Examination of Fentanyl-Related Drug Overdose Using Literal Text on Death Certificate Records from the National Vital Statistics System (NVSS)

Tuesday, June 6, 2017: 2:50 PM
410C, Boise Centre
Holly B. Hedegaard , Centers for Disease Control and Prevention, Hyattsville, MD
Merianne R. Spencer , Centers for Disease Control and Prevention, Hyattsville, MD
Margaret Warner , Centers for Disease Control and Prevention, Hyattsville, MD

BACKGROUND:  Drug overdose deaths have dramatically increased in recent years, largely driven by increases in opioid related deaths. Fentanyl, a synthetic opioid, is of particular interest to public health given its potency and various non-pharmaceutical forms obtained through the illicit drug market. Deaths involving synthetic opioids can be tracked using death data coded using the International Classification of Diseases, Tenth Revision (ICD-10). However, ICD-10 codes do not specifically identify deaths involving fentanyl or the illicitly-produced fentanyl analogs (e.g. acetyl fentanyl). Using a new method of literal text analysis, a recent CDC report cited fentanyl as the ninth most frequently mentioned drug involved in drug overdose deaths in 2013 and the fifth most frequently mentioned in 2014. This study seeks to better identify drug overdose deaths involving fentanyl and to explore the contextual information surrounding these deaths using the detailed written information (i.e. literal text) provided by the medical certifier on death certificates.

METHODS: ICD-10 codes (X40-X44, X60-X64, X85 and Y10-Y14) were used to identify drug overdose deaths in National Vital Statistics System (NVSS) Multiple Cause of Death files linked with the death certificate literal text. Literal text fields (i.e. Part 1, Part 2 and How the Injury Occurred) were examined for mentions of specific drugs. Drug overdose deaths with any mention of fentanyl and its analogs were then subset and analyzed using SAS Contextual Analysis and SAS Base 9.4.

RESULTS:  In 2014, 52,404 drug overdose deaths were identified using ICD-10 codes. Using text analytics, 4337 of these drug overdose deaths were found to have fentanyl involvement. Roughly 8.7% of the fentanyl-involved drug overdose deaths also mentioned illicit usage and 4.4% mentioned prescription usage. Mentions of both illicit and prescription usage could occur on the same record. The majority of fentanyl deaths did not specify illicit or prescription use. Mentions of fentanyl-related analogs included acetyl fentanyl, butyryl fentanyl, beta-hydroxythiofentanyl, and norfentanyl. Deaths involving fentanyl also frequently involved heroin, cocaine, oxycodone, alprazolam or morphine.

CONCLUSIONS:  Drug-specific information on death certificates can offer information about the types of drugs involved in overdose deaths. This analysis identified potential misspellings and terms from which fentanyl and related fentanyl analogues could be identified using literal text on death certificate records. This study highlights the importance of continued efforts to improve the specificity of the drug information provided by the medical certifier on the death certificate for drug overdose deaths.