drug abuse and depression (Monahan, 1995). Of those with bipolar disorder (without drug abuse), 5 percent had a history of violence, compared to more than 12 percent of those comorbid for drug abuse and bipolar disorder.

Some of the most important findings regarding the co-occurrence of psychiatric illness, drug abuse, and violence in the general population come from the Epidemiologic Catchment Area (ECA) surveys. This study of more than 20,000 community and institutional residents in five metropolitan areas found lifetime rates of drug abuse or dependence disorders to be as high as 47 percent among respondents with schizophrenia, 32 percent for those with major depressive illness, 56 percent for persons with bipolar affective disorder, and 87 percent for those with antisocial personality disorder (Regier et al., 1990). In data pooled from three ECA sites, about 2 percent of respondents with no disorder reported some violent behavior occurring within a one-year period. By comparison, the violence rates were 7 percent among those with a major psychiatric disorder only (schizophrenia or affective disorder) and 22 percent among those with co-occurring psychiatric and drug abuse disorders.

In multivariable models that controlled for age, sex, race, socioeconomic status, and marital status, the co-occurrence of psychiatric and drug abuse disorders emerged as one of the strongest predictors of violence toward others. Certain demographic covariates also increased the risk of violence among respondents with co-occurring disorders; among younger adult males of lower socioeconomic status, who reported a history of arrest and hospitalization, the predicted probability of violent acts within one year was 64 percent (Swanson, 1994).

Four mechanisms have been proposed to explain the underlying relationship between co-occurring drug abuse and psychiatric disorders and violence (Smith and Hucker, 1994). The first hypothesis is that violence in this group is linked primarily to the chemical effects of psychoactive drugs (e.g., cocaine may stimulate impulsive and aggressive behavior; alcohol may have a disinhibiting effect, possibly reducing tolerance for frustrating situations). Such effects may occur at lower doses for people with underlying psychiatric disorders (Drake et al., 1990). Antisocial personality traits often underlie both drug abuse and violence, and those antisocial traits may co-occur with psychotic disorders or other major psychiatric disorders as well. A third proposed mechanism is that drug use may exacerbate psychiatric symptoms, such as paranoid delusional beliefs, which can lead to violent actions in response to perceived threats. Finally, it has been proposed that social and economic factors—such as poverty and crime in the surrounding environment—largely account for the increased risk of violence among persons with co-occurring psychiatric and drug abuse disorders (Hiday, 1995). Limited evidence exists for each of