Abstract
This cross-sectional study investigates the association of physician vs nonphysician spousal occupation with full-time work rates among married physicians.
Introduction
With an increasing number of female physicians, dual-physician marriages (ie, physicians married to physicians) are increasingly a topic of discourse in the medical literature.1,2 However, there are limited national-level data on the association of spousal occupation with physician labor force participation.3,4
Methods
The University of Tokyo ethics committee approved this cross-sectional study and granted a waiver of informed consent because we used deidentified data. The STROBE reporting guideline was followed. We used stratified random samples from the Japan Population Census (2005-2015), a quinquennial survey conducted on October 1, collecting information on family structures, demographics, and employment of all Japanese residents (95%-100% response rates).5
Married individuals working as physicians were analyzed. Specialty data were unavailable. We limited analyses to couples aged 25 to 50 years (ie, child-rearing years). Same-sex couples were excluded. The outcome was a full-time work indicator, dichotomized for physicians (1 = “primarily worked,” and 0 = “worked part time besides housework,” “worked part time while studying,” or “on leave”) according to the question, “How did you work during the week of September 24-30?” (eMethods in the Supplement).
We used multivariable modified Poisson regression models separately for male and female physicians to investigate the association between spousal occupation (physician or nonphysician) and full-time work rates. We adjusted for both spouses’ ages, youngest child’s age, number of children, urbanicity of residence, prefecture, and year indicators. Analyses were repeated after (1) further subgrouping physicians married to nonphysicians into those married to individuals in other high-income occupations vs others or (2) stratifying physicians by youngest child age. High-income occupations were determined by mean income per occupation (eMethods in the Supplement). Analyses were conducted using Stata statistical software version 17 (StataCorp) from June 24 to September 21, 2022.
Results
We observed 25 321 married physicians: 20 858 men (82.4%; mean [SD] age, 40.8 [6.2] years) and 4463 women (17.6%; mean [SD] age, 37.9 [5.8] years) (Table 1). Approximately 15% of male physicians and 69% of female physicians had a physician spouse.
Table 1. Characteristics of Married Physicians.
| Characteristic | Physicians, weighted No. (%) (N = 25 321)a | |||||
|---|---|---|---|---|---|---|
| Men | Women | |||||
| Total (n = 20 858) | Physician spouse (n = 3074) | Nonphysician spouse (n = 17 783) | Total (n = 4463) | Physician spouse (n = 3074) | Nonphysician spouse (n = 1389) | |
| Age, mean (SD), y | ||||||
| Physician | 40.8 (6.2) | 39.4 (6.3) | 41.1 (6.1) | 37.9 (5.8) | 37.7 (6.0) | 38.3 (5.5) |
| Spouse | 38.1 (6.0) | 37.7 (6.0) | 38.1 (6.0) | 39.4 (6.2) | 39.4 (6.3) | 39.4 (6.0) |
| Spouse’s occupation | ||||||
| Physician | 3074 (14.7) | 3074 (100) | 0 | 3074 (68.9) | 3074 (100) | 0 |
| Other high-income occupationb | 524 (2.5) | 0 | 524 (2.9) | 594 (13.3) | 0 | 594 (42.8) |
| Non–high-income occupation | 5328 (25.5) | 0 | 5328 (30.0) | 711 (15.9) | 0 | 711 (51.2) |
| Unemployed | 11 931 (57.2) | 0 | 11 931 (67.1) | 84 (1.9) | 0 | 84 (6.0) |
| Age of youngest child, yc | ||||||
| No children | 3548 (17.0) | 718 (23.4) | 2830 (15.9) | 1119 (25.1) | 718 (23.4) | 401 (28.9) |
| 0-3 | 6473 (31.0) | 990 (32.2) | 5483 (30.8) | 1467 (32.9) | 990 (32.2) | 477 (34.3) |
| 4-6 | 3272 (15.7) | 474 (15.4) | 2798 (15.7) | 665 (14.9) | 474 (15.4) | 191 (13.8) |
| ≥7 | 7565 (36.3) | 893 (29.0) | 6672 (37.5) | 1212 (27.2) | 893 (29.0) | 320 (23.0) |
| Children, No. | ||||||
| 0 | 3548 (17.0) | 771 (25.1) | 2830 (15.9) | 1119 (25.1) | 718 (23.4) | 401 (28.9) |
| 1 | 5482 (26.3) | 854 (27.8) | 4628 (26.0) | 1323 (29.6) | 854 (27.8) | 469 (33.8) |
| 2 | 8008 (38.4) | 1090 (35.5) | 6918 (38.9) | 1510 (33.8) | 1090 (35.5) | 420 (30.2) |
| ≥3 | 3820 (18.3) | 412 (13.4) | 3408 (19.2) | 511 (11.5) | 412 (13.4) | 99 (7.1) |
| Living in urban aread | 18 158 (87.1) | 2780 (90.4) | 15 378 (86.5) | 4002 (89.7) | 2780 (90.4) | 1222 (88.0) |
| Year of census | ||||||
| 2005 | 7410 (35.5) | 903 (29.4) | 6507 (36.6) | 1233 (27.6) | 903 (29.4) | 330 (23.8) |
| 2010 | 7008 (33.6) | 1006 (32.7) | 6002 (33.8) | 1500 (33.6) | 1006 (32.7) | 494 (35.6) |
| 2015 | 6439 (30.9) | 1165 (37.9) | 5274 (29.7) | 1730 (38.8) | 1165 (37.9) | 565 (40.7) |
Analysis was conducted using stratified random samples of the Japan Population Census (an 11.7% sample for physicians). To provide nationally representative estimates, census-provided weights were used; therefore, numbers of physicians in each category may not add up to the total number of physicians.
Other high-income occupations were determined based on mean annual income per occupation (eMethods in the Supplement).
Ages at the end of the academic year were used to determine categories for youngest child age.
Urban areas were defined as census-defined densely inhabited districts.
Nearly all male physicians worked full time with or without a physician spouse (Table 2). In contrast, adjusted full-time work rates were lower for female physicians married to physicians vs those married to nonphysicians (68.1% vs 76.3%; P < .001). However, there was no significant difference in adjusted full-time work rates between female physicians married to physicians and those with spouses in other high-income occupations (68.3% vs 71.8%; P = .10).
Table 2. Adjusted Full-time Work Rates by Occupation of Spouse.
| Spouse’s occupation | Male physician | Female physician | ||||||
|---|---|---|---|---|---|---|---|---|
| Weighted No. | Adjusted full-time work rate (95% CI), % | Adjusted rate ratio (95% CI)a | P value | Weighted No. | Adjusted full-time work rate (95% CI), % | Adjusted rate ratio (95% CI)a | P value | |
| Main analysis | ||||||||
| Physician | 3074 | 97.7 (97.1-98.2) | 1 [Reference] | NA | 3074 | 68.1 (67.2-69.0) | 1 [Reference] | NA |
| Nonphysician | 17 783 | 97.9 (97.8-98.0) | 1.00 (1.00-1.01) | .51 | 1389 | 76.3 (74.4-78.3) | 1.12 (1.08-1.16) | <.001 |
| By income level of spouse’s occupation | ||||||||
| Physician | 3074 | 97.6 (97.0-98.2) | 1 [Reference] | NA | 3074 | 68.3 (67.3-69.3) | 1 [Reference] | NA |
| Other high-income occupation | 524 | 97.3 (96.1-98.5) | 1.00 (0.98-1.01) | .70 | 594 | 71.8 (68.3-75.3) | 1.05 (0.99-1.12) | .10 |
| Otherb | 17 259 | 97.9 (97.8-98.0) | 1.00 (1.00-1.01) | .44 | 795 | 78.9 (76.5-81.3) | 1.15 (1.11-1.20) | <.001 |
| By age of youngest child, yc | ||||||||
| No children | ||||||||
| Physician | 718 | 95.7 (94.3-97.2) | 1 [Reference] | NA | 718 | 86.0 (84.5-87.6) | 1 [Reference] | NA |
| Nonphysician | 2830 | 95.7 (95.3-96.1) | 1.00 (0.98-1.02) | .99 | 401 | 87.4 (84.7-90.1) | 1.02 (0.97-1.07) | .52 |
| 0-3 | ||||||||
| Physician | 990 | 96.3 (94.9-97.7) | 1 [Reference] | NA | 990 | 53.9 (52.0-55.8) | 1 [Reference] | NA |
| Nonphysician | 5483 | 96.4 (96.2-96.7) | 1.00 (0.98-1.02) | .89 | 477 | 66.1 (62.2-70.0) | 1.23 (1.12-1.35) | <.001 |
| 4-6 | ||||||||
| Physician | 474 | 98.1 (97.2-99.1) | 1 [Reference] | NA | 474 | 68.5 (65.6-71.4) | 1 [Reference] | NA |
| Nonphysician | 2798 | 98.7 (98.6-98.9) | 1.01 (0.99-1.02) | .28 | 191 | 80.7 (73.4-88.0) | 1.18 (1.03-1.34) | .02 |
| ≥7 | ||||||||
| Physician | 893 | 99.5 (99.1-100.0) | 1 [Reference] | NA | 893 | 68.9 (67.4-70.5) | 1 [Reference] | NA |
| Nonphysician | 6672 | 99.8 (99.7-99.8) | 1.00 (1.00-1.01) | .32 | 320 | 75.5 (71.1-79.8) | 1.10 (1.01-1.19) | .03 |
Abbreviation: NA, not applicable.
Regression analysis was conducted for the full-time work indicator by spouse’s occupation (physicians or nonphysician), adjusting for both spouses’ ages (25-29, 30-34, 35-39, 40-44, and 45-50 years), youngest child’s age (no children and 0-3, 4-6, and ≥7 years), number of children (0, 1, 2, and ≥3), urbanicity of residence, prefecture, and year indicators. Multivariable modified Poisson regression models with standard errors clustered at the prefecture level were used. Census-provided weights were used. Adjusted full-time work rates were calculated according to spouse’s occupation using marginal standardization. P values < .05 were interpreted as statistically significant.
The other occupation group consisted of individuals with non–high-income occupations and those who were unemployed.
Interactions between spouse’s occupation and age of youngest child were tested using the Wald test (P = .83 and P < .001 for male and female physicians, respectively).
For female physicians without children, adjusted full-time work rates of those married to physicians or nonphysicians were similar (86.0% vs 87.4%; P = .52). However, among female physicians with a youngest child aged 0 to 3 years, adjusted full-time work rates were lower for women married to physicians vs nonphysicians (53.9% vs 66.1%; P < .001). Similar differences remained when the youngest child was aged 4 to 6 years or 7 years or older.
Discussion
This cross-sectional study in Japan found that having a physician spouse was associated with lower full-time work rates among female physicians with children. This outcome may be associated with physician spouses’ higher income levels compared with nonphysician spouses. Alternatively, female physicians engaged or married to physicians may choose specialties considered to be parenting friendly and part-time work. Study limitations included unmeasured specialties and individual income and use of cross-sectional data. Generalizability to other countries is also unknown. Nevertheless, these findings aligned with prior US-based studies,3,4 despite US-Japan differences. These include a lower ratio of female to male physicians3 and more generous public childcare support in Japan,6 which legally grants 1-year paid parental leave and provides government-subsidized access to babysitters, childcare facilities, and after-school programs regardless of income level. Nearly 90% of preschoolers (ages 3-5 years) attend government-funded childcare facilities at subsidized rates.6
Full-time physician work often involves on-call responsibilities and long overtime hours. This may be associated with pressure on child-rearing female physicians to move to part-time work based on gender norms and domestic expectations. Our findings suggest that this choice for female physicians was accelerated for those married to another physician. Reducing overwork and providing flexible work environments (eg, shift work) may help child-rearing physicians, especially those in dual-physician marriages, continue their careers.
eMethods
eReferences
References
- 1.Ferrante L, Mody L. Dual-physician households: strategies for the 21st century. JAMA. 2019;321(22):2161-2162. doi: 10.1001/jama.2019.4413 [DOI] [PubMed] [Google Scholar]
- 2.Andersson L, Chudnovskaya M, Shahbazian R, Ghaznavi C, Ueda P. Nationwide study of trends in physician partner choice for childbearing unions. J Intern Med. 2022;292(1):165-167. doi: 10.1111/joim.13464 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ly DP, Seabury SA, Jena AB. Characteristics of U.S. physician marriages, 2000–2015: an analysis of data from a U.S. census survey. Ann Intern Med. 2018;168(5):375-376. doi: 10.7326/M17-1758 [DOI] [PubMed] [Google Scholar]
- 4.Hu X, Dill M. When physicians marry physicians: gender inequities in work hours and income. Womens Health Rep (New Rochelle). 2021;2(1):422-429. doi: 10.1089/whr.2021.0048 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Statistics Bureau of Japan . Population census. Accessed September 25, 2022. https://www.stat.go.jp/english/data/kokusei/index.html
- 6.Organisation for Economic Co-operation and Development . OECD family database. Accessed September 25, 2022. https://www.oecd.org/els/family/database.htm
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eMethods
eReferences