Abstract

Using data taken from the organised mass screening system in Finland, risk indicators of cervical cancer were identified in order to define a high-risk group which could then be used for selective screening of cervical cancer. Single risk factors classified at best 39% of the cases into a high-risk group of 8%. A combination of risk factors by different statistical methods was applied, but the results were essentially the same. In order to find a high-risk group small enough to yield a reduction in costs, the number of cases originating from the low-risk group was increased. Theoretical calculations showed that for selective screening to be effective, the risk of disease in the high-risk group relative to that in the low-risk group must be greater than that implied by current knowledge of cervical cancer epidemiology. It was concluded that selective screening has only a limited applicability.