About one in 20 infants worldwide is subjected to physical abuse by a caregiver in their first two years of life.
That's the central finding of a new study co-led by researchers from the UBC faculty of medicine and Memorial University of Newfoundland (MUN), the first to bring together anonymous reports from caregivers about behaviors like spanking, slapping, shaking and hitting.
About four to five per cent of parents are behaving in physically aggressive ways toward their babies. This should give us all pause and make us think about interventions that could shift behavior."
Dr. Nichole Fairbrother, clinical associate professor in UBC's department of family practice and head of the Perinatal Anxiety Research Lab
Infants are entirely dependent on their caregivers, making them especially vulnerable to physical aggression that can arise in many circumstances-sometimes in moments of stress for overtired caregivers, but also within families facing other challenges such as mental health difficulties or existing violence in the home.
While some still view spanking or wrist slapping as 'discipline', such behaviors can have lasting effects on a child's learning, behavior and mental health. Evidence has linked these actions with harm and with a higher chance of more severe violence in the future. Understanding how common these kinds of behaviors are could help health systems, policymakers and families focus on prevention, getting support to parents before a bad moment turns into injury.
Dr. Fairbrother, co-lead author Cora Keeney, MUN's Dr. Jonathan Fawcett and their collaborators reviewed 20 studies that together encompassed reports on more than 220,000 infants and caregivers across several countries. When caregivers were asked privately and anonymously, 4.8 per cent reported at least one act of physical aggression toward an infant under 24 months. Rates were slightly lower when "milder" actions like spanking were excluded (3.9 per cent), but the picture stayed the same: This is common, and mostly hidden.
The study helps explain why official statistics often look lower. Most child‑protection records capture only the most serious incidents that someone else notices and reports.
"Anonymous surveys let caregivers disclose behavior that never reaches a clinic, police service or social worker," said Dr. Jon Fawcett, an assistant professor at MUN who co-led the research. "By pooling these studies, we can estimate what's really happening at a population level, not just what gets detected."
The researchers also looked at specific actions. Shaking occurred at a low yet noteworthy prevalence, around two to three per cent in estimates. Spanking was more common in some samples.
The authors point to a mix of practical supports and policy steps that could help address the issue. Parenting an infant is stressful-sleep loss, constant crying, money worries-and stress can push even calm caregivers into moments of frustration when they react in ways they normally wouldn't. Simple, well‑timed education for new parents such as what to do when crying peaks, easy‑to‑reach support lines and home‑visiting programs could help.
Clear policies that set the norm against hitting children also matter. Countries that prohibit physical punishment send a consistent message that violence has no place in caregiving.
This area of study still needs better and broader data, especially from regions that are under‑represented. More studies would also allow researchers to explore important details, such as whether rates differ by a baby's age or by caregiver characteristics, so prevention efforts can be targeted where they're most needed.
"Infancy is the most vulnerable stage of life, and many cases of physical aggression never show up in official counts," said Dr. Fairbrother. "Knowing the true scope is the first step toward stopping it."
The research was published Wednesday in eClinicalMedicine, a Lancet journal.
Source:
Journal reference:
Keeney, C. L., et al. (2026). Prevalence of physical abuse by caregivers toward infants aged 0–24 months: a systematic review and meta-analysis. eClinicalMedicine. DOI: 10.1016/j.eclinm.2026.103812. https://www.thelancet.com/journals/eclinm/article/PIIS2589-5370(26)00059-3/fulltext