Abstract
Drs Junod and Spinosa raise the issue of overdiagnosis in the absence of screening. We used the customary definition of overdiagnosis, the diagnosis by screening of cancer which would never have become symptomatic in the lifetime of the host if screening had not taken place. With this definition, overdiagnosis in the absence of screening is impossible. However, Junod and Spinosa's point raises an interesting issue. If we accept for the sake of argument that there is a population of indolent tumours which would never be life-threatening if left untreated, it is by no means out of the question that a proportion of these tumours will give rise to symptoms leading to diagnosis outside of screening.
Our estimates 1 are not consistent with those of Zahl and colleagues.2,3 because we believe their estimates to be inaccurate due to failure to take account of complexities in temporal changes in cancer incidence. 4 Our assumption of a constant rate of overdiagnosed tumours at prevalance screen is reasonable, since the increased harvest of tumours at the prevalence screen of the control group was smaller than one would expect from the increased age compared with the study group at prevalence screen. Drs Junod and Spinosa's assumption is untenable since it would give a negative rate of overdiagnosis at incidence screens (see our equation (1)).
Junod and Spinosa state ‘The number of lives saved based on the two county trial has also been debated’. They have been debated, but the Two-County Trial endpoints were vindicated by a complete review published in 2009. 5 For these reasons, we submit that our results stand. We thank Drs Junod and Spinosa for their interest. It is a pleasure to discuss these points without the name-calling which usually accompanies such debate.