Trends in the geographic distribution of nursing staff before and after the Great East Japan Earthquake: a longitudinal study

Our study indicated that the trend in the ratio of total nursing staff per population after the GEJE was different between the physically damaged areas and those affected by radiation. In coastal SMAs, which were greatly damaged by tsunamis, the ratio of total nursing staff per population greatly decreased immediately after the GEJE, as indicated previously [2-5]. In most SMAs, because the reconstruction of hospitals was progressing after the GEJE, the ratio of total nursing staff per population increased and exceeded the pre-GEJE level for 2 years. In four SMAs (Ryouban, Ishinomaki, Sousou and Iwaki), the ratio of total nursing staff per population had not yet recovered to the pre-GEJE level. In Ryouban, Ishinomaki and Iwaki, the ratio of total nursing staff per population achieved almost 96% and over before the GEJE. In the inland area of Ryouban, the decrease in total nursing staff per population was caused by the post-GEJE increase in population. In Ishinomaki and Iwaki, which were one of the greatest damaged by the tsunami, the decrease in total nursing staff per population was explained by the decrease in associate nurses.

In Sousou, the SMA where the NPP is located and which was also severely damaged by the earthquake and tsunamis, the total nursing staff per population in 2013 was only half of that in 2010. The number of physicians per 100 000 population also decreased by 34.4% (from 120.4 on 31 December 2010 to 79.0 on 31 December 2012) in Sousou [33], which was almost the same as the reduction of nurses per population. These large and continuous decreases of nursing staffs might be explained by the following reasons. First, the decrease in the number of hospitals might cause a decrease in the ratio of total nursing staff per population. After the GEJE, the Japanese government designated three categories of areas to which evacuation orders were issued to prevent further expansion of the damage caused by the NPP accident: Area 1, where it is confirmed that the annual cumulative dose will definitely be ≤20 mSv, is an area in which evacuation orders have already been lifted; Area 2, where the annual cumulative dose from the present moment is expected to be >20 mSv and where residents are ordered to remain evacuated in order to reduce the risk of radiation exposure; and Area 3, where the annual cumulative dose is expected be ≥20 mSv for a long time, especially within 5 years, and the current annual cumulative dose is >50 mSv, and where it is expected that residents will face difficulties in returning for a long time [34,35]. Most parts of Areas 1 and 2 have been included in Sousou SMA, and the hospitals located within the areas have been closed [4]. Health care workers who had worked in those hospitals had to move, even if they wanted to work there. Even in the hospitals that maintained their functions, the increased workload among the nursing staff, due to the initial significant reductions in the number of available staff after the GEJE, might encourage further departure of nursing staff [36] and discourage their return, which could result in a continuous shortage of nursing staff.

Another possible reason is the concern over radiation exposure among the nursing staff, which might have made them move to areas located far from the NPP. As previous studies have indicated, only half of the health care providers were willing to work in areas affected by nuclear disasters [6,37−40]. The most frequently cited reason for employees’ unwillingness to report for duty during a disaster was fear and concern for the safety of their families and themselves [37]. Although the surrounding SMAs have also suffered from high doses of radiation, the relatively higher radiation doses and the shorter distance from the NPP in Sousou might be a reason for the larger reduction and delayed return of the nursing staff. There might be some other factors that could be associated with the workforce distribution after the disaster. Although such factors would be strongly associated with the mortality or radiation doses in each SMA, further investigations are required to find those factors that are associated with the workforce distribution.

The change in the ratio of nursing staff per population differed according to qualifications, and the change in ratio of nursing aides was the lowest among the three types of nurses. Cone et al. indicated that willingness to work in a disaster area was higher for nurses than for nonclinical staff [39], but our findings indicated different tendencies. Because most nursing aides are irregular employees in Japan, they might be more likely to change their workplace than nurses are. Also, there might be differences in knowledge regarding radiation. Lanzonni et al. showed that physicians and nurses with knowledge regarding biological events had greater willingness to work in a biological disaster [40].

From the above findings, policy and actions to secure the different types of nursing staff are required, especially for the areas suffering from a continuous shortage of staff. The Fukushima prefectural government developed a policy for securing nurses after the GEJE: (1) establish a study fund aimed at the security of nurses and associate nurses, (2) support re-employment of latent nurses, (3) set aside a subsidy to assist with part of the personnel expenses in the hospitals forced to reduce operation by the GEJE to maintain employment of nurses and (4) strengthen the in-service training in the hospitals within Sousou [7,15]. These policies might be partly effective and contribute to the relatively higher ratios of nurses and associate nurses per population than nursing aides, even in Sousou SMA. To secure the overall nursing staff, however, more comprehensive approaches to enhance the return of nursing aides as well as nurses and associate nurses to the hospitals are required. As the skill mix of nursing staff has been promoted globally [41], a financial incentive was introduced to promote employment of nursing aides in hospitals, which was expected to reduce nurses’ workload and allow them to spend more time doing their own professional jobs such as bedside care [42]. According to a previous study, the motivation of nursing aides in Japan can be defined by the following factors: salary satisfaction, free time to do one’s own thing, nursing aides as important partners in the job, feeling helpful to patients, participating in decision-making and job-skill improvement [43]. Thus, policies including salary increase, strengthening in-service training and ensuring a pleasant working environment, might be effective in securing nursing aides in Sousou and, consequently, in providing a better work environment by reducing the overall workload of the nursing staff in each hospital. To prevent further reductions in staff and for future preparedness, education programmes regarding nuclear disasters could be effective, since such programmes were found to minimize health care providers’ fears and concerns about radiation [44].

In Minamiaizu, not only the ratio of total nursing staff per population but also the ratio of hospitals and physicians per population has been the lowest and much lower than the nationwide levels [45]. The relative shortage and misdistribution of health care providers in remote areas such as Minamiaizu has been a continuous policy concern in Japan [46-48].

To the best of our knowledge, the longitudinal impact of a disaster on the distribution of health care providers is unknown, especially during a nuclear accident. Therefore, this study is the first to show the longitudinal trend in the ratio of total nursing staff per population in SMAs according to qualifications before and after the GEJE.

There were some limitations to our study. First, the population might have been overestimated in those SMAs that were greatly damaged by the GEJE because of evacuation, without people changing area registration. Accordingly, the ratio of total nursing staff per population might have been underestimated in these areas. The Japanese government reported that there are about 133 309 people actually living around restricted areas and areas for which evacuation orders have been issued, such as in the Sousou SMA, excluding three municipalities (Shinchi Town, Soma City and Hirono Town), as of 31 December 2013 [35]. In our study, the population in the same area was 131 705 as of 1 October 2013. In considering these numbers, a perceived limitation of overestimation is not serious because the government’s calculation of Sousou’s population was almost the same as our estimation. Second, there were data limitations with regard to type of occupation, workplace and time. In our data, the data before 2007 and other health providers’ information did not exist. Trends in other health care providers, such as physicians, dentists and pharmacists, should be clarified in future studies. Total nursing staff included only hospital workers because of data availability. Trends in total nursing staff in other workplaces, such as clinics or long-term care facilities, should be investigated in the future. Third, our findings could not indicate the causal association between the numbers of total nursing staff and effects of radiation because we described the temporal trends using statistical data and did not assess the dynamic flows of nursing staff according to radiation dose. Therefore, further studies are required to investigate the causal association.