Methodological issues in etiologic studies of diet and colon cancer

Abstract

Recently, several hypotheses regarding the role of diet in the etiology of cancer have emerged. Here we review the principal methods available tor studying these hypotheses in human populations. Particular attention is given to the strengths and weaknesses of ecologic analyses, diet interview and other clinical and experimental approaches. Although the contributions of these methods are examined in the context of their application to research on colon cancer, the issues are of general relevance to the study of diet and cancer. For each method examined, potential sources of error are noted and suggestions for improving their reliability and validity are discussed. Since each method is vulnerable to error, a research program utilizing both existing and new methodological approaches is required if the role of diet in cancer etiology is to be understood.