Citation

Volpe FM, Naves Barbosa Lacerda DR. Lancet Reg. Health Am. 2024; 35: e100800.

Copyright

(Copyright © 2024, Elsevier Publishing)

DOI

10.1016/j.lana.2024.100800

PMID

38911347

PMCID

PMC11192789

Abstract

The recent paper by Alves et al. has unveiled a remarkable increase in self-harm notifications in Brazil from 2011 to 2022. Adding to their findings, we propose a reanalysis using a different approach than Alves et al., to improve the growth curve model and assess the impact of the COVID-19 pandemic on self-harm notifications.

Using the same dataset extracted from SINAN/DATASUS, we aggregated data by month instead of annually intending to revise the accuracy of the estimations. Raw frequencies were used rather than rates, since population data are not available on monthly basis and its growth curve cannot be precisely modelled.

We then conducted a segmented regression analysis using an exponential model, with March-2020 dividing the time-series before and after the COVID-19 pandemic. The exponential model showed a better fit than the log-linear equation, particularly in the second period. Segmented regression is superior to joinpoint when the timing of the suspected change is known (the official start of the pandemics in Brazil), because it provides estimates not only for the change in trends, but also for the immediate impact of the event.

Self-harm notifications in Brazil followed an exponential growth between 2011 and 2020...

Language: en