A Hospital Report With Holes

Yesterday's Web posting by the federal government of information that shows how well individual hospitals deal with heart attacks and heart failure was a big deal—in principle. It was the first dip of the federal toe into a pool of statistics called outcomes data. It means what it sounds like—what happens to people. Whether they live or die.

Previously the federal Centers for Medicare and Medicaid Services, which is running a continuing project called Hospital Compare to give the public a peek at how well hospitals are doing their jobs, had only collected "process measures"—information that shows whether a hospital does the right things, like give aspirin to heart-attack patients as soon as possible. But outcome is what counts.

So why is only one hand clapping? Because very few people will learn anything about a hospital they look up. Out of 4,453 hospitals, 17 had death rates for heart attack patients that were better than the national average, and seven—seven!—were rated as below average. The performances of the other 4,429 were labeled "no different than U.S. national rate." In other words, what's posted on the Web shows that 99.9 percent of all hospitals were equal. For heart failure patients, 4,734 hospitals provided data; 38 did better and 35 did worse than the national rate. That's better. Only 98.5 percent were in the "no different than" category.

You can't even look at numbers and reach your own conclusion. The boxes show only a check in one of the three boxes. I picked a dozen hospitals more or less at random within a hundred miles of our offices, some huge and some small, some urban and some suburban or rural, and looked them up. Every one had a check in the "no different than" box.

How is this useful?

What is useful appears on a different part of the CMS site, and was the subject of a more restrained announcement from CMS yesterday that it had been updated. It has spreadsheets that display, for almost every hospital in the country, how many times the hospital performed each of various important elective procedures in the most recent year. Want to see how many heart valve or hernia operations a hospital did during the year? Data are absent for many hospitals, but the page is worth consulting. The number of procedures could be disconcertingly small—I ran across one facility with "heart hospital" in its name that reported doing only 28 valve procedures for the entire year. Volume is a critical clue to quality—the place to have a complicated operation performed is a place that does a lot of them.