[8] JAMA, November 28, 1990 Vol 264(20) pp. 2639-2643 “The Efficacy of Pre-employment Drug Screening for Marijuana and Cocaine in Predicting Employment outcome.”


We recognize that a number of potential confounding factors might be associated both with the risk variables of marijuana and cocaine use and with the outcome variables of turnover, absenteeism, accidents, injuries, and discipline. Potential confounders included age at hire, gender, race, job classification, smoking status, exercise habits and alcohol use or abuse.




Demographic Characteristics Negative Marijuana Cocaine Other


Sex M 65.7% 73.7% 60.0% 65.9%


F 34.3% 26.3% 40.0% 34.1%


Race White 90.1% 88.4% 83.6% 90.9%


Black 5.6% 11.1% 16.4% 7.3%


Asian 4.4% 0.5% 0.0% 1.8%


Smoker 30.7% 47.7% 40.0% 45.4%


Non-smoker 69.3% 52.3% 60.0% 54.6%


In our study, we have not been able to control for the possible confounding effect of alcoholism. A substantial body of literature suggests that alcohol abuse correlates with the abuse of other substances. Other literature suggests that alcoholics have poor employment outcomes. We did not obtain alcohol levels in the test urine samples. Also, we did not administer any validated alcoholism questionnaires because we did not think they would provide accurate information in the context of a pre-employment examination. ...


The findings of this study suggest that many of the claims cited to justify pre-employment drug screening have been exaggerated. Drug users have been reported to be involved in 200% to 300% more industrial accidents, to sustain 400% more compensatable injuries, and to use 1500% more sick leave. We found that those with marijuana positive urine samples have 55% more industrial accidents, 85% more injuries, and a 78% increase in absenteeism. For those with cocaine-positive urine samples, there was a 145% increase in absenteeism and an 85% increase in injuries. These findings could be used to re-evaluate ... the cost-effectiveness of pre-employment drug screening.