Work

Binomial vs. Poisson distribution to determine the weekend effect on completed suicide.

Year: 2026

Type: article

Abstract: Objective To compare weekday and weekend suicide frequency in Colombia using Binomial and Poisson distributions to evaluate the weekend effect. Method A cross-sectional study was conducted using public administrative records of completed suicides (CSs) from 2015 to 2023, analyzing personal, temporal, and geographic variables. The frequency of CSs on weekdays and weekends was measured by comparing proportions and daily rates using Binomial (π, 95% CI) and Poisson (λ, 95% CI) distribution models. Results A total of 23.544 CSs were analyzed; ~75% occurred during adulthood, ~80% occurred in men, and ~ 90% occurred in people with a high-school education or less. Although the weekend effect (WE) was not demonstrated over the nine-year period when applying the Binomial distribution, it was identified in four features: early adulthood, basic or high-school education, the Amazon geopolitical area, and common-law marital status (p < 0.05). Additionally, only five of the 33 departments showed a higher frequency of CSs on weekends (p < 0.05). Using the Poisson distribution, the WE was identified for each year of the study period, for 17 features (personal, temporal, and geographic variables), and for 16 departments (p < 0.05). Conclusions The WE in CSs was easier to detect using the Poisson distribution. This methodology could support the reorganization of mental health services and could be applied to identify the WE in other health outcomes.

Source: F1000Research

DOI: 10.12688/f1000research.179610.1