Standardization of laboratory test in the JPHC study. Japan Public Health Center-based Prospective Study on Cancer and Cardiovascular Diseases.

The standardization committee has carried out standardization of 23 laboratories in the cohort area. They participated in the External Quality Control Survey by the Japan Medical Association. Most laboratories got A or B in evaluation criteria for most test items, but the results of AST, ALT and gamma-GTP were unsatisfactory. As for the lipid standardization, accuracy and precision of all 23 laboratories were satisfactory except for one. Close communication and collaborative study with reference laboratory improved the accuracy control.


INTRODUCTION
Laboratory data which were sent from each study area should be standardized to compare each result. The standardization committee (Chairman: Dr. Iida) has carried out standardization of 23 laboratories in the cohort area. Osaka Medical Center for Cancer and CVD is a member of Cholesterol Reference Method Laboratory Network (CRMLN), and contributed to standardize lipid measurement in Japan for both epidemiology and clinical chemistry.

METHODS OF STANDADIZATION
The subjects for standardization were 10 items that were measured at the health screening program in the local health centers by the Law of Promoting Health for Elderly. These included total protein, glucose, uric acid, creatinine, total cholesterol, AST (GOT), ALT (GPT), gamma-GTP, triacylglycerol, and HDL cholesterol. Accuracy control and standardization among each laboratory in the cohort area were performed by the Division of Mass Screening, Osaka Medical Center for Cancer and CVD.
Basically, External Quality Control Survey by the Japan Medical Association was employed. Accurate reference value was available for total cholesterol, HDL cholesterol and triacylglycerol by the CRMLN Lipid Standardization Program through Osaka Medical Center for Cancer and CVD.
External Quality Control Survey by the Japan Medical Association has been planed in every June, called for participation in July, and 4 samples were sent to laboratories in September. Each participant measured samples and reported the results to the Association in October. The results were evaluated according to the evaluation criteria (Tables la-c), and returned to each participant in the next February. We collected copies of these evaluation sheet from each laboratory to the Osaka Medical Center for Cancer and CVD, and evaluated.
Thirty sample sera for lipid analysis were sent to the laboratories in December or at the time of mass screening, according to the CRMLN Lipid Standardization Program by Osaka Medical Center for Cancer and CVD. Each laboratory should randomly select 3 samples for one day, and repeat duplicate measurement for 10 consecutive days. We collected the data and evaluated the results according to the criteria (

S-82
Standardization of Laboratory Testing    Table 3. Participation in the external quality control survey by the Japan medical association.

RESULTS OF STANDADIZATION AMONG LABORATORIES
The participation of the 28th External Quality Control Survey by the Japan Medical Association in 1994 by each laboratory were shown in Table 3. Sixteen of 24 laboratories sent us the results. Participation status to the CRMLN Lipid Standardization Program was shown in Table 4. Twenty-three laboratories participated in the second study which was carried out during January and February in 1995.
Evaluation of 16 laboratories was summarized in Table 5 . Most laboratories got A or B in evaluation criteria for most test items, but the results of AST, ALT and gamma-GTP failed to be C or D category in many laboratories. As for the 2nd Lipid Standardization Program, accuracy and precision of 23 laboratories are shown in Figure 1. Accuracy of cholesterol measurement was satisfied in 19 out of 23 and precision satisfied in all laboratories. Triacylglycerol measurement was satisfied in both accuracy and precision by all laboratories except one . The same was for HDL cholesterol measurement.

PROBLEMS OF FUTURE RESOLUTION
From the results of External Quality Control Survey by the Japan Medical Association, AST, ALT and gamma-GTP measurement showed the problem in accuracy control . These should be improved. On the other hand, the results of lipid standardization program were satisfactory. This is the result of our close contact in the cohort study since the project had started. The difficult problem for quality control on the laboratories was that the health screening data were indirectly obtained from laboratories that were nominated by the local city , town or village which should carry out the health screening for the residents by the law. At the beginning of this study , direct communication between reference laboratory and each laboratories was not established, but it has been improved recently , as S -85