Cost-effectiveness thresholds in funding decisions: a CART analysis

What is it about?

What determines funding decisions in health care? Is it the type of intervention (e.g. lifestyle or medical) being considered, or perhaps the objective of the intervention (e.g. prevention versus treatment) that matters to decision-makers? What about the cost-effectiveness of the treatment? Is there one common cost-effectiveness threshold applied across all interventions? This new study tackles these questions using a ‘big data’ technique, classification and regression trees (CART). It found that silos exist in funding decisions, where cost-effectiveness thresholds vary dramatically depending on the attributes of the interventions being assessed.

Featured Image

Why is it important?

Two things stand out from this research: 1) In textbook health economics, we compare the cost-effectiveness of an intervention to a given threshold - in Australia we typically use $50,000/QALY. But what we find in this research is that in reality the threshold is dependent on the intervention, with lifestyle and prevention interventions facing a much lower threshold than traditional medical treatments. 2) Methods like CART have advantages over traditional linear regression when analysing decision rules that are likely to be non-linear. We suggest there is much scope for the use of CART in health economics.