search

giabaio / BCEA

Bayesian Cost Effectiveness Analysis. Given the results of a Bayesian model (possibly based on MCMC) in the form of simulations from the posterior distributions of suitable variables of costs and clinical benefits for two or more interventions, produces a health economic evaluation. Compares one of the interventions (the "reference") to the others ("comparators"). Produces many summary and plots to analyse the results
https://gianluca.statistica.it/software/bcea/
GNU General Public License v3.0
40 stars 16 forks source link

Does it have to be bayesian? #19

Open seabbs opened 4 years ago

seabbs commented 4 years ago

My understanding of the analysis is that the output needs to be based on some kind of simulation that gives a distribution of outcomes.

Does this need to be a true Bayesian posterior or could it be a prior distribution or sample arrived at using other means?

Obviously, it would be better if everyone went bayesian but might it be a good idea to point out that this is optional (if it is)?

giabaio commented 4 years ago

Well --- BCEA is agnostic in that respect. It does have a strong prior that things should be done in a Bayesian way, but as long as you do have simulation-based output for the matrices of "effects" e and costs c (however defined --- and BCEA doesn't check that these are correct. Nor it could --- that's the modellers' job!), then you can run all its functionalities.

Essentially, BCEA is a software for post-processing of the outcome of the economic modelling and, as such, it can afford to be agnostic about the method (while silently judging non-Bayesian users)... ;-)

I do think this has been pointed out in several places (including the presentations, books and even in the help functions, I think).

seabbs commented 4 years ago

Awesome that is just as I thought.

Do you think there is any space for making that clearer in the README/package description? Some people that I have spoken to have dismissed the package after just looking at those.

Very much agree on the silent judgment aspect but perhaps easing people into the tool so that they are using it makes them more likely to switch (as they will see all the case studies + book using Bayesian methods).

Talking to some HE last week it seemed like using Excel was still quite common. My thought was that easing them into using something better (like BCEA) could be a good gateway to then improving the actual modeling.

Again happy to have a stab at this if you are okay with that?

giabaio commented 4 years ago

We're doing a big push to move people away from MS Excel and into proper stats software (see for example here) and I too was at a big health economics conference last week harassing people to that effect... ;-)

I'm happy for you to have a go at modifying the README file and if you feel there are other parts (eg help files etc) that need updating, please go ahead and send me a push request! Thanks!

seabbs commented 4 years ago

Awesome - will do :smile:

Sounds like a great idea - I was pretty shocked to hear about all the excel use honestly.

Could you open this up so I can keep track of what I am having a crack at? I have some time blocked out tomorrow for this.

giabaio commented 4 years ago

Just done.