A Review of the New Patient Satisfaction Tool

I've spent an hour or so browsing the overhauled Hospital Compare page (which went live today at 3 p.m.) and its patient satisfaction data.

• You may be frustrated if you search by hospital rather than by city or by a particular condition or procedure. I found that the name of the hospital has to be typed exactly right, including even hyphens, or nothing will come up.

• Be careful when interpreting the patient satisfaction numbers. It's tempting to think that a 65 percent rating in a category is obviously better than a rating of 61 or 62 percent. In fact, a difference between two hospitals of less than 6 percentage points is unlikely to have statistical validity.

• For both the patient satisfaction and the clinical measures, "view graphs" as your choice of how to display the results will convey the message directly and emphatically. Brightly colored bars show at a glance how a hospital compares with the average performance of all reporting hospitals in the United States and with reporting hospitals in the same state. Two more hospitals can be added for comparative purposes.

• The completeness of the data varies from city to city. Every hospital within 10 miles of the U.S. News offices in Washington, D.C., reported and made its data accessible to the public, so all can be compared. The same is true for every hospital within a 10-mile radius of Baltimore and San Francisco. New York City is another story; 24 of 55 hospitals either didn't submit data or chose to hide it. What's up with New York?