Norfolk hospital's Medicare death rates higher than average

NORFOLK

Medicare patients with pneumonia or heart failure at Sentara Leigh Hospital were likely to die at higher rates than national averages, according to recent federal data.

Out of more than 4,600 hospitals evaluated, 40 listed mortality rates worse than the national average for at least two of the three conditions measured: pneumonia, heart failure and heart attack. In Virginia, that group included Sentara Leigh, Danville Regional Medical Center and Memorial Hospital of Martinsville and Henry County.

Sentara Bayside Hospital also showed a death rate for pneumonia higher than the national average. The Virginia Beach facility converted into an outpatient center last month and stopped offering inpatient care.

No other Hampton Roads hospital deviated from the national average on any of the measures.

Dr. Ashish Jha, an associate professor at the Harvard School of Public Health who studies hospital quality, said the rates are an important measure of a hospital's care.

"We know from lots of data that which hospital you go to has a pretty profound impact on whether you live or die from your condition," he said.

However, the American Hospital Association, the national organization that advocates for hospitals, maintains that the measure is flawed because it doesn't correct for patients who choose to end aggressive treatment for terminal illnesses.

"That kind of decision is more often made if the patients are older, frailer, have many complicating medical conditions," said Nancy Foster, the association's vice president for quality and patient safety policy. "For those hospitals that have a very close relationship with nursing homes or older, frailer patients in the community, they're more likely to have a higher rate of mortality."

Sentara Healthcare officials said that is the case with Leigh Hospital, which serves nine nursing homes.

The hospital also has a robust "palliative" care program that focuses on relieving pain and sometimes serves patients who decide to end life-prolonging therapies.

"The population that Leigh serves is very different than a typical hospital," said Terrie Edwards, Sentara Leigh's president.

However, officials from the federal Centers for Medicare and Medicaid Services said its rates correct for a patient population like the one at Leigh, writing in an email that hospitals "treating older, sicker patients can be fairly compared to hospitals with a healthier case mix."

In recent years, Medicare started calculating individual hospitals' mortality rates for pneumonia, heart attack or heart failure. The agency looks at traditional Medicare recipients who died within 30 days of being admitted. Deaths from any cause are included.

The agency chose those three conditions because they're common, they often cause suffering or death, and good treatments exist.

A hospital's care, along with what happens during the early transition to an outpatient setting, can strongly influence how a patient fares in the first 30 days after hospitalization, according to QualityNet, a website established by Medicare to provide information about health care quality.

In fall 2013, Medicare will begin tying monetary incentives and penalties to hospitals' 30-day mortality rates for the three conditions.

Jha said the federal government chose a conservative formula for calculating the rates, leaving no doubt about the outliers.

"If a hospital has outcomes that are worse than average, it really suggests to me there's a problem and that the hospital has to focus on getting better," he said.

Less than 5 percent of the hospitals evaluated were rated worse than the U.S. average for the pneumonia mortality rate, and 2.5 percent were listed as worse than average for heart failure deaths.

The rates are calculated using records for patients 65 and older who were enrolled in traditional Medicare the entire year before their hospital admission. The national average and the hospital's past performance were combined in a complex statistical procedure, which corrected for how sick patients were when they arrived at a hospital and other factors, including age and whether the patient has a history of dementia.

Using the numerical calculations, Medicare rated each hospital as worse than average, average or better than average, for each condition.

However, the hospital association's Foster said the rate calculations don't consider enough details about patients' health that would give more indication of how sick they are upon admission.

"What the measure doesn't really capture is the real clinical sense of how ill the patient is, how close to death that patient might be based on the spread of illness within any particular diagnosis," she said.

Sentara Leigh's heart attack mortality rate was no different from the national average, but the hospital showed a heart failure death rate higher than the U.S. average for the second consecutive annual report. The two reports contain data from overlapping three-year periods, so they include some of the same information. The most recent report covered July 2007 to June 2010.

Sentara officials said Medicare's numbers were surprising because the health system tracks inpatient deaths within their hospitals, and that Leigh's ratio typically was better than average.

A large proportion of the hospital's patients come from nursing homes, including two facilities that serve patients on ventilators, a particularly fragile population, said Dr. Scott A. Miller, Sentara Leigh's vice president of medical affairs.

About 30 percent of the heart failure deaths and 50 percent of the pneumonia deaths were patients admitted from nursing homes, he said. Nearly 70 percent of the heart failure deaths were patients who had chosen comfort care or palliative care, Miller said.

The rates in the Medicare report exclude patients who have enrolled in hospice care - end-of-life care that no longer includes treatment meant to cure - up to the second day of their hospital stay.

It doesn't account for patients who choose comfort care or palliative care because that choice "is not necessarily an indication that a patient is no longer seeking life-sustaining measures," Medicare officials wrote in an email. "It is increasingly used in patients who are not at the end of life."

What's more, a patient might transition to a palliative service because the hospital's care was inadequate, they wrote: "Excluding these patients could mask quality problems."

Sentara officials said they reviewed Bayside hospital's pneumonia deaths, but discontinued monitoring the data for that facility after it stopped providing inpatient care.

About 500 of Bayside's employees moved to the newly opened Sentara Princess Anne Hospital, also in Virginia Beach.

Bayside's high mortality rate could have been related to the administration of the correct antibiotics at the correct times, said Dr. Gene Burke, Sentara's vice president of clinical effectiveness: "We had opportunities around the issues of the antibiotics, and we have been working on that."

Leaders at both Leigh and Bayside hospitals - as well as the other Sentara hospitals - regularly review cases and meet with nurses and physicians to be sure patients received appropriate care, including medications for pneumonia and heart failure, Burke said.

At Leigh, "when a patient doesn't get what we think was the preferred antibiotic, we actually investigate it," Miller said.

The Sentara system has created three teams to find ways to improve efficiency and effectiveness of its heart failure care, and Leigh is one of the sites where the work is happening, Burke said.

For example, several physicians are looking into the best, most consistent approach for administering medications for heart failure, he said.

"Every hospital around the country can do better than it's doing today, no matter what your starting point is," Burke said. "We are part of that."

Amy Jeter, (757) 446-2730, amy.jeter@pilotonline.com