Self-reported drink driving, enforcement, crashes, and crash reporting: A 6-country comparison

Abstract

Objective

The objective of this study was to compare drink driving and related road safety issues in 2 urban areas of 6 countries and develop an equation for estimating the rate of crash underreporting to the police in urban areas of countries that lack this information.

Methods

This study is a secondary analysis of 1 to 2 waves of surveys in pairs of matched medium-sized cities in Belgium, Brazil, China, Mexico, South Africa, and Ohio, United States; the surveys supported evaluation of local alcohol harm reduction efforts. Data were from 2017 to 2019 except 2023 for Mexico. Mailed surveys in Ohio and household interviews elsewhere of quota samples matched to census data yielded 23,240 completed interviews. Relevant questions covered drinking, driving under the influence of alcohol (DUI), DUI enforcement, and, except in South Africa, road crashes. GLM regression provided an equation for estimating police reporting rates of urban injury and no-injury crashes from a country’s purchasing-power parity–adjusted gross domestic product (GDP) per capita.

Results

The percentage of drivers driving unlicensed was 30% in Mexico and South Africa, 15% in Brazil, 8% in China, and <1% elsewhere. Among adults who both drove and drank, self-reported urban DUI rates ranged from 12% in China to 53% in South Africa, with 4 countries between 18% and 26%. Among those reporting DUI, the percentage stopped by police for doing so was 14% in Belgium, 15% in Brazil, 25% in China, 31% in Mexico, 45% in South Africa, and only 3% in Ohio. The surveys yielded data on 380 urban crashes. Past-year crash involvement was 2% to 3% in Belgium and China and 5% to 6% elsewhere. The 10% injury rate in Ohio crashes was significantly below the 24% to 35% rates elsewhere. Injury crashes were almost universally reported except in Brazil (60% reported). Only 49% to 56% of non-injury crashes were reported, except in Ohio (73%). Perceived alcohol-involved crash rates of 18% to 19% in Belgium and Ohio were significantly lower than the 32% reported in Brazil, 41% in China, and 57% in Mexico. In the regression, GDP per capita and injury involvement were positively associated with police crash reporting.

Conclusions

Our equation more closely approximates urban police crash reporting rates than prior studies that assumed that they matched U.S. data. DUI enforcement is weak/ineffective in urban Ohio. With suggested adjustments, our survey questions should be usable in other international road safety and DUI studies.