Overview and implementation of multi criteria decision analysis (MCDA) for benefit-risk assessment
Praveen Thokala
University of Sheffield
Overview/Description of presentation:
Health care decisions are complex and involve confronting trade-offs between multiple, often conflicting, objectives. Multiple criteria decision analysis (MCDA) are a set of techniques that use structured, explicit approaches to decisions involving multiple criteria to improve the quality of decision making.
Value measurement MCDA models are by far the most widely used in health care and although there are many differences in the ways in which these models are used and applied, there are several main elements of the process that are common among these methods. Broadly speaking, any value measurement modelling approach entails defining the decision problem being addressed, selecting the criteria, measuring alternatives’ performance, scoring alternatives and weighting criteria, aggregation, uncertainty analysis, and interpretation of results.
An example of how value measurement MCDA approaches can help benefit risk analysis (BRA) will be illustrated. The key element of preference elicitation i.e. the weighting of the different criteria and scoring of the different alternatives will be highlighted.
There is a need to consider different areas of uncertainty when using MCDA to support BRA. An important consideration is the appropriate choice of endpoints (i.e. criteria), which can be addressed using structural sensitivity analysis. Another source of uncertainty is the importance attached to criteria and performance (i.e. weights and scores captured using preference elicitation), which can be addressed using subgroup analysis or probabilistic sensitivity analysis (PSA). There is also the uncertainty that results from lack of information (for example, lack of trial data), where expert elicitation and Bayesian approaches can be used to support PSA in MCDA for BRA.
Conclusions/Take home message:
MCDA approaches are ideally suited to deal with multiple endpoints featured in benefit risk analysis. They can support transparency by making the criteria and preferences explicit. There are techniques available, such as Bayesian analyses, which can help capture the uncertainty.
