Editorial
The following editorial by Albany Med President and CEO James J. Barba appeared in the Nov. 27, 2005 edition of the Albany Times Union:
Public Must See Methods Behind Hospitals' Grades
In the last half dozen years, socalled health care "report cards" have proliferated. There are now, literally, dozens of these reports, each affording consumers the "opportunity" to select the health care that is optimum for them, merely by following the recommendations of the latest such report. Indeed, a recent story in the Times Union titled "Taking pulse on region's hospitals" (11/12) provides a discussion of one such report card-HealthGrades. But before digesting its recommendations whole, there are a number of factors of which the public should be aware.
The discussion should begin with a simple question: Does a report card, like HealthGrades, offer valid information that should be used by someone in need of health care services to select a provider of those services? The answer is "No," and I can support that position.
The HealthGrades report card is seriously flawed. That fact is established by several of the health care experts who were interviewed for the Times Union article. They cite the need for "caution" or point to certain "flaws" in the HealthGrades approach. Here are some significant problems that various experts have identified with this particular report card:
Fact: The data utilized by HealthGrades is not information that details the medical conditions of patients. In fact, no patient medical records are consulted by HealthGrades to produce the data that it uses. Instead, HealthGrades uses administrative data that is compiled to produce a patient's bill. Can such financial data be an accurate barometer of health care quality? Can it be an accurate barometer of anything, except the bill? Should a layman use this type of information to make a healthcare decision?
Fact: As the article notes, HealthGrades compiles its data primarily on Medicare patients?generally those age 65 or older. For most conditions, it excludes from consideration the majority of patients, who are privately insured, Medicaid insured, or uninsured. Again, if a layman were looking for accurate data, would he or she, exclude the largest segment of the U.S. population, and still think that the sampling was representative? Was reliable?
Fact: The HealthGrades report card does a poor job of tracking preexisting conditions that often can have a profound effect upon a patient's outcome, according to a recent study by New York State. The absence of such information is, by itself, sufficient to call into question the validity of the entire survey.
Fact: HealthGrades is a report card based on complications and mortality. Albany Medical Center is the primary referral hospital for all other hospitals in a vast area of upstate. When another hospital cannot deal with the level of serious illnesses or injury of a patient in its facility, Albany Med gets the call. The lives of most of those patients are saved. However, some do die. And, when they do, and when HealthGrades does not accurately adjust for severity, which is often the case, that counts as a failure of Albany Med. If the Times Union, or anyone else wants to see our HealthGrades "scores" improve, all we need to do is to stop accepting the toughest cases referred. Of course, that is something that we shall never do.
Fact: HealthGrades uses what it calls a "risk adjustment formula" that it believes corrects the deficiencies that I have mentioned, so that the comparison between hospitals is "apples to apples." The formula does not accomplish this. An article appearing in 2002 in the Journal of the American Medical Association assessed the formula, and concluded that HealthGrades approach provides "little meaningful discrimination between individual hospitals' performance." In fact, an official for HealthGrades is cited in the Times Union article as acknowledging issues with its risk adjustment formula, and that the formula has "been getting tweaked for the past eight years." Eight years is a long time to spend trying to get something right, especially when the inaccuracy is being used to predict something as important as outcomes from an individual hospital's treatment of stroke, heart disease, and the like.
Thus, should a consumer or a business use "findings" (1) from nonmedical data; (2) that do not reflect a majority of the population; (3) that do not account for preexisting or secondary conditions; (4) that focus primarily on complications and deaths; and (5) that are adjusted by a formula that has been neither peer reviewed nor thoroughly evaluated, to decide where to go for health care? I expect that the question is rhetorical.
Experts agree that health report cards are an important tool that, when correctly formulated, can challenge providers to understand the strengths and weaknesses of their processes of delivering care. But no report card now available provides definitive and infallible information. Selecting where we go for care, and whom we see, is as complex a decision as any we make. Indeed, it can be a life or death decision. That decision is best made with a critical eye toward the needs of the individual patient and with the counsel of the patient's family and physician. That is not the opinion of Albany Med because we had a low score in some categories; it is not our opinion despite the fact that we had the highest possible score for stroke care. It is our position because it is right.
Before any community accepts the findings of any specific report card, it needs to understand, and rigorously evaluate, the methods and the limitations of that report. Said simply, evaluating "quality" needs to start with the quality of the report card itself. And, assessed by these metrics, the quality grade that I give the HealthGrades report card is "poor."
