Cancer publications using real-world data from the Taiwan National Health Insurance Research Database: Conceptual...

Bibliometric analysis often overlooks study-based components such as study aims, design, and statistical methods. In this study, we propose a conceptual framework incorporating these study-based components with disease-based components for the bibliometric analysis of cancer articles using real-world data. This framework is a significant step forward in cancer research. We then investigated the distribution and temporal trends of these components for cancer articles using the Taiwan National Health Insurance Research Database (NHIRD) published from 2006 to 2022.

Study- and disease-based components were extracted and cross-validated. The distribution and temporal trends of these components were then presented.

We analyzed 1232 articles and found a noticeable increase in the annual publication count from 2011 onward. This upward trend signified the growing momentum in cancer research. Cancer risk factors were the most studied (52%), followed by cancer outcomes (36%) and incidence/prevalence (3%). Among the publications on risk factors and outcomes, most were cohort studies (85%), followed by case-control studies (10.7%). In both study designs, the use of the propensity score method increased steadily from 2.4% in 2011 to 40% in 2022. The most frequently studied cancer site was "all cancers or multiple cancers" (25.6%), followed by breast (9.6%), hepatobiliary (9.2%), and colorectal cancers (8.8%). Among the top 10 cited articles, the first and fourth focused on whether suppressing hepatitis B viral load with nucleoside analogs could reduce hepatocellular carcinoma recurrence and incidence in chronic hepatitis B patients. The remaining eight examined the association between medications and cancer risk.

Beyond citation metrics, our research underscores the importance of considering study-based and disease-based components in bibliometric analysis. These components form the foundation of the real-world data cancer research framework and have practical implications for diseases beyond cancers, providing a broader perspective for researchers and practitioners.