Evolution of the risk of death and hospitalisation in drivers involved in road crashes in spain, 1993-2020: an age-period-cohort analysis.

Background: A prerequisite for understanding temporal changes in road crash severity is an unbiased description of this phenomenon. The aim of this study was to estimate the independent association trends of age, period and cohort with severity, encompassing the risk of death (RD) and the risk of death or hospitalisation (RDH) within 24 h, for drivers of passenger cars involved in road crashes with casualties in Spain from 1993 to 2020.

Methods: The study population comprised 2,453,911 drivers of passenger cars aged 18 to 98 years involved in road crashes included in the registers of the General Directorate of Traffic. Crash- and driver-related variables with sufficient continuity over time were included, establishing RD and RDH as study outcomes. Temporal trends of both outcomes were analysed using multivariable Poisson regression and multivariable age-period-cohort intrinsic estimator models. An additional sensitivity analysis was performed for the subset of single crashes.

Results: Severity estimates showed some variation across strategies. The APC model identified: (1) a J-shaped pattern for the effect of age on severity, (2) a decline in severity between 2001 and 2004 and 2013-2016, and (3) a birth cohort effect for both RD and RDH. In particular, the 1952-1958 cohort had the highest risk (RD = 1.17; 95%CI = 1.11-1.24 and RDH = 1.16; 95%CI = 1.13-1.19), followed by a decreasing trend in subsequent cohorts. Restricting the analysis to single crashes yielded similar results, with the exception of the age effect (severity increased with age). Furthermore, sex differences were observed-female sex was inversely associated with severity, especially for RD.

Conclusions: RD and RDH decreased during the first decade of the 21st century, but seemed to stabilise from 2013 onwards. Evidence from this study support that birth cohort is associated with road crash severity, independent of age and period. This cohort effect might be due, at least partially, to improvements in general and road safety education. Further studies are needed to elucidate the causes of our findings and to identify factors accounting for sex differences.