System targets blood-type mix-ups

A patient in Virginia switched beds to be near the window, then died after the hospital gave her a blood transfusion with her roommate's blood type. A Florida woman died after receiving the wrong blood because a technician mislabeled her blood sample. In St. Louis, a man underwent successful heart surgery only to die in recovery after staff mistook him for someone else and grabbed the wrong blood bag.

Though such mix-ups get relatively little public attention, more than 850 US patients receive transfusions intended for someone else each year, according to projections based on a New York study, and at least 20 people die from the complications. Thousands more never know that only a last-second intervention at the bedside prevented a similar error.

Now, Massachusetts General Hospital, along with a few other institutions, hopes to drastically improve the safety of blood transfusions with electronic monitoring. Sometime in the next few days, patients in two operating rooms will begin wearing wristbands that communicate their blood type by tiny radio signals. If sensors in the bed detect a difference between the signals coming from the wristband and from a microchip on the blood bag, ''Stop! No Match!" will flash in big red letters on the anesthesiologist's computer screen.

''Healthcare workers don't come to work hoping to make a mistake, but systems can defeat good intentions," said Dr. Walter H. ''Sunny" Dzik, codirector of the blood transfusion service at Mass. General, where seven patients received mistaken transfusions from 1999 to 2003, though no one was hurt. The procedures for preventing transfusion mistakes, he said, have not changed much in the last 50 years.

The new monitoring technology, called radio frequency identification or RFID, is old hat for companies such as Gillette, which already attaches chips to boxes to track razors from the warehouse to consumers. But it is a leap for hospitals, which typically rely on medical staff to detect discrepancies by comparing the label on the bag with patient records.

Hospitals have been slow to adopt a 21st century transfusion system, say advocates of the technology, in part because they are reluctant to spend vast sums to prevent such rare events. Medical staff members also often resist safety checks that slow them down. RFID is promising because it works automatically, but the chips, at $1.50 apiece, could cost $150,000 a year at a busy hospital, and thousands more to install sensors.

Still, sensors in the operating room would not prevent other kinds of mistakes such as at Sarasota Memorial Hospital in Florida last June, when a technician applied another patient's preprinted label to a woman's blood sample, causing her to get blood of the wrong type. Such errors, known as WBITs for ''wrong blood in tube," occur in about one out of every 2,000 blood samples, according to a hospital survey, though the vast majority are detected before anyone is hurt.

But Dr. S. Gerald Sandler, director of transfusion medicine at Georgetown University Hospital in Washington, D.C., said the difficulty of preventing transfusion errors is no excuse for inaction.

''In the grocery store, we have . . . electronic devices to make sure when they charge me for orange juice, I'm not getting a bottle of milk. But we don't have in the hospital an analogous protective mechanism to save human life," Sandler said.

Georgetown is planning an experiment in which the maker of blood bags would implant an RFID chip in the label, allowing everyone -- from blood banks to people administering the transfusion -- to electronically match blood to the patient's type.

Since the 1970s, blood collection centers and hospitals have greatly reduced the risk that patients will contract diseases from blood transfusion, but they have paid scant attention to the danger of blood mix-ups. As a result, Dzik calculates that transfusion patients today are 100 to 1,000 times more likely to get the wrong blood than to get the virus that causes AIDS.

Unfortunately, many hospitals have learned how vulnerable patients are by painful experience. Inova Fairfax Hospital in Falls Church, Va., where the patient received the wrong blood in July 2003 after switching beds, now requires transfusion patients to wear a bracelet carrying a number that matches their blood records, and two clinicians must verify the patient's identity and blood type.

In St. Louis, Barnes-Jewish Hospital retrained its staff on proper patient identification and transfusion safeguards after the death of Richard Wood in August 2002 as a result of getting two units of the wrong blood type. Hospital officials said clinicians confused Wood with another patient, then didn't catch the mistake before the transfusion.

Only 35 deaths from transfusion mistakes were reported to the US Food and Drug Administration from 2001 to 2003, but that probably undercounts the total because some go unnoticed.

Sandler estimates the death toll at about 60 for the same period, and other analysts say the number may be higher. ''Most such errors actually go unreported [and undetected] because patients receiving blood are very ill and the complications of a blood mix-up may be difficult to recognize during a massive bleeding episode," Dr. Louis Katz, president of the nonprofit group America's Blood Centers, wrote to the FDA in 2003.

Blood supply officials also know the issue would be far more serious if not for flukes of human physiology. Type O blood, possessed by 43 percent of the population, can be safely transfused to almost everyone, even if by mistake. Type A blood also is so common that even when clinicians give patients the wrong blood, they sometimes get the type right by chance. At Mass. General, such luck averted harm six times despite transfusion errors from 1999 to 2003. The seventh patient suffered chills but recovered.

The key to eliminating errors, argue those in the overhaul effort, is to get away from relying on human attentiveness to catch mistakes. In a nation where 12 million units of red blood cells are transfused annually, even a relatively low error rate translates into thousands of patients at risk. It would be far better, they say, to rely on electronic tracking that is virtually error free.

At Georgetown University Hospital, outpatients who get transfusions already go through a process in which the nurse uses a hand-held scanner to read a barcode on the patient's wrist indicating blood type and compares it to the barcode on the blood bag. The system is cheap -- about 10 cents per wristband -- and effective: The rate of minor errors that can cause delays or require additional blood tests declined from 15 percent of all samples to nearly zero.

But Dzik and Sandler would like a system that works automatically, without the hassle of scanning barcodes in an emergency, which is why both are drawn to radio technology. Dzik delayed the start of Mass. General's experiment for months while technicians refined placement of RFID sensors in the bed and tweaked the system to get a fast reading no matter how the patient is positioned.

Dzik hopes to test the RFID system on 200 surgery patients over the next six months, then publish a report for other hospitals on how well the system works. ''I actually will feel that I failed in some way if it works, but it is only adopted at Mass. General," he said.

Scott Allen can be reached at allen@globe.com.