Tuesday 28 May 2013

Cost effectiveness acceptability curve (CEAC) vs Cost effectiveness acceptability frontier (CEAF)

What is the difference between cost effectiveness acceptability curve (CEAC) and cost effectiveness acceptability frontier (CEAF)?

What is CEAC?
Shortly, CEAC shows you the probability that the option is cost effective if you are willing to pay λ dollars for a QALY.
It represents the uncertainty associated with cost-effectiveness of a particular option at different values of a QALY. If there are several competing treatments, CEAC can be presented for each of them. But, usually it is presented for the option with the highest net benefit. CEAC is what you usually see in a cost-effectiveness study.
CEAC is used only to represent uncertainty, not to identify the best alternative. The best alternative is the one with highest net benefit (unless it has very bad CEAC).

How CEAC is constructed?
Lets say you have a model that counts resource utilization and outcomes of two treatments. If you input means costs and outcomes, it gives you mean incremental cost-effectiveness ration (ICER). But, resource utilization and outcomes in healthcare is not certain. Thus, instead of feeding mean numbers, you allow the model to randomly draw a number for each parameter given its distribution. Repeat 10,000 times, get 10,000 costs, outcomes, and ICERs.
Now, for each λ dollars (x-axis), calculate how many of those 10,000 ICERs are less than λ. That's the probability of being cost-effective.

What is CEAF?
CEAF shows the probability that the option with the highest net benefit is cost effective if you are willing to pay λ dollars for a QALY. Basically, you identify the option with the highest net benefit using simple net benefit calculation for each λ and plot CEAC for that option only at that point of λ. Plotting it for each λ gives you CEAF. CEAF only shows uncertainty associated with the best alternative.

Note, that for each λ there might be different option with highest net benefit. For instance, what is less that λ<$20k option 1, and for λ>$20k option 2 may have higher net benefit. So, CEAF would consist of parts of those two CEACs.

Why do you need CEAF?
To calculate expected value of perfect information (EVPI). And yeah,, NICE requires CEAFs to be submitted in economic-evaluation studies.