Individual and city-level variations in heat-related road traffic deaths in Latin America

Abstract

Road-traffic mortality and extreme heat are two major urban health challenges, increasingly found to be associated. However, few studies have examined this association in Latin America-one of the world's most urbanized, fastest-motorizing region, with a high share of vulnerable road users-and even fewer have analyzed multiple cities across diverse climates and urban settings. Using temperature and road-traffic mortality data (2000-2019) from 272 cities in six Latin American countries, we conducted a time-stratified case-crossover study. The relative risks (RRs) of road-traffic mortality at the 95th and 99th temperature percentiles, compared to the minimum mortality temperature percentile, were 1.16 [95% CI: 1.14-1.19] and 1.18 [1.15-1.21], respectively. Risks were particularly high among adolescents, males, motorcyclists, bicyclists, and in cities with hotter climates and longer commutes. Policymakers in the tropical Global South should prioritize protecting vulnerable road users in peripheral communities, where many endure long, heat-exposed commutes in non-climate-controlled informal transport.

Fig. 1.. Association between temperature conditions and road-traffic mortality (main effects model).

(a) Relative temperature measure (temperature percentiles). (b) Absolute temperature measure (°C). (c) Distribution of temperature percentiles. (d) Distribution of absolute temperatures (°C). Dashed vertical lines indicate temperature percentiles: 1^st, 25^th, 50^th, 75^th, and 99^th from left to right. Shading in each association curve represents the 95% confidence interval.

Fig. 2.. Associations between temperature and road-traffic mortality by temperature cluster.

(a) Temperature distribution (°C) for each cluster. (b)-(g) Association between temperature percentiles (x-axis) and road-traffic mortality for each temperature cluster. N indicates number of cities in each cluster. Dashed vertical lines in figs. b through g indicate temperature percentiles: 1^st, 25^th, 50^th, 75^th, and 99^th from left to right. Shading in each association curve represents the 95% confidence interval.

Fig. 3.. Sub-group analysis results. Relative risk of road mortality at each temperature percentile in reference to the 1^st temperature percentile.

(a) Sex. (b) Age. (c) Mode of transportation. Dashed vertical lines indicate temperature percentiles: 1^st, 25^th, 50^th, 75^th, and 99^th from left to right. Shading in each association curve represents the 95% confidence interval.

Fig. 4.. Effect modification from city-level transportation-related factors.

(a) Average street-segment length. (b) Average peak-hour travel time. High (red) and low (blue) indicate cities at the 90^th percentile (higher values) or 10^th percentile (lower values) of respective variable values. Dashed vertical lines indicate temperature percentiles: 1^st, 25^th, 50^th, 75^th, and 99^th from left to right. Shading in each association curve represents the 95% confidence interval.