查看文章

Background: Data quality (DQ) is important for disease registry and is associated with the values and reliability of collected data. There is no widely accepted practical quantitative data quality assessment framework for single-center disease registry.

Methods: We proposed a quantitative data quality framework which is composed of five dimensions, including Adjusted Total Number of Records (aTNRs), Total Number of Variables (TNVs), Record Completeness, Record Reliability, and Completion Rate of Follow-up (CRFu), with a final Disease Registry Quality Index (DRQI). We used this framework to assess the data quality of the Breast Cancer Registry of Sun Yat-sen Memorial Hospital (BCR-SYSMH) on an annual basis. The annual assessment reports were given back to the disease registry management group with the aim to improve their works.

Results: During the period between 2015 and 2017, the size of registry database has been increasing significantly. The aTNRs showed a steep upward trend of 157%. TNVs increased significantly between 2015 and 2017 and subsequently reached a plateau. Similarly, Record Completeness and Record Reliability also increased significantly between 2015 and 2016 and reached a plateau (Record Completeness: 91.2%-93.9%; Record Reliability: 82.1%-98.1%) since 2017. The CRFu was rising continuously and reached more than 80% in 2019 and up to 97.9% in 2020. A summary of DRQI by year was also calculated.

Conclusions: This study proposes a novel, easily implemented, quantitative data quality assessment framework that would provide guidance for the management of single-center disease …