Back to Annual Meeting

Back to Annual Meeting

APHA Scientific Session and Event Listing

Katherine Hempstead, PhD, Center for Health Statistics, New Jersey Department of Health and Senior Services, P.O. Box 360, Trenton, NJ 08625-0360, 609-292-9353, khempstead@doh.state.nj.us

Purpose There is a widespread belief that a non-trivial fraction of fatal poisonings coded as undetermined are actually suicides. The purpose of this study is to see whether manner of death coding for fatal poisonings can be predicted from characteristics of the decedent and/or circumstances of the death. Additionally we wish to test the hypothesis that there is county-level variations in medical examiner decision-making. Methods Data on fatal poisonings in New Jersey between 1998 and 2003 are used. Death certificates and medical examiner reports are the primary sources of information. Several multivariate models are estimated to analyze the determinants of manner of death coding, with data on the decedent and their family, circumstances of the death and evidence of intent, and county level medical examiner office dummy variables serving as regressors. An ordered probit model is estimated which includes all fatal poisonings, including those coded as accidental, undetermined, and suicides. Several bivariate logistic regressions estimate determinants of being coded as unintentional as compared with all other intents, or suicides compared with all other intents. Results Results suggest that circumstances and intent vary significantly by manner of death coding, even after controlling for decedent characteristics. Preliminary analysis suggests that significant medical examiner effects exist, but more socioeconomic data on counties needs to be included to avoid potential confounding. Conclusions This study suggests that undetermined fatal poisonings and suicides are significantly different with regard to characteristics of decedents, circumstances, and evidence of intent. Yet county level medical examiner effects may exist. Standardization of manner of death coding will improve the quality of data on suicide.

Learning Objectives:

• 1) Examine whether manner of death for fatal poisonings can be predicted from data on decedents and circumstances.
• 2) Assess whether or not there are county level medical examiner effects which affect manner of death coding after controlling for circumstantial and decedent characteristics.

Keywords: Injury, Drug Abuse

Presenting author's disclosure statement:

Not Answered