Prospective confirmation of low arrest rates among intoxicated drivers in motor vehicle crashes

Objective: Several states have legally sanctioned or mandated physician rep orting of drivers who were driving while intoxicated (DWI). Valid prospecti ve evidence demonstrating extremely poor performance of the criminal justic e system seems ethically and scientifically essential if overriding public health considerations are to abrogate the fundamental principles of patient -physician confidentiality. No such evidence is available. The authors reas oned that poor performance of the judicial system would be most evident if drivers who were DWI were not arrested under conditions selected to optimiz e legal intervention. The authors therefore wished to estimate the unbiased proportion of DWI drivers brought to an emergency department (ED) under th ese optimized conditions who escape detection by law enforcement officials. Methods: Prospective, consecutive cohort of drivers transported to an urba n ED following a motor vehicle crash (MVC). Conditions selected to optimize legal intervention included: police at scene; inebriation of driver clinic ally evident to out-of-hospital personnel; and confirmatory blood ethanol l evel greater than or equal to 100 mg/dL (greater than or equal to 22 mmol/L ). Main outcome measure was arrest for DWI. Results: Of 294 drivers in MVCs , 270 had ethanol levels, of whom 18 met criteria for optimum likelihood of legal intervention. Of these, 22% were arrested for DWI (95% CI = 6% to 48 %). Adjustment for missing data, under assumptions designed to maximize arr est frequency for DWI, did not materially alter these findings. No patients were lost to follow-up. Conclusion: These findings prospectively confirm t hat, even under conditions selected to optimize detection by law enforcemen t officials, only about one of every five drivers who were DWI and were bro ught to an ED following an MVC-and almost certainly no more than a minority -comes to the attention of the criminal justice system.