Problems and pitfalls in a clinical resarch data management system

Abstract

The problems and pitfalls encountered in the computerized data bank for the Netherlands Coronary Surgery (NCS) study are reviewed. This study involved 848 patients seen before coronary artery surgery and at 1 and 3 yr after surgery. Nineteen data forms were used resulting in maximally 1142 variables per patient. The importance of quality control is emphasized as well as the efficient transfer of information from data bank to statistical processing.