![[ Visit Client Website ]](images/banner.jpg)
BACKGROUND: Heroin drug-overdose mortality is rising nationally and Nevada has the fourth highest drug-overdose mortality rate in the United States. However, final data from state vital records (VR) can take >1 year and does not include toxicology. We sought to determine how effectively drug-related death information collected by Clark County Office of Coroner/Medical Examiner (coroner) could be used for local surveillance.
METHODS: We merged the coroner’s 2013 drug-related death records with records having underlying cause of death indicative of drug-overdose in 2013 VR data. Cases were linked by gender, birth and death dates. Treating VR as the standard, we calculated sensitivity and positive predictive value (PPV) of using coroner’s data to detect drug-overdose mortality. We conducted sub-group descriptive analyses on cases with heroin included as a multiple cause in VR and cases with heroin listed as the cause of death in coroner data to further compare the datasets. We interviewed six stakeholders to assess coroner surveillance system attributes according to CDC guidelines.
RESULTS: We linked 375 of 534 coroner cases (70.2%) to VR. Sensitivity of coroner data in detecting drug-overdose mortality was 94.2%; PPV was 70.2%. Thirty-two cases included heroin as a multiple cause; 31 were unintentional, and 27 documented a detectable metabolite unique to heroin. Nine coroner cases listing heroin as the cause of death were not linked to VR; 7 had missing underlying-cause codes. Stakeholders reported satisfaction with coroner data (e.g., timeliness and usefulness of toxicology) and recommend enhancements for simplifying population-based analyses (e.g., specification of free-text elements to increase information quality).
CONCLUSIONS: Coroner data will be used to increase timely surveillance. Conversion of free-text data to standard variables is warranted.