OBJECTIVES: Data maturity in observational studies is essential for ensuring reliability and validity of findings. Inherent biases may indeed lead to unexpected results compared with randomized clinical trials. Interim analyses outputs are typically hard to predict, either because all patients are not yet recruited, or because longer follow-up is required. Thus, endpoints stability in real-world evidence (RWE) must be evaluated largely based on the data in hand maturity assessment. This work proposes a comprehensive approach, from protocol to interpretation of results, for assessing data maturity.

METHODS: First, a literature review was conducted to provide exhaustive knowledge on data maturity concepts and considerations. Next, a mock prospective, observational, multi-center study design was developed with time-based recruitment. Patient-level data and associated survival outcomes were simulated for interim and final analyses. Data maturity was assessed at both stages based on the variance of Kaplan-Meier estimate and underlying survival distribution. At each time point, existing data were also used to estimate expected enrolment dynamics, time-to-event distribution, and time to censoring.

RESULTS: From the literature review, conceptual maturity has been further refined along four themes: maturity, stability, validity, and quality. This work provided definitions and tools for measurement to be included in RWE protocols. For the mock study, we showed at each time point how data maturity depended not only on the underlying survival distribution and the current estimate, but also on patient enrolment dynamics and the pattern of censoring. Statistical simulation was repeated 200 times to account for the margin of error of survival estimates as well as trajectory plots to accompany cautious interpretation and decision making.

CONCLUSIONS: We showed the cruciality of evaluating data maturity at various milestones in RWE settings and brought to light useful methodology that can be easily applied. This work highlights the current regulatory gap regarding data maturity in observational studies.