Journal

of Epidemiology

Vol. 9, No. 5 Nov

Life Span

of Japanese

Male

Medical

Doctors

Motoi Nishi, Hirotsugu Miyake, Tatsuya Kato, Masami Yamazoe, Eiji Tanaka, Takao Ishii, and Katsuya Usui

There have been few reports with regard to the life spans of medical doctors. The status of the medical doctors graduating from 1926 to 1974, alive or dead as of October 1996, was ascertained on the basis of the list of graduates from the School of Medicine, Hokkaido University. Excluding data on female doctors and those who died in battle during World War II, data on a total of 3,982 doctors were available for study. Their mortality as of October 1996 decreased in parallel with the graduation year. Their mean future life span at graduation was estimated to be about 52.88 years (95% Cl, 52.45-53.31) through linear regression (r=0.992). Their mean age at graduation was 25.17 years. This was not different from the future life expectancy at 25 years of age of the general population (52.35 years). The future life span of surgeons and gynecologists-obstetricians was shorter than that of the doctors of basic medical sciences and internal medicine. This difference might be accounted for by factors peculiar to each speciality (e.g., exposure to blood) or by the degree of stress from work. J Epidemiol, 1999; 9 : 315-319 medical doctor, life span, occupation, mortality

Mortality is different among occupations. According to the Japanese vital statistics data in 1990 '), the mortality of male agricultural workers and fishers was 11.9 per 1,000 population, while that of clerical workers was 2.6. Since doctors have special knowledge of health and diseases, they appear to be more advantageous than general people concerning the span of life. On the other hand, doctors often have disadvantages that they have to work hard or have the occasions to be infected with various diseases. Several papers have reported the mortality of physicians, but there have been few reports as to their life span. We investigated it, employing the lists of graduates from the School of Medicine, Hokkaido University. SUBJECTS AND METHODS

published by the school every two years, which also has the data on their specialities and deaths. The newest speciality of each doctor was employed for this study. It also has a record of deaths in battle during World War II. The newest list has the data as of October 1996. From 1926 through 1974, a total of 4,192 doctors graduated. Among them there were 101 female doctors and 109 deaths in battle during World War II . Excluding them, data on a total of 3,982 male doctors were available. RESULTS 1. Mortality of graduates of each year Figure 1 shows the mortality by graduation year as of October 1996. Most of the doctors who graduated from 1926 to 1930 have already died. The mortality decreased almost in a straight line until 1960. After 1960 the mortality became almost stable (about 3-5%) . The correlation coefficient between mortality and graduation year from 1926 to 1960 was

The first graduation from the School of Medicine , Hokkaido University was in 1926. Since then there have been about 100 graduates every year. Their names have been

Received October 20, 1998 ; accepted February 14, 1999. Department of Public Health, Sapporo Medical University. Address for correspondence : Dr. Motoi Nishi, Department apporo, Hokkaido, 060-8556, Japan.

of

Public Health

, Sapporo

Medical

University,

S.1, W.17,

Chuo-ku, S

315

M. Nishi, et al.

316

Graduation

year

Figure

1. Actual mortality

(broken line, %) and mortality

by the moving average for 5 years (solid line, %).

0.981. This figure shows the mortality curve after "smoothing" by using the moving average for 5 years, as well. The mortality for 1930 by this method is thus the average of the mortality from 1928 to 1932. The correlation coefficient between 1928 and 1960 was 0.992. The correlation coefficient calculated using a 3-year moving average was 0.990. The contribution of the regression line using the 5-year moving average was, consequently, the highest. 2. Calculation of life span for all subjects Using the coefficient and the intercept of the regression line obtained from the 5-year moving average, the graduation year when 50% of the graduates had died was estimated. The year when the mortality was 50% was 1943.62 (95% CI, 1943.191944.05). Since the present data were as of October 1996 and graduation was at the end of March, their mean future life span at graduation was estimated to be 52.88 years (1996 1943.62+ 0.5). Mean age at graduation was estimated through the data on birth date and year of graduation in the list of doctors in Hokkaido Prefecture published in 1975 by Hokkaido Medical Press. Birth dates of a total of 2,072 doctors (52.0%) were identified. The average age at graduation was 25.19 ( 1.77) years. The future life expectancy of a 25-year-old Japanese male in 1995 was 52.35 years 2).(A 0-year-old male, 76.38 years.) The difference between male doctors and the general population was small. 3. Mean future lifespan by speciality There were 3,775 doctors (94.8%) whose speciality was identified. Table 1 shows the mortality as of October 1996 of

the doctors of basic medical sciences, internal medicine, surgery, and gynecology-obstetrics. The mortalities of surgeons and gynecologists-obstetricians were high, while those for doctors in basic medical sciences and internal medicine were low. The graduation years when mortality was 75%, 50% and 25% were estimated in the same way as above, using 5-year moving average (Table 2). The correlation coefficients between them were sufficiently high. The graduation years of the doctors of basic medical sciences were different from those of gynecologists-obstetriciansby 2-4 years. Their mean life spans at graduation were calculated by the same method as above (Table 3). Doctors of basic medical sciences and internal medicine had longer future life spans than surgeons and gynecologists-obstetricians. 4.The survival rate of professors of the school The lists of graduates of the school include the list of professors. All of their dates of birth and death were clarified. Their birth years ranged from 1880 to 1952, while the vast majority of the present subject were born after 1900. The lifestyle of professors is somewhat different from the general doctors, and all of them lived up to be 40 years of age at least. Though there.are such biases, their survival rates by birth year were calculated, using the Kaplan-Meier's method (Table 4). The survival rates of those who were bom before 1899 are fixed, since all of them had died. The figures of those who were born from 1900 to 1924 may change in the future, since some of them are still alive. But at several ages after 75 years the survival rates of the former were higher than the latter. So far as these two cohorts are concerned, there seems to be no large cohort effect that the survival rate improves according as the birth years become newer. In addition, the

317

Life Span

of Male Doctors

Table

1. Mortalities medicine, graduation

(%; number in parentheses) of doctors in basic medical science, internal surgery, and gynecology-obstetrics at the end of October 1996 by years.

Table

2. Estimation

of graduation

year by linear regression

(95% CI) when mortality

is 75%,

Table

3. Future life span of medical doctors estimated by the year of 50% mortality.

at graduation

50% survival 80 years, 1924. The sciences mean

age

of those was

who

were born before in those born from

1899 was 1900 to

while it

79 years

age at death of the professors years,

of basic medical medicine was

was 75.13

while that of clinical

74.25 years.

M. Nishi, et al.

318

Table 4. Survival rate (%) by birth year of the professors in the School of Medicine, Hokkaido University as of October 1996 (Kaplan-Meier's method).

DISCUSSION

Since the birth dates of about half of the subjects were unknown, and the data of date of death were incomplete, the life table method could not be employed. Therefore, we used the linear regression as the second best method. Similarly, the cohort effect could not be taken into account. However, so far as the data of the professors are concerned, the cohort effect does not seem to be large. Most doctors are healthy at graduation. On the other hand, the general population includes many unhealthy persons. The mortality of jobless males was 26.3 per 1,000 in 1990 in Japan 1). Considering this "healthy worker effect" 3),thesubstantial life expectancy of doctors may be the same as, or even shorter than, that of the general population. In any case, however, their life expectancy may not be very different from the general one, as has been reported by several authors 4, 5) The speciality did bring about different life spans among doctors, though there might be a tendency that those who enter the university directly upon graduation from high school or graduate without staying in the same class become a researcher of the basic medical sciences. In fact, the age at death of the professors of basic medical sciences was older than that of clinical medicine. Mental and physical stress from work may account for the difference in life span among the four specialities reported herein. Surgeons often have to perform difficult operations. Obstetricians must work even at midnight when births are difficult. Occasions of contact with blood may lead to various infectious diseases, e.g., viral hepatitis, in these two specialities. On the other hand, the doctors of basic medical sciences have a higher degree of freedom in work. Their work is similar to that of clerical workers whose mortality is

reported to be low. Exposure peculiar to each speciality or stress from work might influence the life span. It is possible that some other specialists, e.g., radiologists, have hazards peculiar to their occupation e. However, we could not do analysis concerning other specialities, since the number of doctors in them was not large enough. It seems that doctors as a whole do not benefit from their special knowledge, so far as the mean life expectancy is concerned. Moreover, there might be some disadvantages in several specialities. ACKNOWLEDGEMENT The authors thank the staff of the alumni association of the School of Medicine, Hokkaido University for their help with compilation of the data. REFERENCES 1) Health and Welfare Statistics Association. Table 14. Vital statistics by occupation. J Health and Welfare Statistics (Kokumin Eisei No Doukou) 1997;44:435 (in Japanese). 2) Health and Welfare Statistics Association. The 18th Life Tables. Statistics and Information Department Minister's Secretariat, ed. 1998, Tokyo (in Japanese). 3) Araki S, Murata K, Kumagai K, Nagasu M. Mortality of medical practitioners in Japan: social class and "healthy worker effect". Am J Ind Med 1986;10:91-99. 4) Kono S, Ikeda M, Tokudome S, Nishizumi M, Kuratsune M. Cause-specific mortality among male Japanese physicians: a cohort study. Asian Med J 1988;31:453-461. 5) Rimpera AH, Nurminen NM, Pulkkinen PO, Rimpela

319

Life Span

of Male Doctors

MK, Valkonen T. Mortality of doctors: do doctors benefit from their medical knowledge? Lancet 1987; 1:8486. 6) Matanoski GM, Seltser R, Sartwell PE, Diamond EL,

Eliott EA. The current mortality rates of radiologists and other physician specialists: deaths from all causes and from cancer. Am J Epidemiol 1975; 101:188-198.