By Mary Jo Feldstein, St. Louis Post-Dispatch, June 22, 2007

Patients treated for a heart attack at Barnes-Jewish Hospital were more likely to survive than the national average, according to data released Thursday by the Hospital Quality Alliance.

Barnes was one of only 17 hospitals in the country and the only in Missouri to have heart attack mortality rates below the national average. The data also looked at mortality rates of patients with heart failure. Barnes'' heart failure mortality rate was considered within the national average.

All other area hospitals ranked within the national average for both measures.

The study didn''t look at why a hospital did better or worse than others.

The alliance is a public-private partnership of the Centers for Medicare and Medicaid Services, hospitals, clinicians, quality experts and other groups. CMS is the federal agency responsible for administering Medicare, Medicaid and other health-related programs.

The vast majority of hospitals nationwide had very similar scores. Others treated few patients, which could affect the scores.

Facilities considered better or worse than the national average had to score high or low enough that researchers felt the outcomes were statistically significant and would occur again if other patients'' records were pulled from the same hospital.

Climbing such a high statistical hurdle means that some hospitals rated average likely should have been in the worse than average category and some hospitals in the middle likely deserved to be ranked better than average, Herb Kuhn, acting deputy administrator for CMS, said during a conference call.

CMS said it hoped to add other measures and refine the current measures to provide more differentiation between hospitals.

"We''re going to get nothing but better at this," said Health and Human Services Secretary Mike Leavitt, who also participated in the call. "What you''re seeing today is a glimpse of what''s possible."

In part, the agency released these broad, conservative results because of a flub the last time it measured mortality rates 20 years ago. The statistical reliability of that research was criticized.

"We have now the overwhelming majority of hospitals saying we''re fine with this," said Gerry Shea, assistant to the president for governmental affairs, for the AFL-CIO, one of the organizations participating in the alliance.

Best practices

In this report, researchers used Medicare claims to determine what percentage of patients survived 30 days after they were admitted to more than 4,500 hospitals. Thirty-day mortality is used because this is the period when deaths are most likely to be related to the care patients received.

The data were taken from admissions for heart attack and heart failure patients between July 2005 and June 2006. It was adjusted to account for patients'' ages as well as those with conditions such as diabetes or high blood pressure.

The study didn''t look at why a hospital did better or worse than others.

This is the first time since launching new quality improvement measures that CMS has released data on patients'' outcomes. Previous research only looked at whether hospitals followed best practices.

When those measures were first released, Barnes failed to meet state or national averages in several areas. The hospital has since improved in those measures as well. It credits changes in process and more focus on documentation for the better scores.

Dr. Richard Bach, a Barnes cardiologist, said there is some correlation between process or performance measures — whether hospitals consistently follow best practices - and outcome measures - whether patients survive. Examples of best practices include giving patients an aspirin when they arrive at the hospital and making sure an angioplasty or stent procedure is begun within two hours of arrival.

"The performance measures, as limited as they might be, reflect a broader attention to detail and concern for detail," said Bach, who directs the cardiac intensive care unit at Barnes.

This study is part of a broader effort by government health programs and insurers to determine whether the patients they''re paying for are getting good quality care. Many are requiring physicians to document that they''ve followed best practices. Looking at whether patients survive has been more controversial. That''s because adjusting for other factors such as previous health problems requires a more sophisticated analysis.

Quality care

There are a couple of reasons for the emphasis on quality. Research shows patients receive recommended care only about half the time. Not only do patients who get the wrong care suffer with worse outcomes, their care is often more expensive.

"We want to make certain that everyone in the health care system has incentive to drive quality up and costs down," Leavitt said.

Doctors and hospitals, however, have hesitated. They often question the validity of the results. Some believe the measures, particularly those tied to their fees, are designed to pay them less not improve quality.

Bach said he was satisfied with the study''s ability to adjust for patients who are sicker or would be more likely to suffer from complications. But he said it might be difficult to use the same methods to compare mortality rates of other conditions.

"This is a condition where the science is relatively mature," Bach said.

Billing data

Some critics of the data will likely question whether problems with documentation or coding affected the results. For example, a medical billing clerk could enter incorrect information or a nurse might forget to write down that she gave a patient an aspirin.

Researchers used claims or billing data for the study. That information is easier to collect and there''s more of it, but some question whether it is as accurate as clinical data generated for quality review.

"The reason that sometimes the core measures don''t match the care we provide to patients may have more to do with data collection and documentation," Bach said.

Leavitt said he expects "a rapid increase" in the number of measures available for comparison and how they are displayed over the next two to three years. He expects cost information will be added and patients will be able to compare physicians as well as hospitals.

"Ten years from now, this system will be ubiquitous," Leavitt said. "It will just be the way that it works."

In the immediate future, the information will be used most frequently by health care professionals and insurers, but Leavitt hopes it will eventually be in a format patients can easily access and understand.

He joked that someone having a heart attack should go to the hospital, not their computer.

"This is a very important first step, but I think we''re all emphasizing it''s a first step," Leavitt said.