The data also are available via a link posted on Hospital Compare, the CMS' quality site for consumers at hospitalcompare.hhs.gov.
Dr. Michael Rapp, director of the CMS' quality measurement and health assessment group, said the data set contains important safety information and that making it public would drive quality improvements at hospitals.
But providers expressed concern about the data's accuracy and usefulness to consumers, particularly because the CMS used administrative claims data, which many experts say is tailored for determining payment and not for measuring clinical quality.
“I think the literature is pretty complete showing administrative data is not a reasonable mechanism to use to select one hospital over another,” said Dr. Gregg Meyer, senior vice president for quality and patient safety at Massachusetts General Hospital, Boston. “What is striking to me is the number of hospitals that are listed as having zero HACs. For a hospital that has 10,000 discharges not to have a single catheter-associated urinary-tract infection or an instance of poor glycemic control, that's just not credible.”
Meyer did acknowledge the importance of lowering the rates of HACs, and he agreed with Rapp that the publication of the data would likely spur improvement efforts, particularly at hospitals that are just beginning to undertake them. But he cautioned that there is “real danger” in patients using the data to compare and select hospitals for their care.
“I would argue that some of the institutions that have a relatively high number of HACs are safer by virtue of the fact that they are more open and transparent, they have a better reporting system in place and they are doing something about these conditions,” Meyer added.
The CMS originally had planned to publish the data last September, but the agency twice postponed the release of the data when issues with their accuracy came to light during the preview-and-comment period. Specifically, Rapp said, there was a discrepancy in the way the present-on-admission indicator was carried through in claims data. The CMS took its time correcting the glitch, Rapp said, adding he was confident it had been addressed.
The changing release dates seems to have caught some hospitals off guard. Officials at four-hospital Methodist Health System, Dallas, were expecting the data to be released later this month, said Dr. Adam Myers, the system's chief medical officer.
The CMS had initially said it would post the file on its website March 31 and then on Hospital Compare on April 21. But it pushed that second date up when it added a link to the data set on the Hospital Compare site April 6.
Myers also expressed uneasiness with the agency's use of coding data. According to the HAC data set, 247-bed Methodist Charlton Medical Center, Dallas, one of the system's hospitals, had one of the highest rates of vascular catheter-associated infections.
“Financially oriented coding data does not always reflect clinical reality,” Myers said. “I think the emphasis needs to be on providing accountable, understandable information, and I'm not sure that this data is doing that. It might even be misleading.”
The American Hospital Association strongly opposed the posting of the HAC data because of the use of claims data and because the HAC measures were not endorsed by the National Quality Forum. Additionally, AHA officials argued there were problems with the methodology used to calculate rates of each condition.
For instance, in determining blood incompatibility, the CMS used as the denominator the same number of eligible discharges—18,737,512—that it used for nearly all of the other conditions, said Lisa Grabert, the AHA's senior associate director of policy.
“We know that's not right because you can only be at risk of receiving the wrong blood if you have a blood transfusion,” Grabert said. “That denominator should be much smaller than it would be for another condition like pressure ulcers.”
Nancy Foster, the AHA's vice president of quality and patient safety policy, said the association has been a longtime supporter of posting data to Hospital Compare, and it believes the public deserves trustworthy measures.
“But the HAC data does not rise to that level for us,” she said.
This is just the first stage of the HAC data's release, said the CMS' Rapp. Patients and providers can expect to see the information presented in a graphical format on Hospital Compare by January 2012, he said.
“All of the data will be in a much more usable form when we have it graphically displayed,” Rapp said. “This was just an effort to get the data out in advance.”
