Hazardous Use Should Not Be a Diagnostic Criterion for Substance Use Disorders in DSM-5

Dear Editor:

Draft criteria for substance use disorders in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5; www.DSM5.org), retain the DSM-IV symptom of “recurrent substance use in situations in which it is physically hazardous” (hazardous use; American Psychiatric Association, 1994, p. 183). Multiple empirical and conceptual problems with the hazardous use symptom lead us to conclude it should not be a diagnostic criterion for substance use disorders in DSM-5.

Although many activities are physically hazardous when persons are intoxicated, hazardous use is typically endorsed as a result of driving a motor vehicle under the influence (Dawson et al., 2004). Most persons with DSM-IV substance abuse have this diagnosis only because they endorse hazardous use (Dawson et al., 2004; Vergés et al., in press). Yet drunk drivers have less substance involvement than those with other patterns of alcohol abuse (Hasin et al., 1999). In Item Response Theory analyses, hazardous use shows only modest discrimination of alcohol and drug problem severity (Lynskey and Agrawal, 2007). The psychometric properties of hazardous use differ by socioeconomic status, age, gender, country, and culture and over time, introducing a remarkable degree of diagnostic bias. Unlike all other DSM-IV alcohol abuse symptoms, hazardous use is positively associated with income (Keyes and Hasin, 2008) and is more common among adults than adolescents (Martin et al., 2008), likely reflecting differences in access to motor vehicles. Hazardous use is associated with higher levels of substance problem severity in women compared with men (Martin et al., 2006; Saha et al., 2006) and is far more common in the general U.S. population than in Hispanics in the United States and in other countries (Caetano, 2011). The symptom shows much higher severity in Australia (Proudfoot et al., 2006) than in the United States (Saha et al., 2006), likely reflecting Australia's relatively strict enforcement of intoxicated driving. The relative severity of lifetime reports of hazardous use in U.S. adults has been found to decrease with age (Kahler and Strong, 2006), suggesting that greater penalization of intoxicated driving over recent decades may have increased the problem severity level indexed by the symptom.

The poor psychometric performance of hazardous use is not surprising given that intoxicated driving is highly influenced by many contextual factors, including access to an automobile, enforcement efforts and penalties, and behavioral norms and attitudes. It is conceptually problematic that unlike all other substance use disorder symptoms, hazardous use is defined simply by risky behavior and not by substance-related consequences, physiological features, or compulsive use. Risky behavior does not reflect the mainstream concept of mental disorders as syndromes reflecting dysfunction in internal mechanisms (Wakefield and First, 2003). If substance-related behavior merely reflects fool-hardiness or poor judgment, it should not be considered as an indicator of disorder (although it might be assessed as a behavior that influences health status, such as with V-codes in DSM-IV).

Rather than being specific to substance use disorders, hazardous use can reflect a generally incautious approach to a variety of situations (Sher et al., 2009). Even within a driving context, many behaviors, such as speeding or text messaging, dramatically increase risk for physical harm, but these are not singled out as symptoms of a psychiatric disorder, despite risk that is similar to or greater than intoxicated driving. Although intoxicated driving is a massive public health problem, the inclusion of hazardous use as a diagnostic criterion has proven to be a mistake.