Causes of Death Among US Medical Residents

Abstract

Importance: From 2000 to 2014, the leading causes of medical resident death in the United States were neoplastic diseases and suicide.

Objective: To examine whether US medical resident rates of death have changed since 2014 and whether causes of resident death differ by specialty.

Design, setting, and participants: In this cross-sectional study, residents and fellows who were enrolled in Accreditation Council for Graduate Medical Education (ACGME)-accredited training programs and who died from January 2015 to December 2021 were submitted to the National Death Index to obtain causes of death. These decedents were compared with residents and fellows who died between January 2000 and December 2014. Data were analyzed between July 2024 to March 2025.

Exposure: Death while actively enrolled in an ACGME-accredited residency and fellowship training program.

Main outcomes and measures: The primary outcome was the difference in rates of death for US residents and fellows between 2 time periods, 2000 to 2014 and 2015 to 2021. Poisson regression modeling was used to calculate incidence rate ratios (IRRs) with 95% CIs for this comparison. Rates were also compared across specialties. Secondary outcomes included comparing trainee decedents with age- and gender-matched peers in the general population and querying differences in causes of death by specialty from 2000 through 2021.

Results: Between 2015 and 2021, 370 778 residents and fellows participated in 961 755 person-years of training. In that same period, 161 residents (50 [31.1%] female; median [IQR] age, 31 [29-35] years) died during training. Forty-seven residents (29.2%) died by suicide, 28 (17.4%) by neoplastic diseases, 22 (13.7%) from other medical and surgical diseases, 22 (13.7%) from accidents, and 21 (13.0%) from accidental poisoning. The highest number of resident suicides occurred during the first quarter of the first year. The death rate from neoplastic diseases decreased since 2000 to 2014 (IRR, 0.59; 95% CI, 0.38-0.90). Rates of other causes remained unchanged. Resident death rates from 2000 to 2021, including rates of death by suicide, were lower than age- and gender-matched peers across causes. The highest specialty suicide rate was for pathology (19.76 deaths per 100 000 person-years). The highest death rate from neoplastic diseases was psychiatry (9.67 deaths per 100 000 person-years). The highest death rate from accidental poisoning was anesthesiology (15.46 deaths per 100 000 person-years).

Conclusions and relevance: In this cross-sectional study comparing rates of US medical resident deaths from 2000 to 2014 with rates observed in 2015 to 2021, the rate of resident deaths from neoplastic diseases decreased, while the rates of death from all other causes remained unchanged. Nevertheless, the number of residents who died by suicide during their very first academic quarter, observed during both study windows, remains concerning. Future efforts to address trainee well-being must focus on the drivers and mitigating factors of distress, particularly during transitions.

Figure 1.. Resident Death by Suicide by Postgraduate Year and Academic Quarter, 2015 to 2021

The data in this figure include 370 778 residents and 961 755 person-years of training enrolled in more than 13 000 programs accredited by the Accreditation Council for Graduate Medical Education for calendar years 2015 through 2021.

Figure 2.. Number of Resident Deaths by Cause and Academic Quarter, 2000 to 2021

Cause of death data from 2000 to 2014 are from a previously published study. The second study window includes residents enrolled in more than 13 000 programs accredited by the Accreditation Council for Graduate Medical Education for calendar years 2015 through 2021. This figure does not include trainees enrolled in fellowships nor trainees enrolled in Transitional Year programs.

Figure 3.. Causes of Death for Specialties With 15 or More Deaths Between 2000 and 2021

This figure’s data include residents enrolled in programs accredited by the ACGME for calendar years 2000 through 2021. Internal medicine residents were used as the reference population. ^aStatistically significant at the P < .05 level. ^bThere were 1 or no deaths by accidental overdose for emergency medicine, obstetrics and gynecology, pathology, diagnostic radiology, and surgery, precluding reliable incidence rate ratio calculations.