Skip to main content

Mortality Rates at CAHs Ratcheting Higher

 |  By John Commins  
   April 03, 2013

The nation's critical access hospitals have higher mortality rates on several key measures than do urban and rural hospitals without the specia l designation, and the trend steadily worsened over the past eight years, according to a new study by Harvard researchers.

The study was published this week in the Journal of the American Medical Association.

Researchers from Harvard Medical School and the Harvard School of Public Health examined administrative data from nearly 10 million Medicare fee-for-service admissions for acute myocardial infarction, congestive heart failure, and pneumonia between 2002–2010.

The researchers compared 30-day mortality rates of CAHs and non-CAHs in urban and rural settings. They found that while CAHs and non-CAHs had comparable mortality rates for the three conditions in 2002, CAHs gradually fell behind each year to the point where by 2010 CAHs mortality rates were 13.3% and non-CAH mortality rates were 11.4%, a difference of 1.8%.

"What it is telling us is that we have left these hospitals behind because 1.8% of the absolute mortality rate is one in 50 people. That is more than 10% of the actual rate so in clinical trial speak it is a significant relevant risk," Karen E. Joynt, MD, lead researcher of the study, said in an interview.

"But what is important here is not the specific numbers but recognizing that the effort to carve these critical access hospitals into a separate program than the remainder of the hospitals in the country has not done them any favors and it hasn't done any favors for the people seeking care at these hospitals," says Joynt, an instructor at Harvard Medical School and the Harvard School of Public Health.  

When the Critical Access Hospital Program was created by Congress in 1997, qualifying CAHs could have no more than 25 beds and had to be located at least 35 miles from the nearest hospital. Since then states have been given leeway to broaden eligibility and now only 20% of the CAHs current meet the distance requirement and nearly one in four U.S. hospitals is designated as a CAH.

CAHs are exempted from prospective payments but are reimbursed at 101% of costs. They are also exempted from national quality improvement programs.

Joynt says that while the higher reimbursements have allowed scores of CAHs to keep their doors open, it may be time to reconsider their exemptions from reporting and quality improvement programs.

"Part of the program worked terrifically. The closures have dropped tremendously. There really are areas in which there would be no medical care if it weren't for this program. But leaving them on their own and saying 'good luck' has not been a good solution," she says.

"I understand that (the Centers for Medicare & Medicaid Services) was trying to relieve these hospitals of an administrative burden by not including them. But the consequence seems to be that no one knew these outcomes were not improving at many of these CAHs in the way that we were seeing mortality from many inpatient conditions drop over the last decade. That to me in this era of transparency and trying to build better systems and being more patient-centered doesn't make much sense as a long-term strategy."

The American Hospital Association took issue with the findings and said the "data presented now are not a reliable presentation of what is actually happening at critical access hospitals." 

"CAHs and other small hospitals have engaged in a variety of practices designed to identify and rapidly transfer patients who could benefit from more aggressive interventions at nearby hospitals that have the capability of intervening. One reason for seeing a rise in mortality in some CAHs could be because the hospital may tend to keep patients that are too sick for transfer or not stable enough. As a result CAHs' data maybe skewed toward patients who are sicker," AHA said in an email exchange.

The contentious study also prompted an editorial rebuttal in the same issue of JAMA from Stanford University researcher John P.A. Ioannidis, MD, who wrote that "even if the differences in CAH vs. non-CAH mortality rates are genuine, this does not mean that policy makers should necessarily advocate for CAHs to collect and report performance data or to participate in quality improvement programs, change their payment mechanisms, or both."

"The study by Joynt et al can minimally inform such decisions. Trying to impose quality data collection and reporting in such hospitals that have already strained resources may actually do more harm than good. Even for non-CAHs, the evidence is tenuous that performance and quality initiatives do work."

Gary Tiller, CEO of Ninnescah Valley Health Systems, Inc. in Kingman, KS, which operates the 25-bed Kingman Community Hospital, says the study "has renewed my lack of faith in the Harvard School of Public Health."

"This is all much ado about nothing. There are a lot of ways to look at the data depending upon what you want to do with it," Tiller says.

"They said we don't participate in quality reporting or quality initiatives. I don't know where they got that notion. We've done so much of it. We are reporting well over 100 data points now and the great bulk of the CAHs are doing that. They obviously never went out to a CAH."

"I don't need more paperwork. I've had to divert way too many staff hours to that kind of crap already. I am not really thrilled about having to do more of it."

Joynt concedes that the use of administrative data for clinical studies has limitations "but it doesn't make it useless."

"Administrative data can show us a pattern and then we need clinical and qualitative data to understand why," she says. "The wrong response to this paper would be to say there is obviously no problem. This is all a trick of administrative data. That is a little biased."

"It's not fake data. These are real people. They are dead. These are Medicare patients, someone's grandmother who died. It is true that administrative data will give you statistically significant results for very small differences because we have millions of people. That is a sample size issue. But the differences particularly for acute MI are clinically relevant. It is more than a one-third higher mortality over the last decade for heart attacks where the rest of the country has been improving care for heart attacks. That is not write-off-able," Joynt said.

"The data doesn't tell us the details about why and it doesn't tell us how we can fix it. It tells us maybe there is a problem here. Let's stop ignoring these hospitals and see if we can think creatively to help rural patients do better."

John Commins is a content specialist and online news editor for HealthLeaders, a Simplify Compliance brand.

Tagged Under:


Get the latest on healthcare leadership in your inbox.