Geographic characteristics and mortality profiles in the JPHC study. Japan Public Health Center-based Prospective Study on Cancer and Cardiovascular Diseases.

The study areas of the Japan Public Health Center-based Prospective Study on Cancer and Cardiovascular Diseases (JPHC Study) are distributed throughout Japan and represent both rural and urban communities. These geographical differences yield considerable difference in population, culture, and lifestyle. The mortality rates in the study areas were apparently influenced by these factors. The pattern of standardized mortality ratios (SMRs) for all causes of death (cancer, heart diseases, ischemic heart disease, and cerebrovascular diseases) in each area was different. Age-standardized site-specific cancer mortality rates showed large variation even when compared with corresponding figures of prefectures in Japan. The areas of the JPHC study showed different patterns of SMR for major causes of death. The differences in site-specific cancer mortality rates for most of the sites are sufficient for epidemiological analyses.


INTRODUCTION
The Japan Public Health Center-based Prospective Study on and Miyako), which are mostly located in western Japan, made the study area representative of Japan, covering both the northern and southern regions. Variation in mortality and baseline characteristics between the different study areas were large enough to investigate relationships.

MATERIALS AND METHODS
Information on population, size of districts, and composition of industries in each study area was obtained from the 1990

Results
Population numbers; size of areas; and proportion of primary, secondary, and tertiary industries in each area are shown in Table 1. Tertiary industry accounted for more than 60% in Katsushika and Suita, but even in some rural areas it represented more than 50%. Primary industry exceeded 50% in only Gusukube in the Miyako public health center area and a few villages in the Saku public health center area.
Population pyramids in each area are shown in Fig. 1. The proportion of younger people (20-29 age category) was large in Katsushika (Tokyo) and Suita (Osaka). In contrast, the aged category (more than 65) was prominent in Kashiwazaki (Niigata), Saku (Nagano), and Tosayamada (Kochi). A deficit in the age 20-24 category, especially males, suggested their moving toward large metropolitan cities. The number of workers according to industry is shown in Fig. 2. The urban areas of Katsushika and Suita showed a very low proportion of farmers. Instead, manufacturing, trade, and the service industries were dominant. Agriculture accounted for nearly one-fourth of the jobs in Ninohe, Yokote, Kashiwazaki, Saku, Tosayamada, and Miyako, and the fishing industry is prominent in Arikawa. Thus, each public health center district has rather characteristic industrial features.
As for the mortality in cohort I areas, in Ninohe, SMRs for all causes (male), heart disease (male), and cerebrovascular disease (both sexes) were high (Tables 2 and 3, Fig. 3), whereas SMRs for ischemic heart disease (male) and cancer (female) were low. In Yokote, the SMR for cerebrovascular disease (female) was high, whereas SMRs for heart disease (female) and ischemic heart disease (both sexes) were low. In Saku, SMRs for all causes (cancer, heart disease, ischemic heart disease, and cerebrovascular diseases) were all low in both sexes. In Katsushika, SMRs for all causes (female), cancer (female), ischemic heart disease (both sexes), and cerebrovascular disease (both sexes) were high. In Ishikawa, SMRs for all causes were low in both sexes.
As for the areas in cohort II, in Kasama, SMR for cerebrovascular disease (female) was low, as well as SMRs for cancer (male), heart disease (female), and ischemic heart disease (both sexes). In Kashiwazaki, SMRs for all causes (both sexes), cancer (male), heart disease (both sexes), ischemic heart disease (both sexes), and cerebrovascular disease (male) were low. In Tosayamada, SMRs for ischemic heart disease (both sexes) were low.
In Arikawa, SMRs for all causes (male) and cancer (both sexes) were high. SMRs for heart disease (female) and ischemic heart disease (female) were low. In Miyako, SMRs for all causes (female), cancer (female), heart disease (both sexes), ischemic heart disease (both sexes), and cerebrovascular disease (both sexes) were low. In Suita, SMR for cancer (female) was high, although SMR for cerebrovascular disease (female) was low.
Age-standardized site-specific cancer mortality rates are shown in Tables 4 and 5, using the 1985 model population of Japan as the standard, and in Tables 6 and 7, using the world population as the standard. Cancer mortality rates of males were more than twice those of females at most of the sites. Highest and lowest mortality rates for each site were as follows: for stomach, Yokote (male) and Kasama (female) were highest and Ishikawa (both sexes) was lowest; for lung, Arikawa (male) and Ishikawa (female) were highest and Kashiwazaki (male) and Ninohe (female) were lowest; for      Table 4. Area-specific age-standardized mortality rate for male (/100,000) (Japan 1985 model population as standard) .
*maximum of age-adjusted rate between areas . **minimum of age-adjusted rate between areas. ***difference between maximum and minimum rates . ****ratio of maximum to minimum rates. Table 5. Area-specific age-standardized mortality rate for female (/100,000) (Japan 1985 model population as standard).
*maximum of age-adjusted rate between areas. **minimum of age-adjusted rate between areas. ***difference between maximum and minimum rates. ****ratio of maximum to minimum rates. Table 6. Area-specific age-standardized mortality rate for male (/100,000) (world population as standard). Area-specific cumulative risks and rates for site-specific cancer mortality are shown in Tables 8 and 9 (males) and Tables 10 and 11 (females). Patterns similar to age-adjusted mortality rates were observed. .
SMRs for site-specific cancer mortality rates are shown in Tables 12 and 13. Site-specific SMRs varied widely within the areas; more than 70% of participants had SMR over 110 (high) or under 90 (low).

Discussion
The different aspects of each study area may reveal relationships between the different natural environments , individual Geographic Characteristics and Mortality Profiles in JPHC Study S _21 Table 8. Area-specific cumulative risk (0-74) for male. Table 9. Area-specific cumulative rate (0-74) for male.    cultural aspects, or habitual lifestyles. Cross-sectional or ecological studies would be possible across these 11 public health center districts. To investigate the relationship between exposure to environmental factors and mortality, substantial variation is needed in mortality rates between areas, especially when investigating each cancer site separately. Variability was examined by ageadjusted mortality rates and SMR. Although age-adjusted mortality rates are superior to SMRs in terms of the comparability between areas, they may not be stable when the number of site-specific cancer deaths in an age strata is low.
Variability of site-specific cancer mortality rates in the study areas was compared with corresponding rates of prefectures in Japan (4). Most of the maximum and the minimum values in our study areas were of equal size to those of prefectures, and some values even surpassed the prefecture values. The differences and ratios between maximum and minimum age-standardized mortality rates showed that there were relatively large differences in mortality rates due to stomach, lung, and liver cancers in males and stomach cancer in females. Ratios were relatively large for skin, lymphoma, biliary tract, esophagus, and liver cancers in males and skin and uterine cancers and leukemia in females. The pattern of the ratios was similar to the difference of SMRs. Large SMR differences tended to be seen in large mortality rates and large ratios tended to be seen in small rates. The large value of ratio of age-standardized rates and difference of SMR should be cautiously interpreted, however, because they were heavily influenced by minimum values and absolute numbers. Due to these reasons, in addition to their public health impact, absolute rates and differences should be taken into account in etiological studies..
In summary, the study areas of the JPHC Study showed different patterns of SMR for major causes of death. And , for most of the sites examined in this paper, there appear to be sufficient differences in site-specific cancer mortality rates to draw inferences regarding possible environmental, cultural, and lifestyle causal factors..