Older adult drug overdose: an application of latent class analysis to identify prevention opportunities

Older adult overdose death rates have increased significantly in recent years. However, research for prevention of drug overdose death specific to older adults is limited. Our objective is to identify profiles based on missed intervention points (touchpoints) to inform prevention of future older adult unintentional overdose deaths.

We used latent class analysis methods to identify profiles of decedents aged 55 + years in the Illinois Statewide Unintentional Drug Overdose Reporting System. This system collects data on 92.6% of all unintentional overdose deaths in Illinois and includes data from death certificates, coroner/medical examiner, toxicology, and autopsy reports. Data include decedent demographics, circumstances leading up to and surrounding the fatal overdose and details regarding the overdose. Variables in the latent class analysis model included sex, race, alcohol test result, social isolation, recent emergency department (ED) visit, chronic pain, and pain treatment.

We identified three distinct decent profiles. Class 1 (13% of decedents) included female decedents who were in pain treatment, had physical health problems, and had greater likelihood of a recent ED visit before their death. Class 2 (35% of decedents) decedents were most likely to be socially connected (live with others, employed, had social or family relationships) but less likely to have recent healthcare visits. Class 3 (52% of decedents) decedents had higher social isolation (lived alone, unemployed, unpartnered), were mostly male, had fewer known physical health conditions, and more alcohol positivity at time of death. White decedents are clustered in class 1 while Black decedents are predominant in classes 2 and 3.

These profiles link to potential touchpoint opportunities for substance use disorder screening harm reduction and treatment. Class 1 members were most likely to be reachable in healthcare settings. However, most decedents were members of Classes 2 and 3 with less engagement in the healthcare system, suggesting a need for screening and intervention in different contexts. For Class 2, intervention touchpoints might include education and screening in work or social settings such as senior centers given the higher degree of social connectivity. For Class 3, the most isolated group, touchpoints may occur in the context of harm reduction outreach and social service delivery.