What method is commonly used to handle skewed cost data in economic evaluations?

• A. Grouping identical costs
• B. Bootstrapping
• C. Linear regression without transformation
• D. Mean imputation

Answer: (B)

Bootstrapping handles skewed cost data by not assuming any specific distribution for costs. It repeatedly resamples the observed costs with replacement to build many simulated samples and recalculates the statistic of interest (such as mean cost or incremental cost) for each resample. This creates an empirical distribution that reflects the actual skewness and outliers in the data, allowing you to build confidence intervals and quantify uncertainty without relying on normality. In economic evaluations, this means you can compare interventions and derive uncertainty around mean costs and cost differences (or ICERs) directly from the data.

Other approaches don’t address the skewness and uncertainty as effectively. Grouping identical costs loses information and doesn’t fix the underlying distribution. Linear regression without transformation relies on normally distributed residuals and constant variance, which skewed costs violate. Mean imputation can bias results and understate variability, making uncertainty harder to interpret.