[Revisit to observation and experiment: real world study cannot replace randomized controlled trial].

Area of Science:

• Epidemiology
• Clinical Research Methodology

Background:

• Intervention studies involve human activities altering natural conditions.
• The term 'intervention' broadly includes beneficial or harmful actions, past or present, by researchers, subjects, or third parties.
• Distinguishing experiments from observational studies solely by intervention is insufficient.

Purpose of the Study:

• To clarify the definition of experimental studies in epidemiology and clinical research.
• To differentiate between experimental and observational study designs.
• To emphasize the irreplaceable role of randomized controlled trials.

Main Methods:

• Conceptual analysis of 'intervention' in epidemiological and experimental studies.
• Examination of study design features, particularly randomization.
• Comparison of randomized controlled trials with observational and real-world data studies.

Main Results:

• Randomization is the defining characteristic that elevates randomized controlled trials above controlled observational studies.
• Non-randomized intervention studies are considered trials but not experiments.
• Real-world data and big data studies, lacking randomization, are observational and cannot substitute for randomized trials in efficacy confirmation.

Conclusions:

• Randomized controlled trials are the gold standard for experimental research due to randomization.
• Observational studies, including real-world data, have limitations in establishing causality compared to randomized trials.
• The paper strongly advocates for randomized controlled trials and cautions against over-reliance on non-randomized or real-world data for efficacy confirmation.