A few days ago I read Atul Gawande’s article on health care costs and thought his story was interesting enough that I wanted to know the statistics on what factors predict high or low costs.
Commenter Marc pointed me to this recent article by Elliott Fisher, Julie Bynum, and Jonathan Skinner on regional variation in heath costs across the United States.
Commenter Ao pointed to this Congressional Budget Office report which, to me, was a bit disappointing. It had some nice maps and charts but did not seem nearly as serious as Gawande’s article in trying to understand what was going on.
Finally, Alan Zaslavsky, a statistician who specializes in healh-care economics (and who uses multilevel models) wrote:
Atul Gawande’s article in the New Yorker is an excellent review of some of the issues we have been struggling with in health policy research. While a lot of energy has gone into looking at the impact of various incentive schemes (public reporting of quality measures, “pay for performance”) on quality of healthcare, it has been difficult to address the kinds of issues of organizational culture described in the article. The differences in provider structure and culture that are key to success or failure in providing high-quality, efficient care are not that readily brought into analysis — the variables are just not measured and available. So instead we have very convincing case studies — the McAllen market at one extreme and the integrated Kaiser, Geisinger etc systems at the other.
One problem is that many of the analyses take place at the level of the health (insurance) plan, but health plans in most cases are not in the business of providing care, they are in the business of buying it. (Even some of the original staff-model HMOs went through the transition to being insurance companies, like HIP in New York.) The variables that are routinely available for analysis at the health plan level are very crude proxies for the underlying organizational structures and cultures. For example, some economists have told me that they are perplexed by findings that not-for-profit plans provide better-quality care than for-profits, since both types of firms should be subject to similar incentives. What this leaves out is the distinct histories of some of the leading not-for-profits, and consequent differences in organizational cultures.
I do highly recommend the work of the Dartmouth group (including Jon Skinner and Elliott Fisher, mentioned in the article) on area variations to interested readers. However, it does suffer from some of the same limitations — the variations can be found and clearly show that more is not always better, but it is hard to say what actually drives the differences or what can be done to implant the cultural and organizational features that would make more areas look like the best areas. [emphasis added]
While I [Zaslavsky] am a proponent of a universal system of health care, I don’t think that will be enough to solve our problems without some fundamental changes in the incentives and structures under which care providers operate. Paradoxically, “rationalization” has meant squeezing out not the most wasteful aspects of care but some of the unprofitable but essential services that could make care more effective.
Perhaps more could be done to take the quantitative analyses of Fisher et al. and Zaslavsky and his colleagues, and see what is needed to move toward useful recommendations. This is all on top of the difficult political issues; for example, doctors in the U.S. get paid a lot and I don’t think they’d be happy about getting less money.
I think the McAllens of this world are just statistical variation. While it is true that more doesn't necessarily mean better, "more" isn't necessarily bad when there is a distribution about a mean.
I think if serious efficiencies were realised in healthcare doctors could be paid the same or more. For example, many tests ordered every day are either definitely or possibly non-diagnostic. The cost of this is far more than just the doctor's time, so if it stopped, there would be no reason that you couldn't maintain or increase the doctor's take-home in line with the saving. (If you think paying a lot is a good idea/are worried about disincentives… I tend to think they'll get by either way…)
There was a big issue with Gawande's article and it surprised me that nobody noticed. He used Medicare costs as a stand-in for health care costs generally. Everybody does this because we have huge amounts of data about Medicare costs that we don't have on the system overall. There are some issues with this (the Medicare population is older than the general population obviously), but there's a special issue in McAllen, Texas. It's a huge destination for snowbirds and retirees. The Medicare-eligible population is substantively different from the general population in the area.
My theory is that Medicare costs per patient go up dramatically when a critical mass of Medicare-eligible people live in a geographic area. When there is enough demand, facilities are built that make more tests available to Medicare patients. Look at the places with high Medicare costs per patient. They have have high utilization rates of expensive tests. They are also places with high concentrations of the elderly (big metro areas and retirement communities).
Unnecessary tests are equally motivated by the need to manage risk as the desire to increase one's income. If a doctor or hospital gets sued and they haven't done a particular test that could have caught the bad thing that was the impetus for the suit, it can mean the difference between no liability and huge liability in court.
If an analysis is going to be useful, its scope has to be broader than a focus on incentive schemes and organizational culture. Then again, trial lawyers in the U.S. get paid a lot of money and I don't think they'd be happy about getting less money.
Its hard to underestimate the magagerial/political challenges Zaslavsky pointed to "without some fundamental changes in the incentives and structures under which care providers operate"
In a way its all very "one sided" – I want my physician to be my advocate getting me the best treatment without worrying about the cost to others (or I'll find another physician), physicians want to make their patients healtheist while making the best incomes, improved health care usually means doing more things better than doing less of anything, etc.
and incentives often backfire …
but better data analyses can't hurt
The easiest way I can think of to analyze organizational structure is to take the NPI type 2 file (for organizational health care entities) and develop a count of distinct type 2 NPI entities per capita in a given geographic area.
This would leave out a lot of facility locations (not all are listed in NPPES), but would give you a rough estimate of the fragmentation of health care business entities that exists in any area. (different entities enumerated in different ways, so it would require some de-duplication of identical business entities with multiple NPIs)
So I'm not surprised that the statistical results of looking at a large number of these experiments are disappointing. The generator of these experiments is broken.
I have yet to observer a managerial attempt to impose a scheme to improve cost or quality improvement using incentives and feedback that wasn't tainted by combination of lack of knowledge about how the system in question actually functions plus a large dose of disdain for the people involved.
Sometimes this happens because there are always short-term gains to be had at the cost of a long-term damage. Those instituting these are usually on very short tenures in a given role and have bonuses tied to short-term metrics. But, in a surprising number of cases those setting up the system take pride in both; their lack of knowledge and their adversarial relationship. This is a legacy of both labor/management dialectic of the last century and the way that social was bled out of the social sciences of economics and managerial training.
Unnecessary tests are also done because one's health record is not available when you are admitted in an acute situation. Without an electronic medical record many tests need to be done during the acute admit that could be avoided simply if the information on your primary care physician's chart were available.
On a separate point, there is a good paper by Bertsimas et al, "Algorithmic Prediction of Health Care Costs" in Operations Research, vol 56, #6, Nov-Dec 2008 pp 1382-1392.