In CPR, fast, regular chest compressions move blood around the body of a person whose heart has stopped beating, carrying oxygen to the tissues to keep them alive. For cardiopulmonary resuscitation to work, you have to push down hard on the patient's heart—the chest should go down 1 1/2 to 2 inches, 100 times a minute. It doesn't sound easy, and it's not. Studies on mannequins have shown that a few months after passing the CPR test, people already don't do it as well as they did when they were certified. For two studies published this week in the Journal of the American Medical Association, researchers looked at how well CPR actually gets executed in and out of the hospital.

What the researchers wanted to know: How well do real people in the real world perform CPR?

What they did: One study was carried out in a municipality near Oslo, in Stockholm, and in London. Six ambulances in each region were outfitted with special defibrillators that were attached to a chest pad that measured compressions. Paramedics or, on some Swedish ambulances, a nurse anesthetist, carried out CPR on the patients. The researchers got usable data on 163 CPR episodes. The other study looked at CPR in hospitals at the University of Chicago. They also used a defibrillator that recorded data on chest compressions, and got information on 67 patients.

What they found: In the study on ambulance crews, the crew members did chest compressions for only about half the time that they could have been doing them. That meant patients went for long periods with no blood flow. The researchers say that other activities, such as using the defibrillator and checking for a pulse, accounted for only 15 to 20 percent of the missing time. Also, more than half of the chest compressions were too shallow, making them less effective at moving blood around the body. In the in-hospital study, chest compressions were too shallow 37 percent of the time. The chest compressions often happened too slowly, too.

What the study means to you: Cardiopulmonary resuscitation often isn't performed as well as it should be. And these are people who do CPR all the time—think how well the rest of us would probably do. In an accompanying editorial, two doctors argue that it's time to stop blaming people for not learning to do CPR properly, and time for fixing the way they're trained; instead of a long list of complicated guidelines, for example, they suggest emphasizing the message "push hard and push fast on the center of the chest without interruption, defibrillate promptly, and don't provide too many rescue breaths per minute." (Other studies have found that people give too many breaths when they perform CPR.)

Caveats: Neither study looked at how the sub-par CPR performance affected a patient's chance of survival, but other studies on animals have shown that better CPR means a better chance of recovery from cardiac arrest.

Find out more: The American Heart Association wants you to get certified in CPR

Read the articles: Wik, L., et al. "Quality of Cardiopulmonary Resuscitation During Out-of-Hospital Cardiac Arrest." Journal of the American Medical Association. Jan. 19, 2005, Vol. 293, No. 3, pp. 299-304.Abella, B.S., et al. "Quality of Cardiopulmonary Resuscitation During In-Hospital Cardiac Arrest." Journal of the American Medical Association. Jan. 19, 2005, Vol. 293, No. 3, pp. 305-310.

Abstracts online: Out-of-hospital: http://jama.ama-assn.org/cgi/content/abstract/293/3/299

In hospital: http://jama.ama-assn.org/cgi/content/abstract/293/3/305