Enhancing knowledge of treatment effects by analysing observational data as a randomised trial

Area of Science:

• Epidemiology
• Biostatistics
• Causal Inference

Background:

• Randomized controlled trials (RCTs) are the gold standard for establishing causal treatment effects.
• Ethical and practical constraints often limit the feasibility of RCTs for certain research questions.
• Observational data presents an alternative for investigating causal relationships when RCTs are not viable.

Purpose of the Study:

• To introduce and explain the framework of target trial emulation.
• To demonstrate how target trial emulation can address confounding and treatment complexities in observational studies.
• To highlight the utility of causal per-protocol analysis within this framework.

Main Methods:

• Target trial emulation involves designing observational studies to mimic RCTs.
• Methods focus on minimizing bias from baseline confounding and time-varying confounders.
• Causal per-protocol analysis is employed for unbiased treatment effect estimation.

Main Results:

• Target trial emulation provides a structured approach to causal inference from observational data.
• The framework facilitates the estimation of treatment effects comparable to RCTs.
• A worked example on hip fracture rehabilitation illustrates the application and benefits.

Conclusions:

• Target trial emulation is a valuable tool for answering causal questions using observational data.
• This approach is particularly important when RCTs are not feasible.
• Investigators are encouraged to adopt target trial emulation for robust causal inference.