Substance Use Disorder and Violence Among People with Severe Mental Illness in the United States.

Abstract

In studying substance use disorder (SUD) and violence in severe mental illness (SMI), researchers account for presence of SUD or addictions to specific substances. However these studies fail to comprehensively capture solitary drug use versus specific combinations in a single exhaustive variable with more nuance (e.g., opioids only, alcohol only, both alcohol and opioids only, and so on). Using logistic regression to predict past-year violence, this study compared conventional SUD measurement (Model I: presence versus absence of SUD or specific SUDS) to a newer and more holistic approach (Model II: a single exhaustive variable with both solitary addictions [e.g., opioids only] and specific combinations of addictions [e.g., both opioids and alcohol only]) among 10,551 people with SMI in the National Survey of Drug Use and Health (20,015 - 2019). After adjusting for a wide variety of factors in Model II, people with (1) alcohol use disorders only were 2.24 times more likely to be violent (CI = 1.46-3.45, p <.001); (2) opioid use disorders only were 3.45 times more likely (CI-1.48-8.05, p,>01); (3) both alcohol and cocaine use disorders or cocaine only were 5.85 times more likely (CI = 2.63-13.05, p <.001); and (4) both alcohol and opioid use disorders only were 4.28 times more likely (CI = 1.34-13.71, p <.05). These more nuanced findings in Model II differed substantially from those using conventional SUD assessment in Model I, and the newer and more holistic approach can better reflect the complexity of addiction in relation to violence in SMI. Therefore studies, practices, and policies that address SUD and violence in SMI could be beneficially revisited with this greater comprehensiveness and detail.