused on all age groups, and much of the available evidence is outdated. Moreover, up-to-date studies specifically focusing on female youths aged 15–24, a group at high risk for early childbearing, are limited. Therefore, this study aimed to identify the determinants of time to first birth among female youths in Ethiopia using the 2019 Ethiopia Mini Demographic and Health Survey. A total of 3,691 female youth aged 15–24 were included in the analysis. The Kaplan-Meier method was used to estimate the time to first birth, a log-rank test was used to compare the difference in survival curves. AFT models with gamma and inverse Gaussian frailty distributions was applied to identify significant determinant of time to first birth, with model selection based on the Akaike Information Criterion. The Weibull inverse Gaussian shared frailty model was the best fit model since it had a low Akaike Information Criterion. Overall, median survival time was 20 years (IQR: 19, 21). The significant determinants of time to first birth were richest wealth index (ϕ = 1.552, 95% CI: 1.211–1.989), knowledge of at least one method (ϕ = 1.284, 95% CI: 1.163–1.625), use of contraceptives (ϕ = 2.673, 95% CI: 1.809–3.951), rural residence (ϕ = 0.765, 95% CI: 0.616–0.951), secondary educational (ϕ = 1.242, 95% CI: 1.028–1.501) and higher educational (ϕ = 1.286, 95% CI: 1.122–1.475). Addressing these issue through targeted policies and improved access to education and reproductive health services is key to reducing early pregnancies and enhanceng youth well-being in Ethiopia.

Introduction

Youth is a stage marked by significant physical, emotional, and psychological changes, during which sexual thoughts and feelings become more prominent¹. According to the World Health Organization (WHO), youth are defined as individuals aged 15–24 years². This age group is particularly vulnerable to early childbearing, as many young women transition from school to marriage and reproductive life³. The age at which a woman has her first child plays a crucial role in shaping her fertility and future well-being⁴.

Globally, approximately 16 million adolescent girls aged 15–19 give birth each year, accounting for nearly 11% of all births, with 95% occurring in developing countries and most of young women pregnancy and childbirth are unplanned and unwanted⁵. Several Asian countries, including China and Vietnam, have successfully reduced fertility rates through government policies discouraging early and arranged marriages⁶. In contrast, sub-Saharan Africa has experienced a delayed fertility transition, with fertility rates remaining high compared to other regions⁷. Sub-Saharan Africa is the only region where fertility decline has been slower and occurred later⁸.

In Ethiopia, 21.1% of the population aged 15–24 years and 13% of women aged 15–19 have already begun childbearing, indicating a high burden of early pregnancies⁹. A previous study reported that the median age at first birth is 18 years, highlighting the early transition to motherhood¹⁰. Ethiopia’s low economic growth, widespread poverty, and poor and insufficient health services, along with early marriage particularly in rural areas where over 40% of girls marry before 18 and low education, remain key contributors to early childbearing^11,12. To address these challenges, the country has implemented a national adolescent and youth health strategy focusing on sexual and reproductive health, including the prevention of early childbearing. Despite these efforts, early childbearing continues to pose serious health, social, and economic consequences¹³. Understanding the current nationwide status of age at first birth among female youths is therefore crucial for guiding effective interventions.

The economic and social consequences of youth’s pregnancy and early child bearing are severe in developing countries, which is associated with maternal mortality, preterm birth, low birth weight, malnutrition, poor developmental outcomes, low school attainment and productivity, and consequently intergenerational transmission of poverty^14,15. Young mothers face an increased risk of complications such as hemorrhage, obstructed labor, puerperal endometritis, eclampsia, and systemic infections^16,17. Evidence also suggests that children born to young mothers are more likely to experience cognitive impairments, academic difficulties, and behavioral problem¹⁸. Conversely, Delaying pregnancy and child bearing among young women may contribute to higher school attainment, which might translate into better human capital out comes for their children¹⁴.

Several factors influence the timing of first birth, including socio-demographic, economic, and reproductive health-related determinants, Educational status, residency, wealth index, Religion, Current contraceptive method, age at first sexual intercourse and age at first marriage were identified as predictors for early age at first childbirth in several studies^10,19–23.

A few studies in Ethiopia have examined age at first birth, but most focused on women aged 15–49 and were limited to specific districts. While these studies provide useful insights, nationally representative evidence that focuses specifically on female youths remains limited. Research also suggests that the optimal age for pregnancy is in the late 20 s to early 30²⁴. Highlighting the need to focus specifically on younger women. At the same time, Early childbearing remains a persistent challenge in developing countries like Ethiopia, whereas delayed first birth is more common in high-income settings²⁵. This underscores the importance of national level, youth specific evidence to inform targeted interventions for female youths aged 15–24.

To address this gap, the current study focuses on female youths aged 15–24 using the recent Ethiopia Demographic and Health Survey (EDHS 2019), which is nationally representative. Understanding the timing and determinants of first birth in this age group is critical for addressing youth pregnancy, reducing early childbearing, and mitigating its adverse consequences. The findings will provide evidence to guide targeted interventions aimed at reducing early motherhood, promoting education, improving reproductive health, and supporting youth development in Ethiopia.

Therefore, the aim of this study was to identify the determinants influencing the timing of first birth among female youths in Ethiopia.

Methods

Study setting, study design and period

The study was conducted in Ethiopia using Ethiopia’s latest Mini Demographic Health Survey 2019 (EMDHS 2019). Permission to access the data sources was requested online by providing a detailed explanation of the study’s purpose. The survey employed a population-based cross-sectional design and was conducted between March 21, 2019, and June 28, 2019. The Ethiopian Public Health Institute (EPHI) implemented the survey in collaboration with the Central Statistical Agency (CSA) and the Federal Ministry of Health (FMoH), with technical support from the ICF and financial and technical assistance from development partners.

Sampling method

This study used data from the 2019 Ethiopia Mini Demographic and Health Survey (EMDHS), which employed a stratified two-stage cluster sampling design. Detailed information on the sampling procedure can be found in the EDHS 2019 report²⁶. A total of 8,885 women aged 15–49 years were interviewed at the time of the survey. For this study, only female youth ages 15 and 24 were included, resulting in a weighted sample of 3691 participants. Women older than 24 were excluded as they fall outside the target age range. An overview of the sampling procedure used for this study is illustrated in (Fig. 1).

Fig. 1.

Sampling procedure of time to first birth and its determinants among female youths in Ethiopia, 2019 EDHS.

Study variables

The outcome variable in this study was time to first birth, measured in years from a woman’s birth until her first childbirth. In this survival analysis, the event of interest was giving first birth, while women who had not yet given birth by the time of the survey were considered censored. The independent variables included marital status (married, unmarried), religion (Orthodox, Protestant, Muslim, other), place of residence (rural, urban), wealth index (poorest, poorer, middle, richer, richest), Sex of household head: categorized as male if the respondent lived in a male-headed household at the time of the survey, and female if otherwise, Media exposure was categorized as “yes” when there was exposure to at least one of reading newspapers, listening to the radio, or watching television, and “no” when there was no exposure to any of the three, Knowledge of contraceptive methods was grouped as knowing no method or knowing at least one method, while current contraceptive use was classified as “not using” or “using”, with both modern and traditional methods categorized under contraceptive use. Educational status was categorized as primary, secondary, and higher.

Data processing and analysis

Stata version 17 was used for data extraction, cleaning, coding, and analysis. Sample weighting was applied prior to further analysis, and descriptive statistics, including frequencies, percentages, medians, and interquartile ranges (IQR), were presented through tables, figures, and narratives. The median age at first birth was obtained from the Kaplan-Meier survivor curve, with comparisons across categorical variables conducted using the log-rank test. Multicollinearity was assessed before running the selected survival model. The mean variance inflation factor (VIF) was 1.45, indicating that there was no multicollinearity between the covariates. The Schoenfeld residual test was employed, and the proportional hazards assumption was found to be violated (χ² = 56.20, p < 0.001). Therefore, the Cox model was excluded. Given these results, alternative modeling approaches, such as the Accelerated Failure Time (AFT) model, were considered to better account for the dynamic effects of covariates on survival outcomes.

To account for clustering in the data, a shared frailty model was utilized to predict the time to first birth among female youth in Ethiopia, under the assumption that time to first birth remains consistent within the same clusters (EAs). Accelerated failure time (AFT) models with baseline distributions (Weibull, exponential, log-logistic, and lognormal) and frailty distributions (gamma and inverse Gaussian) were fitted, treating enumeration areas (EAs), represented by v001 (cluster number), as a random effect to identify predictors of time to first birth. In the AFT model, a positive regression coefficient indicates a longer survival time (delay in the occurrence of first birth), whereas a negative coefficient indicates a shorter survival time (earlier occurrence of first birth). Model adequacy was evaluated using the Akaike Information Criterion (AIC). The Weibull inverse Gaussian model was identified as the best fit, given its lowest AIC value (4317.524). Variables with a p-value less than 0.20 in the univariable analysis were included in the multivariable analysis of the Weibull inverse Gaussian shared frailty model (Table 1).

Table 1.

Comparison of Akaike information criterion among different accelerated failure time model and frailty distributions.

Information criteria	Models	No frailty	Gamma frailty	Inverse Gaussian frailty
AIC	exponential	4393.63	4373.339	4361.655
	Log logistic	4467.026	4451.578	4445.662
	Weibull	4352.899	4340.042	4317.524
	Lognormal	4586.516	4575.243	4548.119

Ethics approval and consent to participate

This study is based on secondary data analysis, for which we obtained permission from the MEASURE DHS program to download and utilize the data for research purposes. Therefore, ethical approval and participant consent were not needed. The dataset utilized in this study is publicly accessible on the MEASURE DHS program website https://dhsprogram.com and does not contain any personal identifiers.

Result

Socio-demographic and reproductive health characteristics of Ethiopian women

The study included a total of 3,691 weighted women to examine the time to first childbirth in Ethiopia. Of this total, 1,055 (28.5%) women had experienced their first birth (event), while the remaining 2,635 (71.4%) women had not undergone their first childbirth (censored). Weighted frequency analysis showed that 14.87% women were in the poorest household wealth index. 65.28% of the respondents resided in rural areas. 43.57%) of women were married. 79.84% of women not used contraceptive method. 94.25% of the women know of at least one contraceptive method (Table 2).

Table 2.

Socio-demographic and reproductive health characteristics of Ethiopian women (EDHS, 2019).

Variable	Categories	unweighted frequency (%) 3,678	Weighted percentage 3,691
Marital status	Married	1,661 (45.16)	43.57
	unmarried	2,017 (54.84)	56.43
Religion	Orthodox	1,371(37.28)	41.38
	Protestant	709 (19.28)	28.58
	Muslim	1,538 (41.82)	28.61
	others	60(1.63)	1.43
Media exposure	No	2,038(55.41)	61.41
	yes	1,640 (44.59)	38.59
Knowledge of any method	Knows no methods	317(8.62)	5.75
	Knows at least one methods	3,361(91.38)	94.25
Current contraceptive used	Not using	3,067 (83.39)	79.84
	Using	611 (16.61)	20.16
Place of residence	Urban	1,290 (35.07)	34.72
	Rural	2,388 (64.93)	65.28
Higher educational level	No education	663 (18.03)	14.20
	Primary	2,002 (54.43)	59.06
	Secondary	683 (18.57)	19.36
	Higher	330 (8.97)	7.39
Sex of household	Male	2,631 (71.53)	78.42
	Female	1,047 (28.47)	21.58
Wealth index	Poorest	789 (21.45)	14.87
	Poorer	514 (13.97)	17.18
	Middle	495 (13.46)	18.16
	Richer	581 (15.80)	21.53
	Richest	1,299(35.32)	28.27

Survival time of first childbirth

The overall median time to the first birth was 20 years (IQR: 19, 21). This indicates that 50% of women survive without giving birth until 20 years of age, while the other 50% experience their first birth by this age. The total follow-up period for all women in this study was 66,861.9068 person-years of observation. (Fig. 2)

Fig. 2.

Overall Kaplan–Meier survival curve of time to first childbirth among female youths in Ethiopia (EDHS 2019).

Comparisons of survival functions of different categorical variables

To compare and estimate the survivor function across different respondent characteristics, we used the Kaplan–Meier survival curve and log-rank test. In the Kaplan–Meier curve, a survivorship function curve located below another indicates that the group represented by the lower curve has a lower survival status compared to the group represented by the upper curve. The significance of the graphically observed difference was assessed by log rank test and it is indicated all the predictor variables showed significant survival differences at p = 0.0000 (Fig. 3).

Fig. 3.

Kaplan–Meier survival curves and log rank tests of time to first birth among female youths by their characteristics, EDHS 2019.

Determinants of time to first birth Weibull inverse Gaussian shared frailty

The frailty in this model is assumed to follow an Inverse-Gaussian distribution with a mean of 1 and a variance equal to theta (θ). There is evidence of heterogeneity across clusters, as indicated by the frailty term θ = 0.1321129. The significant p-value obtained from the likelihood ratio test for the hypothesis θ = 0 yields significant p-value of 0.000, indicates that the frailty component made a significant contribution to the model, suggesting the presence of unobserved heterogeneity within clusters. In the multivariable Weibull inverse Gaussian shared frailty model place of residence, wealth index, Knowledge of any method, Current contraceptive method and highest education were found to be determinants of time to first birth.

The acceleration factor for time to first childbirth among mothers in the richest wealth quintile was 1.552 (ϕ = 1.552, 95% CI: 1.211, 1.989) compared with the poorest quintile, indicating that higher household wealth delays the occurrence of first birth. Women who lived in rural areas had an acceleration factor of 0.765 (ϕ = 0.765, 95% CI: 0.616–0.951) compared to those in urban areas, suggesting that rural residence accelerates first birth. Women who knew at least one contraceptive method had an acceleration factor of 1.284 (ϕ = 1.284, 95% CI: 1.163, 1.625) compared to women who did not know any method, showing that knowledge of contraceptives delays first birth.

Women who used contraceptives had an acceleration factor of 2.673 (ϕ = 2.673, 95% CI: 1.809, 3.951) for time to first childbirth compared with women who did not use contraceptives. Indicating that contraceptive use substantially delays first birth.

Women who attended secondary educational levels had an acceleration factor of 1.242 (ϕ = 1.242, 95% CI: 1.809, 3.951) compared to uneducated women. Suggesting that secondary education delays first birth. Furthermore, women who attended higher educational levels had an acceleration factor of 1.286 (ϕ = 1.286, 95% CI: 1.122, 1.475) compared to uneducated women, highlighting that higher education delays first birth (Table 3).

Table 3.

Multivariable Weibull inverse Gaussian shared frailty model for determinants of time to first among female youths in ethiopia: a shared frailty model based on EDHS 2019.

Variable	Categories	Coef	Acceleration factor (φ)	95% CI for φ	P-value
marital status	Married	0.0660	1.068	0.944,1.936	0.294
	Unmarried	1
Wealth index	Poorest	1
	Poorer	−0.0576	0.943	0.774,1.150	0.997
	Middle	−0.1263	0.881	0.729,1.064	0.190
	Richer	0.0003	1.000	0.837,1.194	0.568
	Richest	0.4396	1.552	1.211,1.989	0.001
Media exposure	No	1
	Yes	0.0704	1.073	0.939,1.225	0.297
Knowledge of any method	Knows no method	1
	Knows at least one method	0.2505	1.284	1.163,1.625	0.037
Current contraceptive method	Not using	1
	using	0.9835	2.673	1.809,3.951	0.000
place of residence	Urban	1
	Rural	−0.2669	0.765	0.616,0.951	0.016
Highest educational level	No education	1
	Primary	−0.2432	0.784	0.591,1.039	0.091
	Secondary	0.2170882	1.242	1.028,1.501	0.025
	Higher	0.2521	1.286	1.122,1.475	0.000
Sex of household head	Male	0.0764	1.079	0.942,1.235	0.268
	Female	1

Discussion

This study assessed the time to first birth and its determinants among female youths using a AFT shared frailty analysis method. Weibull Inverse Gaussian shared frailty model was selected based on AIC. The clustering effect was significant (p-value = 0.000) in Weibull-inverse Gaussian shared frailty model. This showed that there was heterogeneity between the regions on the time to first childbirth in Ethiopia. The result of the Weibull Inverse Gaussian shared frailty model shows place of residence, wealth index, Knowledge of any method, Current contraceptive method and highest education were significantly associated with time to first childbirth.

In this study, the median age at first birth was 20 years (IQR = 19–21), reflecting an early transition into motherhood. This early timing is closely linked to limited educational opportunities for girls, particularly in rural areas, where social and economic pressures often encourage early marriage²⁷. In addition, limited access to reproductive health information and services for young women can further increase the likelihood of early childbearing¹⁰. Entering motherhood at a young age can interrupt schooling, limit future economic prospects, and elevate health risks for both young mothers and their children²⁸. This finding highlights that the median age at first birth remains below the recommended threshold²⁴, placing young women at substantial risk and emphasizing the need for targeted interventions to reduce early childbearing and improve youth outcomes in Ethiopia.

The results of this study suggested that women from the richest households had a longer time to their first birth compared with women from the poorest households. This may be explained by young women from wealthier households often experiencing prolonged school enrollment, which leads them to pursue higher education²⁹. Higher education, in turn, empowers women with greater autonomy in reproductive decision-making, increased participation in career or income-generating activities, and improved access to sexual and reproductive health information and services, all of which facilitate the postponement of first birth^30,31. In contrast, women from poorer households often experience early school dropout and face economic constraints that limit their access to sexual and reproductive health services, thereby increasing their risk of early childbearing^32,33. Furthermore, financial dependence may lead young women to engage in transactional sexual relationships or early marriages with older, wealthier men as a means of economic survival, which in turn contributes to early childbearing³⁴. This finding suggests that addressing early childbearing requires an integrated approach that combines reproductive health education with programs promoting economic empowerment for young women from low-income families.

Women who knew at least one contraceptive method are less likely to give birth to their first child at an early age than those who did not know any method. this may be due to the increased likelihood of contraceptive use among women who are well-informed about contraceptives, Access to and utilization of contraceptive methods often lead women to delay maternity at an early age³⁵. In Ethiopia, adolescents and young women primarily acquire knowledge about modern contraceptive methods through health facilities, particularly via health extension workers and youth-friendly services, as well as through peer education programs, mass media, and non-governmental organizations^36–38. However, comprehensive sexuality education (CSE) within the formal school system remains limited and inconsistently implemented, and discussions about sexual and reproductive health within families are often constrained by cultural norms^39,40. As a result, many young women rely on informal or fragmented sources of information, which may contribute to superficial or incomplete knowledge of contraceptive methods³⁸. The finding underscores the importance of reproductive health education in preventing unintended pregnancies and delaying childbirth, highlighting that enhancing women’s understanding of contraceptive methods is crucial for improving maternal and child health outcomes in Ethiopia.

Women who used contraceptive had a longer time-to-first birth than the non-users. This is because a lower interest in using contraceptive methods may arise from a strong desire to have children, increasing the likelihood of early childbirth⁴¹. On the other hand, the appropriate utilization of contraceptives enables sexually active women to delay unintended pregnancies and childbirth⁴². In Ethiopia, family planning services are available through the Health Extension Program, but access and consistent use remain limited, particularly among youth among youth in rural areas and those of lower socio-economic status^43,44. These challenges are compounded by embarrassment and stigma when young women attempt to use family planning services, as well as difficulty discussing reproductive health with family members⁴⁵. These factors place young Ethiopian women at a high risk of early childbirth. This finding underscores that contraceptive use plays a crucial role in delaying first births. Improving young women’s knowledge of and access to contraceptive methods, as well as addressing stigma around contraceptive use for female youth, could increase utilization and help delay first births.

This study also found that women living in rural areas were more likely to experience early childbirth compared to women living in urban areas. The reason for this might be because those women residing in rural areas are less knowledgeable, have less accessible to health information and could not easily access and utilize family planning services compared with urban area³¹. In Ethiopia, the majority of the population resides in rural areas where educational opportunities, infrastructure, and employment prospects are limited⁴⁶. Girls in urban areas are more likely to continue schooling, gain employment, and receive guidance from better-educated parents, which together delay first birth, whereas rural girls are more likely to be influenced by cultural practices such as early marriage, making them vulnerable to adolescent childbearing⁴⁷. This finding suggests the need to strengthen reproductive health services and to expand educational and livelihood opportunities in rural communities in order to reduce early childbearing in Ethiopia.

This study found also that women’s educational level is a significant predictor of the timing of their first childbirth. Women who attained secondary and higher education delayed their first childbirth compared to those with no formal education. The result was consistent with the studies conducted in Ethiopia^10,42, Kenya⁴⁸, Ghana⁴⁹Bangladesh⁵⁰. This is may be due to that educated women are more aware about information related to reproductive health, including the benefits of family planning, the risks associated with early pregnancy, and the proper use of contraceptive methods. As a result, educated women delay first childbirth, reduce the risk of youth pregnancy^51,52. Although Ethiopia has improved girls’ enrollment in primary schools, the majority of youth aged 15–24 have not completed primary education (54%) or have no education at all (16%), and only a small proportion complete secondary school (1%) or attain post-secondary education (5%)⁵³. This indicates that many young women remain at risk of early childbirth, highlighting the critical role of completing secondary education in delaying first births and improving maternal and child health outcomes.

The main strength of this study was using nationally representative data, which has a high response rate and is generalizable to all Ethiopian female youths. The study analyzed national survey data, providing policymakers with valuable information for developing effective intervention strategies at both national and regional levels. However this study has some limitation, one of the main limitations of this study is that recall bias and social desirability bias may have been caused by the reliance on self-reporting. Furthermore, due to the use of secondary data, some potentially important variables were not available in the mini DHS data set. As a result, the analysis may not have accounted for all relevant factors that could influence the outcomes of interest.

Conclusion

The study aimed to examine time-to-first childbirth and its associated factors among Ethiopian female youth (15–24) using the AFT Shared Frailty Model, based on data from the 2019 Ethiopia Mini Demographic and Health Survey. The median age at first birth was 20 years. Indicating that half of young women have their first child before this age. This period is generally considered crucial for education and personal development. Therefore giving birth before 20 year will prevent female youths from continuing their schooling and pursuing other personal growth opportunities, in addition to the adverse consequences of early childbearing. Young women from wealthier households, those who had knowledge of and used contraceptives, and those with secondary or higher education were more likely to delay childbirth, while those residing in rural areas were more vulnerable to early childbearing. These findings highlight disparities in early childbearing across wealth, residence, education, knowledge of any method and contraceptive use, emphasizing the need for targeted interventions that improve educational attainment, reproductive health knowledge, and access to contraceptives, particularly for young women in rural and low-income communities.

Abbreviations

AFT

Accelerated Failure Time

AIC

Akaike Information Criterion

CI

Confidence Interval

CSA

Central Statistical Agency

DHS

Demographic and Health Survey

EAs

Enumeration areas

EMDHS

Ethiopia Mini Demographic and Health Survey

EPHI

Ethiopian Public Health Institute

FMOH

Federal Ministry of Health

IQR

Interquartile ranges

PH

Proportional hazard

SNNPR

Southern Nations, Nationalities, and Peoples’ Region

WHO

World Health Organization

Author contributions

EGM is involved in conception, the data extraction, analysis and interpretation of the finding. HAK, AE, YT and SAK assisted in the analysis of the study. EGM writing the original draft. EGM, HAK, AE, YT and SAK writing the review and editing the manuscript. The final manuscript has been read and approved by all authors.

Data availability

The dataset utilized in this study is publicly accessible on the MEASURE DHS program website [https://dhsprogram.com].

Declarations

Competing interests

The authors declare no competing interests.