Estimation of overdiagnosis by detection to incidence ratio in cancer screening

Objective We aimed to introduce a concept of overdiagnosis with actual formulae. In screening for invasive cancer, the detection proportion to incidence rate (D/I) ratio is used as an intermediate indicator of effect. D is the detection rate (proportion) at screen and I is the background risk (i.e. the expected incidence rate if there were no screening). We sought to determine a proper use of the D/I ratio in the estimation of overdiagnosis in screening. Methods We propose the difference between the D/I ratio per year and incidence-based sensitivity as an estimate of overdiagnosis. As an example to evaluate this, we used screening for colorectal cancer, run as a randomized public health policy in Finland. Conclusion The use of the D/I ratio is inappropriate in predicting effect, and we recommend that the D/I ratio be used as a component in the estimation of overdiagnosis in screening for invasive cancer.