Substance use among emergency room patients: Is self-report preferable to biochemical markers?

To explore the validity of self-reported substance use among emergency room populations and the processes of sample selection bias, to establish their influence on the prevalence rates found.

Self-reported alcohol and illicit drug use of patients in the emergency room is compared with results from an alcohol breath analyser and urine toxicology.

Variations in reported substance use occur when comparing self-report measures with alcohol breath analyser results and urine toxicology. Self-reported alcohol use was found among 7.5% of the patients compared with 4.7% based on alcohol breath analysers. Illicit drug use was reported by 9.0% of the patients whereas urine toxicology resulted in 30% patients positive for illicit drug use. Patients that voluntarily participate in the study differ from those that do not participate. Patients who refuse an alcohol breath analyser report slightly more alcohol use prior to the injury (difference not significant), and patients who provide a urine sample report more illicit drug use prior to the injury compared to those that refuse.

Differences in prevalence rates can be explained partly by the measurements used and partly by sample selection bias. Self-reported alcohol use and self-reported illicit drug use are preferable to the "gold standard" when used among emergency room patients, because both measures provide more accurate information on the actual use. Sample selection bias also influences the prevalence rates.