Electronic supplementary material
The online version of this article (doi:10.1186/s12889-015-2609-1) contains supplementary material, which is available to authorized users.
The authors declare that there are no competing interests, financial or non-financial.
This project was conceptualized by CS, VS, EK, JV and LA. Data gathering was performed by VS, EK, SR, BM. Data analysis and interpretation was performed by CS, VD, EK, LA, KJ, JV and TO. Manuscript development was performed by CS, VD, LA, SR, BM, JV and TO. All authors edited and approved the final manuscript.
Road traffic crashes (RTCs) are a leading cause of death. In low and middle income countries (LMIC) data to conduct hotspot analyses and safety audits are usually incomplete, poor quality, and not computerized. Police data are often limited, but there are no alternative gold standards. This project evaluates high road utilizer surveys as an alternative to police data to identify RTC hotspots.
Retrospective police RTC data was compared to prospective data from high road utilizer surveys regarding dangerous road locations. Spatial analysis using geographic information systems was used to map dangerous locations and identify RTC hotspots. We assessed agreement (Cohen’s Kappa), sensitivity/specificity, and cost differences.
In Rwanda police data identified 1866 RTC locations from 2589 records while surveys identified 1264 locations from 602 surveys. In Sri Lanka, police data identified 721 RTC locations from 752 records while survey data found 3000 locations from 300 surveys. There was high agreement (97 %, 83 %) and kappa (0.60, 0.60) for Rwanda and Sri Lanka respectively. Sensitivity and specificity are 92 % and 95 % for Rwanda and 74 % and 93 % for Sri Lanka. The cost per crash location identified was $2.88 for police and $2.75 for survey data in Rwanda and $2.75 for police and $1.21 for survey data in Sri Lanka.
Surveys to locate RTC hotspots have high sensitivity and specificity compared to police data. Therefore, surveys can be a viable, inexpensive, and rapid alternative to the use of police data in LMIC.