Partner violence against women is tied to higher child mortality, review finds

Intimate partner violence against women is one of the most common forms of violence worldwide, and a systematic review published on 2 September in JAMA Network Open adds a downstream harm to the count: the children of women who experience it appear to die before their fifth birthday at somewhat higher rates [s1]. The association is real but, for the broadest measure, modest — a distinction the paper is careful to keep in view.

What the researchers did

The authors searched seven databases from inception to 22 April 2026 for studies linking any form of intimate partner violence (IPV) — physical, sexual, emotional, or controlling behaviour — to death before age five, in infancy, or in the neonatal period [s1]. Twenty-eight studies covering 684,065 participants met the criteria for the systematic review, and 25 were pooled in random-effects meta-analyses [s1]. Risk of bias in the underlying observational studies was assessed formally, a step that matters because none of this evidence comes from experiments — you cannot randomise women to abuse [s1].

What they found

Overall, IPV was associated with higher odds of under-5 mortality (odds ratio, 1.10; 95% CI, 1.05–1.16), infant mortality (OR, 1.32; 95% CI, 1.11–1.57) and neonatal mortality (OR, 1.36; 95% CI, 1.19–1.55) [s1]. The under-5 figure is the smallest of the three, and its confidence interval sits close to 1.0 — a 10% relative increase in odds, not a dramatic one.

The signal was sharper the earlier the death and the more severe the violence. For neonatal death, physical IPV carried an odds ratio of 1.37 (95% CI, 1.17–1.61), sexual IPV 1.70 (95% CI, 1.33–2.17) and emotional IPV 1.29 (95% CI, 1.07–1.56) [s1]. Physical IPV was also associated with higher odds of infant death (OR, 1.20; 95% CI, 1.01–1.44) [s1]. Notably, when the analysis was narrowed to specific IPV subtypes and the broadest outcome — under-5 mortality — no statistically significant association survived [s1]. That inconsistency is a caution against over-reading any single number here.

How to read this

An association of this kind can arise through several routes at once, and the study design cannot separate them. Violence during pregnancy is linked to poor antenatal care, injury, stress and preterm birth; households where a mother is abused may also be poorer, with less access to safe delivery and newborn care. The meta-analysis pools findings from very different countries and health systems, and pooled estimates around a small effect can hide wide variation between settings. Observational data can establish that two things move together; they cannot, on their own, prove that one causes the other.

What the consistency across the neonatal estimates does suggest is that the earliest weeks of life are where any effect concentrates — the period most sensitive to a mother's health, safety and circumstances during and immediately after pregnancy.

Why it matters

Roughly one in three women worldwide is estimated to experience physical or sexual partner violence in her lifetime, a burden this site has covered through the World Health Organization's global estimates. This review reframes that burden as, potentially, a child-survival issue as well as a women's-health one — the authors argue that addressing IPV belongs inside child-health strategies, not only in violence-prevention programmes [s1].

The practical implication is about where to look, not what any individual should do. If the association is causal even in part, then antenatal and newborn-care contacts — moments when health systems already reach pregnant women — are also opportunities to identify and support women at risk. Whether acting on that actually lowers child mortality is a question this study raises rather than answers.

This article is informational and does not constitute medical advice. Anyone affected by partner violence can seek help from local support services.

Sources

• [s1] Intimate Partner Violence Against Women and Mortality Among Children Younger Than 5 Years: A Systematic Review and Meta-Analysis. JAMA Network Open, 2 September 2026. https://doi.org/10.1001/jamanetworkopen.2026.31794