Abstract
Aims: This study aims to assess the reliability of New Zealand Cancer Registry data on colon cancer.
Methods: Data from a review of the clinical records of 642 people diagnosed with colon cancer between 1996 and 2003 were used to audit the data held on these individuals by the New Zealand Cancer Registry (NZCR). The record review data were treated as the "gold standard".
Results: Age at diagnosis (measured in years) recorded by NZCR was 96% accurate, and date of diagnosis was within 6 weeks of the clinical date of diagnosis in more than 97% of cases. Overall tumour site was recorded with more than 95% accuracy, with 86% accuracy for tumour sub-site within the colon. Tumour grade was only recorded consistently by the NZCR from 1999 onwards, from which time the NZCR was 83% accurate for tumour grade. Tumour stage was the least accurate variable studied, with 80% accuracy. The NZCR data quality improved over the period of this study.
Conclusions: The accuracy of the NZCR appears to be similar to that found in comparable audits of cancer registries, with stage being the hardest variable for registries to collect accurate information on. NZCR data could be improved by improving the quality of information provided to the registry.