Related Experiment Videos
A comparison of multivariable regression models to analyse cost data.
Susanna Dodd1, Asish Bassi, Keith Bodger
1Centre for Medical Statistics and Health Evaluation, University of Liverpool, UK. s.r.dodd@liv.ac.uk
Journal of Evaluation in Clinical Practice
|January 21, 2006
Summary
Analyzing healthcare cost data is crucial for budgeting. Gamma modeling with a log link is the most suitable method for analyzing skewed cost distributions, outperforming other regression models.
Area of Science:
- Health Economics
- Biostatistics
- Medical Informatics
Background:
- Accurate cost data analysis is vital for effective healthcare budgeting and decision-making.
- Cost data often exhibits highly skewed distributions, complicating the estimation of arithmetic means and total costs.
- Multivariable regression analysis is essential for predicting future patient costs and controlling for confounding variables in observational studies.
Purpose of the Study:
- To compare the suitability of various multivariable regression models for analyzing skewed healthcare cost data.
- To identify the most appropriate statistical model for cost data analysis in the context of inflammatory bowel disease treatment.
Main Methods:
- Comparison of normal and bootstrapped multiple linear regression, median regression, and a gamma model with a log link.
- Evaluation of model appropriateness using regression diagnostics on inflammatory bowel disease treatment cost data.
Main Results:
- The gamma model with the log link demonstrated superior suitability for analyzing the cost data.
- Bootstrapping had a minimal impact on the conclusions drawn from the normal linear regression (NLR) model.
Conclusions:
- Gamma modeling with a log link is recommended for analyzing skewed healthcare cost data.
- The choice of regression model significantly impacts the reliability of cost estimations and budget predictions.