Suicides and Deaths of Undetermined Intent By Poisoning: Reexamination of Classification Differences By Race/Ethnicity and State

Background and Purpose: US suicide rates have continuously increased over the past two decades; however, reported suicide rates are considered underestimates due to misclassification of intent/manner of death. A significant proportion of poisoning deaths of undetermined intent (UnD) have been considered misclassified suicides. With increasing drug overdose deaths, UnD cases may also have been misclassified unintentional overdose deaths. This study examined racial/ethnic differences in UnD versus suicide classification of poisoning deaths and the potential impact of state variations on UnD rates for Black and Hispanic decedents.

Methods: Data were from the 2005-2015 National Violent Death Reporting System that links data from death certificates and coroner/medical examiner and law enforcement reports from 25 participating states. We used χ2 tests to compare 16,208 poisoning UnD decedents and 13,384 poisoning suicide decedents aged 14+ who did not leave a suicide note. We fit logistic regression models to examine associations of the odds of being classified as UnD versus suicide with race/ethnicity (Black vs. White in the first model; Hispanic vs. White in the second model), state (five highest UnD states vs. others), and interaction effects between race/ethnicity and state. We applied weighted coarsened exact matching (CEM; Blackwell, Iacus, King, & Porro, 2010) to establish equivalent racial/ethnic groups of decedents using all 12 precipitating/risk factors (i.e., so that these 12 factors are distributed as would be expected under random assignment).

Results: Overall, Blacks and Hispanics were slightly more likely than Whites to be classified as UnD than suicide when other demographic factors and toxicology results were also controlled. The odds of UnD classification was 7-10 times higher in the five highest UnD states (KY, MD, MI, UT, and RI) than in other states. Blacks had 2.71 (95% CI=2.07-3.56) higher odds of UnD classification in these five states, but had lower odds in other states. Hispanics were also more likely than Whites to be classified as UnDs than suicides without a note, but their rates did not vary by state. In addition to race/ethnicity, opiate and cocaine positive toxicologies were significant factors for UnD classification.

Conclusions and Implications: Coupled with their high rates of substance abuse problems, the high proportions of opioid and cocaine positive toxicologies among UnD cases suggest the possibility of unintentional overdose among these decedents. The extremely high UnD rates in the five highest UnD states likely reflect their consistently higher opioid misuse problems compared to other states.

Implications of the findings are: (1) the high occurrence of UnD deaths by poisoning among Blacks underscores the need for affordable, accessible substance abuse treatment and targeted overdose prevention programs in minority communities; (2) efforts to curb the opioid epidemic and related deaths should emphasize nonpharmacological pain treatment for people who suffer from chronic pain; and (3) more comparative research on suicide, UnD, and unintentional overdose deaths and standardizing manner-of-death determination processes for poisoning deaths across states may lead to improving prevention approaches nationally.