Nearly 5% of Infants Face Physical Aggression From Caregivers, Global Review Finds

One in 20. That is the proportion of infants worldwide who experience physical aggression from a caregiver before their second birthday, according to a new meta-analysis published in eClinicalMedicine, a Lancet journal.

The number comes not from hospital records or child protection files, but from anonymous surveys of caregivers themselves -- the first systematic effort to pool these self-reports across multiple countries and estimate what is actually happening behind closed doors.

What the official numbers miss

Child protection records capture only the most severe cases -- the ones serious enough that someone else notices and reports them. The vast majority of physical aggression toward infants, from spanking and slapping to shaking and hitting, goes undetected by any official system.

Anonymous surveys change that equation. When caregivers are guaranteed confidentiality, a clearer picture emerges. Across 20 studies encompassing reports on more than 220,000 infants and caregivers in several countries, 4.8% of caregivers reported at least one act of physical aggression toward an infant under 24 months.

When milder actions like spanking were excluded, the rate dropped to 3.9% -- still a substantial figure for a population entirely dependent on its caregivers. Shaking, which carries particular neurological risks for infants, occurred at a rate of roughly 2 to 3%.

Stress, not malice, drives many incidents

The researchers, co-led by Nichole Fairbrother of the University of British Columbia and Jonathan Fawcett of Memorial University of Newfoundland, emphasize that physical aggression toward infants often arises in moments of extreme caregiver stress rather than from deliberate cruelty. Sleep deprivation, persistent crying, financial pressure, and lack of support can push even calm caregivers past their limits.

That context matters for intervention design. If most incidents stem from overwhelmed parents rather than abusive ones, the most effective responses are likely to be supportive rather than punitive: education about infant crying patterns and how long they last, accessible support lines, and home-visiting programs that provide practical help.

The evidence on harm is clear

While some cultures still view certain forms of physical contact -- a slap on the wrist, a spank -- as acceptable discipline, the evidence on consequences is consistent. Physical aggression toward infants has been linked to lasting effects on learning, behavior, and mental health. It is also associated with a higher probability of escalation to more severe violence over time.

Countries that have enacted explicit prohibitions on physical punishment of children send a clear normative signal, and the researchers point to such policies as one component of a prevention strategy.

Major data gaps remain

The study comes with significant limitations. Many regions of the world are underrepresented in the available data. The 20 studies varied in methodology and definitions. The researchers could not reliably break down rates by infant age or caregiver characteristics due to inconsistent reporting across studies.

These gaps matter. Without broader geographic coverage and more granular data, it is difficult to target prevention efforts where they are most needed. The 4.8% figure is a global estimate, and actual rates likely vary considerably across countries, cultures, and socioeconomic contexts.

Fawcett noted that pooling anonymous surveys provides a population-level estimate that complements -- rather than replaces -- clinical and child protection data. Neither source alone captures the full picture.

A starting point, not an endpoint

Infancy represents the most vulnerable period of human life, and physical aggression during this window can have consequences that reverberate across development. The study does not claim to have identified every contributing factor or the most effective intervention. What it provides is a baseline: a defensible estimate of how widespread the problem is, grounded in the one source of information -- anonymous self-reporting -- most likely to capture behaviors that otherwise remain invisible.