Thursday, May 05, 2005

Cost effectiveness of X, , how can a doctor possibly interpret what that means?

A typical cost-effectiveness article appears in the May 2, 2005 Annals of Internal Medicine. The summary states in part "alendronate ... is not cost effective."
When I was doing research on pulmonary function testing in diffuse lung disease, I could read an article , plow into the methods section and learn what technique of certain tests were done, what prediction equation was used, etc etc and could understand to a reasonable level of knowledge what it was all about.
Try that with the Annals article.
You will read that " We used a QALY value estimated with the EuroQul questionnaire and ...we derived the QALY value ...from direct prospective estimate of Kanis and colleagues" And " we constructed a Markow cost-utility model that contained 8 health states and compared 5 years of treatment with alendronate with no drug therapy..." and "We assumed relative risk for incident vertebral fractures of 0.54 and 0.82 for those with femoral neck T-scores of -2.0 t0 2.4 and -1.5 respectively..." And on and on the methods section goes through an elaborate mathematical exercise.
How could a hypothetical well informed general internist-the hypothetical ideal target of Annals articles- possible analyze what was done procedurally.How could she know if the assumptions are reasonable or biased towards some particular outcome?.How would he know if the Markow model was done correctly or more basically how valid is that approach or any of the particular details are in the first place?
You wonder who is the targeted audience of this type article. Is it an internist who would, armed with this latest research,inform his patient that they should not take X because it has been "shown to be" not cost effective. Is it aimed that a benefits manager who who would love some more quasi-justification to save money?
And even if you could understand- or give up and just accept- what they actually did, then you have to deal with the following statement."Assuming a social willingness to pay $ 50,000 per QALY gained...our results indicate that alendronate is not cost effective..."
I don't claim to be expert in Markow chain nuances but I do know that the authors or any one else cannot "determine" societal willingness " because that is not a thing you can determine.It is a meaningless abstraction.
Cost effectiveness articles should not be considered Evidence based medicine. They are calculations based on stacks of assumptions and typically end by concluding that something is or is not cost effective based on whether the indicator of interest is above or below an arbitrary dollar value that other authors have decided will be the threshold value.(The magic number,typically,is $50,000 per QALY,which is treated as a constant of nature rather than a contingent construct based on a stream of guesses) I ranted about this general topic few blogs ago but keep returning to it as articles appear and I see so little criticism of these pseudo-scientific papers.

No comments: