In solid tumor oncology studies, analyses of endpoints such as progression-free survival (PFS) are based on information detected by imaging assessments. When using real-world data (RWD) in comparative effectiveness studies however, surveillance bias may be introduced due to differential imaging assessment frequency. This study ultimately provided evidence that the frequency of imaging assessments to detect disease progression can differ by treatment type in real-world patients with cancer and researchers should remain cognizant of the potential influence of surveillance bias when working with RWD.
Why this matters
To leverage the full value that RWD brings to clinical research, it is critical to understand and address some of the relevant limitations associated with this type of data. Surveillance bias, namely, distortion in the frequency of observed events introduced by variability in patient monitoring, could have a crippling effect on the validity of certain data. This report tackles that topic in oncology RWD and RWE, where well-established endpoints, such as those related to progression outcomes in solid tumors, depend on the frequency of imaging evaluations. The study illustrates an approach to better understand this confounder, by leveraging actual RWD to characterize the originating issue, and using simulations to estimate the potential impact in downstream analyses.