Propensity score matching for comparative studies: a tutorial with R and Rex

Area of Science:

• Medical research methodology
• Biostatistics
• Health technology assessment

Background:

• Advancements in medical technology, including minimally invasive surgery, necessitate robust evaluation methods.
• Randomized controlled trials are ideal for treatment evaluation but not always feasible.
• Observational studies are often used but are prone to selection bias from confounders.

Purpose of the Study:

• To demonstrate propensity score matching for bias reduction in two-group comparisons.
• To provide a practical example using R programming.
• To introduce Rex, an Excel add-in for users less familiar with R.

Main Methods:

• Propensity score matching was applied to a two-group scenario.
• R code was developed for the propensity score matching process.
• An Excel add-in (Rex) was created as a user-friendly alternative.

Main Results:

• The study successfully illustrated the application of propensity score matching.
• The R code and Rex tool facilitate bias reduction in observational data.
• The tutorial provides a foundation for applying propensity score methods.

Conclusions:

• Propensity score matching is a valuable technique for mitigating bias in observational studies.
• Accessible tools like R code and the Rex add-in can aid researchers in applying these methods.
• Further complex techniques like multi-group matching and imputation are summarized for advanced applications.