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

Several states have legally sanctioned or mandated physician reporting of drivers who were driving while intoxicated (DWI). Valid prospective evidence demonstrating extremely poor performance of the criminal justice 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 reasoned that poor performance of the judicial system would be most evident if drivers who were DWI were not arrested under conditions selected to optimize legal intervention. The authors therefore wished to estimate the unbiased proportion of DWI drivers brought to an emergency department (ED) under these optimized conditions who escape detection by law enforcement officials.

Prospective, consecutive cohort of drivers transported to an urban ED following a motor vehicle crash (MVC). Conditions selected to optimize legal intervention included: police at scene; inebriation of driver clinically evident to out-of-hospital personnel; and confirmatory blood ethanol level > or =100 mg/dL (> or =22 mmol/L). Main outcome measure was arrest for DWI.

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 arrest frequency for DWI, did not materially alter these findings. No patients were lost to follow-up.

These findings prospectively confirm that, even under conditions selected to optimize detection by law enforcement officials, only about one of every five drivers who were DWI and were brought to an ED following an MVC-and almost certainly no more than a minority-comes to the attention of the criminal justice system.