[Skip to Content]
[Skip to Content Landing]
Article
November 28, 1925

WHAT DO CANCER STATISTICS REVEAL?

JAMA. 1925;85(22):1730-1731. doi:10.1001/jama.1925.02670220048017
Abstract

Medical statistics, especially cancer statistics, are notoriously open to various interpretations. A lay statistician, to whom a recorded diagnosis means that death was undoubtedly the result of the disease to which the diagnosis attributes it, will draw conclusions different from the analysis made by a physician untrained in the methods of the statistician. The latter is likely to avoid pitfalls that trap the non-medical statistician and to fall into errors that the statistician avoids. A still different result is likely to be reached by a pathologist studying the same figures. He, with his daily observation of the margin of errors in diagnosis, especially of cancer, is likely to discard many figures that would be accepted by both the statistician and the clinician, or give to them an entirely different interpretation. The German pathologist Lubarsch, after considering the statistics of the German hospitals for the years 1920 and 1921, as part

×