audit of the quality of data in the National Cancer Registry, King Khaled University Hospital, Riyadh, Saudi Arabia, 2000.: a statistical approach

ABSTRACT

The completeness of case ascertainment was assessed through the utilization of multiple data sources including medical records, pathology reports and death certificates by employing a capture- recapture method and building up several log linear models. The validity of data by re-abstracting random sample [10%] of records to assess the minor and major disagreement of certain items and timeliness of data through calculation of reporting delay time [RDT] was investigated. A total of 991 cancer cases was reported in the registry during the study period, 71.4% of them had medical records, 77.5% had pathology reports and 28.1% had death certificates. A total of 571 missed cases was identified from the overlapping sources of reporting and fitting log linear models which assured the dependency of the reporting methods. The adjusted rate of ascertainment was 63.4%. The stage of cancer showed the highest major disagreement [43.6%], followed by the histology and primary site [25.6% and 10.3%, respectively] and the date of diagnosis [23.1%]. On the other hand, 250 cases showed RDT less than 6 months