Annual prevalence of non-communicable diseases and identification of vulnerable populations following the Fukushima disaster and COVID-19 pandemic

Disasters, pandemics, and their response measures can have secondary effects on the physical and psychological health of affected populations. Using health insurance receipt data from 2009 to 2020, we assessed changes in the prevalence of major non-communicable diseases (NCDs), including hypertension, hyperlipidemia, diabetes, and mental disorders, among affected populations before and after the Fukushima disaster and coronavirus disease (COVID-19) outbreak in Japan. Furthermore, age and sex groups with the largest increases in prevalence after these events were identified. The participants of this study were employees of large companies and their dependent family members who were insured by health insurance societies (HIS). The dataset was provided by JMDC Inc. The annual age-adjusted prevalence of each disease belonging to the HIS scheme was used to calculate the ratio of disease prevalence before and after the events. After the Fukushima disaster, hypertension, hyperlipidemia, and diabetes generally increased over a 9-year period in Fukushima Prefecture. The increase in the prevalence rate of these three NCDs and mental disorders was the highest among females aged 40–74 years compared to males and the other age groups. The prevalence of all four diseases increased after the COVID-19 outbreak in Japan, with a marked increase in males aged 0–39 years. Changes in prevalence ratios of NCDs after the COVID-19 outbreak among the areas affected by the Fukushima disaster were lower than in the whole of Japan. It is important to provide tailor-made public health support among populations in accordance with the type of disasters and pandemic.


Introduction
Disasters, pandemics, and their countermeasures impact the lives of affected people with secondary effects on their physical and psychological health. Major non-communicable diseases (NCDs), such as hypertension, hyperlipidemia, diabetes, and mental disorders, have been reported to increase among people who have experienced natural or man-made disasters, such as the Chornobyl nuclear power plant accident in 1986 and Hurricane Katrina in 2005 [1,2]. Several studies have also revealed that the Great East Japan Earthquake and Tsunami (GJET) and the Fukushima Daiichi Nuclear Power Station accident in March 2011 (hereafter referred to as estimated that more than 140,000 people were evacuated from these radii [19]. The affected people were evacuated to municipalities in Fukushima Prefecture outside the evacuation areas, such as Iwaki City (Hamadori), as well as to areas outside Fukushima Prefecture.
In 2012, the evacuation order was lifted in Kawauchi Village and Hirono Town, where the evacuation was decided by each municipality soon after the disaster. Furthermore, since 2014, the government has gradually lifted the evacuation order, and by the end of 2019, the evacuation order was lifted for the following municipalities: Tamura City, Kawauchi Village, Naraha Town, Katsurao Village, Minamisoma City, Iitate Village, Kawamata Town, Namie Town, Tomioka Town, and Okuma Town, excluding the difficult-to-return areas where radiation levels remain high located around the Fukushima Daiichi Nuclear Power Station.

COVID-19
The World Health Organization (WHO) declared COVID-19 a pandemic on March 11, 2020 [20]. In Japan, a declaration of a state of emergency was issued in seven prefectures on April 7, 2020 and extended to all prefectures on April 16, 2020. The state of emergency included requests to stay at home, instruction for schools to close temporarily, closure of businesses and stores that are not essential to daily life, and cancellation or postponement of events and exhibitions, but they were not mandatory. The state of emergency ended in 39 prefectures on May 14, 2020, in three more prefectures on May 21, 2020, and in all prefectures on May 25, 2020. This was the only emergency declaration for 2020, and the reported number of COVID-19-positive cases in Japan in 2020 was 234,109 [21].

Datasets and target population
In this study, we used health insurance receipt data provided by JMDC Inc. The target population covered in the JMDC data were employees of companies and their dependent family members who belonged to Employees' Health Insurance (EHI) schemes. Please note that EHI can be divided into several types, depending on the size and type of organization to which the employee belongs. Of these, JMDC Inc. covers only employees working at large companies and their dependents who are insured by health insurance societies (HIS). Approximately 29 million people in the country are insured by HIS [22], and JMDC Inc. contracts a part of these insurers, covering 9 million people (as of 2022), and collect insurance claims containing completely anonymized receipt information. Other insurance schemes that are part of the EHI include those for employees of small and medium-sized companies and their families, and those for civil servants, private school teachers and seafarers, which JMDC Inc. does not cover.
In addition to the EHI schemes, Japan's insurance system has two major schemes: National Health Insurance (NHI) and Long-term Care Insurance (LTCI). All persons, regardless of citizenship, who have been recognized as having resided in Japan for more than three months are obliged to join one of three. NHI targets individuals who do not belong to the EHI scheme, including farmers, freelancers, non-regular employees, retired company employees, etc., and LTCI includes individuals aged 75 and above.
We used data on the annual number of patients for each targeted disease and other diseases (i.e., all patients visiting medical institutions regardless of their diagnosis) and the estimated prevalence belonging to the HIS schemehereinafter, it is simply described as "prevalence," unless otherwise notedby age and sex groups in each prefecture or municipality from January 2009 to December 2020. The prefectures and municipalities were based on the location of the medical institutions. The target diseases included hypertension, hyperlipidemia, diabetes, and mental disorders, based on the ICD 10 classification: I10-I15, E78, E10-E14, and F00-F99, respectively. The diseases were diagnosed by a physician, with no distinction made between primary and secondary diseases. The total number of patients in the JMDC database for hypertension, hyperlipidemia, diabetes, mental disorders, and other diseases in Japan from 2009 to 2020 is provided in Table S1.
Other studies have used the JMDC database to calculate the prevalence of specific diseases for the whole of Japan [23,24]. In this study, the prevalence (P d,t,r,a,s ) of the target disease (d) in each prefecture or municipality (r) by age (a:0-19, 20-39, 40-64, and 65-74 years) and sex (s) groups for each year (t) were calculated according to eqs. (1) and (2).
where A d,t,r,a,s is the actual number of patients with the target disease by each prefecture or municipality among the target age and sex population in the target year, B t,r,a,s is the actual number of patients who visited the medical institution regardless of their diagnosis in each prefecture or municipality among the target age and sex population in the target year, C d,t,a,s is the estimated number of patients nationwide among the target age and sex population in the target year and was estimated by multiplying the prevalence of each disease (the ratio of the number of patients to the number of people in the mother population covered by JMDC Inc.) by age (1-year interval) and sex groups and by the national population, D d,t,r,a,s is the estimated number of patients with the target disease in each prefecture and municipality in the target age and sex population in the target year, and E t,r,a,s is the number of people in each prefecture and municipality in the target age and sex population in the target year. The Basic Resident Register data were used to determine the number of people in the prefectures for each year [25]. For the municipal population data, the national census performed in 2010, 2015, and 2020 was used for the years 2009-2010, 2011-2015, and 2016-2020, respectively [26]. Selection bias among the target population was corrected by C d,t,a,s and D d,t,r,a,s , as described in eq. (1).
Furthermore, using the commonly used 1985 model population as a reference population (i.e., distribution of the population by age and sex groups in 1985) [27], the age-adjusted prevalence was calculated for the 0-74, 0-39, and 40-74 years of age groups for males, females, and both males and females.
Diagnostic guidelines have been revised for various diseases, including the establishment of borderline hyperLDL-cholesterolemia in 2012 and the diagnosis of diabetes using HbA1c in 2010. Although these revisions may have impacted the prevalence, the bias was reduced by standardizing the assessment of the impact before and after the Fukushima disaster by normalizing the prevalence with that for the whole of Japan (see "2.5. Analysis" for details).

Analysis
Two analyses were performed in this study. First, to evaluate changes in prevalence after the Fukushima disaster, we considered the whole of Japan, Iwate, Miyagi, and Fukushima prefectures, as well as three regions within Fukushima Prefecture (i.e., Hamadori, Nakadori, and Aizu (mountainous area); Fig. 1, Table S2). These three regions, which are closer to the nuclear power station in the order Hamadori, Nakadori, and Aizu, divide Fukushima Prefecture into three parts from east to west. In other words, Hamadori was the region most affected by the disaster, and Aizu was the region least affected among regions in the prefecture. The prevalence of targeted diseases in the Hamadori, Nakadori, and Aizu regions was calculated by weighting data by the municipality based on B t,r,a,s . The arithmetic means and standard errors of the prevalence for each of the four time periods, such as 2009-2010, 2011-2013, 2014-2016, and 2017-2019, were calculated, and the prevalence for each time period was assumed to follow a normal distribution. To ensure that the prevalence was within the range of 0-1, two-tailed censoring was performed so that the arithmetic mean was at the center of the distribution. Monte Carlo simulations were performed with 10,000 iterations to determine the ratio or difference in prevalence (males and females, 0-74 years) between 2011 and 2013, 2014-2016, or 2017-2019 and 2009-2010 in the whole of Japan, Iwate, Miyagi, Fukushima prefectures, and Hamadori, Nakadori, and Aizu regions. These prevalence ratios and differences were further normalized to the values for the whole of Japan. Similarly, prevalence ratios and differences by age and sex groups were calculated for the whole of Japan and Fukushima Prefecture.
Second, to assess changes in the prevalence after the COVID-19 outbreak, data from 2015 to 2020 were used (Tables S2 and S3). For simplicity, we defined 2020 data as the period after the COVID-19 outbreak. We calculated the ratios and differences between the prevalence in 2020 (the period after the outbreak) and the prevalence in 2020 estimated from the 2015-2019 dataset. A linear regression equation was modeled using the dataset as a baseline to estimate the hypothetical 2020 prevalence in the absence of COVID-19. Based on the values and their 95% confidence intervals, calculations of the ratios and differences between the measured and estimated prevalence were performed with Monte Carlo simulations in 10,000 runs, assuming that the prevalence followed a normal distribution; two-tailed censoring was performed so that the prevalence fell within the range of 0-1, and the arithmetic mean was at the center of the distribution. The prevalence ratios and differences were estimated for age and sex groups in the whole of Japan: males and females (0-74 years), males and females (0-39 years), males and females (40- . Furthermore, the prevalence ratios and differences in 47 prefectures were calculated for males (0-39 years), the vulnerable population group identified in the above analysis, to assess the correlation with population density [25,28] or the reported number of COVID-19 infection cases per 100,000 people (COVID-19 infection rate) in 2020 [21] (Table S4). Spearman's rank correlation coefficient was estimated. The prevalence ratios (males and females, 0-74 years) in Iwate, Miyagi, Fukushima prefectures, and Hamadori, Nakadori, and Aizu regions and the values normalized to those for the whole of Japan were then calculated. The prevalence ratios and the values normalized to those for the whole of Japan by age and sex groups were calculated for Fukushima Prefecture.

Age-adjusted prevalence over time
Fig . 2 shows the annual changes in age-adjusted prevalence (male and female, 0-74 years) of NCDs from 2009 to 2020 in the whole of Japan; Iwate, Miyagi, and Fukushima prefectures; and Hamadori, Nakadori, and Aizu regions. The age-adjusted prevalence of hypertension, hyperlipidemia, diabetes, and mental disorders in the whole of Japan was 0.102, 0.107, 0.061, and 0.075 in 2009 and 0.113, 0.121, 0.068, and 0.098 in 2020, respectively. The age-adjusted prevalence of hypertension, hyperlipidemia, diabetes, and mental disorders in Fukushima Prefecture was 0.106, 0.098, 0.044, and 0.072 in 2009 and 0.117, 0.119, 0.056, and 0.082 in 2020, respectively.
Age-adjusted prevalence by age and sex groups for each disease in the whole of Japan and Fukushima Prefecture from 2009 to 2020 is shown in Figs. 3 and 4. Age-adjusted prevalence of hypertension, hyperlipidemia, and diabetes were generally higher in males than in females in the 40-74 years age group, and the differences tended to increase over time in the whole of Japan. In contrast, mental disorders were higher in females than in males in the 40-74 years age group, and the prevalence in the 0-39 years age group has been increasing in recent years. In Fukushima Prefecture, mental disorders showed an increasing trend in 2011-2012 for both age and sex groups, followed by a decreasing trend. Table 1 shows the ratios of the age-adjusted prevalence of NCDs in the 0-74 years group in the whole of Japan; Iwate, Miyagi, and Fukushima prefectures; and Hamadori, Nakadori, and Aizu regions before and after the Fukushima disaster. These differences are listed in Table S5. The table also shows the values normalized to those for the whole of Japan. After the Fukushima disaster, the increases in the prevalence of hypertension, hyperlipidemia, and diabetes were significantly higher among the Fukushima Prefecture        Tables 3 and S7 show the ratios and differences in age-adjusted prevalence of each disease by sex and age group among the whole of Japan before and after the COVID-19 outbreak. The prevalence of all four diseases generally increased, with the ratios among males and females aged 0-74 years being 1.058 (95% UI: 1.032-1.087) for hypertension, 1.075 (1.013-1.143) for hyperlipidemia, 1.038 (1.012-1.066) for diabetes, and 1.033 (1.027-1.039) for mental disorders. Among the age and sex groups, the ratios of actual to estimated values for these four diseases were particularly high among males aged 20-39 years and 0-39 years (ratio among males aged 0-39 years: 1.098 (1.072-1.126) for hypertension, 1.143 (1.009-1.309) for hyperlipidemia, 1.060 (1.040-1.080) for diabetes, and 1.041 (1.032-1.050) for mental disorders). In contrast, the differences in prevalence were large for males and females aged 40-64, 64-74, and 40-74 years.

Age-adjusted prevalence before and after the COVID-19 outbreak
The ratios and differences in the prevalence of mental disorders before and after the COVID-19 outbreak by prefecture in males aged 0-39 years were significantly negatively correlated with the population density and COVID-19 infection rate in 2020 ( Fig. 5 and Tables S8 and S9). In contrast, the population density and COVID-19 infection rates were not significantly associated with the prevalence ratios and differences in other diseases. Table 4 shows the prevalence ratios in age-adjusted prevalence of each disease among Iwate, Miyagi, and Fukushima prefectures and three regions within Fukushima before and after the COVID-19 outbreak and standardized values for the whole of Japan. There were no significant changes for any of the diseases in males and females aged 0-74 years, except for a significant decrease in the prevalence of diabetes in Fukushima Prefecture: 0.969 (0.945-0.995). The prevalence of hypertension and diabetes by sex and age group decreased significantly in Fukushima Prefecture, especially among males (hypertension: males aged 0-74 years, males and Table 3 Ratio of observed to estimated prevalence in 2020 by age groups and sex (the whole of Japan). Arithmetic mean (95% uncertainty interval). The prevalence represents those belonging to the health insurance societies scheme.

Ages and sexes
Hypertension Hyperlipidemia Diabetes Mental disorders In terms of standardized values for the whole of Japan, the change in the prevalence of hypertension was significantly lower among some sex and age groups in the areas affected by the Fukushima disaster than the whole of Japan: males and females aged 0-74 years in Fukushima Prefecture, Hamadori, and Aizu; males aged 0-74 years, males and females aged 0-39 years, males and females aged 40-74 years, males aged 0-39 years, males aged 40-74 years in Fukushima Prefecture. The changes in the prevalence of diabetes were also significantly lower: males and females aged 0-74 years in Iwate Prefecture, Fukushima Prefecture, and Nakadori; males aged 0-74 years, males and females aged 0-39 years, males and females aged 40-74 years, males aged 0-39 years, and males aged 40-74 years in Table 4 Ratio of observed to estimated prevalence in 2020 (Iwate, Miyagi Fukushima prefectures, Hamadori, Nakadori, Aizu). Arithmetic mean (95% uncertainty interval). The prevalence represents those belonging to the health insurance societies scheme. Fukushima Prefecture. The change in prevalence of hyperlipidemia was significantly lower among males and females aged 0-74 years in Iwate and Miyagi prefectures, males aged 40-74 years in Fukushima Prefecture than the whole of Japan. The change in prevalence of mental disorders was significantly lower in males and females aged 0-74 years in Iwate, Miyagi, and Fukushima prefectures and Aizu 0-74, females aged 0-74 years, and males and females 40-74 years in Fukushima Prefecture than the whole of Japan.

Discussion
In this study, we evaluated the annual changes in the prevalence of hypertension, hyperlipidemia, diabetes, and mental disorders among employees of companies and their dependent family members belonging to the HIS scheme before and after the Fukushima disaster and the COVID-19 outbreak in Japan. Furthermore, we identified the vulnerable age and sex groups of people who experienced significant secondary health effects after the disaster and pandemic.
The analysis of the prevalence of these diseases after the Fukushima disaster by age and sex groups revealed that hypertension, hyperlipidemia, and diabetes generally increased in Fukushima Prefecture over the 9-year period following the disaster. Significant increases in hypertension, diabetes, and mental disorders were observed in Hamadori, hyperlipidemia and diabetes in Nakadori, and hypertension and diabetes in Aizu. Among the four diseases, the increased prevalence of diabetes was higher than that of the other diseases. Diabetes increased consistently in all regions.
After the Fukushima disaster, a significant increase in the prevalence of diabetes was consistently reported in multiple previous studies targeting various municipalities where evacuation orders were issued [3][4][5]8]. The findings of our study are consistent with those of previous studies. For hyperlipidemia, previous studies reported an increase among evacuees within 4 years after the disaster but no significant increase in the 7-year data [3,8]. For hypertension, some studies that used a definition based on blood pressure alone have reported a risk decrease [3,8], whereas a study that used a definition based on blood pressure as well as the use of antihypertensive medication reported an increased risk [5]. Regarding psychological distress, a 5-fold risk increase was reported in the first year after the disaster, followed by an improvement trend [29]. Notably, this study highlighted that the worsening prevalence of these NCDs was not observed only in Hamadori, which was significantly affected by the Fukushima disaster, but also in Nakadori and Aizu. This is consistent with previous findings on non-evacuees [3,8]. The results suggest that the Fukushima disaster may have secondary physical health effects not only on evacuees but also on the wider population. The association between evacuation and increased prevalence of NCDs may be related to the exemption of medical expenses for the evacuees, in addition to behavioral changes related to lifestyle habits, such as exercise and diet, as well as to occupational and community changes [30]. However, the widespread increase in the prevalence of NCDs in Fukushima Prefecture cannot be explained by the exemption of medical expenses. The reasons for secondary physical health effects among non-evacuees are not clear; however, social support for non-evacuees is more likely to be overlooked in the aftermath of a disaster than that for evacuees. Considering that these diseases generally did not change in the Iwate and Miyagi prefectures that were affected by the GJET, but not largely by the Fukushima disaster, it is possible that the increase in the prevalence of these diseases observed over a wide area in Fukushima Prefecture was related to the long-term evacuation, hosting of evacuees, radiation anxiety, and socioeconomic changes in the residents of Fukushima Prefecture following the disaster, possibly resulting in lifestyle changes among residents. Hamadori, where a significant increase in the prevalence of mental disorders was observed, includes not only municipalities whose residents were ordered to evacuate, but also those that accepted the majority of evacuees. It has also been suggested that psychological distress among residents of host communities may be caused by the perception that evacuees are a threat to their social identity [9]. Strong radiation anxiety was observed not only in Hamadori but also Nakadori or Aizu, especially just after the Fukushima disaster [31], possibly resulting in behavioral changes (e.g., avoidance of radiation) and worsening mental health [32].
The increase in the prevalence of these four diseases in the post-disaster period was most pronounced among females and those aged 40-74 years. Psychological distress risk has been reported to be higher among elderly females [33]. Most of the previous studies after the Fukushima disaster have assessed the impact of evacuation on NCDs [3][4][5]8], and there are limited studies that specifically discuss which age and sex groups experienced significant changes in prevalence after the disaster based on the same dataset. The strength of this study is that, based on the same dataset, changes in the prevalence after the disaster were assessed by age and sex groups to identify populations vulnerable to secondary health effects.
In contrast, after the COVID-19 outbreak, the prevalence of all four diseases increased nationwide, especially in males aged 20-39 and 0-39 years. Among the four diseases, the prevalence increase post-disaster in hyperlipidemia was particularly higher than that in the other diseases. To date, previous findings on NCDs risks after COVID-19 have not been coherent. While some population have improved their lifestyle habits [11,12], others have decreased their physical activity levels due to lockdown and other behavioral restrictions [34]. Molarius et al. [35] reported no statistically significant changes in the prevalence of high blood pressure before and after the COVID-19 outbreak in Sweden. A systematic review regarding the impact of the COVID-19-related lockdown on glycemic control among patients with type 1 and type 2 diabetes reported that glycemic values in patients with type 1 diabetes improved, but those in patients with type 2 diabetes worsened during the COVID-19 lockdown [14]. An increasing burden of mental disorders due to the COVID-19 pandemic has been reported among the public worldwide [15]. The increased prevalence of mental disorders among the young generation after the COVID-19 outbreak found in this study was consistent with a previous study [36] reporting that the increase in suicide after the COVID-19 outbreak was also risen especially among young generations in Japan possibly due to a complex reasons including unsecure working conditions and psychological depression associated with the reopening of school. In Japan, since the early days of the COVID-19 pandemic, active efforts have been made to raise awareness of exercise to promote physical and psychological health [37,38]. Despite these social awareness efforts, our study showed that the prevalence of NCDs were increasing in Japan, especially among young males.
Interestingly, the age and sex groups that showed large increases in the prevalence of these diseases differed between the Fukushima disaster and COVID-19 outbreak. After the Fukushima disaster, the affected people had often experienced family separation, fragmentation of communication, job loss, and behavioral changes due to radiation anxiety [32,39,40]. In contrast, after the COVID-19 outbreak, not only were there risk factors such as lack of exercise and worsening of dietary and drinking habits associated with measures such as stay-at-home requests, but also preventive factors such as reduced drinking opportunities and lifestyle modification. These differences between the Fukushima disaster and the COVID-19 outbreak may have made a difference in the high-risk group regarding NDCs. The pandemic may have resulted in polarization, with some individuals experiencing improved health and others experiencing worsened health. The young Japanese male population tends to have lower health practices than those of other age and sex groups [41], possibly resulting in significant secondary health effects due to the pandemic. In addition, the prevalence ratios of mental disorders before and after the outbreak were significantly negatively correlated with population density and COVID-19 infection rates. This suggests that the increased prevalence of mental disorders was not mainly attributed to the burden among patients with COVID-19 but also attributed to other social factors. Interestingly, the results of this study were in contrast to that of a previous study reporting that increases in major depressive and anxiety disorders in each country were significantly positively correlated with the rate of COVID-19 infection [15]. Although there were some differences in the timing of the declaration of the state of emergency, it was issued in all prefectures in 2020 in Japan. Therefore, the differences in the impact of social activity restrictions associated with these regulations were probably small among the prefectures. There are two possible reasons for the negative association between population density and the increased rate of mental disorders found in this study. First, the availability of mental health-promotion services and social infrastructure may differ among prefectures; urban areas may have had better access to mental health-promotion services and social infrastructure, even under restricted social activities during the COVID-19 pandemic. Second, there may be differences in the cultural worldviews of infection and related preventive behaviors among prefectures. It is possible that the social environment toward COVID-19 infection, shaped by this cultural worldview, had a negative impact on mental health among the population. Differences in strict cultural worldviews among prefectures in Japan and their relationship with mental disorders remain unclear and require further analysis.
The secondary effects of the COVID-19 outbreak in areas affected by the Fukushima disaster showed different trends from the whole of Japan. The prevalence of hypertension and diabetes in Fukushima Prefecture (especially among males) after the COVID-19 outbreak showed an improving trend compared to the pre-outbreak period. The changes in the prevalence of hypertension in Fukushima Prefecture (especially among men), diabetes in Fukushima Prefecture (especially among males) and Iwate Prefecture, hyperlipidemia in Iwate and Miyagi prefectures and Fukushima Prefecture (males aged 40-74 years old), and mental disorders in Iwate, Miyagi, and Fukushima prefectures were lower than those in the whole of Japan. This represents that the areas that experienced the Fukushima disaster were either less affected overall than the whole of Japan during the subsequent COVID-19 outbreak, or showed improvement in hypertension and diabetes. A previous study reported that after the COVID-19 outbreak, psychological distress was not worsened and problem drinking behavior was improved among residents who lived in municipalities spanning evacuation order areas during the Fukushima disaster [11]. There are two interpretations of the fact that areas that experienced the Fukushima disaster were less affected by the COVID-19 outbreak than other areas. First, NCDs had already been rising after the Fukushima disaster, so the impact of the COVID-19 outbreak was small. Disasters can cause fatalities among vulnerable populations and subsequently reduce survivors' mortality rate (i.e., a harvesting effect) [42]. There may be similar effects on secondary health impacts. Second, the people affected by the Fukushima disaster may have unique and resilient characteristics and be more conscious about their health [43]. This study demonstrates the importance of providing support for secondary health effects on NCDs in the aftermath of disasters and pandemics. Since significantly affected age and sex groups are unique to each disaster or pandemic, tailor-made public health support is considered important. Specifically, it would help the central and local governments develop strategies such as creating a point of contact to increase support for vulnerable populations, publicizing information about that support to the populations, and establishing a support system of local health workers. Furthermore, while COVID-19 infections have a high fatality rate among the elderly [44], NCDs are more likely to affect young people. This means that infection control measures, including stay-at-home requests and other restrictions on socioeconomic activities, may be beneficial in reducing infection among the elderly under 75 years of age but may pose greater risks for the young regarding secondary health effects. Comprehensive public health measures in addition to infection control are required based on the characteristics of the population while considering the differences in the balance between risk-benefits and non-infectious secondary health effects among the populations.
This study has some limitations. First, the target population was employees of large companies and their dependent family members. The findings of this study cannot be generalized to the entire population and other countries. Furthermore, only the population aged 74 years and below were included in this study. Although COVID-19 isolation measures are known to worsen cognitive and mental health among people with dementia [45], this study was unable to assess the secondary health effects among the population aged 75 years and older. Second, we could not fully examine the possibility of changes in accessibility to healthcare facilities after the Fukushima disaster or the COVID-19 outbreak compared to accessibility to healthcare before the disaster and outbreak. However, it is unlikely that NCDs were overlooked before the disaster or outbreak since employees of companies, the target population of this study, generally receive regular health checkups. Third, we cannot rule out the possible discrepancies between the diagnosis obtained from the receipt data and the diagnosis in actual clinical settings. We reduced this bias by comparing the ratios of the prevalence of NCDs before and after the Fukushima disaster and COVID-19 outbreak, as well as the values in the whole of Japan. Fourth, this study was based on an ecological study design and did not identify the causes of changes in the prevalence of various diseases. Long-term cohort studies based on individuals exposed to disasters and pandemics are expected to assess the impact of disasters and pandemics on secondary health and identify the factors that influence this impact.
Despite these limitations, this study provided unique insights into the secondary health effects of the Fukushima disaster and