Estimation of occupational mortality rates of workers in Japan

The occupational mortality rates of the workers in each data sources are shown in Fig. 1. For the RWC, the mortalities per 100,000 workers were 4.0, 3.1, and 18.0 in 1991 and 1.4, 1.5, and 6.2 in 2019 overall and for the manufacturing and construction industries, respectively. The mortality rates for the three industry categories in 2019 were two to three times lower than those in 1991. The number of deaths increased in 2011 due to the Great East Japan Earthquake (March 11, 2011). These data were excluded in the RWC but not in the ABR. The ASR data trends were the same as those of the RWC data. The mortality rates overall in the ABR showed a slightly declining trend. The Ministry of Health and Labour Standards launched an occupational safety and health program in 1958 that aimed to improve the occupational environment using RWC data as the number of deaths. Thus, we would expect that the mortality rates reported in the RWC would show a slightly declining trend. In contrast, the mortality rates in the manufacturing and construction industries increased after 2000. This increasing trend may have been caused by deaths due to asbestos since the Act on Asbestos Health Damage Relief was enforced in 2006 in Japan. Furthermore, there has been a large increase in mesothelioma cases due to occupational asbestos exposure compensated after 2004 [10]. It should also be considered that the latency period of mesothelioma is about 30–40 years. The effect of the latency period on the mortality rates is discussed in another subsection. The mortality rates in 2019 in the ABR were 3.1, 4.8, and 3.1 times higher overall and for the manufacturing and construction industries, respectively. The main reason for this could be related to the rules for recording deaths. Deaths in the RWC were based on reports submitted immediately after the occurrence. In other words, deaths after long-term medical treatment for occupational diseases were not included in the number of deaths in the RWC. The number of deaths in the ASR is calculated by subtracting the number of cases in which medical treatment benefits were received for more than 18 months from the total number of funeral expense claims. The reason why the trends of mortality rates are also the same may be that the data collection rule in the RWC is similar to that in the ABR and that in most cases of death after medical treatment, medical treatment benefits have been received. In contrast, the number of deaths in the ABR is taken from the total number of funeral expense claims. Thus, the difference in mortality rates seems to be due to the number of cases in which medical treatment benefits were received, namely, the number of patient receiving long-term medical treatment. To the best of our knowledge, this is the first time that the difference between these three Japanese statistical data sources has been estimated from the point of view of mortality rates.

Fig. 1

Trends of mortality rates in Japan for overall (left), and for the manufacturing (center) and construction (right) industries in the RWC, ABR, ASR and previous studies Iwasaki et al. [5] and Iwasaki and Nishizawa [6]. The symbol (*) means the values calculated using some industry’ data

The number of deaths per 100,000 workers in the RWC by age group and industry category is shown in Fig. 2. The mortality rates of those aged 60 + and 50–59 years were higher than those aged 30–39 and 40–49 years, and the mortality rate of those under 19 years was highly variable. This is because the number of workers under 19 years is relatively small in Japan.

Fig. 2

Trends of mortality rates in Japan in the RWC by age groups overall (left), and the manufacturing (center) and construction (right) industries. Upper figures show the mortality rates from 1991 to 1994 and lower figures show those from 2015 to 2019. Plots and bars mean the values of mortality rate in each single year and the average for each age group, respectively

The average ages at death and of the workers in the RWC and ASR are shown in Fig. 3. The five-year averages for the average ages at death and of the workers in Fig. 3 are shown in the Supplement (Table 2). The dashed lines for the estimated average age of the workers appear stepped because these values were estimated using data from the national census, which is conducted every five years. Both average ages overall are increasing, and the differences are almost constant from 1991 to 2019 (approximately five to six years). The average age in the manufacturing and construction industries is also increasing; however, the differences between the average ages at death and of the workers decreased to 3.5 and 2.7 years in 2019, respectively. Furthermore, the average age of the workers in the ASR is one year and two years higher than in the RWC, respectively. These differences could be due to the differences in the age group categories. In the ASR, there were two age groups of elderly workers (60–69 and 70 years and over) and younger workers (less than 17 and 18–19). However, the age group of elderly workers in the RWC is only 60 years and over, and that of younger workers is only 18–19. This explanation is consistent with our findings that when the average age at death in the ASR was recalculated using the age group categories in the RWC, the results approximately agreed with that in the RWC (Figure S1).

Fig. 3

Trends of average ages at death and of workers in Japan overall (left), and the manufacturing (center) and construction (right) industries

Table 2 List of five-year averages for average ages at death and of the workersOccupational mortality rate trends in industrialized countries

This study makes it possible to compare recent occupational mortality rate trends in Japan with those in Western countries. The number of deaths per 100,000 workers in Western countries (United States, Germany, Netherlands, France, and the United Kingdom) and Japan (RWC) is shown in Fig. 4 and Table S1. The mortality rates have declined over time, except for overall and those in the construction industry in the United States. The current mortality rates in Germany, the Netherlands, and the United Kingdom are lower than in other countries. The ranges of mortality rates in the last ten years (2009–2018) are roughly estimated as 0.5–5, 0.5–2.5 and 1–15 per 100,000 workers overall and for the manufacturing and construction industries, respectively. This indicates that the upper mortality rates are at least five to 15 times higher among industrialized countries than the lowest mortality rates. In 2018, the highest mortality rates overall and for the manufacturing and construction industries were 10.8, 6.5, and 18.1 times higher than the lowest rates, respectively. Further analysis of the data sources is needed to compare these countries’ mortality rates. As described above, the mortality rates in Japan differ according to the data sources. The data sources recorded in ILOSTAT also differ between countries. The records of ILOSTAT in Japan and the United Kingdom are based on labour inspection reports; in Germany, the Netherlands and France, they are based on insurance records; and in the United States, they are based on an established survey. Additionally, the data source in the United States for ILOSTAT was the Current Employment Statistics Survey until 1991 and the Census of Fatal Occupational Injuries since 1992. Interestingly, the mortality rate overall in the United States increased two-fold in 1992 (Fig. 4) due to data source change. This indicates that the differences among the reference data sources may not be unique to Japan but a problem common to other countries.

Fig. 4

Trends of mortality rates in the United States, Germany, the Netherlands, France, the United Kingdom and Japan overall (left), and the manufacturing (center) and construction (right) industries

Considerations for the possible underestimation and time lag between occurrence and certification

This study aimed to estimate the occupational mortality rates in Japan using open-source data. However, this study showed what likely affects mortality rates; underestimation and time lag.

The occupational mortality rates were calculated using statistical data related to occupational accidents. The number of reported cases tends to be underestimated because of the nature of the occupational accidents. Hämäläinen [1, 2] used a factor of 1.126 to compensate for the underestimation of fatal occupational accidents. When including deaths due to occupational diseases, Hämäläinen [1] estimated that the number of deaths due to occupational diseases in Japan was more than ten times larger than that due to occupational accidents. In this study, the mortality rates, except for those in 2011, calculated using data from the ABR were roughly 1.2 to 6.1 times higher than those calculated using data from the RWC. However, no calculation to compensate for the underestimation was performed in this study because no open data sources were identified that analyzed the number of unreported cases in Japan.

The mortality rates were defined as the ratio of deaths to the number of workers. When the number of deaths is estimated from insurance claims, a time lag between the occurrence of occupational accidents and the certification of claims is inevitable. In this study, the mortality rates from the ABR were particularly affected because the mortality data from the ABR included deaths due to occupational asbestos exposure, which has a long latency.

The results of this study should be interpreted in consideration of these factors. Occupational mortality rates can be used as a reference to determine the magnitude of risk. However, the estimation results of occupational mortality rates differ according to the data source, whether labor inspection reports or insurance records are used, and the method of collection.