The descriptions you give match a number of studies I’ve read and already evaluated. E.g. dozens of papers investigating various aspects of the Oregon randomized Medicaid trial, with substantially varying conclusions in this area.
This is just the summary given in The Elephant In the Brain, I haven’t read the original papers and I’m sure that you know more about this than me. Here’s what TEITB says about the Oregon Medicaid trial (screenshotted from my Kindle version):
If you think this misrepresents what we should take away from this study, I’m keen to hear it!
It’s mixed. As far as it goes for the original study, it’s mostly accurate but I do think that the use of the phrase “akin to a placebo effect” is misleading and the study itself did not conclude anything of the kind. There may be later re-analyses that do draw such a conclusion, though.
Most objective health outcomes of medical treatment were not measured, and many of those that were measured were diagnostic of chronic conditions that medical treatment cannot modify, but only provide treatment that reduces their impact on daily life. There are objective measures of outcomes of such treatment, but they require more effort to measure and are more specific to the medical conditions being treated. This is relevant in that a large fraction of medical expenditure is in exactly this sort of management of conditions to improve functionality and quality of life without curing or substantially modifying the underlying disease.
It should also be borne in mind that the groups in this study were largely healthy, relatively young adults. The vast majority of health service expenditure goes to people who are very sick and mostly older than 65. It seems unwise to generalize conclusions about overall effectiveness of health expenditure from samples of much healthier younger adults.
Would you characterize the Oregon Medicaid study as poorly designed, or perhaps set up to make Medicaid look bad? From your description, it sounds like they chose a population and set of health metrics that were predictably going to show no effect, even though there was probably an effect to be found.
Do you have links?
The descriptions you give match a number of studies I’ve read and already evaluated. E.g. dozens of papers investigating various aspects of the Oregon randomized Medicaid trial, with substantially varying conclusions in this area.
This is just the summary given in The Elephant In the Brain, I haven’t read the original papers and I’m sure that you know more about this than me. Here’s what TEITB says about the Oregon Medicaid trial (screenshotted from my Kindle version):
If you think this misrepresents what we should take away from this study, I’m keen to hear it!
It’s mixed. As far as it goes for the original study, it’s mostly accurate but I do think that the use of the phrase “akin to a placebo effect” is misleading and the study itself did not conclude anything of the kind. There may be later re-analyses that do draw such a conclusion, though.
Most objective health outcomes of medical treatment were not measured, and many of those that were measured were diagnostic of chronic conditions that medical treatment cannot modify, but only provide treatment that reduces their impact on daily life. There are objective measures of outcomes of such treatment, but they require more effort to measure and are more specific to the medical conditions being treated. This is relevant in that a large fraction of medical expenditure is in exactly this sort of management of conditions to improve functionality and quality of life without curing or substantially modifying the underlying disease.
It should also be borne in mind that the groups in this study were largely healthy, relatively young adults. The vast majority of health service expenditure goes to people who are very sick and mostly older than 65. It seems unwise to generalize conclusions about overall effectiveness of health expenditure from samples of much healthier younger adults.
That’s helpful information, thanks.
Would you characterize the Oregon Medicaid study as poorly designed, or perhaps set up to make Medicaid look bad? From your description, it sounds like they chose a population and set of health metrics that were predictably going to show no effect, even though there was probably an effect to be found.