Abstract

PURPOSE
We evaluated the accuracy of death certificates for persons who registered in the Central Cancer Registry in 1993 and died from 1993 to 1995.
MATERIALS AND METHODS
The study consisted of 27,058 cases which were classified into five groups. according to the possibility of accuracy of the underlying causes of death. We compared the distribution of five groups according to several demographic factors. Also we calculated the detection rate and the confirmation rate for a selected 23,858 persons reported to die of a cancer.
RESULTS
Among the 27,058 deaths, only 64.4% was included in the group which had reported the same cancer site with registry as underlying cause of death. The accuracy decreased with increasing age and was worse for women and rural residence. And physicians certification was an important factor to improve the accuracy. Cancers of stomach, lung, esophagus and breast were included into the high accuracy group, cancers of the colon, rectum and gallbladder and extrahepatic biliary tract into the low accuracy group. Cancers of the colon, pancreas, liver and lung were overreported, varied from 1.2 to 1.4 times, and cervical cancer was severely underreported, about 0.5 times.
CONCLUSION
These results suggest that the caution in the use and interpretation of cancer certificate data would be required.

• Keywords: Cancer registry;Cancer death;Medical certificate;Accuracy