Basic issues in population screening for cancer

The value of population screening as an approach to cancer control is controversial because the objectives, benefits, costs, and potential adverse effects of screening programs are not widely agreed upon, nor are the many difficulties involved in evaluating a cancer screening program well recognized. A review of the issues pertinent to cancer screening and their interrelationships is presented. The characteristics of a particular cancer that should be considered before it becomes the target of a screening program are described; emphasis is placed on the prevalence of the detectable preclinical phase. The features of a good screening test are reviewed and the importance of specificity is indicated. Methods for evaluation of screening programs are considered, and the need for measurement of the effects of screening on the mortality rate is emphasized. The biases inherent in some commonly used evaluation procedures, especially lead time bias and length bias, are discussed. The possible adverse effects of cancer screening are pointed out, especially its capacity to increase morbidity. Several approaches to improving cancer screening programs are presented. Finally, the relevance of knowledge regarding the optimum interscreening interval to cancer control and to research issues is described.