Author: Naoki Okada, M.D. Junior Resident
Editor: Tetsuya Tanimoto, M.D.
In spite of the increasing number of physicians, a lot of Japanese hospitals still suffer from physician shortage. What causes physician shortage? We often find that in some prefectures, hospitals have a lot of physicians, while many others do not have enough physicians. That is, there is misdistribution of physicians in Japan. It is said that domestic migration is usually from rural to urban areas, which might partly explain the misdistribution. However, geographical differences and the magnitude of physicians’ migration in Japan are unclear. We created a model to learn how physicians migrate and why they move, and our analyses have been published in a peer-reviewed international medical journal, Medicine, in June 2018.
To achieve an adequate supply of health professionals is one of the most challenging issues all over the world. Misdistribution of physicians due to international migration among low-, middle-, and high-income countries is an area of concern. This is also true for domestic migration from rural to urban areas. Japan is categorized as a high-income country and suffers from an undersupply of physicians in some of its 47 prefectures. With a relatively large population of approximately 127 million in 2016, within a wide variety of geographic areas covering approximately 378,000 km, each prefecture has its own characteristic background. Furthermore, in Japan, the aging population has been increasing rapidly, which has led to a significant increase in physicians’ workload.
Historically, the number of physicians in Japan has been increasing over the past 50 years. In 1961, when universal health care coverage was established, the country had about 100,000 physicians, representing 103.6 physicians per 100,000 population (PPR). In 2016, the country had 319,480 physicians, representing 240.1 PPR throughout the country. The government also decided to establish two new medical universities, one of which opened in 2016 and one in 2017. Although the total number of physicians in the whole country has been increasing gradually, as mentioned above, and the lack of physicians has been mitigated, the inter-prefectural inequality within the country still remains a significant problem.
Among all 47 prefectures, the three highest practicing PPRs were 315.9 in Tokushima, 314.9 in Kyoto, and 306.0 in Kochi. On the other hand, the three lowest ratios were 160.1 in Saitama, 180.4 in Ibaraki, and 189.9 in Chiba. Why does this happen? Most Japanese physicians can choose their workplace freely. Under these circumstances, some advocates believe that a mitigation policy, such as obligatory service in rural areas, should be implemented to reduce the inequality of physicians’ distribution among prefectures. However, there is no official tracking system for noting their workplaces after graduation, nor is there quantitative data on the migration of physicians. No one knows how and why the physicians migrate precisely.
Therefore, we aimed to estimate the extent of migration of physicians among prefectures and explore possible factors associated with physicians’ migration patterns. Using the Japanese public database from the Ministry of Health, Labor and Welfare (MHLW), a model was constructed to ostensibly estimate the migration of physicians after graduation among prefectures over a 20-year period. The outflow and inflow ratio of the physicians was calculated in each prefecture. In addition, publicly available background sociodemographic data from the government were analyzed to explore potential influences on migration.
We show the results below. The maximum outflow and inflow ratio of physicians was 68% outflow in Ishikawa prefecture and 245% inflow in Chiba prefecture, with a maximum flow difference of 313%. Most of the outflow prefectures were in the rural regions facing the Sea of Japan or the Pacific Ocean (Ishikawa, Shimane, Kochi, Tottori, and Akita), while most of the inflow prefectures were peripheral zones of the metropolitan cities in more densely populated regions (Chiba, Saitama, Shizuoka, Hyogo, and Hiroshima). The metropolis of Tokyo, the most densely populated area, had an outflow ratio of 13%. Other urban prefectures, such as Aichi, Osaka, and Fukuoka, had modest inflow ratios, ranging from 7.7 to 22.8%.
Compared to the high inflow prefecture (HIP) group, the high outflow prefecture group (HOP) had a larger median practicing PPR (218.1 vs. 250.9) and a larger median annual newly licensed PPR (3.7 vs. 11.5). The HOP also had a larger median aged population ratio (23.8% vs. 27.9%) and a higher average age of physicians (50.0 years vs. 51.4 years). On the other hand, compared to the HIP, the HOP had a smaller median population density in inhabitable land areas (1,564/m2 vs. 646/m2), lower average income of the general population (JPY 3.22 million vs. JPY 2.74 million), and a lower unemployment ratio of the general population (3.3% vs. 2.8%).
Background sociodemographic data used for the multiple linear regression analyses (one of the most common methods of analyses in medicine) were as follows: practicing PPRs, newly licensed PPRs, population density in inhabitable land areas, average income of the general population, and aged population ratio. Among them, only the newly licensed PPR was significantly associated with physicians’ migration (p < 0.001) (i.e., a high ratio led to increased outflow, and a low ratio led to increased inflow).
As above, significant differences among prefectures in the pattern of migration were revealed. In some prefectures within the HOP, more than half of the newly licensed physicians ostensibly moved to another prefecture after graduation. On the other hand, in some prefectures within the HIP, more than twice the number of newly licensed physicians arrived from other prefectures. Because migration of physicians can be seen as a redistribution mechanism to influence inequality in the number of physicians throughout the country, recognizing and deciphering the pattern of migration would be useful in policy making to mitigate the misdistribution. As our data indicate, the magnitude of physicians’ migration would have implications on the underlying mechanism concerning the development of unequal distribution of physicians, which should be considered in policy decision-making for underserved areas.
Arguably, a trend whereby physicians move from the rural to urban areas appears to exist. However, it should be noted that even the metropolis of Tokyo, which has 13 medical schools with an annual newly licensed PPR of 10.4 (9th among 47 prefectures), had an outflow ratio of 13%. Data also showed that specific urban prefectures neighboring Tokyo, such as Chiba and Saitama, had extremely high inflow ratios of over 200%. Actually, the multivariate analyses suggest that the newly licensed PPRs, and not rural-urban migration, might be one of the keys in explaining the migration ratios of physicians.
In addition, despite having an inflow ratio of greater than 200%, Chiba and Saitama still had very low PPRs of 180.4 and 160.1, respectively, in 2016. Therefore, the inequality in PPR could not be sufficiently mitigated by migration alone. Consistent with this assumption, a previous Japanese study showed that an increase in the number of physicians between 1998 and 2008 did not lead to a more equal geographical distribution of physicians. Namely, inequality remained even though the policy for increasing physicians mitigated the scarcity of physicians in medically underserved areas. Such findings are not fully recognized among the public because there have been a paucity of studies as well as discussions about the issue.
The government has spent a large amount of money on medical education; hence, Japanese physicians are generally considered by the public to make more effort to serve in community health care. However, the absolute numbers of physicians are not high enough in some prefectures, despite the relatively high newly licensed PPR (for example, Yamanashi prefecture has a PPR lower than the national average PPR of 230.2 and an outflow ratio of 52.9%, despite the high newly licensed PPR of 10.9). This illustrates the difficulty in mitigating the current misdistribution.
If the inequality in PPR remains unsolved in the future, it might become a significant political agenda for the central as well as local governments. Nevertheless, it should be noted that health equality is distinct from health equity, and it might be better to set a minimum requirement for health outputs and/or health outcomes rather than pursue equality haphazardly. Regardless, the current study could contribute toward accelerating discussions within the Japanese government, and these findings might also be useful for policy making throughout global health communities.