Document Type: Research Articles
Department of Internal Medicine, School of Medicine, Tulane University, 1430 Tulane Ave., New Orleans, LA, USA.
Section of Hematology and Medical Oncology School of Medicine , Tulane University, 1430 Tulane Ave., New Orleans, LA, USA.
Parkview Health, Parkview Cancer Institute, 11050 Parkview Circle, Fort Wayne, IN, 46845, U S A.
Introduction: The National Cancer Database (NCDB) is a clinical oncology database utilized by many researchers and clinicians internationally. We sought to investigate the various trends in data of two of the most common cancers, breast and lung, published using the NCDB. Materials and Methods: We selected a multitude of pre-determined variables for analysis. We then performed two separate literature searches using an advanced PubMed search builder, and the data was combined to determine each variables’ association with journal impact factor (IF) using both univariate and multivariate analyses. Results: A total of 191 published studies were identified. We found that a journal IF > 5 was associated with a publication year prior to 2017 (univariate analysis OR 2.68, 95% CI 1.38-5.21, p-value 0.004 and multivariate analysis OR 3.47, 95% CI 1.62-7.42, p-value 0.001) and a sample size > 10,000 (univariate analysis OR 3.27, 95% CI 1.43-7.50, p-value 0.005 and multivariate analysis OR 4.68, 95% CI 1.89-11.6, p-value 0.0008). Variables such as number of authors, region, cancer type, stage, treatment outcome and treatment incidence were not significant for an association with an IF >5. Conclusion: Based on our data, studies published after 2017 using the NCDB were associated with a lower IF. This could suggest that the quality of the NCDB data may be declining over time, or NCDB is becoming more widely used.