The Effects of Substance Use on Workplace Injuries

Evaluating Drug Testing Prospectively
The use of preemployment and random drug testing at businesses opens the door for research 
that could prospectively investigate the effect of drug testing on substance use, injuries, and 
other work-related performance measures. In our review, we found no studies that conducted 
such an analysis. Instead, studies have been conducted at the company level (Gerber and Yacou-
bian, 2002) or have employed pooled, cross-sectional time-series designs (Mehay and Pacula, 
1999; Ozminkowski et al., 2003). A prospective design may be used to follow individuals and 
inquire about drug-using behaviors before and after random drug tests occur at the workplace. 
Such a study could identify the effectiveness of the intervention on substance use, try to control 
for any possible selection bias, and provide potentially valuable information on the causal influ-
ence of substance use on occupational injuries. Other studies on workplace drug testing should 
examine variability across employers with respect to sanctions imposed and whether these vary 
by the laws governing drug testing across states. 
Designing New Workplace-Based Interventions
Regardless of their impact on occupational injuries, heavy substance use and substance-use 
disorders are serious public-health issues that are prevalent among workers, and interventions 
centered in workplaces hold great promise for recognizing and treating these problems (Roman 
and Blum, 2002). Researchers should continue to design and evaluate the effectiveness of these 
strategies. For example, how might social-norm campaigns affect substance use and occu-
pational injuries? Might they affect off-the-job substance use and nonoccupational injuries? 
Could training supervisors or altering a managerial structure influence substance use? Work-
place interventions have also recently been established to prevent smoking among employed 
youth and should be extended to encompass other substance-using behaviors among this par-
ticularly vulnerable group (Stoddard et al., 2005). 

---

35
CHAPTER SEVEN