One in Twenty Babies Died in Tigray’s War, Landmark Cohort Study Reveals

In one of the most comprehensive assessments of child survival ever conducted in an active conflict zone, researchers have documented the devastating toll that two years of war took on the newest residents of Ethiopia’s Tigray region. A retrospective cohort study published in BMC Pediatrics followed more than 30,000 live births across the war-torn territory and found that roughly one in every twenty children died before reaching eighteen months of age. The findings, drawn from a large community-based mortality survey spanning 121 local administrative units across 31 districts, offer the clearest quantitative picture yet of how armed conflict reshapes the odds of survival during the most vulnerable window of human life, and they identify with unusual precision which children faced the greatest danger.

The scale of the data collection alone makes the study remarkable. The research team, led by Brhane Ayele of the Tigray Health Research Institute together with colleagues from Mekelle University, the Tigray Regional Health Bureau, and the University of South Carolina, analyzed records for 30,935 live births. The first eighteen months of life is widely recognized as a critical period for long-term survival, yet robust evidence on mortality during this window in conflict-affected settings has been persistently scarce. Health systems in Tigray collapsed under the strain of fighting that began in November 2020, leaving clinics nonfunctional, supply chains severed, and families displaced into camps or hiding. Against that backdrop, the study provides something rare: a statistically rigorous, population-wide measurement of what happened to children born into catastrophe.

The headline number is stark. Of the 30,935 live births tracked, 1,430 children died before turning one and a half years old, a mortality proportion of 4.6 percent with a 95 percent confidence interval of 4.4 to 4.9 percent. Even more striking is the timing of those deaths: three-quarters occurred during the neonatal period, the first four weeks of life. This concentration in the earliest days points directly to the machinery of safe childbirth as the broken link. When hospitals and health centers close, when skilled birth attendants disappear, and when roads become impassable, the moment of delivery itself becomes the deadliest passage a child faces. The mean survival time among the cohort was 524.69 days, with a confidence interval of 523.41 to 525.97 days, a figure that conceals the sharp clustering of deaths in the opening weeks.

To understand which children were most at risk, the researchers deployed the standard analytical toolkit of survival analysis. Kaplan-Meier curves were used to estimate survival probabilities over time, while log-rank tests compared differences between groups defined by maternal, household, and geographic characteristics. The team then fitted a Cox proportional hazards model, the workhorse of epidemiological time-to-event analysis, to estimate adjusted hazard ratios for each potential predictor while controlling for the others. A hazard ratio above one indicates elevated risk of death at any given moment; the study declared statistical significance at a p-value below 0.05, and all analyses were performed in SPSS. This layered approach separates genuine risk factors from mere correlations, and the results are sobering in their clarity.

The single most dramatic predictor was where and how a child was born. Babies delivered at internally displaced persons camps or on the way to a health facility faced a 4.27-fold higher hazard of death compared with those born in safer conditions, with a confidence interval of 3.08 to 5.92. Children born in the bush, away from any medical infrastructure, carried a 1.74-fold higher hazard, ranging from 1.36 to 2.23. These numbers translate an abstract statistic into a concrete picture: a mother in labor fleeing her home, delivering in a tent or a field, without sterile equipment, emergency obstetric care, or neonatal resuscitation, faces odds that her child will not survive that are several times worse than a facility birth would have offered. The finding underscores that the collapse of institutional delivery services was not a peripheral casualty of the war but a central driver of infant death.

Reproductive and household characteristics also carried enormous weight. Multiple pregnancies, meaning twins or higher-order births, were associated with a 4.91-fold increase in the hazard of death, with a confidence interval of 4.02 to 6.00, reflecting the intrinsic biological vulnerability of multiple births compounded by the absence of specialized neonatal care. Short birth intervals, the spacing of pregnancies too closely together, emerged as the strongest single predictor in the entire model, raising the hazard of death by 6.96 times, with a confidence interval of 5.89 to 8.24. This finding aligns with decades of demographic research showing that depleted maternal recovery time between pregnancies compromises both fetal development and newborn resilience. In a war zone where contraception and family planning services evaporated, the biological consequences of uncontrolled fertility were amplified into a lethal risk multiplier.

Geography and social structure left their own fingerprints on the survival curves. Children living in rural areas faced a 1.37-fold higher hazard of death than their urban counterparts, with a confidence interval of 1.16 to 1.62, a gap that reflects both the greater destruction of rural health infrastructure and the longer distances to any functioning care. Districts classified as having an extremely very high level of maternal mortality transition showed a 1.33-fold elevated hazard, from 1.04 to 1.70, indicating that the places where mothers were dying in greatest numbers were also the places where their babies were most likely to die. Residence outside Mekelle, the regional capital, was independently associated with higher mortality, as was male infant sex and membership in smaller families, the latter possibly reflecting reduced household support networks and caregiving capacity during displacement and scarcity.

The methodological foundation of the study deserves attention because it determines how much weight the findings can bear. As a retrospective cohort built on a large community-based survey of under-five mortality, the study captured births and deaths reported by household members rather than relying on hospital records that would have missed the vast majority of deliveries occurring outside facilities. Ethical clearance was obtained from the Mekelle University College of Health Sciences Research and Community Service ethical review board, and informed verbal consent was collected from participants, with adult household caregivers responding when mothers were absent. The primary survey received partial funding from UNICEF and UNFPA, though the funders played no role in data collection, analysis, or interpretation. The breadth of coverage, spanning 121 tabiyas across 31 districts, gives the estimates a population-level authority that smaller facility-based studies in conflict zones have lacked.

The authors frame their conclusions as a call to immediate, targeted action rather than a lament. Their recommendations include expanding community-based delivery services so that skilled care reaches women where they are, strengthening referral systems to move obstetric emergencies to functioning facilities, and promoting safe facility-based births through maternity waiting homes and transport support for mothers in remote areas. They further argue that conflict-sensitive health planning should be institutionalized, with equity-driven resource allocation that prioritizes rural and conflict-affected districts rather than defaulting to urban centers. Integrating culturally appropriate family planning services addresses the short birth interval finding directly, while establishing robust mortality surveillance systems would ensure that future crises are measured in real time rather than reconstructed afterward.

Beyond its immediate regional significance, the study contributes to a growing scientific literature on how armed conflict reshapes demographic risk. The concentration of deaths in the neonatal period, the outsized hazards attached to place of birth, and the powerful effect of birth spacing together sketch a coherent causal architecture: war kills newborns primarily by destroying the systems that make childbirth safe. For humanitarian agencies, the message is that neonatal survival in conflict settings is not primarily a nutrition problem or a vaccination problem, at least not in the first weeks of life, but an obstetric access problem. For epidemiologists, the demonstration that a Cox proportional hazards framework can extract actionable risk profiles from community survey data in a war zone offers a template for future studies in Ukraine, Sudan, Gaza, and other theaters where children are being born into violence. The 1,430 children whose deaths anchor this analysis represent a measured, verifiable accounting of loss, and the predictors identified give the region, and the world, a precise map of where to intervene first when the next crisis comes.

Subject of Research: Child survival and predictors of time-to-death among live births during the Tigray war in Ethiopia

Article Title: Survival status and predictors of time-to-death among live-birth children during war time in Tigray: A retrospective cohort study

Article References: Ayele, B., Teka, H., Legesse, A. Y., Tsadik, M., Gebrekurstos, G., Abraha, H. E., Berhe, B., Fisseha, G., Ebrahim, M. M., Hailu, A. G., Gebremeskel, T., Gebremariam, T., Hadush, M. Y., Hagos, T., Muez, K., Tsegay, H., Mulugeta, A., & Godefay, H. (2026). Survival status and predictors of time-to-death among live-birth children during war time in Tigray: A retrospective cohort study. BMC Pediatrics. https://doi.org/10.1186/s12887-026-07797-0

Image Credits: AI Generated

DOI: 10.1186/s12887-026-07797-0

Keywords: child mortality, Tigray war, Ethiopia, neonatal mortality, survival analysis, Cox proportional hazards, retrospective cohort, maternal health, conflict health, birth interval, internally displaced persons, BMC Pediatrics