Key points are not available for this paper at this time.
In an autopsy series of 8571 cases from 1975 to 1984, cancer deaths increased significantly, particularly in those cases over 60 years of age. A similar trend was seen in the mortality statistics from the district. Study of the degree of correlation between the frequency for the different types of cancer and the total deaths from that cancer at autopsy in the population supplies guidelines for use in assessing the relevance of autopsy findings to epidemiological research. It is stressed that without autopsy control monitoring of cancer in a population is incomplete and unreliable. An autopsy frequency of 40 per cent or more for the cancer in question in the population as a whole is required combined with series of 5-600 cases if satisfactory studies are to be initiated. Such conditions are difficult to fulfil for the less common types, but offer a valid approach to the monitoring of those types of cancer that are numerous enough to merit such attention.
Karwinski et al. (Wed,) studied this question.