Estimating uncertainty ranges for costs by the bootstrap procedure combined with probabilistic sensitivity analysis

Area of Science:

• Health Economics
• Biostatistics
• Clinical Trial Analysis

Background:

• Economic evaluations often use patient-level data, leading to uncertainty from data variation and model assumptions.
• Existing methods like bootstrapping and sensitivity analysis address these uncertainties separately.
• A unified approach is needed to provide a total uncertainty range for economic variables.

Purpose of the Study:

• To present a combined method using bootstrapping and probabilistic sensitivity analysis.
• To estimate a total uncertainty range for incremental costs.
• To illustrate the method with cost data from a clinical trial on dyspeptic patients.

Main Methods:

• The study combines bootstrapping with probabilistic sensitivity analysis.
• This integrated approach is applied to cost data from a randomized controlled trial.
• The method calculates a combined uncertainty range for incremental costs.

Main Results:

• Endoscopy for H. pylori negative dyspeptic patients showed no clinical benefit and was more costly.
• The combined 95% uncertainty range for incremental costs was wider than ranges from individual methods.
• The combined range was -£236 to £931, compared to £43-£592 (PSA) and -£95-£667 (bootstrap).

Conclusions:

• The combined method provides a more complete uncertainty estimate for incremental costs.
• This approach can be extended to uncertainty ranges for incremental cost-effectiveness ratios.
• Accurate uncertainty quantification is crucial for robust health economic decision-making.