Using the Literal Text on Death Certificates to Identify the Drugs Involved in Drug Overdose Deaths

Tuesday, June 6, 2017: 2:40 PM
410C, Boise Centre
Holly B Hedegaard , Centers for Disease Control and Prevention, Hyattsville, MD
Margaret Warner , Centers for Disease Control and Prevention, Hyattsville, MD
James P. Trinidad , Food and Drug Administration, Silver Spring, MD
Brigham A. Bastian , Centers for Disease Control and Prevention, Hyattsville, MD
Arialdi M. Minino , Centers for Disease Control and Prevention, Hyattsville, MD

BACKGROUND:  From 1999 through 2015, the age-adjusted rate of drug overdose (OD) deaths more than doubled from 6.1 to 16.3 per 100,000. To better understand this public health epidemic, the National Center for Health Statistics and the U.S. Food and Drug Administration collaboratively developed a method to analyze death certificate literal text (the written information provided by the medical certifier) to identify the drugs involved in OD deaths.

METHODS:  National Vital Statistics System mortality data from 2010 through 2014 were linked to electronic files containing literal text information from death certificates. OD deaths were identified based on International Classification of Diseases, Tenth Revision(ICD-10) underlying cause of death codes X40-X44, X60-X64, X85 and Y10-14. A list of search terms from existing drug classification systems, drug lists and manual review of the literal text was used to identify drug mentions. SAS Version 9.3 software programs automated the identification of drug mentions by applying these search terms to the literal text from 3 fields of the death certificate. To determine the involvement of the drug in the death, the context of the drug mention was also considered. Age-adjusted death rates were calculated using the direct method and the 2000 standard U.S. population.

RESULTS:  The percent of OD deaths mentioning at least one specific drug increased from 67% in 2010 to 78% in 2014. From 2010 through 2014, the top ten drugs involved in OD deaths included opioids (heroin, oxycodone, methadone, morphine, hydrocodone and fentanyl), benzodiazepines (alprazolam, diazepam) and stimulants (cocaine and methamphetamine). In 2014, of the 36,667 OD deaths with mention of at least one specific drug, 52% mentioned only one drug and 48% mentioned more than one drug. For each of the top ten drugs (the “referent drug”), the percent of OD deaths that mentioned additional drugs varied by referent drug. For example, the majority of OD deaths involving methamphetamine did not involve other drugs. In contrast, OD deaths involving diazepam or alprazolam had on average more than 2 additional drugs involved in the death. One in five OD deaths involving heroin also involved cocaine. Nearly a third of the deaths involving diazepam also involved oxycodone. Twenty-three percent of the OD deaths involving fentanyl also involved heroin.

CONCLUSIONS:  This study demonstrates the ability of a new method for abstracting data from the death certificate to enhance national monitoring of drug overdose deaths and to better identify the combinations of drugs involved.

Handouts
  • Literal text on death certs HEDEGAARD handout.pdf (317.6 kB)