Deaths Attributed to Psychiatric Disorders in the United States, 2010–2018

Abstract

Objectives. To quantify deaths in the United States from 2010 through 2018 that were reported with an underlying cause of death as a psychiatric diagnosis, which do not indicate a clear mechanism of death, and that may be misclassified suicide and overdose deaths.

Methods. We used national vital statistics data to identify rates and circumstances of deaths by specific underlying cause of death categories in the US population.

Results. There were 115 442 deaths attributed to psychiatric diagnoses and 834 763 deaths attributed to suicide or overdose. The population rate of deaths attributed to psychiatric diagnoses increased from 3.26 to 4.96 per 100 000 US persons between 2010 and 2018.

Conclusions. Psychiatric diagnoses may represent a fairly substantial number of misclassified overdose and suicide deaths. Improving mortality surveillance requires improving the accuracy of diagnoses reported on death certificates.

Accurate mortality surveillance is essential to public health practice and policy.¹ Mortality surveillance is based on the underlying cause of death (UCOD) classifications derived from death certificates. The UCOD is a diagnosis intended to reflect the single most important factor leading to death. However, because coroners, medical examiners, and other physicians completing death certificates can use the full range of International Statistical Classification of Diseases and Related Health Problems, 10th Revision² diagnosis codes, the UCOD reported on a certificate may not actually describe a clear mechanism of death. Such diagnoses include psychiatric diagnoses, such as depression, alcohol abuse, and psychosis. Many deaths attributed to psychiatric diagnoses may be misclassified as suicide or overdose deaths, given the associations of psychiatric conditions with these mortality outcomes.

The purpose of this analysis was to quantify deaths reported with a psychiatric diagnosis UCOD in the United States during the years 2010 to 2018. For context, we also characterized rates and changes in rate over time for suicide, overdose, and other common causes of death during the same period.

METHODS

Death certificates are completed by a coroner or medical examiner when an investigation into the cause of death is warranted. Deaths certificates for deaths occurring in a health care facility are typically completed by physicians in that facility.¹ The deaths are then reported by state vital statistics offices to the Centers for Disease Control and Prevention (CDC).

The CDC’s National Center for Health Statistics (NCHS) uses an automated coding system for deriving UCODs from death certificates. The single UCOD per record indicates “the disease or injury which initiated the train of morbid events leading directly to death, or the circumstances of the accident or violence which produced the fatal injury.”³ For death certificates with multiple diagnoses, the NCHS applies modification rules to select the UCOD.⁴ This process selects a UCOD based on the location of each code on the death certificate and the known pathophysiology of each condition. Thus, it can be inferred that deaths attributed to psychiatric diagnoses could not be reclassified to a UCOD that more clearly explained the mechanism of death.

We queried the WONDER system⁵ to ascertain the frequency and rates of deaths attributed to specific UCODs, as well as the type of location of deaths and the percentage of deaths for which an autopsy was performed. We defined the category of psychiatric diagnoses from codes F06.0 to F06.6, F10 to F19, F20 to F29, F30 to F39, F40 to F48, F51.0 to F51.9, F53 to F55, F59, F60 to F64, and F68. This definition excluded diagnoses for cognitive impairment or developmental and childhood disorders. Consistent with CDC reporting,⁶ we identified deaths classified as suicide through UCODs X60 to X84, Y87.0, and ∗U03 and identified unintentional overdose through UCODs X40 to X49. We identified undetermined intent overdoses through UCODs Y10 to Y14.⁷ Codes for other common causes of death also followed CDC reporting conventions.⁸

We used Prais–Winsten regression with the Cochrane–Orcutt transformation and semirobust standard errors to test for temporal trends in population rates using Stata version 10 (StataCorp LP, College Station, TX).

RESULTS

There were 115 442 deaths attributed to psychiatric diagnoses and 834 763 deaths attributed to suicide or to unintentional or unknown intent overdose from 2010 through 2018. The population rate of psychiatric deaths increased from 3.26 to 4.96 per 100 000 US persons between 2010 and 2018, which was a similar degree of increase as seen for unintentional overdose and suicide (Figure 1). Prais–Winsten regression models indicated that for psychiatric deaths, suicide, and unintentional overdoses, the rate was increasing before 2014 and increased faster after 2014 (Table A, available as a supplement to the online version of this article at http://www.ajph.org). We did not consistently observe a similar pattern in regression results for other common causes of death, such as heart disease and cancer (Table A). Deaths attributed to a psychiatric diagnosis were more likely to occur in a health care facility (33.8%) compared with unintentional overdose (11.1%) or suicide (7.5%). These deaths were also less likely to have resulted in an autopsy (30.6% for psychiatric deaths vs 75.8% for unintentional overdose and 54.7% for suicide).

FIGURE 1—

Mortality Rates by Death Certificate–Listed Underlying Cause of Death: United States, 2010–2018

Among psychiatric UCODs, diagnoses related to alcohol and drugs were the most common (Table B, available as a supplement to the online version of this article at http://www.ajph.org). In total, 74 297 of deaths had a UCOD of “mental and behavioral disorders due to use of alcohol,” with 90.4% of these under the subcategories for “harmful use” or dependence. An additional 2685 deaths (2.3% of all psychiatric deaths) were attributed to “mental and behavioral disorders due to use of opioids,” and 23 392 deaths (20.3% of all psychiatric deaths) were attributed to use of other, multiple, or unspecified substances as the UCOD. No deaths had a UCOD for acute intoxication, consistent with CDC instructions that such deaths be classified as overdose.⁹ Of non–substance-related psychiatric deaths, schizophrenia and related disorders were the most common (4871 deaths) and mood disorders the next most common (4806 deaths).

DISCUSSION

In this study, we report that 10 000 to 16 000 deaths each year in the United States were attributed to psychiatric diagnoses as the UCOD between 2010 and 2018, despite the fact that these codes do not indicate a mechanism of death. For comparison, the number of deaths attributed to psychiatric diagnoses is similar to the number of firearm-related homicide deaths in the United States each year.⁶ The majority of psychiatric deaths were given alcohol- and drug-related diagnoses as the UCOD.

The increasing rates of psychiatric deaths were parallel to increases in overdose and suicide deaths. The study data do not allow for a conclusive evaluation of the degree to which psychiatric deaths were misclassified as suicide or overdose deaths. It may be the case that only a relatively small proportion of psychiatric deaths are actually suicide or overdose deaths, and that the increase in these deaths is driven by shared underlying causes, such as substance use and economic factors.¹⁰ Alternatively, the parallel trends could be coincidental. An important next step to addressing the misclassification identified here is an examination of deaths with a psychiatric UCOD using other data sources that provide more information about the mental health of decedents and the circumstances of their deaths.

A third of psychiatric deaths occurred in a health care facility, which was a higher percentage than for suicide and overdose deaths. The death certificates of these cases were probably completed by physicians working in these facilities, for whom death certificate completion is not a routine part of their work. The deaths given a psychiatric UCOD were also less likely to have had an autopsy than suicide and overdose deaths. This may indicate that the use of these diagnoses on death certificates is more common in jurisdictions with limited resources.

PUBLIC HEALTH IMPLICATIONS

There are a substantial number of deaths each year attributed to psychiatric diagnoses, even though these diagnoses do not indicate a clear mechanism of death. Given the close connection of psychiatric conditions with suicide and overdose, an implication of this study is that suicide and overdose are likely underestimated to a greater degree than previously believed. Accurate measurement of the rate of suicides and unintentional overdoses in particular is critical to determining if the large investments in public health to address these problems have been effective.