Abstract
Background
Having highly educated spouses may influence physicians' choice of practice location.
Methods
We analyzed a representative physician sample in the 2020 Japan Population Census.
Results
Regression analyses adjusting for physicians' characteristics showed that physicians married to highly educated (with a graduate degree) spouses were less likely to practice in low physician supply areas (LPSAs) than other married physicians (adjusted, 7.8% vs. 9.1%; adjusted odds ratio, 0.84; 95% CI, 0.72–0.98).
Conclusion
Having highly educated spouses was associated with a lower likelihood of practicing in LPSAs. Qualitative research is needed to explore how spousal academic attainment influences physicians' practice location choices.
Keywords: dual‐career marriages, health policy, homogamy, physician distribution, physician work‐life balance
Physicians’ spouses are often highly educated, with independent careers, and having a highly educated spouse may influence physicians’ choice of practice location and exacerbate physician maldistribution. In this study using a representative physician sample in the 2020 Japan Population Census, physicians married to highly educated (with a graduate degree) spouses were less likely to practice in low physician supply areas than other married physicians. Our findings suggest that policymakers addressing physician geographical maldistribution should consider the careers and work‐life balance support for physicians’ family members.
1. BACKGROUND
Low physician supply in rural, medically underserved areas is an important public health issue. 1 In Japan, the geographic maldistribution of physicians has not improved dramatically over the past two decades. 2 Rural origin, age, and gender have been linked to physician choice of rural settings. 3 , 4 Another potential factor may be that many physicians have highly educated spouses with independent careers. 5 Since research has shown that highly educated couples tend to cluster in urban areas seeking job opportunities commensurate with their academic attainment, 6 having a highly educated spouse may influence physicians' choice of practice location and exacerbate physician geographical maldistribution. However, there is little national‐level data on the academic attainment of physicians' spouses and its association with physicians' choice of practice location globally. 3 Therefore, we examined whether having highly educated spouses was associated with working in areas with a low physician supply.
2. METHODS
2.1. Data
We analyzed stratified random samples from the 2020 Japan Population Census (conducted on October 1, 2020). The details of the Census were described elsewhere. 5 We analyzed data on physicians aged ≥25 years who were identified by self‐reported occupation, an approach used and validated in a prior study. 5 To generate nationally representative estimates, all the analyses were conducted using census‐provided weights.
2.2. Exposure
Physicians were categorized by marital status as “single,” “married to a highly educated spouse,” or “other married.” Being highly educated was defined as having a graduate degree or self‐reporting an occupation that requires a 6‐year college education (i.e., physicians, dentists, and veterinarians). 3 Pharmacists were not classified as highly educated because as of 2020, the majority of pharmacists had a 4‐year college education (graduates in 2012 or later had a 6‐year college education).
2.3. Outcome
The outcome was physician's practice at low physician supply areas (LPSAs). Japan contains 335 secondary medical areas (SMAs), which represent regional healthcare markets for general inpatient care. LPSAs, also known as “ishi syosuu kuiki” in Japanese, are SMAs in the bottom 33.3% of the “physician maldistribution index,” which was developed and published by the Ministry of Health, Labour and Welfare (MHLW) in 2020. 7 The MHLW calculated the physician maldistribution index based on the number of physicians per population in the SMA. In their calculations, to account for differences in the demographic composition of physicians across SMAs and the resulting variations in hours worked, the number of physicians was standardized by an indirect method using the average hours worked according to gender and age for physicians nationwide, which was obtained from a past survey by a research group funded by the government. 8 Furthermore, to account for differences in the demographic composition of inhabitants across SMAs and the resulting variations in demand for medical care, the population in each SMA was also standardized by the indirect method using the nationwide average number of visits to medical facilities per year according to gender and age, which was obtained from the Patient Survey, 2017 (a government fundamental statistical survey).
2.4. Statistical analysis
We compared physician characteristics (gender, age, nationality [Japanese or foreign], presence of children [aged <18]) and practice at LPSAs by marital status. We also used a multivariable logistic regression model to investigate the association between marital status and practice in LPSAs, adjusting for physician gender, age (in 5‐year increments), nationality, and presence of children. Standard errors were clustered at the household level. Second, based on the a priori hypothesis that the choice of practice location may vary by physicians' demographics and family structure, we conducted analyses stratified by physician's gender, age (<65 or ≥65 years) or presence of children. Finally, given the high frequency of dual‐physician couples, 5 we subgrouped physicians married highly educated spouses as those married to highly educated non‐physicians vs. those married to physicians and repeated analyses. Two‐tailed p values <0.05 were interpreted as statistically significant. We used Stata version 17 (StataCorp).
3. RESULTS
Of the 23,658 physicians analyzed (mean [SD] age 51.0 [15.0] years, 23.0% female), 4101 (18.2%) were single, 5815 (26.4%) were married to highly educated spouses, and 13,742 (55.5%) were other married physicians (Table 1). Among married physicians, those who married to highly educated spouses were younger than other married physicians (mean 47.3 vs. 55.1 years) and were more likely to be female (48.3% vs. 6.7%).
TABLE 1.
Characteristics of Physicians, Japan Population Census 2020.
| Physicians, weighted no. (%) a | ||||||
|---|---|---|---|---|---|---|
| Total (n = 23,658) | By physician's marital status b | p Values c | ||||
| Single (n = 4101) | Married to a highly educated spouse (n = 5815) | Other married (n = 13,742) | Single vs. other married | Married to a highly educated spouse vs. other married | ||
| Female | 5440 (23.0) | 1717 (41.9) | 2806 (48.3) | 918 (6.7) | <0.001 | <0.001 |
| Age, mean (SD) years | 51.0 (15.0) | 42.8 (16.8) | 47.3 (12.6) | 55.1 (13.9) | <0.001 | <0.001 |
| Age category, years | ||||||
| <65 | 18,816 (79.5) | 3595 (87.6) | 5165 (88.8) | 10,056 (73.2) | <0.001 | <0.001 |
| ≥65 | 4842 (20.5) | 507 (12.4) | 650 (11.2) | 3686 (26.8) | <0.001 | <0.001 |
| Foreign nationality | 149 (0.6) | 38 (0.9) | 30 (0.5) | 81 (0.6) | 0.06 | 0.59 |
| Having children aged <18 years | 9299 (39.3) | 166 (4.1) | 3396 (58.4) | 5736 (41.7) | <0.001 | <0.001 |
| Practice in a low physician supply area | 2098 (8.9) | 339 (8.3) | 408 (7.0) | 1351 (9.8) | <0.001 | <0.001 |
Analyses were conducted with stratified random samples of the Japan Population Census (an 8.5% sample for analytic physicians). To generate nationally representative estimates, census‐provided weights were used; therefore, the number of physicians in each category may not add up to the total number of physicians.
Married physicians whose spouses had graduate degrees or self‐reported occupations requiring 6 years of college education (i.e., physicians, dentists, and veterinarians) were categorized as “married to a highly educated spouse;” otherwise, they were categorized as “other married.”
p Values were estimated with the use of Student's t‐test for means or the z‐test for comparison of proportions.
Compared with other married physicians, those married to highly educated spouses were less likely to practice in LPSAs (adjusted, 7.8% vs. 9.1%; adjusted odds ratio [aOR], 0.84; 95% CI, 0.72–0.98; p = 0.02) (Table 2). In stratified analyses, this association was observed for male physicians (aOR, 0.79; 95% CI, 0.68–0.93; p = 0.005) and those with children (aOR, 0.78; 95% CI, 0.63–0.96; p = 0.02), but not for female physicians or those without children. The association was observed regardless of age category, although it was no longer statistically significant because of reduced sample size among physicians aged ≥65. Finally, the differences were particularly pronounced when comparing physicians married to highly educated non‐physicians with other married physicians.
TABLE 2.
Association between having a highly educated spouse and practicing in low physician supply areas.
| Marital status of the physician | Weighted no. | Adjusted percentage a | Adjusted odds ratio (95% CI) a | p Value |
|---|---|---|---|---|
| Main analysis | ||||
| Single | 4101 | 9.3 (8.2–10.5) | 1.02 (0.87–1.20) | 0.77 |
| Married to a highly educated spouse | 5815 | 7.8 (6.8–8.8) | 0.84 (0.72–0.98) | 0.02 |
| Other married | 13,742 | 9.1 (8.6–9.7) | Reference | |
| By physician's gender | ||||
| Male | ||||
| Single | 2384 | 10.3 (8.7–11.8) | 1.06 (0.88–1.28) | 0.55 |
| Married to a highly educated spouse | 3009 | 7.9 (6.8–9.0) | 0.79 (0.68–0.93) | 0.005 |
| Other married | 12,824 | 9.8 (9.2–10.3) | Reference | |
| Female | ||||
| Single | 1717 | 6.7 (5.3–8.1) | 0.92 (0.64–1.31) | 0.64 |
| Married to a highly educated spouse | 2806 | 6.6 (5.5–7.6) | 0.90 (0.66–1.23) | 0.50 |
| Other married | 918 | 7.2 (5.5–9.0) | Reference | |
| By physician's age category | ||||
| <65 years | ||||
| Single | 3595 | 8.8 (7.5–10.1) | 1.04 (0.86–1.26) | 0.68 |
| Married to a highly educated spouse | 5165 | 7.3 (6.3–8.3) | 0.85 (0.72–1.00) | 0.048 |
| Other married | 10,056 | 8.5 (7.9–9.1) | Reference | |
| ≥65 years | ||||
| Single | 507 | 11.1 (8.3–14.0) | 0.95 (0.70–1.30) | 0.75 |
| Married to a highly educated spouse | 650 | 9.7 (6.6–12.7) | 0.81 (0.56–1.17) | 0.26 |
| Other married | 3686 | 11.6 (10.6–12.7) | Reference | |
| By presence of children | ||||
| With children | ||||
| Single | 166 | 9.7 (4.0–15.3) | 1.20 (0.62–2.33) | 0.60 |
| Married to a highly educated spouse | 3396 | 6.5 (5.4–7.6) | 0.78 (0.63–0.96) | 0.02 |
| Other married | 5736 | 8.2 (7.5–9.0) | Reference | |
| Without children | ||||
| Single | 3935 | 9.8 (8.6–11.0) | 1.01 (0.85–1.20) | 0.91 |
| Married to a highly educated spouse | 2418 | 9.0 (7.4–10.6) | 0.92 (0.74–1.13) | 0.42 |
| Other married | 8006 | 9.7 (9.0–10.4) | Reference | |
| By spouse's physician status | ||||
| Single | 4101 | 9.3 (8.2–10.5) | 1.03 (0.87–1.21) | 0.75 |
| Married to a highly educated non‐physician | 1132 | 6.9 (5.3–8.5) | 0.74 (0.57–0.96) | 0.02 |
| Married to a physician | 4683 | 8.0 (6.9–9.2) | 0.87 (0.73–1.03) | 0.10 |
| Other married | 13,742 | 9.1 (8.6–9.6) | Reference | |
Logistic regression analyses were conducted for practicing in low physician supply areas by marital status, adjusting for physician gender, age in 5‐year increments, nationality, and the presence of children aged <18 years. Census‐provided weights were used; therefore, the number of physicians in each category may not add up to the total number of physicians. The adjusted percentage of individuals practicing in low physician supply areas was calculated according to physician's marital status using marginal standardization.
4. DISCUSSION
This study using a nationally representative physician sample showed that physicians with highly educated spouses were less likely than other married physicians to practice in LPSAs. This finding aligns with a previous US study 3 ; the current study further indicates that the difference is particularly evident for male physicians. This may be explained by the fact that the spouses of female physicians are predominantly physicians, 5 who can easily continue their careers in rural areas, while the spouses of male physicians are largely non‐physicians, who may have difficulty finding jobs commensurate with their educational background in rural areas (especially if they have children). 4 Aligned with this argument, our findings are particularly evident when focusing on physicians married to highly educated non‐physicians. Our findings suggest that policymakers addressing physician geographical maldistribution should consider the careers and work‐life balance support for physicians' family members. In addition, highly educated parents may be more inclined to seek urban areas for the sake of their children's education.
Although support for physician spouses in rural physician retention programs is not always adequate in many countries, the importance of spouse's job opportunities and educational resources for children is widely recognized. 9 Recently, Japan has promoted the planned allocation of physicians (in 2020, a resident cap by region and specialty was introduced, and, starting in 2023, more than 6 months of practice experience in LPSAs were required for hospital directors of community medical support hospital [i.e., acute care hospital with a large proportion of patients being referrals and emergency patients]), but this plan exhibited limited consideration for physicians' family members. 10 Future longitudinal studies, analyses by specialties, and in‐depth qualitative study are warranted to understand the association between spousal academic attainment and physicians' practice location.
Study limitations included the use of cross‐sectional data and the inability to fully account for unmeasured confounders (e.g., physicians' specialties and rural origins 4 ). For example, working in an urban area may increase the possibility of marrying a highly educated spouse (reverse causation). Also, physicians in certain specialties may be more likely to work in urban areas and also more likely to marry highly educated spouses. If these were true, the association should be observed even among physicians without children. However, the fact that the likelihood of practicing in LPSAs was similar regardless of spousal education level among physicians without children indicates that these possibilities should minimally explain our findings. Additionally, while some highly educated spouses might have been misclassified as non‐highly educated, this would bias our estimate toward the null, and the true difference in outcome would be larger than what we have estimated. Finally, the “single” group included both never‐married and divorced physicians, but we could not distinguish between them.
CONFLICT OF INTEREST STATEMENT
The authors have stated explicitly that there are no conflicts of interest in connection with this article.
ETHICAL APPROVAL
Ethics approval statement: The University of Tokyo Ethics Committee approved this study. Patient consent statement: The requirement for informed consent was waived by the Ethics Committee because all data were blinded. Clinical trial regitration: None.
ACKNOWLEDGMENTS
We appreciate the enumerators of the Japan Population Census. This study was funded by the Japan Society for the Promotion of Science Grant‐in‐Aid for Scientific Research (B) (Grant No. 22H03325: 23K24583). The funder had no role in the design and conducting of the study; the collection, management, analysis, or interpretation of the data; the preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.
Miyawaki A, Tomio J. Practice locations of physicians with highly educated spouses in Japan: A cross‐sectional study using National Census Data. J Gen Fam Med. 2024;25:284–288. 10.1002/jgf2.710
REFERENCES
- 1. OECD . Geographic distribution of doctors. Health at a glance 2021: OECD indicators. Paris: OECD Publishing; 2021. [Google Scholar]
- 2. Ikesu R, Miyawaki A, Kobayashi Y. Physician distribution by specialty and practice setting: findings in Japan in 2000, 2010 and 2016. Tohoku J Exp Med. 2020;251(1):1–8. [DOI] [PubMed] [Google Scholar]
- 3. Staiger DO, Marshall SM, Goodman DC, Auerbach DI, Buerhaus PI. Association between having a highly educated spouse and physician practice in rural underserved areas. JAMA. 2016;315(9):939–941. [DOI] [PubMed] [Google Scholar]
- 4. Fritsma T, Henning‐Smith C, Gauer JL, Khan F, Rosenberg ME, Clark K, et al. Factors associated with health care professionals' choice to practice in rural Minnesota. JAMA Netw Open. 2023;6(5):e2310332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Miyawaki A. Full‐time work rates of physicians with physician spouses vs nonphysician spouses in Japan. JAMA Netw Open. 2022;5(11):e2242143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Costa DL, Kahn ME. Power couples: changes in the locational choice of the college educated, 1940–1990. Q J Econ. 2000;115(4):1287–1315. [Google Scholar]
- 7. Ministry of Health Labour and Welfare . The 35th subcommittee on supply and demand of physicians, materials. 2020. Available from: https://www.mhlw.go.jp/stf/newpage_13283.html. Accessed 25 Mar 2024.
- 8. Ministry of Health Labour and Welfare . Survey on working status of physicians and their working style intentions. 2016. Available from: https://www.mhlw.go.jp/file/05‐Shingikai‐10801000‐Iseikyoku‐Soumuka/0000161146.pdf. Accessed 25 Mar 2024.
- 9. AAMC . Attracting the next generation of physicians to rural medicine. Available from: https://www.aamc.org/news/attracting‐next‐generation‐physicians‐rural‐medicine. Accessed 25 Mar 2024.
- 10. Ministry of Health Labour and Welfare . 5th interium summary. Subcommittee on supply and demand of physicians. 2022. Available from: https://www.mhlw.go.jp/stf/newpage_23302.html. Accessed 25 Mar 2024.