Blinded and uniform cause of death verification in a lung cancer CT screening trial
Horeweg, N × van Klaveren, R J Groen, H J M Lammers, J-W J Weenink, C Nackaerts, Kristiaan Mali, W Oudkerk, M de Koning, H J #
Lung Cancer vol:77 issue:3 pages:522-525
Disease-specific mortality is the final outcome of a lung cancer screening trial, therefore cause of death verification is crucial. The use of death certificates for this purpose is debated because of bias, inaccurate completion and incorrect ante mortem diagnoses. A cause of death evaluation process was designed to ensure a uniform and unbiased determination of the graduation of certainty that lung cancer was the underlying cause of death. An independent clinical expert committee will review the medical files of all deceased participants once diagnosed with lung cancer and will make use of a flow chart and predetermined criteria. A pilot study of fifty cases was conducted to determine the performance of this process and to compare the outcome with the official death certificates. The independent review has shown an agreement of 90% (kappa 0.65), which demonstrates a uniform classification. The sensitivity and specificity of the death certificates for lung cancer specific mortality were 95.2 and 62.5%. This demonstrates a limited distinctive character of the death certification process in lung cancer patients. Our results imply that the final outcome of a lung cancer screening trial cannot reliably be established without predetermined criteria and an independent review of blinded cases.