By J. Alexander Short
All too often, the modern opioid epidemic is reduced to numbers. Over 70,000 drug overdose deaths occurred in the United States in 2017. This marked a substantial increase from the more than 63,000 deaths reported in 2016. So many news articles, books, and even policymakers depend on these numbers as an accurate measure of the opioid crisis. However, can we rely on their accuracy?
Unfortunately, there are surprising inconsistencies in the reporting of drug overdose deaths that warrants further investigation.
Public Health Reports, the journal of the Office of the U.S. Surgeon General, recently published a study by the University of Pittsburgh School of Public Health. The study revealed an alarming fact: as many as 70,000 opioid-related deaths have gone unreported since 1999.
Pointing to a lack of specificity utilized by coroners and medical examiners, the study highlights important weaknesses in how public health officials currently measure the impact of the opioid crisis. Since national statistics on drug intoxication deaths are based solely on data derived from death certificates, this lack of specificity has serious implications. Indeed, this underreporting prevents public health officials from clearly understanding the magnitude of the opioid crisis.
Without accurate information, how are public health officials able to have an clear picture of which regions are most seriously affected by this health epidemic?
In the United States, medical examiners and coroners use ICD-10 codes (International Classification of Diseases) to classify causes of death. In addition to these codes, “T-Codes” are often used to identify specific drugs contributing to overdose deaths. For example, all codes between T40.0 and T40.6 are opioid related. A code of T40.1, for example, indicates the presence of heroin. Not all codes are so specific, however. A T50.9 code, for example, indicates the presence of “other and unspecified drugs, medicaments, and biological substances.” Typically used to represent obscure drugs (i.e., bath salts), the T50.9 code has also been utilized as a catch-all classification, even when greater detail is available.
According to the findings of the University of Pittsburgh study, overuse of the T50.9 code varies significantly from state to state. For example, the study found 17 states had less than 5 percent of overdose deaths coded as unspecified. On the other hand, five states had over 35 percent of drug overdose deaths coded as unspecified. Pennsylvania, the country’s leading lack-of-specificity offender, had over half (50.8 percent) of all drug overdose deaths coded as unspecified.
Reasons for these inconsistencies are not entirely clear. Variations in classification of accidental drug overdoses may reflect underlying geographical discrepancies. Rural counties tend to have less resources than urban counties for investigating suspicious deaths. This difference in resources may affect the availability of toxicology testing and thus may impact the specificity of cause-of-death reporting. For example, the five states with highest use of the T50.9 codes are largely rural: Pennsylvania, Louisiana, Alabama, Indiana and Mississippi. Other explanations, however, may shed more light.
The type of system in place for investigating suspicious deaths may play a role in these inconsistencies. States with centralized medical examiner systems saw higher percentages (92 percent) of drug specification than did states with decentralized county coroner systems (62 percent). While medical examiners must be physicians, county coroners need not having medical training as they are an elected office.
In addition to these variations in the consistency of death-reporting, there are other shortcomings worth mentioning in the current cause-of-death reporting system. For example, there is currently no reporting mechanism in place for medical examiners and coroners to identify the method of drug administration. So while we know over 70,000 people passed away from a drug overdose death in 2017, we do not know how many of them died from administering a drug intravenously.
The differences in drug administration reflect related but differing health policy concerns. Developing a mechanism to collect this data may serve to further illuminate the opioid health epidemic. This would allow public health officials to react to the opioid epidemic more appropriately.
These explanations overlook some important philosophical questions, too.
Michael J. Young’s 2014 piece on this blog addresses philosophical implications of causation involved in death reporting with relation to Alzheimer’s Disease. These questions relate similarly to reporting of drug overdose deaths. Oftentimes, drug overdoses reflect a combination of numerous drugs, none of which would have proven fatal individually. Medical examiners and coroners have limited space on death certificates in which to explain complex medical situations that lead to death. Indeed, this problem expands beyond accidental drug overdose deaths and likely impacts ‘cause-of-death’ reporting across the board.
The University of Pittsburgh study highlights an alarming fact: the opioid epidemic may be more urgent than data suggests. In order for public health officials to appropriately understand and react to the widespread societal impact of the opioid epidemic, the data must first be reliable. To ensure accurate reporting of opioid-related deaths, there needs to be greater standardization among medical examiners and coroners when reporting accidental drug overdose deaths. Granted, such reforms will not solve this health crisis. However, more standardized cause-of-death reporting systems may allow policymakers to respond to the opioid epidemic in a more efficient and appropriate manner.
J. Alexander Short is a student working with the Addiction Legal Resources Team at Penn State University’s Dickinson School of Law.