To extend quantitative bias analysis (QBA) methodology specifically to multivariable time-to-event analyses using data-driven simulations.
Key Findings:
The proposed simulation approach allows for direct assessment of how imperfect data diverges from true parameter values, highlighting the significance of unmeasured prognostic factors and imprecise timing of events.
Interpretation:
Data-driven simulations provide a flexible and relevant method for evaluating the effects of data imperfections in time-to-event analyses, thereby enhancing the robustness of findings in practical applications.
Limitations:
The approach relies on accurate assumptions for simulations, which may not always reflect real-world complexities, potentially impacting the validity of results.
The number of repetitions required for desired precision may vary based on specific applications.
Conclusion:
This simulation-based QBA approach complements traditional methods and offers valuable insights into the effects of data limitations in epidemiological studies.
by Michal Abrahamowicz, Marie-Eve Beauchamp, Anne-Laure Boulesteix, Tim P Morris, Willi Sauerbrei, Jay S Kaufman, on behalf of the STRATOS Simulation Panel
Damon B. Dixon, MD, at Phoenix Children’s Cardiology, is the author to this EndoText chapter. Dr. Dixon brings nationally recognized expertise in pediatric cardiovascular risk assessment and non?invasive vascular imaging.
