Objectives: Use of Real-World Data (RWD) to augment evidence in clinical trials is often constrained by the type of endpoints considered. Particular limitations arise for Time-to-Event (TTE) endpoints where there are differences in missing events between trial data and RWD, definition of “time zero”, reporting bias, immortal time bias, among others. This research focuses on the problem of time zero in RWD, reviews currents methods to address this challenge, and makes recommendations.

Methods: Uncertainty in defining time zero arises when it is not clear if baseline of therapy initiation is the same for subjects in the clinical trial and those in RWD, especially the control arm. Not knowing the common baseline makes comparison of events difficult, and subsequent results less reliable.

Results: Drawing upon experience of TTE endpoints across therapeutic areas, we (Methodology Team of PSI Special Interest Group on Real-World Data) propose leveraging RWD for ‘softer’ endpoints for time zero such as time to ‘first’ events in neuroscience (multiple sclerosis relapse), respiratory (asthma exacerbations), infectious diseases (viral suppression), atopic dermatitis (change in treatment), oncology (treatment discontinuation, next therapy).

Conclusions: We make recommendations to mitigate the time zero uncertainty in RWD for TTE by adjusting the baseline to ‘first’ events for multiple diseases.