Drug Use, Workplace Accidents and Employee Turnover
Hoffmann, John, Larison, Cindy, Journal of Drug Issues
A number of studies presume that illicit drug use and heavy alcohol use increase the risk of work-related accidents and employee turnover. The results of these studies are inconsistent, however, with several reporting a significant association between employee drug use and accidents, and others finding no association. A more consistent result is that drug use is associated with employee turnover, especially an increased risk of termination and resignation. A significant limitation is that most of this research has relied on regional or industry-specific samples. Thus, whether their results generalize to the U. S. workforce is unknown. Using data from a large, representative sample of the U.S. population, we provide a detailed analysis of the relationships among drug use, work-related accidents, and employee turnover. The results indicate that various measures of drug use are not associated with work-related accidents. However, several types of drug use are related to the risk of being fired or resigning from a job in the previous year. Moreover, the risk of being fired varies by occupation.
Drug use among the U.S. workforce is believed to attenuate productivity, increase the risk of accidents, and cause workers to become loosely attached to the labor market. Several studies have sought to estimate the impact of drug use on lost productivity (Rice et al. 1990; Normand et al. 1994), but much of this research uses broad measures of risk and productivity. Studies using individuallevel data provide little evidence of lost productivity in terms of depressed wages due to drug use but suggest, nevertheless, that drug use may lead to riskier work behaviors or loose attachment to the workplace. For example, illicit drug use and heavy alcohol use have been linked to an increased risk of workplace accidents and injury (Dawson 1994; Hingson et al. 1985; Pollack et al. 1998), while illicit drug use appears to be associated with a greater number of absences and a higher risk of being fired or resigning from a job (Kandel and Yamaguchi 1987; Newcomb 1995; Zwerling et al. 1990). Much of this research relies on regional or industry-specific samples to estimate the impact of drug use on work-related outcomes. Therefore, the question of whether these results generalize to the U.S. workforce has not been answered.
The purpose of this paper is to examine the relationship between drug use and the risk of three work-related outcomes-past-year accidents, being fired, or resigning from a job-using data from a nationally representative survey that includes information on a large sample of full- and part-time workers. A key advantage of these data is that they allow an unbiased, nationally representative picture of the potential link between drug use and several measures that underlie economic productivity in the United States. Moreover, the large number of observations permits precise estimates of the impact of various measures of drug use.
The literature on the association between drug use and work-related outcomes is replete with studies that focus on wages, job instability (e.g., absences, termination), and on-the-job accidents and injuries. Based on several wellpublicized incidents and the known risks of drinking and driving (Ross 1992), a common perception is that there is a strong relationship between drug use and accidents in the transportation industry (Normand et al.1994). Nonetheless, the empirical evidence of an association between drug use and work-related accidents is inconsistent. Studies of construction workers in Washington state, workers residing in New England, or select workers in high-risk municipal jobs show a heightened risk of work-related accidents and injuries among illicit drug users (Hingson et al. 1985; Holcom et al. 1993; Pollack et al. 1998; Zwerling et al. 1990); however, a study of several companies from three areas of the United States and a national study of postal workers fail to substantiate this result (French et al. …