Female homicides in Brazil before and during the COVID-19 pandemic: an interrupted time-series analysis

TL;DR

This study examines how the COVID-19 pandemic affected female homicide rates in Brazil, finding a significant increase in the Northeast region followed by a decline.

Contribution

The study provides novel insights into the impact of the pandemic on female homicides in Brazil, using an interrupted time-series design.

Findings

Female homicide rates increased by 16.0% after correcting for misclassification in Brazil from 2017 to 2022.

The Northeast region experienced a significant abrupt increase in homicides during the pandemic followed by a gradual decline.

Domestic homicides were more common than those in public spaces, with firearms being the primary method.

Abstract

The COVID-19 pandemic has influenced violence against women worldwide, but its impact on female homicides in low- and middle-income countries remains poorly understood. Brazil, which records some of the highest female homicide rates globally, provides a critical setting to examine this association. This study assessed the temporal association between the pandemic and monthly female homicide rates in Brazil from January 2017 to December 2022.

We applied an interrupted time series (ITS) design with quasi-Poisson regression to estimate changes in homicide levels and trends after the pandemic onset, adjusting for serial autocorrelation and seasonality. Pre-pandemic trend linearity was tested, and sensitivity and placebo analyses were performed. To address underreporting and deaths classified as undetermined intent, homicide counts were corrected for misclassification.

From 2017 to 2022, 23,727 female homicides were recorded, corresponding to an adjusted mortality rate of 5.09 per 100,000 women, a 16.0% increase after correction. Rates were highest in the North and Northeast. Domestic homicides exceeded those in public spaces (1.50 vs. 1.37 per 100,000 women), and firearms were the predominant method. The Northeast showed a significant level change with an abrupt increase (RR = 1.236; p = 0.002), followed by a progressive decline (RR = 0.9923; p < 0.001). In other regions, across age groups, and in blunt-related cases, no significant level change occurred (p > 0.05), although downward trends emerged during the pandemic (p < 0.05).

Findings warrant cautious interpretation due to ITS constraints, sensitivity to the observation window, and omitted variables. Nonetheless, persistently high female homicide rates in Brazil, particularly in the Northeast, highlight the need to strengthen mortality surveillance, improve misclassification corrections, and adopt region-specific prevention strategies, including firearm control, protective services, and targeted social policies.

The online version contains supplementary material available at 10.1186/s12889-025-24814-6.

Figures