Study of CPR quality reveals frequent deviation from guidelines

January 18, 2005

New technology has allowed researchers from the University of Chicago to measure, for the first time, how closely well-trained hospital staff comply with established guidelines for cardio-pulmonary resuscitation (CPR).  The results reveal room for improvement.

In the 19 January 2005 issue of JAMA, the researchers show that, even in the hospital setting, chest compressions during CPR are often too slow, too shallow and too frequently interrupted, and ventilation rates are usually too high.  A second study assessing out-of-hospital CPR by paramedics and nurse anesthetists in three European cities found even greater deviation from the guidelines, suggesting that the problem is endemic.

"CPR has been around for 50 years but until now we haven't had a precise, reliable way to assess how well it's being done," said study author Lance Becker, M.D., professor of emergency medicine and director of the Emergency Resuscitation Research Center at the University of Chicago. "Now we find that it's not being done very well."

The two JAMA studies "document a major problem in the treatment of cardiac arrest," notes an editorial that accompanies the papers, adding, "this conclusion is not surprising."

"You can't fix what you can't measure," Becker said. "Performing CPR was like driving a car without a speedometer, based more on feel than on feedback.  Now, with a device that tells us how fast we are going, we think we can rein in the speeders and speed up those who fall behind."

The key to this study was an investigational monitor/defibrillator, developed by Laerdal Medical Corporation and Philips Medical Systems. The device records the rate and depth of chest compressions, the rate and volume of ventilations, and the presence or absence of a pulse.  It also notes when no compressions are being performed and calculates total "no-flow" time, as well as the fraction of time during a cardiac arrest when there is no blood flow.

In this study, the cardiac arrest response team at the University of Chicago Hospitals used the device to measure the quality of CPR during the first five minutes of each resuscitation attempt on 67 patients who suffered a cardiac arrest at the Hospitals between Dec. 11, 2002, and April 5, 2004. The researchers then compared the results, broken down into 30-second segments, to guidelines developed by the American Heart Association.

The guidelines recommend 100 chest compressions per minute to a depth of at least 38 millimeters (about 1.5 inches).  They call for a ventilation rate of 12 to 16 breaths a minute, and they advise keeping the no-flow fraction -- the fraction of time in cardiac arrest without chest compressions -- under 0.17 (about 10 seconds out of every minute.)

"We did not expect to find perfect compliance," said Benjamin Abella, M.D., assistant professor of medicine and lead author of the study.  "A cardiac arrest is an emergency and the response -- even in the hospital setting -- is a hectic, tumultuous, chaotic event.  To perform well in such a high-pressure situation," he said, "is difficult at best."

And absolute compliance was not what they found.  In 28 percent of the cases, chest compression rates fell below 90 per minute.  Thirty-seven percent of the chest compressions were too shallow.  Ventilation rates were usually too high; in 61 percent of the 30-second segments ventilations rates were over 20 per minute.  In 40 percent of cases the no-flow fraction rose above 0.20.

Multi-site studies by the same team, although preliminary, have found similar rates in other teaching hospitals, community hospitals and among paramedics in the field. The European study found even greater deviation.

The device that uncovered the problem, however, may help solve it.  A follow-up study is already under way at the University of Chicago, in which a second-generation device not only measures compliance but also provides immediate feedback, prompting the resuscitation team how to adjust compression speed or depth and ventilation rate.

"This immediate feedback," Becker said, "may help us improve the quality of CPR and potentially increase survival rates."

"Unfortunately, we cannot yet say how much of a difference that will make," he added.  "The official guidelines define what we consider optimal, but this was the first effort to study compliance with those guidelines in a real-life setting and no one has ever demonstrated a difference caused by deviation from the guidelines."

This pilot study was too small to show a detectable difference in results between optimal and flawed CPR.  Survival rates were just as good for patients receiving imperfect CPR as for those where performance exactly matched the guidelines.  Overall, 40 percent of the patients in this study regained spontaneous circulation after CPR and 10.4 percent were discharged from the hospital.

"We do know," Becker insisted, "that CPR makes a difference and it only makes sense that better CPR would make more of a difference."

CPR is a crucial component of what emergency physicians call the "chain of survival" for patients who have a cardiac arrest.  CPR "buys you time," explained Abella.  "It protects your heart and brain until we can get your heart pumping again," which usually requires shocking it back into a normal rhythm.

The study authors have used their data to launch an effort to increase the quality of CPR delivered at the hospital, combining additional training for CPR teams with widespread use of the measurement device, which provides text and spoken prompts for the user on how to improve performance during the process, pointing out any deviation from the guidelines.

"Learning how to measure quickly and accurately showed us that we had a problem," said Becker, "but it also provides us with at least part of the solution, a real-time critique that can catch and correct common mistakes before they do any harm.  This should help tighten up one link of the chain."

The consistency of poor CPR performance also suggests the need to revise the guidelines, Arthur Sanders and Gordon Ewy of the University of Arizona College of Medicine argue in their editorial. The guidelines "are too complex," they wrote, and need to be simplified "so that all patients who sustain cardiac arrest can receive optimal treatment."

Additional authors of the paper include Jason Alvarado, Dana Edelson, Anne Barry, Nicholas O'Hearn and Terry Vanden Hoek from the University of Chicago and Helge Myklebust from Laerdal Medical Corp., Stavanger, Norway.  The study was funded by a grant from Laerdal, but the company had no role in data collection, interpretation of results or drafting of the manuscript.