Skip to main content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.


The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Strengthening the Data Infrastructure for Outcomes Research on Mortality Associated with Opioid Poisonings: Final Report

Publication Date
National Center for Health Statistics
Centers for Disease Control and Prevention

Below is a summary of the problems addressed, the solution, and the accomplishments and products.

Problems Addressed:

Researchers and public health officials frequently rely on the cause of death information made available from the coding and processing of death certificate data for critical programmatic, policy, and research needs.

  • The Mortality Medical Data System (MMDS), a subsystem of NCHS’ National Vital Statistics System (NVSS), was used to electronically code all death certificate records in the United States at the time this project began in 2018. Its ability to automatically code varied greatly by cause of death. Certificates that could not be electronically coded were submitted to a queue of records that required human coding, a process that added days to weeks to the processing time. Drug overdose deaths, which had become a national public health priority in the U.S. at the time of the project’s initiation, were generally not electronically coded. In 2018, while 79% of all death certificate records were coded electronically only 33% of records with a drug overdose death were coded electronically. Timely and accurate reporting of these deaths is critical for surveillance and research efforts.
  • In data year 2016, about 15% of death certificates indicating a drug overdose did not include the names of the specific drug involved. Previous research demonstrated that specificity on death certificates improves when physicians complete death certificates using their electronic medical records. In the case of drug overdose deaths, medical examiners and coroners (ME/Cs) are typically the certifying officials. Thus, providing ME/Cs the ability to integrate their case management systems with electronic death registration system (EDRS) was explored for its potential to improve the quality of data on drug overdose deaths and other deaths of interest to health researchers studying patient outcomes.
  • NCHS/DVS created the Vital Statistics Rapid Release (VSRR) Program to support more timely dissemination of vital statistics data. The VSRR supports surveillance of deaths associated with drug overdoses through: 1) quarterly provisional estimates of national death rates for leading causes of death including drug overdoses; and 2) monthly provisional counts of drug overdose deaths by state for a limited set of drugs identifiable from ICD-10 codes. The provisional monthly and quarterly releases provided timely information, but their utility was reduced by the limited demographic and geographic detail in the reports, as well as limited detail on specific drugs involved in the death.


The purpose of this project was to strengthen the mortality data produced by the National Center for Health Statistics’ (NCHS) National Vital Statistics System (NVSS) infrastructure for outcomes-based research on deaths associated with opioid poisoning.

Accomplishments and Products:

There are several distinct processes that were modified and enhanced to address the problems above.

  • NCHS redesigned the Division of Vital Statistics (DVS) cause of death IT systems to electronically code the multiple cases of death listed as text strings on death certificates using natural language processes and machine learning algorithms. NCHS enhanced the incorporation of details about drugs named on death certificates onto supplemental data files for use by approved researchers by pulling these details for use on provisional data. This information is being obtained from the death certificate literal text fields and provides details on specific drugs that cannot be conveyed solely through ICD-10 coding.
    • The redesigned system is documented here:
    • The report entitled National Vital Statistics Reports Volume 68, Number 3 March 21, 2019 Drug Overdose Deaths Involving Fentanyl, 2011-2016 ( documents the impact of fentanyl on drug overdose deaths. Mortality data previously was limited to ICD-10 codes for analytic work and fentanyl is coded as "synthetic opioids other than methadone."
  • The timeliness of death record reporting from vital record offices for drug-involved deaths as well as the specificity of the drug information reported on death certificates has increased through the creation of HL7 Fast Healthcare Interoperability Resources (FHIR) standards-based interoperability between systems containing mortality information, with a focus on improving data exchange between medical examiners’ and coroners’ offices, state vital records offices, and the NVSS maintained by NCHS.
    • For more information, see GitHub - nightingaleproject/Reference-Client: A reference implementation for a client application that uses the NVSS FHIR API
  • The NCHS DVS Vital Statistics Rapid Release (VSRR) program has been restructured to produce and release more in-depth information regarding drug overdoses including demographic and geographic data monthly for public health surveillance and research.
  • NCHS also worked to ensure utility of mortality data infrastructure improvements with end-users’needs, especially researchers, with studying drug overdose deaths involving opioids using the NVSS. On November 29, 2018,the NCHS Board of Scientific Counselors (BSC) convened a meeting of the Patient Centered Outcome Research Trust Fund Drug Work Group at the NCHS headquarters in Hyattsville, Maryland. The Work Group meeting gathered a broad range of stakeholders to explore approaches to releasing more detailed drug-related information from death certificates to researchers and public health officials. Meeting participants considered and discussed what supplemental drug information should be extracted from death certificate literal text and how it should be classified. Topics of discussion included (1) gaps in the current drug classification system, (2) current innovations to extract and use information from the literal text, and (3) alternative classification schemes that would provide additional, useful information to end users. In July 2019, NCHS held a second meeting of the PCOR Workgroup of the NCHS BSC. The participants were asked to create recommendations on two topics: How to disseminate data files which include details on specific drugs mentions and what future enhancements should be made for these drug files. The meeting generated several useful suggestions on both topics.
  • The BSC took the findings from the PCOR Work Group and developed a report and set of recommendations for NCHSafter each of the two workshops. A summary of the recommendations are available in the minutes from the BSC meetings
  • DVS undertook two efforts to investigate how the opioid crisis was affecting birth outcomes in the US.
  • The first effort was to identify if data collected on birth certificates could be used to accurately estimate the incidence of neonatal abstinence syndrome (NAS). Three states added this condition to their birth certificate. NCHS worked with those states to option data files linking hospital discharge records to birth certificates for infants identified at discharge with NAS. NCHS found that the states each included a different definition of NAS with their birth certificate instructions. Furthermore, there was variable reporting of NAS on birth certificates. Based on these finding, DVS did not move forward to recommend reporting NAS on birth certificates, but rather worked with the Council of State and Territorial Epidemiologists to create a NAS Standardized Case Definition (Council of State and Territorial Epidemiologists (
  • The second effort looking at birth outcomes related to drug use was a National Vital Statistics Report entitled Drug-involved Infant Deaths in the United States, 2015-2017 (National Vital Statistics Reports Volume 70, Number 7 June 3, 2021 Drug-involved Infant Deaths in the United States, 2015–2017 ( This report used the Drug Involved Mortality data files and the linked birth/infant death files from 2015-2017. Of the 442 drug-involved infant deaths identified in this time period, drugs were the underlying cause of death for 163 (37%) infants and a contributing cause of death for 279 (63%) infants.

Related Products:

A National Vital Statistics Report entitled Drug-involved Infant Deaths in the United States, 2015-2017 (National Vital Statistics Reports Volume 70, Number 7 June 3, 2021 Drug-involved Infant Deaths in the United States, 2015–2017 (