Emulating a Novel Clinical Trial Using Existing Observational Data. Predicting Results of the PreVent Study

Area of Science:

• Medical Research Methodology
• Observational Studies
• Clinical Trials

Background:

• Target trial emulation is a proposed observational method for comparative effectiveness questions.
• Its concurrent application with randomized clinical trials (RCTs) is rare.
• This study evaluated the predictive accuracy of target trial emulation against an RCT.

Purpose of the Study:

• To test if blinded analysts using target trial emulation on observational data could predict RCT outcomes.
• To assess the agreement between observational analysis and RCT results for a specific intervention.

Main Methods:

• Emulation of the PreVent RCT's eligibility criteria, randomization, and analysis using existing patient-level data.
• Blinded analysts performed the emulation unaware of the RCT's results.
• Comparison of difference-in-differences estimates between the observational analysis and the RCT.

Main Results:

• Observational data emulation yielded balanced groups, similar to the RCT.
• Both methods indicated higher lowest oxygen saturation with positive-pressure ventilation.
• Rates of severe hypoxemia were reduced by positive-pressure ventilation in both analyses, with similar absolute risk reductions.

Conclusions:

• Target trial emulation applied to observational data produced results comparable to a concurrent RCT.
• This supports the utility of target trial emulation for comparative effectiveness research.
• The method demonstrates potential for evaluating novel interventions using real-world data.