Improving Data Validity at Population-Based Cancer Registry through Trace-Back of Death Certificates: A Concrete...

The validity of data provided by population-based cancer registries (PBCRs) is a key aspect in cancer surveillance. Tracing back cases initially reported by death certificate or death-certificate-notified (DCN) cases, improves data quality and has an especially significant impact on survival estimates. The present study performed in the Mendoza PBCR describes the trace-back procedure of cancer cases notified by death certificate for selected cancers (liver, lung, and stomach cancers) with the aim of reducing the percentage of cases diagnosed by death certificate only (DCO). The study was performed in 2018 using cancers diagnosed between 2006 and 2012 in the framework of a survival project (SURVCAN-3). Among the 822 cases that have been traced back, only 32.1% had an identified source of information. Of these, 70.3% had medical records available for review. Of the reviewed medical records, 86.9% of cancer diagnoses were confirmed. The DCN and DCO cases were much higher among older age groups. With the trace-back, the overall percentage of DCO was reduced from 23.8% to 19.9%. We conclude that DCN trace-back could improve data quality by reducing DCO diagnoses, which directly impacts survival estimates. Trace-back should be performed routinely and in a timely manner.