Assessing the Cause of Death for Men with Prostate Cancer Using Official Mortality Statistics or a Dedicated Cause of Death Committee: Results from 30-year ERSPC Rotterdam Data

Abstract

For men with prostate cancer (PCa) within the European Randomized Study of Screening for Prostate Cancer (ERSPC), the cause of death is determined by a Cause of Death Committee (CODC) that evaluates all medical records using a fixed algorithm. The aim of this study was to compare the classification of PCa-specific mortality (PCSM) between the CODC and Statistics Netherlands. We calculated the sensitivity (PCSM agreement divided by total PCSM deaths according to the CODC) and specificity (agreement for other-cause mortality [OCM] divided by total OCM deaths according to the CODC) using the last 21-yr follow-up data from ERSPC Rotterdam. For the core age group (age 55–69 yr at randomization; n = 1732), the sensitivity was 86% (95% CI 83-89) and specificity was 93% (95% CI 91-94), with no statistical difference between the youngest ages and the oldest ages. Extrapolation of our findings to 30 yr of follow-up would result in an expected risk reduction of PCSM of 30% using data from the CODC and 33% using official statistics in favor of screening. In conclusion, our results support the use of official statistics in determining the cause of death, without compromising the main outcome of ERSPC Rotterdam.

Patient summary

We compared the classification of prostate cancer death between a dedicated trial committee and official statistics in the Netherlands. We found that official statistics are an accurate representation in determining the cause of death.