What It Takes to Be the Best
Just 173 medical centers made the cut. Here's how this select handful did it.
Each specialty had its own eligibility requirements. A hospital either had to have seen a specified minimum number of Medicare inpatients during 2003, 2004, and 2005 who had certain conditions or underwent certain procedures—803 patients with a particular set of respiratory disorders, for example—or the facility had to be named among the best in the specialty by at least one physician in the latest three annual surveys.
Hospitals that got through the second gate then received a score combining three equally weighted elements: reputation, death rate, and care-related factors such as nursing and advanced services. The rankings consist of the 50 highest-scoring hospitals.
Here is how the three elements break down. More detail is available in the glossary.
Reputation. For each of the 16 specialties, a randomized sample of 200 board-certified specialists was selected from the American Medical Association's Masterfile of more than 850,000 U.S. doctors, and those physicians were mailed a survey form. They were asked to list the five hospitals they think are best in their specialty for difficult cases, without taking location or expense into account (or naming their own hospital). The numbers in the "reputation" column are the combined percentages of responding physicians in the 2005, 2006, and 2007 surveys who listed the hospitals. Nearly half of the 3,200 doctors surveyed this year responded.
Mortality index. What is more important than a hospital's ability to keep patients alive? The number shown is a ratio—a comparison of the number of deaths of Medicare inpatients with certain conditions that occurred with the number that was expected (after adjusting for severity of condition) during 2003, 2004, and 2005. An index number below 1.00 therefore means the hospital did better than expected; above 1.00 means worse than expected. Deaths were included if they occurred within 30 days from the date of admission except in cancer, in which only deaths from admission to discharge were included. Severity adjustments were derived from 3M Health Information Systems software (All Patient Refined Diagnosis Related Group).
Other care-related factors. The remaining third of the score reflects quality indicators such as patient volume, relative availability of nurses, advanced technology, and credentialing by professional bodies. Much of the information came from the American Hospital Association's 2005 survey of its members. This year, some measures were reorganized and placed in different categories.
Reputational specialties. The reason for ranking ophthalmology, psychiatry, rehabilitation, and rheumatology only by reputation is that mortality data are irrelevant or unreliable in these specialties. Ranked hospitals were cited by at least 3 percent of responding physicians.
The 2007 rankings were produced by RTI International, a leading research organization based in Research Triangle Park, N.C.
advertisement
