Need for Improved Timeliness of Reporting on Drug Overdose Fatalities: The HEALing Communities Study

Abstract

Objective:

Timely data on drug overdose deaths can help identify community needs, evaluate the effectiveness of interventions, and allocate resources. We identified variations in death investigation and reporting systems within and between states that affect the timeliness and accuracy of death certificate information.

Methods:

The HEALing Communities Study (HCS) is a community-engaged, data-driven approach to combating the opioid crisis in 67 communities in 4 states: Kentucky, Massachusetts, New York, and Ohio. HCS conducted a survey of coroners and medical examiners to understand variability in drug overdose death data. We compared survey results in Massachusetts, New York, and Ohio with national data to investigate the completeness of provisional death counts by type of death investigation system.

Results:

Communities in each HCS state had different ways of collecting and reporting mortality data. Completion of death certificates for drug overdoses ranged from <2 weeks in 23% (7 of 31) of those surveyed to more than 3 months in 10% (3 of 31) of those surveyed. Variabilities in the timeliness of reporting drug overdose deaths were not associated with type of coroner or medical examiner office in each state, urban versus rural setting, or specificity of drug information on the death certificate.

Conclusion:

Having specific drug information on the death certificate may increase death certificate quality, comparability, and accuracy. We recommend the following: (1) all coroners and medical examiners should be trained on conducting death investigations, interpreting toxicology reports, and completing death certificates; (2) 1 office in each state should oversee all coroners and medical examiners to increase data consistency; and (3) communities should identify and address barriers to timely death certification.

Data on drug overdose fatalities are obtained primarily from death certificates. Jurisdictions use death certificates to detect changes in patterns among drug overdose deaths, by calculating the number, type, and location of overdose deaths and collecting demographic information on decedents. These data have shown a steady increase in drug overdose deaths since 1999 to more than 107 000 predicted deaths occurring in the United States in 2022, ¹ an increase in involvement of cocaine in drug overdose deaths since 2013, ² a rise in rates of drug overdose deaths since 2014 among non-Hispanic Black people,^2,3 and a higher risk in rural counties than in urban counties in the rate of drug overdose deaths involving psychostimulants in 2019. ⁴ To our knowledge, these examples are the most current data available, yet these data are ≥2 years old, emphasizing the need for more timely overdose mortality data. Timely data on drug overdose mortality can allow communities to identify their current needs, evaluate whether interventions are reducing drug overdose deaths, and allocate resources effectively to address the drug epidemic.

The Centers for Disease Control and Prevention’s (CDC’s) National Vital Statistics System provides annual national data on drug overdose deaths based on data received from each state. Although data on drug overdose deaths are presented as final, a lag of 1 year can exist between death and data availability, and data may be still preliminary from states that have not finalized their data. ³ To provide more timely information, CDC also publishes provisional drug overdose death counts, which have a lag of 4 months from the date of death to data availability. ¹ However, completeness reports for 2019 have shown that, after 6 months, New York (excluding New York City) had death certificate completion rates of 69.7% to 86.3%, Massachusetts had completion rates of 97.0% to 100%, and Ohio had completion rates of 93.7% to 98.9%. ¹

Timeliness of death certificate completion differs by cause of death, with overdoses having the longest lag. ⁵ Because death certificates for overdoses should contain all drugs that contributed to the death, ⁶ these death certificates are often initially filed with a pending cause of death while waiting for autopsy and/or toxicology results. The number of overdose deaths is likely to be underestimated when a high percentage of records indicate “pending investigation.” Death certificate completeness concerning the type of drug in the drug overdose death varies across and within states depending on the death investigation system. ⁷ Compared with counties with coroners, counties with medical examiners tend to have a lower rate of unspecified drug information on death certificates and, thus, tend to have higher-quality and more consistently documented data. ⁸

We initiated this analysis out of concern that issues that may arise from varying death investigation and reporting systems across states may be magnified in states that have a decentralized county-based system, which may cause investigation and certification of drug overdose deaths to become both geographically and demographically inconsistent. ⁹ Requirements for coroners and medical examiners vary by state legislation. Coroners are typically locally elected officials who are sometimes, but not always, required to have a medical degree, whereas medical examiners are often appointed and are almost always physicians. States also vary as to whether there is a centralized state medical examiner office, a state office that oversees the local coroners or medical examiners, or only local oversight, all of which may affect the manner in which data are collected and shared on death certificates. ¹⁰

We hypothesized that states with a centralized death investigation office would have higher 12-month completion rates than states without a centralized office. To examine factors that may influence death certificate completion times and their associations with the type of death investigation system, we assessed local data from the HEALing Communities Study (HCS) for 3 of the 4 states included in the HCS. To assess whether the type of death investigation system affects death certificate timeliness on a national level, we used publicly available data from CDC.

Methods

The HCS, which is funded by the National Institute on Drug Abuse and the Substance Abuse and Mental Health Services Administration, is testing community-engaged, data-driven approaches to combating the opioid crisis. HCS is a multisite, parallel-group, cluster-randomized, wait list–controlled trial that engages 67 urban and rural communities highly affected by the opioid crisis in 4 states—Kentucky, Massachusetts, New York, and Ohio—with the overarching objective of implementing evidence-based strategies that are known to reduce the number of opioid overdoses. A key component of HCS is to provide communities with current data so that communities can review, select, and adjust the implementation of their strategies, as appropriate. If the HCS model is effective, this intervention could be adopted by other US communities to address the current drug epidemic and other public health crises. ¹¹ We used information from Massachusetts, New York, and Ohio to highlight issues on death certificate timeliness related to state regulations. Advarra Inc, the HCS single institutional review board, approved the study (Pro00038088).

The death investigation system in each HCS state is unique. Massachusetts has a state medical examiner with a centralized death investigation system, whereas New York and Ohio have decentralized county-based coroner/medical examiner systems with no state medical examiner or state office oversight. Massachusetts appoints the chief medical examiner, and the Office of the Chief Medical Examiner serves under the Executive Office of Public Safety and Security. In Ohio, coroners and medical examiners are required to be licensed physicians who have completed at least 16 hours of Ohio State Coroners Association–approved continuing education before the commencement of each 4-year term, and deputy coroners must be either licensed physicians or pathologists. The Ohio State Coroners Association oversees education requirements for coroners and medical examiners. Only 2 counties in Ohio have medical examiners. In New York, coroners are not required to have a medical degree, but medical examiners must have one. In New York, 24 counties (including 5 in New York City) have medical examiners and 38 counties have coroners; no single governmental or private sector organization or agency is responsible for coordinating and supporting a death investigation system.

To understand the details and structure of data collection in each HCS community, HCS developed a survey that asked about toxicology testing, autopsies, and death certificate completion for drug overdoses, which was implemented from March through October 2020. Massachusetts provided the survey to the Office of the Chief Medical Examiner for completion. New York combined the survey with a concurrent survey being conducted in conjunction with CDC as part of an Overdose Data to Action grant, to reduce survey fatigue among coroners and medical examiners. The New York State Association of County Coroners and Medical Examiners identified coroners and medical examiners; the New York State Department of Health, which administered the online survey, added questions that were relevant to the Overdose Data to Action grant and emailed a link to identified participants. HCS consulted with community groups in Ohio to identify coroners and medical examiners, who were emailed a link to the online survey.

We used survey results to understand differences in death investigations at the state and county level that may be associated with the timeliness of filing the final death certificate. The survey asked about death scene investigations, which deaths the office was responsible for, frequency of autopsies and toxicology testing, and turnaround time for completed death certificates to be filed.

We analyzed publicly available CDC data to understand various types of death investigation systems ¹⁰ and the completeness of 12-month provisional counts of overdose deaths relative to final counts from 2019 after a 6-month lag, ¹ by state, for the entire United States. We calculated the 12-month provisional count completeness as the total number of overdoses counted in the state in the 12 months before the reporting month as the numerator and the number reported for the same period 6 months later (considered more complete) as the denominator. For this analysis, we compared distributions of 12-month completeness between states that do and do not have a centralized death investigation oversight by using the Wilcoxon rank sum test with continuity correction using R (R Core Team). We determined the significance of 12-month completeness and centralized death investigation oversight as P < .05.

Results

Coroners and medical examiners in 15 of 16 HCS counties in New York and in 15 of 18 HCS counties in Ohio completed the survey; in Massachusetts, 1 survey was completed at the state level, which covered the entire state and included the 16 HCS communities. In New York and Ohio, we found variability in to whom at the county level the coroner or medical examiner reports, which deaths are investigated, whether information is shared with the local public health agency, and whether autopsies and toxicology screens are conducted for all possible drug-related deaths. In general, most medical examiner offices reported to or interacted with the county-level public health department, which included sharing information on the number and cause of death; however, not all coroners share this information. The lack of sharing was due to county-level laws, concerns about data confidentiality, or concerns that data sharing could affect police investigations. Coroners in New York and Ohio are responsible for completing the death certificate, although many coroners in New York reported relying on a neighboring county’s medical examiner or forensic pathologist to conduct the autopsies and toxicology tests.

In Ohio, 8 HCS counties reported using another county’s toxicology or crime laboratory, 5 counties reported using their own county coroner’s office or crime laboratory, 8 counties reported using independent toxicology laboratories, and 2 counties reported using state laboratories. Seven HCS counties in Ohio reported using more than 1 laboratory. In New York, 6 of the 15 responding HCS counties reported using their own toxicology laboratory, and 9 counties reported that they contracted these services to another county (Table).

Table.

Location of toxicology investigation and laboratory used for suspected drug overdose deaths and turnaround times for toxicology results and death certificate completion in New York, Ohio, and Massachusetts, March through October 2020 ^a

Variable	New York	Ohio	Massachusetts
Surveys completed	15	15	1
Location of toxicology investigation, no.
In county	6	5	0
In another county or state	9	8	1
Type of laboratory used b
State or county	6	10	1
Hospital	1	0	0
Private	9	8	0
Other	0	0	0
Turnaround time for toxicology results, wk
<2	5	2	0
2-<4	4	0	1
4-<8	5	6	0
8-12	0	5	0
>12	1	1	0
Turnaround time for completing death certificate, wk
<2	7	1	0
2-<4	2	1	0
4-<8	1	2	0
8-12	4	9	1
>12	1	2	0

^aResults show the number of counties in New York and Ohio that responded to the survey from the HEALing Communities Study, which was conducted from spring through fall 2020. In Massachusetts, the survey was completed by the Office of the Chief Medical Examiner.

^bCommunities could use >1 laboratory.

Turnaround times for receipt of toxicology results ranged from <2 weeks to 3 months but were typically 2-12 weeks in Ohio and New York and 2-4 weeks in Massachusetts (Table). Respondents from 5 HCS counties in New York and 2 HCS counties in Ohio reported needing ≥12 weeks to finalize and report the cause of death. Respondents in Ohio and Massachusetts reported death certificates being finalized typically in 8-12 weeks. In New York and Ohio, some counties offered preliminary suspected overdose counts to assist the locality in understanding current trends, but this practice was not consistent.

States with centralized medical examiner offices had 3 of the lowest 5 average completion rates (70% for New Hampshire, 78% for North Carolina, and 84% for West Virginia) and 3 of the 5 highest completion rates (100% for Maine, Rhode Island, and Vermont). We found no significant difference in mean 12-month completion rates by death investigation system when comparing centralized state medical examiner offices, county/district–based medical examiner offices, a county-based mixture of coroner and medical examiner offices, and county- or district-based coroner offices (Figure). Likewise, we found no significant differences in death certificate completeness when we analyzed for drug specificity on death certificates reported by region (West, South, Northeast, and Midwest; data not shown).

Figure.

Twelve-month completion rates for death certificates, by type of coroner/medical examiner system, United States, 2021. The horizontal bar inside the box indicates the median monthly completion rate, the whiskers indicate the first and third IQR, and the dots indicate outliers. Data source: National Center for Health Statistics. ¹

Discussion

A review of death investigation and reporting systems used by coroners and medical examiners from 3 HCS states showed variability in data collection and reporting of mortality within and between HCS states. This inconsistency in data collection can result in an unreliable and delayed understanding of the overdose crisis in communities. Our examination of national data disproved our hypothesis that communities with centralized death investigation systems would have more timely completion than communities without centralized death investigation systems; the variability appeared to be specific to the idiosyncrasies of each state’s data collection system and was not associated with the type of coroner or medical examiner office in each state, region of the country, or specificity of the drug information provided on the death certificate.

The findings from the HCS survey indicated that many communities use services of other localities for their autopsies and toxicology testing, which can lead to delays that affect the turnaround time for a finalized death certificate. The delays in conducting autopsies and toxicology testing can affect the turnaround time for a finalized death certificate. Increasing rates in drug overdose deaths can contribute to financial strains in communities that must outsource their autopsies and pay for the costs of both transporting the decedent and conducting the autopsy. These increasing rates in drug overdose deaths can also affect communities that conduct their own autopsies, resulting in strains on personnel and equipment that can ultimately affect the timeliness of death certificate completion; these personnel issues were further exacerbated by the COVID-19 pandemic. ¹² Limited availability of forensic pathologists and medical examiners can add to these strains.^12,13 Although these personnel issues have been recognized for more than a decade, little progress has been made in responding to multiple recommendations, which have included offering state-centered initiatives such as financial incentives, making salaries competitive with other medical specialties, and raising awareness of forensic pathology and death investigations in the medical school curricula.^13,14

Geographic Variability

Although we did not report timeliness among the HCS communities by rural versus urban status, other studies have shown urban–rural differences in the type of opioid involved in drug overdose deaths. ⁴ Rural counties and counties where coroners perform death investigations are more likely than urban counties and counties that use medical examiner offices to have missing T-code information, which identifies the specific drug classes in overdoses. ¹⁵ This disparity by urban versus rural area can affect the quality of the death certificate information.^7,8,16 If the quality of information from counties with only coroners, which are primarily in rural areas, is lacking, then these communities are at a disadvantage in their ability to advance interventions to reduce drug overdose fatality rates in their communities.

Improving Timeliness and Quality of Death Certificates

States now use electronic death registration systems (EDRSs), which should streamline the death registration process, improve the timeliness of data by reducing the time it takes to file death records, and improve communication among those responsible for filing the final death certificate.^17,18 The EDRSs are used by coroners and medical examiners to create death certificates and to streamline submission of mortality data to the National Vital Statistics System. The use of EDRSs also allows for increased communication and feedback on death certificate completeness and accuracy through internal data checks. EDRSs went into effect in 2005 in Ohio, in 2014 in Massachusetts, and in 2019 in New York. Although EDRSs are useful, users of these systems need to maintain vigilance to ensure that drug-related terms, such as drug abuse, drug use, drug exposure, dependency, and addiction, are not misused and that instead the preferred terms from the International Classification of Diseases, Tenth Revision (ICD-10), such as toxicity, overdose, and poisoning, are used. ¹⁹ Ambiguous and incomplete data, along with misuse of these terms, can result in undercoding and misclassification of overdose deaths.^15,20,21

Linking data from EDRSs and toxicology data to coroner and medical examiner case management software along with funeral home software can also improve the timeliness and quality of data. To accurately collect information, all coroner and medical examiner offices should use electronic case management software, which can highlight missing data fields, store all information related to a case in one location, assist in completing the death certificate, and aid in generating rapid reports of numbers of deaths from a specific cause for reporting purposes.^22,23 Linking systems can prevent double entering data into 2 systems, thus streamlining data collection and reducing the possibility of data entry errors.

Providing Preliminary Suspected Overdose Data

With the understanding that a lag will always exist in tabulating data on overdose mortality because of the time required for toxicology testing, communities can use death investigation information to conduct rapid surveillance and estimate the current burden of overdose mortality. Data sharing between coroners and medical examiners with offices of public health, law enforcement, and community coalitions is critical but is currently not being done. ²³ In New York, some counties offer preliminary suspected overdose counts to assist localities in understanding current trends, but this practice is inconsistently applied. Many communities use a small number of variables collected at the death scene to implement a structured algorithm that can better predict whether a death was an unintentional overdose.^24-26 These data can be shared in near–real time to help communities identify new trends, target areas at high risk for drug overdose deaths, and evaluate current public health interventions. Because small communities may have confidentiality issues with potential identification of decedents, data sharing may have to be limited in scope to protect the decedents’ identities.

Limitations

This study had some limitations. First, the HCS survey did not collect information on coroner and medical examiner training; in addition, we did not have this information in the national data when examining the timeliness of provisional data. Second, the HCS survey was conducted during the height of the COVID-19 pandemic, and results may reflect nonstandard trends caused by the increased number of deaths occurring in communities. Of note, the national data reflected pre–COVID-19 trends.

Third, among communities in New York that responded to the HCS survey, questions on the turnaround time for death certificate completion may have been misinterpreted. If coroners and medical examiners are relying on toxicology results to complete information on cause of death, then the turnaround time for death certificate completion should always be longer than the turnaround time for testing; however, this was not always true. Coroners and medical examiners in New York may have responded to this question by considering the original submission of the death certificate, which may have included a pending cause of death. In contrast, respondents in Ohio were specifically asked about the time to death certificate completion, including the time for the postmortem toxicology testing, thus eliminating this confusion.

Conclusion

Some HCS states noted delays in receiving information about the current number of drug overdose deaths. Inconsistencies in death investigations, toxicology testing, and death certification may affect the timeliness, quality, and identity of drugs reported on the death certificate. States should review where delays occur and identify corrective solutions, including potentially having one statewide office that acts as both a centralized data collection site and/or provides training and assistance to coroners and medical examiners.^22,27 As identified in the Association of State and Territorial Health Officials meeting on improving drug specificity and completeness on death certificates, it is essential for coroners and medical examiners to have training and education. ²² States should provide training to help coroners and medical examiners understand the importance of timely completion of death certificates. This training should explain how the death certificate is used to provide data, how to complete the death certificate to improve drug specificity, and how ICD-10 codes are assigned based on the provided cause of death information. When new coroners are elected to office, training on completing death certificates, using EDRSs, conducting death investigations, and interpreting toxicology reports should be mandated.

Delays in reporting the cause of death can affect other programs that rely on death certificates. These include the Fatal Analysis Reporting System, the National Violent Death Reporting System, and the State Unintentional Drug Overdose Reporting System. Addressing factors involved in these delays will result in future benefits by creating consistent, useful, and timely data that can help address the current drug epidemic and future epidemics.