Monday, August 28, 2006

Houston V.A. physicians examine P4P data and find evidence of gaming and treating the chart

The possibility that health care providers (HCP) might try to game the system and treat the chart when offered or forced to play the P4P game is well recognized and quite consistent with basic human nature.

Researchers from the Houston Veteran's Administration Hospital have published an article indicating that behavior of that type does in fact occur. Dr. Laura S. Peterson and her associates write in the August 15 2006 issue of the Annals of Internal Medicine.

They reviewed 16 articles and in 4 studies they found evidence of unintended effects-adverse selection and gaming or treating the chart to achieve financially rewarded goals. Importantly, for those of us who believe this entire P4P movement is motivated by a desire to save money, they found no evidence that these efforts are cost effective. Of course, lack of evidence does not necessarily mean evidence that they are not cost effective but the third party payers and CMS may at some point pull back if they are continue to be unable to show they there are cost savings to the payers. I hope so.

3 comments:

Anonymous said...

Is the goal simply to be cost-effective, or is the goal to also be clinically effective? Is there any relationship between cost-effectiveness and clinical effectiveness?

Intuitively, one would think that clinical effectiveness would, over the long term, tend to decrease costs on average.

I think the trouble is, this is very hard to measure, especially if you are looking at clinical practices that might not pay off in ways that are immediate and/or obvious.

Anonymous said...

Intuitively, one would think that clinical effectiveness would, over the long term, tend to decrease costs on average.

Not necessarily. Say someone gets treated and lives a long life but takes a lot of drugs. Someone else doesn't get treated and dies. Clinically effective, but hardly saves money.
Or consider something like mammograms. You need to screen a thousand for 10 years to save one life; you can assume you may save on treatment of this one person and a couple of others (in some cases early diagnosis will not translate into savings). During the same 10 years over 50% of these 1000 women will have a recall because of false positive (cumultive, 10% after 1st mammogram, slightly less for subsequent one),around 10% will have biopsies (or something like it, could be 7% or 18, I don't remember). Moreover, because of overdiagnosis you will have more women diagnosed with breast cancer, so you end up treating more women at least with surgery and radiation. So the costs of screening and treating extra women will be way over savings. So clinically you get some lives saved, but at a cost.

Not sure about other preventive things like statins. What if someone still ends up having a heart attack but later? What about all of the extra people who would've never had a heart attack even with treatment.

Sid Schwab said...

gaming and treating the chart: no big surprise. P4P is a worthy concept at some level. Implementation of meaningful parameters seems literally impossible. Of necessity, it must involve relatively simple, black-and-white data points that say little about overcall care provided by a given doc.