OHS - Investigation of the Accuracy of Alcohol and Drug Involvement Reporting

Through an analysis of the fatal crashes in 2016, it was found that of the 114 fatal crashes where a police officer indicated alcohol as a contributing human factor, only 95 were shown to have a driver that tested positive for alcohol in the FARS database. The FARS database shows 160 fatal crashes with drivers testing positive for alcohol, meaning police officers failed to identify the remaining 65 crashes. A similar trend is seen with the reporting of drug involvement. FARS shows 266 fatal crashes in 2016 where drivers tested positive for drugs, however police officers only identified 52 crashes with drugs as a human factor. Of those 52 crashes, only 37 showed a driver-testing positive for drugs in the FARS database. This large inconsistency between the alcohol and drug indication in fatal crashes between the FARS and KSP databases indicates an inaccuracy of the reporting of alcohol and drugs as a contributing human factor for not just fatal crashes, but crashes of all severities. It is important that the KSP crash database accurately reflect the number of crashes involving alcohol and drugs in order for researchers to determine the most effective crash countermeasures. When coding crashes, it is not always apparent to officers if alcohol or drugs are a factor in the crash. Certain combinations crash elements besides human factors could indicate a crash likely involved alcohol or drugs. This project proposes to investigate the crash types (run off road, head-on, etc.), crash narratives, and contributing factors (time of day, distraction, etc.) associated with known alcohol and drug related crashes to develop an algorithm to assist in determining an accurate indication of all drug and alcohol crashes. The algorithm would be based on FARS crashes with known alcohol and drug use, and applied to all crashes to confirm the likelihood of alcohol or drugs being involved. In addition, the project will identify locations/roadway features where alcohol and drug crashes are most prevalent (intersections, rural 2 lanes, etc.) to identify countermeasures to combat drunk/drugged driving crashes. 1. To identify the percent of crashes fatal correctly identified to have alcohol/drug involvement using the FARS database 2. To develop an algorithm based on crash codes to indicate drug/alcohol involvement in crashes of all severities with 95% accuracy 3. To identify countermeasures to reduce alcohol and drug involvement in crashes by 10% a year 4. To list the top 10 counties most suitable for drunk/drugged driving countermeasure implementation.