Present status and research activities of cancer registries in Japan.

Population-based cancer registries in Japan were first established in the cities of Hiroshima and Nagasaki in 1957-1958 for assessing radiation effects and in Miyagi in 1959 for promoting epidemiological researches, while most other prefectures in Japan set up cancer registries as a part of their own cancer programs. This resulted in the broader use of registry data in Japan. In 1975, the Research Group for Population-based Cancer Registration in Japan was first organized with a research grant under the National Cancer Research Promotion Program. Since then, continuous efforts have been made by the Group to improve the quantity and quality of registry data and to develop methodologies to utilize it. Studies being conducted using registry data cover (1) descriptive epidemiology, (2) analytical epidemiology, (3) evaluation of screening programs, and (4) evaluation of regional cancer medical care. In 1992, 32 regional cancer registries which were operating in Japan set up the Japanese Association of Cancer Registries. However, there are still many difficulties to overcome in order to achieve completeness of reporting in registries. Further improvement of reporting rate, together with standardization of registry data are left for future efforts.

at of the Research Group, analysed these data and reported the results to the Ministry. Through the Survey, the importance of the population-based cancer registry was identified. Ten years later, the 5th National Cancer Survey was conducted') again with the cooperation of all registries belonging to the Research Group in the same way as in the 4th Survey. The Rojin Hoken Hoo (Health Promotion Act for the Aged) was enacted in Japan in 1983 to promote health and welfare for the aged. As a part of the act, the director of the Bureau of Health, Ministry of Health and Welfare issued a Notice in which he recommended that cancer registration be set up by prefectural governments for planning and evaluating cancer control programs in each prefecture. Thus, prefectural cancer registries have developed in a more systematic administrative basis) since then, a number of prefectures have opened prefectural cancer registries, even though a national scheme for cancer registration had not yet been established.
In 1992, the Japanese Association of Regional Cancer Registries was established with 32 members comprised of existing prefectural and city cancer registries. The Association holds a scientific meeting once a year and encourages younger registries to improve data reliability, while the Research Group has promoted standardization and utilization of registry data with participation of 13 qualified registries in Japan with the capability of conducting epidemiological studies using their own registry data.

Registration System in Osaka
Cancer registration systems vary according to prefecture. The system in Osaka4) is described below as an example. The Osaka Prefectural Department of Health prepares a budget and the Osaka Prefectural Medical Association collects cancer reports from all medical institutions located in the prefecture. The Association requests physicians and hospitals to mail cancer reports to the Association. Cancer reports are prepared in hospitals by medical doctors or by medical record librarians, and sent to the Association. Those are transported once a month from the Association to the central registry. The registry is located in the Department of Cancer Control and Statistics, Osaka Medical Center for Cancer and Cardiovascular Diseases. The Osaka Prefectural Department of Health periodically sends copies of cancer death certificates of Osaka residents to the central registry with the permission of the Ministry of Health and Welfare. Cancer patients are registered from these two sources. After five years from the first diagnosis of cancer, the prognosis of all aliving cases are surveyed by health center staff. They match these files with the citizen's register in the city office and confirm whether they are alive or dead. Some patients are found to have moved outside the prefecture. These cases are further surveyed and accurate survival rates are calculated). Proportion of lost cases was less than 5%.
The Osaka Cancer Registry started a cancer information service program since 1975 in which prognostic information concerning each reported patient as well as the hospital's cancer statistics were provided free at the request of participating hospitals or hospital doctors. The registry has held conference annually on cancer registration in Osaka, inviting all large and middle sized hospitals and clinical departments of medical university to report cancer registration activities. Through these activities, the cancer registry has greatly contributed to the progress and assessment of cancer medical care in hospitals.

Registries in Japan
Prefecture-wide cancer registries have generally been organized by prefectural governments.
Exceptions are the Hiroshima and Nagasaki city registries, which are organized by the City Medical Association and the Radiation Effect Research Foundation.
Type of institution in which central registries are located varies according to prefecture. They are usually established in a prefectural cancer center or in a prefectural general hospital. If there are no such institutions in the prefecture, it is often set up within a prefectural medical association, cancer society, cancer detection center, research institute on health and environment, or medical university, etc.
Cancer registration in Japan has been operating on a voluntary basis. Data on patients are collected actively in three registries (Hiroshima, Nagasaki and Miyagi) and passively in all other registries because of budgetary limitations and a lack of registry personnel. Medical record librarian is not posted in most hospitals, therefore it is extremely difficult for central registries to achive a complete reporting. In many registries the staff visit hospitals to help with completing cancer reports when necessary.
A one-week training program for cancer registry personnel has been held twice a year at the National Cancer Center in Tokyo under the sponsorship of the Ministry of Health and Welfare since 1987 with technical support from the Research Group.
One fourth of existing registries have conducted periodical matching with all death certificates of residents in the prefecture to obtain patient prognostic information. Only a few registries have had further active follow-up surveys. These registries have reported 5-year survival rates every year'"') Half of the exsisting registries in Japan have published annual registry reports and distributed them to the authorities concernedlo.13.14) Additionally, cancer incidence rates are published yearly by the Research Group for the 13 member registries.

Descriptive Epidemiology
Cancer statistics should be prepared for each level of nation, Cancer Registires in Japan S-39 prefecture, city, town and village. The group has estimated cancer incidence in Japan on a yearly basis using data from selected registries whose reporting rates match a certain standard"). Based on these trends, the magnitude and trends of cancer occurence in Japan in the future have also been periodically estimated 16) In the field of descriptive epidemiology, besides studies on cancer incidence and the trends1-19), studies on multiple cancers have been conducted according to site. Incidence rates of multiple cancers in cancer patients were calculated and compared with their expected incidence rates to clarify risk factors for occurrence of multiple cancers"'). Cancer registries have been collecting data on the histological diagnosis of tumors. Incidence rates by histological group were calculated for selected sites. The magnitude of risk and the time trends of specific histological groups were studied in relation to changes in etiological factors25-28).
The figures for cancer prevalence are important indices in planning cancer medical services and in evaluating the effectiveness of the services. A method estimating the numbers of long-term and short-term survivors was developed using registry data. Short-term cancer survivors with shorter survival period than 5 years were regarded as an approximation of cancer prevalence29, 30>

Analytical Epidemiology
In the field of analytical epidemiology, accurate and effective cohort follow-up studies have become feasible by using a method of record linkage between two files; the one is composed of data on each person with well-defined epidemiological features and the other is a registry's cancer patients' file31.32) Many studies have been conducted using this method to identify causal factors, and to assess risks33. 34.35) or to observe natural history of cancers 36,37).

Evaluating Cancer Screening Programs
Historically, population-based cancer registration has been conducted mainly to estimate of cancer incidence in many countries. Since it was started, the Research Group has strived not only to use cancer registry data for cancer statistics and for epidemiological studies but also to actively utilize this data for planning and evaluating cancer programs in a region, and has long promoted the development of methodologies for these objectives.
In the field of assessment of validity of cancer screening programs, an accurate identification of cancer cases newly diagnosed after screening (interval cases or false negative cases) had been regarded as nearly impossible until the method of record linkage was developed for matching between screenees and cancer patients. Since then, validity has been easily estimated for specific screening procedures"',"'. Clinical diagnostic methods introduced in hospitals have also be assessed quantitatively, using the same method40'41~.

Evaluation of Regional Cancer Medical Care
Trends of cancer statistics obtained from population-based cancer registries, such as cancer incidence, mortality, stage distribution, medical examination rates, treatment rates and survival rates of cancer patients, were analysed to assess cancer prevention and the medical care programs conducted42-a5) CONCLUSION Population-based cancer registries were operating in two cities and 33 prefectures out of 49 prefectures in Japan as of 1995 which cover 70% of the entire population in Japan, while fewer than 10 registries have achieved roughly satisfied completeness of reporting15). When the registries adopt a method of passive collection of data, it is essential that hospials have a central medical record library. This is, however, not always realized under the present service system in hospitals.
The Research Group for population-based cancer registries has carried out various activities to promote utilization of registry data for more than 20 years. The results reported from the Group are now being highly evaluated in every fields of cancer related medical sciences. It has been four years since the Japanese Association of Cancer Registries was established. The Association will help registries to improve reliability of their own data.
Achievement of sufficient completeness of reporting and standardization of the methods will be the focus of future efforts. It is expected that the methodology developed thus to utilize registry data will be broadly practiced to bring new and positive results in cancer epidemiology.