Blecker S, Johnson NJ, Altekruse S, Horwitz LI. Association of Occupation as a Physician With Likelihood of Dying in a Hospital. JAMA. 2016;315(3):301-303. doi:10.1001/jama.2015.16976
Although most people report a preference to die at home vs at a medical facility,1,2 most deaths occur in a hospital or nursing home.3 Some articles have proposed that physicians die after receiving less intensive medical care and in a manner more consistent with end-of-life preferences than the general population,4 although studies on this topic are lacking.5 This study compared location of death for physicians with that of other clinicians, non–health care professionals with similar education levels, and the general population.
We used the National Longitudinal Mortality Study, a prospective random national sample of noninstitutionalized individuals based on US Census Bureau surveys matched to the National Death Index.6 We included individuals aged 30 to 98 years who died between 1979 and 2011 and excluded those missing the location of death (n = 12 205). This study was deemed not to involve human subjects research by the New York University School of Medicine institutional review board.
We categorized decedents into 4 mutually exclusive categories based on self-report of occupation or education: physician, other health professional (dentist, veterinarian, optometrist, podiatrist, nurse, pharmacist, dietician), other higher education, and all others. Other higher education included decedents not employed in health care who completed 6 or more years of postsecondary education and were therefore comparable with physicians in this marker of socioeconomic status. We assessed 2 outcomes: death in an inpatient hospital and, more broadly, death in a facility (ie, hospital, skilled nursing facility, professional center, physician office, or clinic).
Location and cause of death were obtained from death certificates. We developed logistic regression models to determine the association of occupation with location of death after adjusting for age, sex, race/ethnicity, and year of death. We performed subgroup analyses on males given the high number of male physicians, and male dentists, which is the subgroup thought to be most similar to physicians. Counts of decedents were presented as raw numbers; all other results incorporated survey weights.
Analyses were performed using SAS version 9.0 (SAS Institute Inc). Significance was prespecified with a 2-sided α level of .05.
Of the 471 243 decedents in the study, 815 were physicians, 2635 other health professionals, 15 308 other higher education, and 452 485 all others. Physicians were more likely to be male and less likely to be black compared with other groups, but the top 10 causes of death were similar among groups (Table 1).
Of deaths, 40.3% occurred in an inpatient hospital and 72.1% occurred in any facility. After adjusting for covariates and compared with physicians, other health occupation and other higher education were not associated with in-hospital death (Table 2). However, those in the all others category were more likely than physicians to die in a hospital (40.4% vs 38.3%; adjusted odds ratio [AOR], 1.10 [95% CI, 1.08-1.12]).
Rates of death in a facility were 63.3% for physicians, 65.4% for other health professionals, 66.1% for other higher education, and 72.4% for all others. Compared with physicians, the other 3 groups had higher likelihoods of dying in any facility (other health professionals: AOR, 1.14 [95% CI, 1.12-1.17]; other higher education: AOR, 1.12 [95% CI, 1.11-1.14]; all others: AOR, 1.34 [95% CI, 1.32-1.37]). Results were similar for subgroups of male decedents and male dentists, except both male health professionals and male dentists were less likely to die in a hospital than male physicians (Table 2).
Physicians were slightly less likely to die in a hospital than the general population, but equally as likely to die in a hospital as others in health care or with similar educational attainment. In addition, physicians were the least likely group to die at any facility.
Our results suggest that familiarity with health care (supported by the subgroup results) and educational attainment may have a small association with experience of death. These results may also be related to socioeconomic differences besides education, which we could not measure, or to differential treatment by clinicians.
Study limitations include that occupation or education may have changed between time of survey and death. Furthermore, death location may not reflect individual choice and was based on the death certificate, which may be subject to misclassification.
Corresponding Author: Saul Blecker, MD, MHS, New York University School of Medicine, 227 E 30th St, 648, New York, NY 10016 (firstname.lastname@example.org).
Author Contributions: Dr Johnson had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Blecker, Altekruse, Horwitz.
Acquisition, analysis, or interpretation of data: Blecker, Johnson, Horwitz.
Drafting of the manuscript: Blecker, Altekruse.
Critical revision of the manuscript for important intellectual content: Johnson, Horwitz.
Statistical analysis: Johnson, Altekruse.
Obtained funding: Altekruse.
Administrative, technical, or material support: Blecker, Altekruse, Horwitz.
Study supervision: Horwitz.
Conflict of Interest Disclosures: The authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Funding/Support: Dr Blecker was supported by grant K08HS23683 from the Agency for Healthcare Research and Quality.
Role of the Funder/Sponsor: The Agency for Healthcare Research and Quality had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: Any views expressed on statistical, methodological, technical, or operational issues are those of the authors and not necessarily those of the US Census Bureau.