Re: Overdiagnosis from non-progressive cancer detected by screening mammography: stochastic simulation study with calibration to population based registry data
Seigneurin et al define ‘Overdiagnosis as the incidence in the screened population minus the incidence in the unscreened population, divided by the incidence in the screened population’.1
In the accompanying figure (figure 2) , there was no spike in the incidence of cancers following introduction of screening in 1991 which would be expected from lead time bias.2 This absence of an expected spike suggests not only poor uptake as acknowledged by the authors but possibly considerable opportunistic screening before mammography screening introduced in 1991and also possibly until 2006.
Thus their model (based on a biased data set) could have underestimated ‘over diagnosis’ by overestimating baseline incidence in the unscreened population.
Lead time bias i.e. early diagnosis (and consequent inevitable overdiagnosis) would be the major factor which would be responsible for improvements in cancer survival attributed to a screening test. If this is not predicted by a model, then the entire model might have an inherent flaw.
1. Seigneurin A, François O, Labarère J, Oudeville P, Monlong J, Colonna M. Overdiagnosis from non-progressive cancer detected by screening mammography: stochastic simulation study with calibration to population based registry data. BMJ 2011;343:d7017
2. Jorgensen KJ, Gotzsche PC. Overdiagnosis in publicly organised mammography screening programmes: systematic review of incidence trends. BMJ 2009;339:b2587
Competing interests: No competing interests