Tuesday, April 14, 2009

Comparative effectiveness research-more warnings

Those, including several medical bloggers, who seem to have placed more than a little faith in the notion of a government funded and run comparative effectiveness research program (CER) should read the following sobering posting from Health Care Renewal.

The commentary in part expresses concerns about the following recent comments from Governor Sebelius from her congressional hearing for head of HHS.

“The goal,” Sebelius said, “is to provide every American with a safe, secure electronic health record by 2014." The nominee also endorsed efforts to use data gleaned from electronic medical records to conduct “comparative effectiveness research" (the authors bolding)(CER) to provide information on the relative strengths and weaknesses of alternative medical interventions to health providers and consumers.”

The author of the Health Care renewal entry is properly alarmed that use of the often garbage type data from EMR have a high likelihood of generating detailed printouts of elaborately presented ( replete with very low p values) results that could be misleading at best and more likely harmful. To expect such " highly uncontrolled" data bases to meaningfully determine if treatment X is better than treatment Y raises unrealistic expectations to a new level.

It seems to me that some pro government CER advocates make two mistakes in their advocacy of the government funded CER.

First, they maintain that we have no or at best very little comparative effectiveness data at all now so the government must provide it. Dr RW in this posting proves that is not the case. A recent example of non-government funded CER is the SYNTAX trial.

Secondly, they conflate the desire to have really good data on what works with the assumption that such will be the product from a government funded CER. Yes,it would be very nice if there were someway to have realms of great CER without bias and without spin but should we really trust a government agency who would somehow be populated by selfless, brilliant, dedicated public servants who would bring no biases to the table and who somehow will be immune to the outside influences that seemed to have plagued every governmental agency ever created and that the output from analysis would not be put to the purpose of limiting government payment for such treatment that they found inferior.
Limiting payment for medical services deemed inferior by data dredging exercises is not what we need to improve medical quality but may well be what government sponsored CER will bring.

1 comment:

refluxman said...

thank you....nice topic for long discussion and interesting