Where Does Your State Rank?
The State of Health Care Spending, by Andrew Rettenmaier and his colleagues:
The compendium is divided in two main parts. Part 1 summarizes the four ways by which the geography of health care spending is described. Health care spending as a percent of the states’ GDP is the first way in which the geography of health care is presented and is separated between Medicare, Medicaid, and non-Medicare/Medicaid spending. These data allow for analysis that extends back to 1980 for each state. Next, health care spending is analyzed on a per capita basis and is again divided between Medicare and Medicaid per enrollee in the programs, and average non-Medicare/Medicaid spending for the states’ population who are not enrolled in the programs. The second part of the compendium comprises 50 state summaries. The two-page summaries are based on the four ways of viewing geographic variation in health care spending and the health care markets. The first page summarizes the key health care spending indicators in each state, and provides graphical representations of how the state compares to the national average now and in the past. Also depicted is the variation in county level Medicare spending. The second page of each state’s summary presents all of the recent metrics in tabular form. Medicare spending in four large or geographically dispersed counties is also presented at the end of each table. (Summary by Michael Ramlet)
Tom Saving and I discuss some of the results at the Health Affairs blog.
Perfect analysis, not a single word about health statistics.
Spending by one payer often substitutes for another; spending on one provider often substitutes for another. Thus, hospital spending substitutes for physician spending in areas where physicians are poorly reimbursed by Medicaid. Medicare substitutes for Medicaid (especially among duel eligibles); and Medicare (and Medicaid) spending is sometimes higher in areas with high uninsured populations.
One thing about the graph I found most interesting is the upper Midwest, which is often touted in the Dartmouth Atlas reports as having lower Medicare spending per beneficiary than other (purportedly more wasteful) areas – such as Florida. Yet, the Upper Midwest areas spend far more in total care than other areas (except for costly Northeastern states). It’s likely that spending by private insurers on people who are not yet Medicare eligible substitutes for Medicare spending later (i.e. uninsured patients often put off seeking care until eligible for Medicare is they’re close to Medicare age).
“In both cases, commentators used these facts to infer that we could save an enormous amount of money if doctors in the high-spending areas practiced medicine the same way as doctors in the low-spending areas.”
Seriously? Ah, the New Yorker, still a titan distinguished publisher of fiction.
And… errors — this is what I get for posting before my first cup of coffee.
That is a perfect example of the Fallacy of Composition – that what will work in one place or at one time will automatically work elsewhere.
While many “fallacies” result in poor policies, the more important thing to notice is that the multidimensional model in empirical research is the only way to ensure better decisions for society. Of course everything will have fallacies but simply pointing them out is not helping anything or anyone.
Where are the statistics within the field on this?