Types of Traumatic Experiences in Drug Overdose-Related Deaths: An Exploratory Latent Class Analysis

Drug overdose related-deaths in the US are increasing, with over 100,000 deaths occurring in 2020, an increase of 30% from the previous year and the highest number recorded in a single year. It is widely known that experiences of trauma and substance use very often co-occur, but little is known about the role of trauma in the context of drug overdose-related deaths. Latent class analysis (LCA) was used to classify drug overdose-related deaths based on type of traumatic experiences and individual, social, and substance use characteristics.

Psychological autopsy data were obtained from the University of Texas Health Science Center at Houston (UTHealth) Brain Collection. A total of 31 drug overdose-related deaths collected from January 2016 through March 2022 were included in this study. LCA was used to identify latent factors via experience of four trauma categories (illness/accidents, sexual/interpersonal violence, death/trauma to another, other situations where life was in danger). Generalized linear modeling (GLM) was used to explore differences on demographic, social, substance use, and psychiatric variables between the latent classes in separate models.

LCA identified 2 classes: C1 ( n =12; 39%) was characterized by higher incidence of overall trauma exposure as well as variation in trauma type; C2 ( n =19; 61%) had lower levels of overall trauma exposure with sexual/interpersonal violence as the most frequent. GLMs indicated that C1 membership was associated with higher incidence of polysubstance use, being married, and having suicidal ideation compared to C2 membership ( p s<0.05).

Among individuals who died by drug overdose, the exploratory LCA identified two distinct subgroups that differed in type of trauma experienced and substance use pattern, the first group having more "typical" characteristics of drug overdoses cases, the other group less typical. This suggests that those at risk of drug overdose may not always exhibit high-risk characteristics.