What explains the differences between centres in the European screening trial? A simulation study

•Differences in the screening interval and PSA cut-off did not explain differences in screening impact by ERSPC centre.•Less intensive screening algorithm may be preferred.•The small screening impact in the Finnish centre was largely due to a low prostate cancer mortality in the Finnish control arm.•Contamination complicates the assessment of maximal efficiency of prostate cancer screening in ERSPC as a whole.

BackgroundThe European Randomised study of Screening for Prostate Cancer (ERSPC) is a multicentre, randomised screening trial on men aged 55-69 years at baseline without known prostate cancer (PrCa) at randomisation to an intervention arm invited to screening or to a control arm. The ERSPC has shown a significant 21% reduction in PrCa mortality at 13 years of follow-up. The effect of screening appears to vary across centres, for which several explanations are possible. We set to assess if the apparent differences in PrCa mortality reduction between the centres can be explained by differences in screening protocols.MethodsWe examined the centre differences by developing a simulation model and estimated how alternative screening protocols would have affected PrCa mortality.ResultsOur results showed outcomes similar to those observed, when the results by centres were reproduced by simulating the screening regimens with PSA threshold of 3 versus 4 ng/ml, or screening interval of two versus four years. The findings suggest that the differences are only marginally attributable to the different screening protocols.ConclusionThe small screening impact in Finland was not explained by the differences in the screening protocols. A possible reason for it was the contamination of and the unexpectedly low PrCa mortality in the Finnish control arm.