Completeness of Reporting for Breast Cancer Data in the National Pathology-Based Cancer Registry in South Africa.

BACKGROUND: It is important for a cancer registry to have adequate coverage of the catchment area to accurately estimate the cancer burden. This study aimed to determine the pathology-based South African National Cancer Registry's (NCR's) catchment rate of breast cancer cases using a hospital-based cancer registry as reference. METHODS: Using 2 record linkage approaches, a combination of deterministic record linkage (DRL) and probabilistic record linkage (PRL), we linked a breast cancer hospital registry (n = 398) from 2015 with breast cancer registry data from the NCR (n = 16,642). Firstly, using DRL, we matched and linked records using the unique laboratory report number. Records that were not matched using DRL were linked using PRL. Manual reviews of both data sources were then performed to evaluate records that did not match using either DRL or PRL. The NCR's catchment rate was calculated using the total number of matched records from the hospital registry to the NCR breast cancer registry. RESULTS: Of 398 records from the hospital registry, 397 were matched to the NCR breast cancer registry, giving the NCR a catchment rate of 99.75%. A total of 291 records were matched with NCR records by DRL; 95, by PRL; and 11, by manual review. Only 1 record did not match. CONCLUSION: Nearly all hospital breast cancer cases were found in the NCR database. This suggests that the workflow used by the NCR for the identification, collection, and registration of breast cancer cases diagnosed histologically is adequate for this hospital.