By Uwe E. Reinhardt
Uwe E. Reinhardt is an economist at Princeton. For previous posts in his series on why America pays so much for health care, click here, here, here, here and here.
After a lengthy discourse on health policy with a physician, I asked him to describe the ideal health system from a physician’s perspective. “Everyone in society should have access to needed health care,” he responded. “Only the physician and the patient should decide how to respond to a given medical condition. And someone should reimburse providers of health care at reasonable rates.”
Leave aside the fact that providers of health care are “paid” for their services, not “reimbursed,” the latter method always breeding bad managerial habits. Leave aside also that the definition of “needed” health care is highly elastic. Focus instead on the core of the ideal: complete clinical autonomy for physicians and their patients to throw whatever resources they wish at given medical conditions, usually at someone else’s expense, with little or no accountability for their preferred treatments’ quality or success.
Most physicians, and I would suspect most of their patients, probably subscribe to this ideal. Alas, a mounting body of research leads experts to doubt that physician autonomy actually serves society well and that it will be affordable much longer.
One strand of this research is typified by a landmark study conducted by Elizabeth A. McGlynn and her research associates at the Rand Corporation. The researchers found that, on average, American patients receive the recommended treatment for their condition only slightly more than 50 percent of the time. There are numerous other studies of this genre.
Another strand of research can be found in the famous and equally widely quoted Dartmouth Atlas of Health Care published by John E. Wennberg and his research associates at the Dartmouth Medical School. These researchers have for two decades now alerted Congress, the medical profession and private insurers to huge geographic variations in per-capita health spending that do not seem to be related to commensurate differences in disease rates, the quality of health care processes, clinical outcomes or patient satisfaction.
Table 6.1 below, taken from the Final Report of the New Jersey Commission on Rationalizing Health Care Resources, illustrates these inexplicable geographic variations in per-capita spending. As the chair of that commission, I had asked Dr. Wennberg to develop for the commission data on Medicare spending per beneficiary in New Jersey’s various hospital market areas.
Shown in the table are the average payments for inpatient care that Medicare made per beneficiary residing in a selected number of hospital-market areas during these beneficiaries’ last two years of life. The averages have been statistically adjusted to account for differences in the gender and age of patients who come to different hospitals and also for differences in the fees Medicare pays hospitals. (Although in principle Medicare uses one uniform fee schedule nationwide, some regional adjustments are made to fees to account for local differences in labor and other practice costs). It follows that the data in the table reflect differences in the use of real resources (hospital days, procedures per hospital day, etc.) and not price differences.
The numbers in the chart are all being compared to how much Medicare reimburses hospitals nationwide, on average, for taking care of a patient in the last two years of his or her life. In technical terms, this means that if a number in the table is equal to 1, then the actual spending per beneficiary, or the actual use of hospital days, etc., in the corresponding New Jersey market area is exactly equal to the national average.
For example, the number 3.21 for the St. Michaels Medical Center hospital market area in the table means that, during their last two years of life, Medicare beneficiaries in this area were 3.21 more costly to the taxpayer times for inpatient care than were similar patients nationwide. These patients received 2.34 times more hospital inpatient days than the national average, and each hospital day they received cost 1.37 times as much as the comparable national average. By contrast, for patients in the Atlantic Medical Center market area, the statistics are very close to the national average, because the numbers for that area are all close to 1.
The insight one gains from the table is that Medicare spending per beneficiary varied by a factor of three across New Jersey, depending on the practice style preferred by the physicians caring for these patients. As the right-most column in the table suggest, there does not appear to be any systematic relationship between Medicare spending and Medicare’s hospital-specific quality index.
Hospital executives, confronted with these numbers, have explained that they have little control over the utilization of health-care resources within their hospitals, as that is the prerogative of the private, usually self-employed physicians who have privileges at these hospitals and can look upon the latter as their free workshops. Furthermore, these executives would be loath to reproach the physicians responsible for using the most Medicare dollars, lest those physicians shift their patients to rival hospitals. Most health-care experts accept these explanations.
Dr. Wennberg and his associates at Dartmouth Medical have published similar data for California and for the nation as a whole. In their most recent publication, they show that these enormous variations in per-capita health spending are confined not only to the traditional, government-run, fee-for-service part of Medicare, but are apparent also for Medicare beneficiaries served by the private “Medicare Advantage” plans as well as among the commercially insured patients served by these private plans. They are a feature of all medical practice in America, regardless of who pays for the care.
The wonder is that neither Congress, nor the medical profession, nor private health insurers has so far felt obligated to come to grips with these enormous variations in health spending nor even to evince any curiosity about them.
According to the Dartmouth researchers, if physicians with relatively higher cost preferred practice styles could be induced to embrace the preferred practice styles of their equally effective but lower-cost colleagues, overall per-capita Medicare spending probably could be reduced by at least 30 percent without harming patients, and similarly for commercially insured younger Americans. How can a nation that routinely wails over its high cost of health care ignore such important research?
In a future post to this blog, I shall explore what might possibly be done to address this issue. In the meantime, it would be fascinating to learn from their comments what readers would do, were they American health-care czars.
No comments:
Post a Comment