Assessing the use of observational methods and real-world data to emulate ongoing randomized controlled trials

Area of Science:

• Health Services Research
• Clinical Trials Methodology
• Real-World Evidence

Background:

• Growing interest in observational research methods for medical product evaluation.
• Previous emulation of two randomized controlled trials (RCTs) using linked claims and electronic health records.
• Need to compare emulated trial populations and results with published RCTs.

Purpose of the Study:

• Compare populations and results of emulated trials with published PRONOUNCE and GRADE RCTs.
• Assess feasibility of using real-world data for baseline participant characteristics.
• Evaluate concordance of baseline characteristics and primary/secondary endpoints.

Main Methods:

• Emulated two RCTs (PRONOUNCE, GRADE) using OptumLabs Data Warehouse structured data.
• Assessed ascertainability of baseline characteristics (ascertainable, partially, not).
• Compared emulated vs. published RCTs for baseline characteristics (SMD < 0.20) and endpoint results (direction, significance).

Main Results:

• PRONOUNCE: 9/31 (29%) baseline characteristics ascertainable; 6/9 (66.7%) discordant. Primary/secondary endpoints concordant.
• GRADE: 8/34 (23.5%) baseline characteristics ascertainable; 4/8 (50%) discordant. Primary endpoint mostly concordant.
• Emulated trials varied in participant numbers compared to published RCTs.

Conclusions:

• Real-world data and observational methods show potential for timely drug evaluation.
• Challenges in data ascertainment exist but do not preclude utility.
• These methods can evaluate diverse and representative patient populations.