Course description
Dates for 2012: 5 - 6 July 2012
Overview
Health economic evaluations often make use of observational data. A major concern is that cost-effectiveness results may be subject to selection bias. While sophisticated methods for addressing selection bias are routinely used in other literatures, their use in economic evaluation is limited.
This course offers an in depth description of methods for addressing treatment selection bias. These methods include regression, propensity score matching and a novel approach that extends propensity score matching, Genetic Matching. The course will highlight underlying assumptions and the pros and cons of each method.
There is a strong emphasis on applying the methods in practice, with practical sessions illustrating how to implement each technique using readily available software (STATA and R).
The course is aimed at health economists, or statisticians with an interest in health economic evaluation. It is envisaged that participants will be interested in undertaking or interpreting cost-effectiveness analyses that use observational data.
Prerequisites
This is an advanced course focusing on statistical methods for economic evaluation. Participants would be expected to have some familiarity with STATA, and relevant statistical concepts such as OLS regression.
Course Objectives
By the end of the course participants will be able to:
- critique the use of observational data in economic evaluation
- estimate regression models for addressing selection bias in economic evaluation
- estimate propensity score models, and use them for matching
- use a computationally intensive matching method, Genetic Matching, that matches on individual confounders
- assess the plausibility of the major assumptions behind each method
- Implement matching methods in both STATA and R.
Teaching Methods and Course Materials
The course consists of lectures and computer practical sessions. Comprehensive course notes will be provided at the start of the course. Computer practicals will use STATA and the R package, which is freely available with documentation. The course is limited to 15 participants.
Course Fee
£675 (commercial sector)
£350 (public sector)