Cancer epidemiological database linkage study of China (CEDLISC): Design, methods and quality evaluation

Connecting health related data from multiple databases is a novel approach for carrying out medical researches. Presently, the linkage of large medical databases in China is still underexplored. The Chinese Chronic Disease and Risk Factor Surveillance (CCDRFS) databases of years 2004–2015 with information on baseline risk factors for common chronic diseases were linked to the Population-based cancer registry (PBCR) database of China with information on cancer diagnosis and outcome. We used resident identification card numbers and combination of personal information variables as unique index variables for exact matching and fuzzy matching, respectively. Strict quality control procedures were performed based on the quality of databases and the logics of matched records. The 35–64 truncated incidence rate (TIR) and mortality rate (TMR) for all-cause of cancer in merged databases were used to select regions. 547,963 baseline records from the CCDRFS database were matched with 9,263 cancer diagnosis records from the PBCR database. Through quality control process, we created an epidemiology database of cancer incidence (EDCI) and an epidemiology database of cancer mortality (EDCM). The EDCI included 368,470 baseline records and 8,049 matched cancer incidences from 230 regions, with a 35–64 TIR of 309.54/100,000. The EDCM included 293,477 baseline records and 3,026 matched cancer mortalities from 183 regions, with a 35–64 TMR of 123.44/100,000. The database linkage and quality control methods were feasible in this study. The merged databases of cancer incidence and mortality were of high quality, which can provide scientific foundations for further cancer epidemiological studies.