Does the Senate Finance Committee version of the health-care bill threaten to cripple evidence-based medicine?

Harry Selker and Alastair Wood say yes.

P.S. The answer is no. The offending language is no longer in the bill (perhaps in response to Selker and Wood’s article).

P.P.S. Somebody checked again, and the offending language is still there!

4 thoughts on “Does the Senate Finance Committee version of the health-care bill threaten to cripple evidence-based medicine?

  1. I don't think there's any question about it, and last night's 60 Minutes story on end-of-life expenditures made the point quite well.

    http://www.cbsnews.com/stories/2009/11/19/60minut

    The money quote comes towards the end:

    "I think you cannot make these decisions on a case-by-case basis," Byock said. "It would be much easier for us to say 'We simply do not put defibrillators into people in this condition.' Meaning your age, your functional status, the ability to make full benefit of the defibrillator. Now that's going to outrage a lot of people."

    "But you think that should happen?" Kroft asked.

    "I think at some point it has to happen," Byock said.

    "Well, this is a version then of pulling Grandma off the machine?" Kroft asked.

    "You know, I have to say, I think that's offensive. I spend my life in the service of affirming life. I really do. To say we're gonna pull Grandma off the machine by not offering her liver transplant or her fourth cardiac bypass surgery or something is really just scurrilous. And it's certainly scurrilous when we have 46 million Americans who are uninsured," Byock said.

    Scurrilous it may be; offensive it may be; antirational it may be; but it is also correct.

  2. Scarry yes, but I do like Sander Greenland's arguments that researchers should just publish what they did in their studies, why and report clearly what happened – and leave it to others to make inferences using all the other relevant reported data and being in a personally less prone to bias situation to overvalue their own results.

    Also think it would be very hard to ever clearly assess what the evidence actually is in most cases – but any "lurking threat" seems like a really bad idea for evidence generation and dessimination.

    Keith

  3. As I wrote somewhere (I can't remember where), "a Bayesian wants everyone else to be non-Bayesian." (Then, said Bayesian can apply his own priors and analyze the data directly.)

  4. Andrew: Yes, you did, pointing out that other's priors are just nuisances at least if you can't separate off the likelihood from the posterior.

    And one of my biggest disappointments with Fisher was his suggestion that investigators just publish their (log) likelihoods (so others could simply add them up) – because then you can’t separate off the results from the particular probability model being assumed.

    Keith

Comments are closed.