Generational disparities in maternal healthcare: a life-course perspective on social determinants in Kersa, Eastern Ethiopia Authors: Afework Tadele, Merga Dheresa, Muluemebet Abera Published: 2026 ABSTRACT AbstractIntroductionMaternal mortality remains persistently high in low- and middle-income countries, driven by inequities in access to and use of maternal healthcare. A life-course perspective offers critical insights into the underlying drivers of these disparities, linking historical, social and structural factors across women’s lives to health outcomes. By contextualising maternal health within lifelong trajectories, the analysis aims to inform targeted strategies to reduce inequities and improve care access across generations.ObjectivesTo analyse the generational differences in the maternal health services utilisation across the maternal life course and its social determinants in Kersa, Eastern Ethiopia.MethodsThis study involved secondary analysis of pre-existing longitudinal health and demographic surveillance system (HDSS) data, comparing mothers born during three distinct historical periods. Reproductive-age women from three different generations, specifically those born in the 1970s, 1980s and 1990s. A linear mixed-effects modelling approach was used, employing the logit link function for the binomial family. The statistical analysis was done using RStudio. A 5% p value threshold and a 95% confidence level were used to assess the statistical significance of the findings.ResultsPrenatal care coverage was lowest among the 1970s cohort (29.6%), rising to 33.2% for the 1980s cohort and 36.2% for the 1990s cohort. Skilled birth care utilisation showed a similar generational gap: 27.2% (1970s), 26.4% (1980s) and 30.6% (1990s). Adjusted analyses revealed maternal education (adjusted OR (AOR) 1.66 (95% CI 1.443 to 1.916), husband’s education and husband’s unemployment (AOR 3.44 (95% CI 1.936 to 6.112)) as key predictors of prenatal care use. Younger age at first marriage (eg, <16 years: AOR 0.747 (95% CI 0.594 to 0.872)) and unintended pregnancy reduced prenatal care utilisation. For skilled birth attendance, household wealth (highest vs lowest wealth: AOR 2.28 (95% CI 1.926 to 2.708)), prior prenatal care use (AOR 2.89 (95% CI 2.514 to 3.33)) and birth cohort (1980s vs 1990s: AOR 0.79 (95% CI 0.681 to 0.914)) were significant determinants. Early marriage (eg, <24 years) and unintended pregnancy similarly decreased skilled care utilisation.ConclusionThe persistent generational disparities in maternal healthcare utilisation, with later-born cohorts (1990s) demonstrating improved coverage compared with earlier generations (1970s–1980s). These disparities are driven by intersecting generational differences in educational opportunity, employment patterns and technological modernisation. Key determinants include maternal and spousal education, household wealth, age at first marriage, pregnancy intention and generational differences in acceptance of modern healthcare. Addressing these disparities requires targeted policies to enhance educational access, economic equity and reproductive health autonomy, ensuring equitable maternal healthcare utilisation across generations.Keywords: Causality, Reproductive History, Health Services Accessibility, Public Health, Public Health Practice WHAT IS ALREADY KNOWN ON THIS TOPIC Maternal healthcare inequities in low-income countries are driven by individual factors like education and wealth. However, research rarely examines how generational context and historical events collectively shape these disparities across a woman’s life course. Maternal healthcare inequities in low-income countries are driven by individual factors like education and wealth. However, research rarely examines how generational context and historical events collectively shape these disparities across a woman’s life course. WHAT THIS STUDY ADDS The study reveals significant generational disparities in maternal care among Ethiopian women born in the 1970s, 1980s and 1990s. It uniquely shows that the 1980s cohort experienced a decline in maternal health service utilisation, highlighting how social determinants impact generations differently. The study reveals significant generational disparities in maternal care among Ethiopian women born in the 1970s, 1980s and 1990s. It uniquely shows that the 1980s cohort experienced a decline in maternal health service utilisation, highlighting how social determinants impact generations differently. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY The findings underscore the need for targeted, multisectoral policies that address generation-specific barriers rather than uniform interventions. They also call for strengthening the continuum of care and guiding future research into the mechanisms behind these generational health inequities. The findings underscore the need for targeted, multisectoral policies that address generation-specific barriers rather than uniform interventions. They also call for strengthening the continuum of care and guiding future research into the mechanisms behind these generational health inequities. Introduction Maternal mortality remains a critical public health challenge in low- and middle-income countries (LMICs),1 with Ethiopia bearing one of the highest burdens globally at 412 deaths per 100 000 live births.2 Over 90% of these deaths are preventable through increased utilisation of maternal health services.1,3 Yet persistent inequities, shaped by generational disparities where older cohorts of mothers are affected more4 and life-course adversities, hinder progress, particularly among marginalised communities.2,5 To unravel these disparities, this study adopts a generational life-course perspective, contextualising maternal healthcare access within Ethiopia’s historical adversities and evolving socioeconomic landscapes.6,7 Maternal health can be seen in the context of the life course. It highlights the three temporal dimensions: historical time, generational time and individual time.8 Historical time refers to the circumstances during a person’s birth and life, shaping their opportunities and challenges.9 Generational time captures the shared experiences and traits of people born in the same period, influenced by historical events and policy environments. Each generation may exhibit health-seeking behaviours due to unique experiences with educational systems, economic opportunities and technological diffusion. Recent generations entered adulthood amid rapid technological advancements, including mobile phone diffusion, which improved health information access and service coordination. However, these advantages are not uniformly distributed; earlier cohorts, having been raised in contexts of limited educational and technological exposure, may continue to face barriers in adopting modern healthcare practices, perpetuating generational inequities. For instance, the Baby Boomer generation in the USA witnessed significant societal changes like the civil rights movement and the sexual revolution,10 the demographic transitions after the Great Irish Famine11 and in South Asia (Bengal, Bangladesh, Bihar and Maharashtra).12 Individual time encompasses personal experiences, changes and life trajectories. It considers decisions, relationships, and the cumulative impact of prior events.13 Even within the same generation, individuals can have distinct educational, professional and family experiences based on their choices and circumstances.14 In addition, recent generations entered adulthood amid relative stability and technological advancements, including mobile phone diffusion, which improved health information access and service coordination. Yet, this generation navigates distinct challenges: urbanisation pressures, shifting gender norms and uneven adoption of modern care. Such generational disparities likely shaped inequities in maternal healthcare utilisation.15 While prior research ties adverse childhood experiences to pregnancy complications,16 few studies examine how early-life famine exposure or generational contexts shape maternal healthcare utilisation. Individual factors (age, education and wealth) and structural determinants (geography, healthcare access) are well-documented,17–21 yet analyses comparing cohorts across Ethiopia’s volatile 20th-century trajectory remain scarce. Such comparisons are critical to identifying high-risk generations and tailoring interventions.22 This study investigates generational gaps in prenatal and skilled birth care utilisation among women born in the 1970s, 1980s and 1990s in Eastern Ethiopia’s Kersa Health and Demographic Surveillance System (KHDSS). By framing maternal health within life-course adversities, famine, displacement and technological shifts, we elucidate how socioeconomic, educational and reproductive determinants differentially impact care access across generations. Our findings aim to inform targeted interventions addressing the unique barriers faced by historically marginalised cohorts while leveraging opportunities in evolving contexts. Method Study context KHDSS center was established in 2007 in 12 kebeles (the smallest administrative unit in Ethiopia) by defining a baseline population census, and then started longitudinal data collection to monitor vital events in the population on a regular basis every 6 months by resident enumerators. In the year 2012, the health and demographic surveillance system (HDSS) field site was expanded to the Harari region (Harar town) and was conducted in six kebeles. HDSS site expanded its study area again by doubling the number of kebeles to 36 kebeles (24 from Kersa district and 12 kebeles from Harar town), which includes 41 056 households and a total population of 197 268.23 The study period of this particular study is between 2015 and 2021, during which pregnancy observation was carried out for maternal health service utilisation. This study represents a secondary analysis of existing longitudinal data from the HDSS. The HDSS team, comprising trained resident enumerators, conducted primary data collection through routine household visits every 6 months. Study design This study involved secondary analysis of pre-existing longitudinal HDSS data, comparing mothers born during three distinct historical periods. This design involves multiple single cohorts, starting each at a different age. As shown in table 1, three single cohorts were followed for 7 years. Cohort 1 (born 1993–1996): at baseline in 2015, these women were 19 years old and were followed to ages 25–29 years. Cohort 2 (born 1982–1986): at baseline, these women were 29 years old and were followed to ages 35–39 years. Cohort 3 (born 1973–1976): at baseline, these women were 39 years old and were followed to ages 45–49 years. The overlapping ages at 29 and 39 years allow us to link the cohorts into a single life-course trajectory spanning 19–49 years. Seven annual measurements were taken from 3708, 2860 and 718 pregnant women in the first, second and third cohort, respectively, with a total of 7286 pregnant mothers included in the study. Overall, there were 7023 observations in the first cohort, 5236 observations in the second cohort and 1040 observations in the third cohort, totalling 13 299 complete observations. Table 1Sampling design of the study participants recruitment of pregnancy observation data in Kersa, Eastern EthiopiaCohortsBirth cohortFollow-up periodMaternal age(in years)Sample sizeNumber of observationsCohort 11993–19962015–202119–2937087023Cohort 21982–198629–3928605236Cohort 31973–197639–497181040Total19–49728613 299 Sampling design of the study participants recruitment of pregnancy observation data in Kersa, Eastern Ethiopia Variable definition and measurement Outcome variables Coverage of maternal health service utilisation was the primary outcome of interest. We examined two elements separately: the utilisation of prenatal care and skilled birth attendance. Outcomes were coded as binary, with 1 if a mother used at least one prenatal care service or if she had a skilled birth attendant for at least one birth, otherwise as 0. Explanatory variables/confounders The sociodemographic variables include maternal education, which is categorised as either never having attended formal education or having attended formal education. Husband education is also categorised as either never having attended formal education or having attended formal education. Household wealth status is classified as either the lowest (poor), medium or highest (rich). Maternal occupation can be unemployed, a housewife or employed in private or public organisations. Similarly, husband occupation can be a farmer, employed or unemployed. The reproductive health variables consist of the age at first marriage, which is categorised as ≤15 years, 16–18 years, 19–24 years or ≥24 years old. The age at first birth is categorised as ≤15 years, 16–18 years or ≥19 years. The number of live children can be ≤2 children, 3–4 children or ≥5 children. The total number of pregnancies can be ≤2, 3–4 or ≥5. The history of stillbirths can be 0, 1–2 or ≥3. Pregnancy intention is categorised as either a planned or unplanned pregnancy. Occupation categories were defined as follows: for mothers: ‘unemployed’ (not engaged in any income-generating activity or household work), ‘housewife’ (engaged in unpaid domestic work/homemaking) and ‘employed’ (engaged in paid work in the private/public sector, self-employment or daily labour). For husbands: ‘farmer’ (engaged in agricultural work, subsistence or commercial), ‘employed’ (engaged in paid work, salaried employment, daily labour or trade) and ‘unemployed’ (not engaged in any income-generating activity). Age at first marriage and age at first birth are critical reproductive health determinants in LMICs. Early marriage is consistently associated with lower educational attainment, limited reproductive autonomy and reduced healthcare-seeking behaviour. These variables capture women’s social status, decision-making power and reproductive trajectories, factors shaped by generational contexts, and are included as potential confounders in our analysis. Maternal nutrition is assessed using the mid-upper arm circumference (MUAC), which is categorised as <21 cm, 21–23 cm or ≥23 cm. MUAC measurements were taken by trained resident enumerators as part of routine KHDSS surveillance protocols, using standardised MUAC tapes following WHO-recommended procedures. Statistical analysis A generalised linear mixed-effects model with a logit link function and binomial distribution was used to analyse the binary response variable. This approach accounts for within-subject correlation and variability over time. To address the potential bias resulting from multiple cohorts in the longitudinal designs, both fixed and random cohort effects were employed, and parameter estimates were obtained for both individual-specific and time-specific variables.24 In addition, the model is able to handle missing values at one or several time points. Potential confounding variables were selected using the forward selection method. Model fit was examined using the lowest Akaike’s information criterion value.25 Intra-class correlation coefficients26 were used to compare the final models. We assessed multicollinearity using variance inflation factors (VIF) and found no issues, as all VIF values were below 5, indicating low correlation.27 Significance was set at the 5% level for all analyses.28 Analyses were carried out using RStudio for the generalised linear mixed-effects model family (binomial), link (logit). Patient and public involvement It was not appropriate or possible to involve patients or the public in the design, conduct, reporting or dissemination plans of our research. Result Sociodemographic characteristics of study participants Over a 7-year study period, 7286 participants contributed to a total of 13 299 observations. The cohort comprised women predominantly born in the 1990s (3708; 50.9%), followed by those born in the 1980s (2860; 39.2%) and 1970s (718; 9.9%). A majority of participants (9548; 71.8%) reported no formal education, whereas nearly half of their husbands (6222; 50.5%) had received formal schooling. Only 4% of mothers (536) were engaged in paid employment, with 83.8% (11 151) identifying as homemakers. Additionally, 80.7% of women (9964) indicated that their husbands worked in agriculture (table 2). Table 2Socioeconomic characteristics of the pregnancy observation and birth outcome data from 2015 to 2021, in Kersa, Eastern EthiopiaVariablesBirth cohorts of 1970s# (%)Birth cohorts of 1980s# (%)Birth cohorts of 1990s# (%)TotalEthnicity13 296 Oromo1022 (98.3%)5138 (98.1%)6904 (98.3%)13 064 (98.3%) Amhara and Gurage18 (1.7%)98 (1.9%)116 (1.7%)232 (1.7%)Religion13 295 Muslim1019 (98.0%)5128 (97.9%)6900 (98.3%)13 047 (98.1%) Orthodox21 (2.0%)102 (1.9%)119 (1.7%)242 (1.8%) Others0 (0.0%)5 (0.1%)1 (<1%)6 (0.0%)Maternal education13 299 No formal education974 (93.7%)4498 (85.9%)4076 (58.0%)9548 (71.8%) Attended formal education66 (6.3%)738 (14.1%)2947 (42.0%)3751 (28.2%)Husband education12 333 No formal education190 (19.3%)1678 (34.0%)4243 (66.2%)6111 (49.5%) Attended formal education796 (80.7%)3255 (66.0%)2171 (33.8%)6222 (50.5%)Maternal occupation13 299 Unemployed9 (0.9%)168 (3.2%)1435 (20.4%)1612 (12.1%) Housewife998 (96.0%)4890 (93.4%)5263 (74.9%)11 151 (83.8%) Employed33 (3.2%)178 (3.4%)325 (4.6%)536 (4.0%)Husband occupation12 353 Farmer958 (97.2%)4620 (93.7%)4386 (68.2%)9964 (80.7%) Employed26 (2.6%)208 (4.2%)1961 (30.5%)2195 (17.8%) Unemployed2 (0.2%)105 (2.1%)87 (1.4%)194 (1.6%)Household Wealth13 053 Lowest355 (34.2%)1819 (35.2%)2382 (34.8%)4556 (34.9%) Medium376 (36.3%)1632 (31.5%)2459 (35.9%)4467 (34.2%) Highest306 (29.5%)1723 (33.3%)2001 (29.2%)4030 (30.9%) Total observations10405236702313 299 Socioeconomic characteristics of the pregnancy observation and birth outcome data from 2015 to 2021, in Kersa, Eastern Ethiopia Maternal health service utilisation across birth cohorts Figure 1 highlights distinct generational disparities in maternal health service utilisation across three birth cohorts (1970s, 1980s and 1990s). The 1970s cohort exhibited the lowest utilisation rates for both prenatal care (29.6% (95% CI 26.9% to 32.5%)) and skilled birth care (27.2% (95% CI 24.5% to 29.9%)). The 1980s cohort showed moderate improvement in prenatal care (33.2% (95% CI 31.2% to 34.5%)) but paradoxically experienced a decline in skilled birth care (26.4% (95% CI 25.2% to 27.6%)), suggesting shifting priorities or inequities during their childbearing years. In contrast, the 1990s cohort achieved the highest coverage in both services: prenatal care (36.2% (95% CI 35.1% to 37.3%)) and skilled birth care (30.6% (95% CI 29.5% to 31.6%)), likely driven by cohort-specific factors. These contrasts, supported by non-overlapping CIs, underscore how maternal care utilisation diverged sharply across generations, shaped by distinct historical, economic and policy contexts rather than uniform progress over time. Prenatal and skilled care utilisation among the three cohorts of mothers in Kersa, Eastern Ethiopia. Determinants of prenatal care service utilisation In the multivariable generalised linear mixed-effects model, social determinants (such as maternal education and husband occupation) and reproductive health-related factors (such as age at first marriage and pregnancy intention) were found to be significant predictors of prenatal care utilisation in Kersa, Ethiopia. After controlling for potential confounders, maternal education was associated with 66% higher odds of antenatal care utilisation (adjusted OR (AOR) 1.66 (95% CI 1.44 to 1.92)), corresponding to an approximately 6–8 percentage point increase in coverage. This effect size is clinically and programmatically meaningful. Attendance at formal education by women and their husbands is also a significant factor in improving prenatal care utilisation. Women who attended formal education were 1.66 times more likely (AOR 1.663 (95% CI 1.443 to 1.916)) to use prenatal care compared with their counterparts. Husband occupation was also a significant social determinant of prenatal care utilisation. A woman having her husband unemployed was more than three times (AOR 3.44 (95% CI 1.936 to 6.112)) more likely to use prenatal care services compared with those practising traditional farming. In addition, reproductive health characteristics such as age at first marriage and pregnancy intention were also found to be significant predictors of prenatal care utilisation. Women who were married before 16 years were 25.3% (AOR 0.747 (95% CI 0.594 to 0.872)), those married between 16 and 18 years of age were 30% (AOR 0.7 (95% CI 0.562 to 0.872)) less likely, those married between 19 and 24 years of age were 29% (AOR 0.71 (95% CI 0.579 to 0.872)) less likely to use prenatal care compared with those married at and after 24 years of age. Women who planned their pregnancy were found to be about one and a half times more likely (AOR 1.427 (95% CI 1.269 to 1.603)) to use prenatal care services compared with those who had no pregnancy intention (table 3). Table 3Generalised linear mixed-effects model for the social determinants of prenatal care utilisation in Kersa, EthiopiaVariablesPrenatal care utilisationAOR95% CIP valueYesNoMaternal education No formal education2954 (64.4%)6594 (75.7%)1 Has formal education1633 (35.6%)2118 (24.3%)1.6631.4431.9160.000Husband occupation Unemployed133 (3.1%)61 (0.8%)3.441.9366.1120.000 Employed915 (21.6%)1280 (15.8%)1.150.9751.3550.097 Farmer3189 (75.3%)6775 (83.5%)1Pregnancy intention Planned2122 (65.7%)3744 (54.5%)1.4271.2691.6030.000 Unplanned1109 (34.3%)3127 (45.5%)1Age at first marriage (in years) ≤15765 (18.1%)1565 (19.3%)0.7470.5940.8720.012 16–181064 (25.1%)2077 (25.6%)0.70.5620.8720.001 19–241854 (43.8%)3637 (44.8%)0.710.5790.8720.001 ≥24553 (13.1%)833 (10.3%)1Stillbirth ≥350 (1.4%)107 (1.6%)1.1630.7491.8050.500 1–2313 (9.0%)770 (11.4%)0.8630.7131.0440.129 03117 (89.6%)5850 (87.0%)1Cohort effect Cohort of 1970s308 (6.7%)732 (8.4%)0.8680.6761.1150.268 Cohort of 1980s1738 (37.9%)3498 (40.2%)0.9420.8191.0840.404 Cohort of 1990s2541 (55.4%)4482 (51.4%)1.00Constant1285 (28.7%)3271 (38.2%)2.2481.9852.6040.000ICC=0.292 AIC=8826.2BIC=8915.7R2=0.6563AOR, 1 reference group.AIC, Akaike’s information criterion; AOR, adjusted OR; BIC, Bayesian information criterion; ICC, intra-class correlation coefficients. Generalised linear mixed-effects model for the social determinants of prenatal care utilisation in Kersa, Ethiopia AOR, 1 reference group. AIC, Akaike’s information criterion; AOR, adjusted OR; BIC, Bayesian information criterion; ICC, intra-class correlation coefficients. Determinants of skilled delivery care service utilisation Social determinants such as household wealth status, reproductive health-related factors, such as age at first marriage, pregnancy intention and attendance at prenatal care services, as well as being born in the 1980s, were found to be significant predictors of skilled birth attendance in the multivariable generalised linear mixed-effects model. After taking into account potential confounders, it was discovered that a statistically significant association existed between household wealth status and skilled birth attendance. Women with the highest wealth status were more than two times (AOR 2.284 (95% CI 1.926 to 2.708)) more likely to use skilled delivery services than women with the lowest wealth status. Furthermore, various reproductive health factors such as the age at first marriage, pregnancy intention and attendance of prenatal care services were found to be statistically significant factors influencing the utilisation of skilled delivery services. Women who got married before the age of 16 showed a 36.7% lower likelihood (AOR 0.633 (95% CI 0.488 to 0.821)) of using skilled birth care compared with those who married at or after 24 years old. Similarly, women who married between the ages of 16 and 18 had a 32.2% reduced likelihood (AOR 0.678 (95% CI 0.529 to 0.869)), while those who married between 19 and 24 years old had a 25.9% decreased likelihood (AOR 0.741 (95% CI 0.588 to 0.934)) of using skilled birth care compared with those who married at or after 24 years old. Prenatal care service utilisation was also found to be a significant factor influencing skilled birth care utilisation. Mothers who used prenatal care were about three times (AOR 2.893 (95% CI 2.514 to 3.33)) more likely to use skilled birth care services. It was also shown that there was a significant association between using skilled delivery care and mothers born in the 1980s. In comparison to women who were born in the 1990s, those who were born in the 1980s were 31.1% less likely (AOR 0.789 (95% CI 0.681 to 0.914)) to use skilled birth care services (table 4). Table 4Generalised linear mixed-effects model for the social determinants of skilled delivery service utilisation in Kersa, EthiopiaVariablesSkilled deliveryAOR95% CIP valueUnskilledSkilledHousehold wealth status Lowest3551 (38.1%)1005 (26.9%)1.00 Medium3313 (35.5%)1154 (30.9%)1.1570.9831.3630.080 Highest2456 (26.4%)1574 (42.2%)2.2841.9262.7080.000Age at first marriage (in years) <161728 (19.5%)602 (17.3%)0.6330.4880.8210.000 16–182320 (26.1%)821 (23.7%)0.6780.5290.8690.002 19–243914 (44.1%)1577 (45.4%)0.7410.5880.9340.011 ≥24915 (10.3%)471 (13.6%)1Attended prenatal care Yes6853 (72.2%)1859 (48.8%)2.8932.5143.330.000 No2634 (27.8%)1953 (51.2%)1Stillbirth (in number) ≥36604 (88.6%)2363 (85.9%)2.031.2713.2430.003 1–2744 (10.0%)339 (12.3%)1.3731.1171.6890.002 0107 (1.4%)50 (1.8%)1Mid-upper arm circumference (cm) <233658 (50.3%)1474 (51.1%)1.0580.9281.2060.396 ≥233615 (49.7%)1409 (48.9%)1Pregnancy intention Yes4024 (55.4%)1842 (64.8%)1.2881.1291.4690.000 No3237 (44.6%)999 (35.2%)1Cohort effect Cohort of 1970s757 (8.0%)283 (7.4%)0.7940.6061.0400.094 Cohort of 1980s3854 (40.6%)1382 (36.3%)0.7890.6810.9140.001 Cohort of 1990s4876 (51.4%)2147 (56.3%)1.00Constant2.6662.2883.1960.000ICC=0.328 AIC7639.9 BIC=7736.2R2=0.9617AIC, Akaike’s information criterion; AOR, adjusted OR; BIC, Bayesian information criterion; ICC, intra-class correlation coefficients. Generalised linear mixed-effects model for the social determinants of skilled delivery service utilisation in Kersa, Ethiopia AIC, Akaike’s information criterion; AOR, adjusted OR; BIC, Bayesian information criterion; ICC, intra-class correlation coefficients. Discussion This study is the first to analyse the utilisation of maternal health services from a life course perspective using longitudinal data from three distinctive birth cohorts, analysed to compare how generational disparities could impact equity in maternal healthcare utilisation. The observed generational disparities in maternal health service utilisation in Kersa align with existing literature highlighting cohort effects in Ethiopia’s maternal health outcomes. Similar to findings from age-period-cohort analyses, younger cohorts—such as women born after 1990—tend to have better maternal health indicators, including lower maternal mortality risk and increased service utilisation, reflecting improvements in health policies and system strengthening over time. The 1970s cohort’s low utilisation mirrors earlier periods of limited healthcare access and infrastructure, consistent with historical accounts of Ethiopia’s health system prior to major reforms.4 The 1970s cohort’s low utilisation mirrors earlier periods of limited healthcare access and infrastructure, consistent with historical accounts of Ethiopia’s health system prior to major reforms.29 The 1970s cohort’s low utilisation mirrors earlier periods of limited healthcare access and infrastructure, consistent with historical accounts of Ethiopia’s health system prior to major reform. The paradoxical decline in skilled birth attendance among the 1980s cohort despite increased prenatal care suggests discontinuities in service delivery, a phenomenon also reported in studies emphasising gaps in the continuum of care in Ethiopia.6 The 1990s cohort’s higher utilisation rates likely reflect the impact of intensified maternal health initiatives, such as the Health Extension Programme, which has been linked to improved access and uptake of maternal services in recent years.6,21 However, persistent socioeconomic inequalities remain significant barriers, as documented in national-level analyses, underscoring the need for targeted interventions addressing both generational and contextual disparities to accelerate equitable maternal health gains in rural Ethiopia. The study found that generational differences have a significant impact on skilled birth care utilisation. Women who were born in the 1980s were less likely to use skilled delivery care services than those who were born in the 1990s. This generational disparity reflects multiple intersecting factors. First, educational attainment differed markedly: the 1990s cohort benefited from the 1994 education reform, with formal education rising from 14.1% (1980s) to 42.0% (1990s). Second, employment patterns shifted substantially: non-farm employment among husbands increased from 4.2% (1980s) to 30.5% (1990s), providing greater income stability, urban proximity and flexible time for facility attendance. Third, technological modernisation followed generational patterns: the 1990s cohort entered adulthood during mobile phone expansion, enabling health information access and service coordination. Earlier cohorts, raised in contexts of limited technological exposure, exhibit lower digital literacy and reduced readiness to adopt modern healthcare practices. Fourth, shifting gender norms evolved across generations, with later cohorts experiencing greater media exposure and reproductive autonomy. These intersecting factors—educational disadvantage, employment precarity, technological unfamiliarity and constrained autonomy—collectively explain the 1980s cohort’s lower skilled birth care utilisation.30,31 In this study, socioeconomic factors significantly affect skilled birth service utilisation. Women in the highest wealth category are more likely to use skilled birth care compared with those with lower wealth positions. Previous studies have also emphasised the influence of socioeconomic status on maternal healthcare, including skilled birth attendance.32–35 Lower socioeconomic background leads to challenges like limited education, lack of awareness about skilled birth care, cultural beliefs and societal norms discouraging healthcare access during pregnancy.6,36 These factors contribute to lower utilisation rates of skilled birth care services, necessitating attention to broader socioeconomic and cultural aspects, improved access to healthcare facilities, transportation, education and equitable utilisation of skilled birth care services in Ethiopia. This study found that maternal education and husband’s occupation significantly influenced prenatal care service utilisation. Educated women were more likely to use these services, while those without formal education were less likely to do so. Prior research supports the positive impact of education on maternal health service usage.37 This could be attributed to the negative effect of famine on literacy levels. For instance, the 1980 famine in Karamoja, Uganda demonstrated a significant adverse effect on the educational achievement of adults who were exposed to the famine while in utero or infancy.38 Education enhances knowledge about the importance of prenatal care and skilled birth attendance, motivating women to seek appropriate healthcare during pregnancy and labour. It also empowers women to make informed decisions about their healthcare needs39 and improves their health literacy, enabling them to understand health-related information, navigate healthcare systems and effectively engage with providers.40 Additionally, education often leads to improved socioeconomic status, granting women greater access to healthcare facilities. These factors contribute to the strong association between formal education and prenatal care use in Ethiopia.41–43 The utilisation of prenatal care attendance is significantly influenced by the occupation of husbands. Even women whose husbands are unemployed show higher rates of prenatal care utilisation compared with women in traditional farming occupations. This could be due to the fact that unemployed husbands may have more time to accompany wives to health facilities, whereas farmers face seasonal labour demands limiting mobility. Additionally, non-farm employment offers income stability and urban proximity, while farming households face seasonal income precarity. Importantly, generational employment patterns show a marked transition: among the 1990s cohort, 30.5% of husbands were in non-farm occupations compared with only 2.6% in the 1970s cohort. This employment transition, reflecting urbanisation and economic diversification, contributes to improved maternal health service utilisation in later generations.44–49 Finally, this study showed that reproductive health characteristics influence the use of prenatal and skilled birth care services. Delaying age at marriage to 24 years or older and having a planned pregnancy have a positive impact on both prenatal and skilled birth care service utilisation. This shows that getting or seeking prenatal care may be difficult for younger women. This tendency might be attributed to issues such as a lack of knowledge, inadequate resources or cultural norms. This study emphasises the need to delay early marriages, promote family planning and encourage spaced pregnancies in maintaining effective prenatal and skilled birth care. It could imply that women with fewer children may have more means and time to seek and participate in prenatal care, resulting in improved health outcomes for both mother and child.50 Strengths and limitations of the study The inclusion of women from three distinct generations provides a comprehensive life-course perspective on maternal health. However, it is important to acknowledge that this approach also comes with its own set of strengths and limitations. First, applying life course perspectives allows for an examination of how individual and societal factors shape maternal health experiences and decisions across women of different generations, as well as a holistic understanding of how early life exposures influence maternal health outcomes over time. Second, it is capable of capturing the long-term consequences of early-life exposure to the famines and political instabilities of the 1970s and 1980s, offering unique insights into the intergenerational implications for maternal health. Finally, this study provides a comprehensive understanding of the contextual factors influencing healthcare access by taking into consideration the social determinants of prenatal and skilled delivery care service consumption in Kersa, Eastern Ethiopia. Nevertheless, potential limitations should be acknowledged. First, data on maternal health service utilisation, pregnancy intention and reproductive history were self-reported during household interviews, which may introduce recall bias. However, KHDSS’s routine data collection every 6 months likely reduces recall error compared with single cross-sectional surveys. Second, social desirability bias may have led women to over-report utilisation of maternal health services, potentially inflating coverage estimates, particularly among later cohorts. Third, participant attrition may arise, as maintaining long-term engagement over an extended period can be challenging. Despite these limitations, this longitudinal study provides important insights into the complex interaction between early-life exposures, societal variables and maternal health outcomes in Eastern Ethiopia. Yet, this longitudinal study provides pivotal evidence for policymakers and healthcare practitioners seeking to design targeted interventions and enhance maternal health services. While our findings highlight generation-specific barriers, we do not advocate for exclusively cohort-based interventions. Instead, we recommend policies that combine both perspectives: addressing generation-specific historical and socioeconomic barriers while ensuring age-appropriate maternal health services. Birth cohort captures shared historical experiences due to modernisations, while age classification reflects biological and life-stage needs. The two concepts are complementary rather than interchangeable. Conclusion This study highlights entrenched generational inequities in maternal healthcare utilisation, driven by intersecting disparities in educational opportunity, employment patterns and technological modernisation. Key determinants include maternal and spousal education, household wealth, age at first marriage, pregnancy intention and generational acceptance of modern healthcare. The 1980s cohort’s paradoxical decline in skilled birth care highlights how intermediate generations face unique barriers despite overall policy improvements. Addressing these inequities requires multisectoral policies that prioritise girls’ education, livelihood diversification beyond agriculture and bridging the digital divide through technology-inclusive health interventions. Interventions should combine digital health innovations with community-based approaches suitable for earlier cohorts. By addressing these determinants, Ethiopia can advance toward equitable maternal healthcare utilisation across generations, ensuring progress benefits all women regardless of birth cohort. Future research should explore longitudinal mechanisms linking generational shifts to healthcare behaviours and evaluate interventions tailored to generation-specific barriers. Acknowledgements The Kersa Health and Demographic Surveillance System (HDSS) data managers, along with other supporting staff, deserve recognition for their efforts in maintaining this longitudinal data capturing vital events for many years. Footnotes Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors. Patient consent for publication: Not applicable. Ethics approval: The National Ethical Review Commission and the College of Health at Haramaya University in Ethiopia have already granted Kersa Health and Demographic Surveillance System (KHDSS) its ethical approval. Furthermore, Jimma University’s Institutional Review Board in Ethiopia provided ethical permission for this specific research project (reference No: JUIH/IRB/113/22). Participants gave informed consent to participate in the study before taking part. In line with the International Ethical Guidelines for Health-related Research Involving Humans (CIOMS/WHO, 2016) and applicable national regulations, informed consent was not required for this study, as it relied on surveillance and anonymized data. Provenance and peer review: Not commissioned; externally peer reviewed. Patient and public involvement: Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research. Data availability statement Data are available upon reasonable request and from the corresponding author. References Associated Data Data Availability Statement Data are available upon reasonable request and from the corresponding author.