How We Selected And Reviewed Hospital Data For This Story

Published: Aug 26, 2007

At the request of The Tampa Tribune, three insurers allowed a reporter access to hospital cost and quality information they post on password-protected Web sites for their members. The companies were Humana, UnitedHealthcare, and Blue Cross and Blue Shield of Florida. One insurer declined our request.

All three sites allow searches to compare the outcomes for procedures, such as knee-replacement surgery, at hospitals in defined geographical regions. The software used to sift the data was developed by medical data analysis companies, which draw information for hospital patients from claim forms for all insurers, both public and private.

Each company has its own formulas for data sorting. They look at complication and mortality rates and other measures, adjusting for hospitals that have an unusually high number of sick, frail or otherwise risky patients.

The Tribune studied the sites to determine which provide the most comprehensive information for consumers on cost and quality of care. The Tribune's review found that the UnitedHealthcare and Blue Cross sites offered little specific information on the actual cost of procedures (what is paid by an insurer, rather than what is billed by a hospital).

As a result, the Tribune selected data on Humana's Web site, myHumana.com.

WebMD Quality Services rates hospitals on medical procedures with one, two or three stars, with three stars for top performers. Additional information on complications and other measures is available on the site.

To compare, a user selects how much importance to assign to geographical distance, patient volume, mortality rates and other factors. Or the user can use the settings that experts at WebMD recommend. After some experimentation, that is the process the Tribune selected.

We analyzed cost and outcome information for four common nonemergency procedures: hip replacement, natural childbirth, heart bypass surgery and heart catheterization.

We conducted our search by sorting for data on hospitals within several geographical boundaries, ranging from 10 miles to 50 miles from downtown Tampa. We did this to ensure we reviewed a broad, representative sampling of hospitals in the Bay area. One caveat: On WebMD, a hospital's rating depends, in part, on the hospitals with which it is compared.

The results are included in the charts published with this article.

For more information on the interactive tools and an online demonstration, go to apps.humana.com/egroups/misc/transdemo/index.htm#.

Carol Gentry