Table 3.

Summary of key study observations

Key observation
1. Factors such as variability in data source, alignment of RCT and RWD eligibility criteria and study design, choice of available prognostic factors for inclusion in the statistical analysis, and varying analytic assumptions can impact the comparability of clinical trial control arms and RWD cohorts.
2. Real‐world external comparator cohorts may produce similar outcomes as observed in clinical trial control arms, however, comparability may be context‐dependent.
3. Transparent accounting and documentation of the clinical rationale behind decisions to apply or not apply certain clinical trial eligibility criteria to real‐world external comparators is essential.
4. Certain barriers to implementation of specific clinical trial eligibility criteria when constructing real‐world external comparators may be overcome with customized abstraction and curation of tailored data models, while others remain as known limitations of RWD. Of note, some eligibility criteria may not be meaningful outside of a clinical trial setting (e.g., ability to sign consent).
5. Successful replication of clinical trial control arms using RWD requires careful selection of a comprehensive set of disease‐ and trial‐specific prognostic factors.
6. Access to clinical trial patient‐level data is important to inform appropriate RWD study design and cohort construction.
7. Prospective applications of real‐world external comparators for ongoing or future trials performed in collaboration with the study sponsor should take advantage of the opportunity to use patient‐level data and proactively incorporate analytic considerations such as comprehensive prognostic variable data capture.

RCT, randomized clinical trials; RWD, real world data.