We feel that it is important to correct serious misinterpretations brought forward by André N. Vis1 in his commentary on the article by Linda Perron and colleagues.2
First, it is wrong to say that the Quebec prostate cancer study had low statistical power without pooling groups. Contrary to this statement, the first analysis performed at 8 years of follow-up showed a 69% decrease in the death rate from prostate cancer (p < 0.03) in the screened versus the unscreened groups of men.3 Thus, in complete contradiction to Vis, the statistically significant difference mentioned above was obtained without any pooling of data.
It is also false to say that men in the unscreened group were at risk for a longer period. Deaths were expressed per 100 000 person-years of exposure, thus avoiding such bias.
Fernand Labrie Bernard Candas Prostate Cancer Research Unit Oncology and Molecular Endocrinology Research Centre Laval University Medical Centre (CHUL) and Laval University Quebec City, Que.