Evaluation of road infrastructure in Yerevan, Armenia through the lens of pedestrian and child safety: a school-based analysis.

INTRODUCTION: Traffic-related injury is a leading cause of morbidity and mortality globally, especially among children. In Armenia, traffic-related mortality rates are the second highest among European Union and former Soviet countries. Currently, there are no published studies on pedestrian safety in Armenia and no notable examinations of pedestrian safety infrastructure. This manuscript describes an audit of Armenia's infrastructure using a modified Inventory for Pedestrian Safety Infrastructure (IPSI) in the Yerevan city center, focusing on areas surrounding schools. METHODS: Intersections (n = 379) and roadways (n = 623) shared by pedestrians and drivers in Yerevan's center were audited using a 36-item modified ISPI and paired with cloud-based smartphone application QField between October 2023 and February 2024. Visual analysis of data was conducted using the open-source software QGIS with statistical analysis in R. Inter-rater reliability of the audit was calculated using Cohen's Kappa and intra-class correlation coefficients. Incidence of each feature surrounding schools was compared to incidence overall using a one-sample z-test for continuous variables and a two-sample test of proportions for binary variables. RESULTS: Many of the features audited were not observed consistently across the audited area and were not more prevalent near schools. Lower inter-rater reliability was found with roadway and midblock data, resulting in analysis of primarily intersection features. Speed limits surrounding schools varied from 20 to 60 km/hr (12.5-37 mph). When compared to the overall audited area, crosswalks near schools demonstrated similar occurrence for traffic lights with crosswalk markings (23.3%, 95% CI: 0.157-0.309), speed bumps (11.9%, 95% CI: -0.139- 0.0338), and pedestrian signals (12.9%, 95% CI: 0.0709-0.194). Of the analyzed items, only bus stops were observed significantly more often near schools (19%, 95% CI: 0.00930-0.220). CONCLUSIONS: The audit revealed an inconsistent presence of safety infrastructure, with infrastructure near schools largely reflecting the overall audited area. However, analysis was limited by low inter-rater reliability for many roadway and midblock features audited. Based on these results, infrastructure changes targeting lower speed limits and increased traffic calming measures are recommended to improve safety in school zones. Given a paucity of studies in LMICs, further work must be done to identify which high-yield features should be targeted in LMICs to prevent injury.