July 2005
Feature Story
What do poker chips and blood transfusions have in common?
Both can be tracked using radio frequency identification tags, which
are sometimes called next-generation bar coding. An RFID reader emits
a radio signal that’s modified by an RFID chip exposed to that signal,
and it’s the change in the signal that the RFID reader picks up. Casinos
are using gambling chips embedded with RFID tags to keep tabs on players
who are beating the house odds and to root out counterfeit chips before
anyone cashes them in.
The U.S. blood banking community and hospitals are eyeing RFID technology
to tag everything from patient identification wristbands to blood containers
in an effort to stamp out the estimated one in 14,000 chance that a patient
will receive the wrong blood due to human error. The odds of dying from
such an error run much lower, at one in 600,000, "partly because erroneously
transfused blood may randomly match the ’wrong’ patient and partly because
of prompt and aggressive medical intervention," says S. Gerald Sandler,
MD, director of transfusion medicine at Georgetown University Hospital.
Dr. Sandler is spearheading research at Georgetown on the role of RFID
as a complement to bar coding to identify patients at each step in the
transfusion loop. He presented the study’s findings to date and plans
for future research at the recent Executive War College in New Orleans,
sponsored by The Dark Report.
The reported instances of someone receiving the wrong blood may be the
tip of the iceberg, some experts warn. "The FDA receives notices of deaths
caused by patients receiving the wrong blood at the rate of about 10 to
15 patients a year, but the belief is that many such errors go undiscovered,"
says Jim MacPherson, CEO of America’s Blood Centers. The British Serious
Hazards of Transfusion, or SHOT, study (www.shotuk.org)
found a much higher rate of deaths resulting from patients being transfused
with the wrong blood through human error, he adds. "And we don’t have
comparable data in the U.S. to know how much under-reporting of deaths
from blood transfusion reactions is occurring," MacPherson says. "For
example, a very ill person’s death may be attributed to his condition
rather than a transfusion reaction, or a patient in surgery may receive
multiple units of blood and no one sorts out the cause of his death."
And unlike errant laboratory values that clinicians may detect by viewing
them within the context of a patient’s symptoms, "a unit of blood that
will save a patient—or kill him—looks exactly the same," Dr.
Sandler says.
Yet blood transfusion errors caused by misidentification can be all but
eliminated by using bar coding or RFID technology, which is the focus
of the Georgetown University Hospital research. Georgetown’s study in
an outpatient oncology-hematology unit, which began in 2004, is comparing
the time required and pros and cons of three different ways of ensuring
the right patient receives the right blood product:
- A visual inspection method that requires two nurses to complete.
- A bar-code method whereby nurses use a personal digital assistant
to scan a patient’s bar-coded wristband and bar-code label on the blood
container.
- Use of RFID tags on patient wristbands and the blood container that
the nurses scan in lieu of the bar codes.
It was the findings of two previous studies at Georgetown that prompted
Dr. Sandler to test the use of RFID as a complement to bar coding. Those
studies showed that bar-coding systems worked well in an outpatient oncology-hematology
setting for transfusions—but ran into obstacles in the inpatient
arena that RFID technology might resolve.
One of the two Georgetown studies, conducted from 1998 to 2000, used
bar-coded wristbands, sample tube labels, and blood containers (and the
I-TRAC Plus system) for more than 1,000 outpatient transfusions. The study
resulted in 100 percent accuracy of labels for request forms and sample
tubes, eliminating the need for repeat submissions, which had been running
as high as 10 percent on some nursing units. For example, a label for
patient "James Smith" might read "Jimmy Smith," a small discrepancy but
one that doesn’t fly in the blood-banking arena, Dr. Sandler says. The
nurses also preferred using the bar code as a double check of the patient’s
identification before transfusing rather than having to find, and sometimes
wait for, a second nurse to perform a visual inspection.
Dr. Sandler and his colleagues next performed a study using the MedPoint-Transfusion
Bar Code system at each step of the transfusion process on an inpatient
oncology-hematology unit. The study procedure called for phlebotomists
and nurses on the research units to follow these steps:
- When the physician ordered a transfusion, the nurse printed a bar-coded
wristband that also had three eye-readable identifiers: patient’s name,
date of birth, and hospital identification number.
- The phlebotomist collecting a blood sample from the transfusion recipient
scanned his or her ID card with a handheld PDA to create an electronic
signature.
- The phlebotomist used the PDA to scan the patient’s bar-coded wristband.
- The PDA beamed the scanned information to a small, portable battery-operated
printer that printed the specified number of specimen labels. (If the
printer wouldn’t work, the phlebotomist was required to label the blood
specimen at the patient’s bedside before leaving the room to get a new
printer or battery.)
- The hospital unit sent the blood sample to the blood bank where a
technologist logged it into the blood bank information system by using
a small scanning device.
- The blood bank performed compatibility testing on the blood sample
using a blood typing analyzer that crossmatched the sample to a blood
product unit in inventory.
- The blood bank technologist placed a bar-coded crossmatch label on
the back of the unit of blood.
- Once the unit of blood component arrived at the patient’s bedside,
the nurse scanned his or her ID badge and the patient’s wristband.
- The nurse scanned the bar-coded crossmatch label on the back of the
unit and the Red Cross bar-coded whole blood number label on the front
of the unit. (By scanning both front and back of the blood bag, the
nurse could detect whether the blood bank technologist accidentally
put the crossmatch label on the wrong unit of blood product.)
- If the nurse had done everything correctly, the PDA screen would
say it was safe to transfuse the patient.
The inpatient study uncovered two unforeseen barriers to nurses accepting
the PDA and bar-code technology. For one, some of the older nurses on
the oncology-hematology inpatient unit questioned whether they could maintain
their competency using the PDA program because they administered relatively
few transfusions, unlike the nurses in the outpatient setting. Also, the
older nurses were not as intuitively in tune with the PDAs as the younger
outpatient nurses in the first study.
Another problem: While the PDA scanning device easily read the bar codes
on crisp, clean wristbands donned by outpatients, it was unable to do
the same for the wristbands worn by inpatients for several days, which
tended to get wrinkled and bent, smudged, or stained. They didn’t scan
well and in some cases not at all, Dr. Sandler said in his War College
presentation.
That unexpected clog in the inpatient study surprised Dr. Sandler and
others who initially assumed the PDA could beep its way through bar codes
at the same speed as a grocery store check-out scanner. But the research
team later discovered that grocery stores use an omni-direction laser
beamed at various angles through the surface to the counter. And bar codes
on cans and boxes are bold and clean, unlike those on wristbands worn
by hospital inpatients. "So you can hold the bar code on a box of cereal,
for example, at any angle or even upside down, and the scanning device
will read it," Dr. Sandler says.
By contrast, the handheld PDA used in the hospital study has a horizontal
laser beam 2.5 inches long that the health care staff has to align with
the bar codes, which can get tricky when scanning the occasional bent
or faded wristband.
Dr. Sandler and his team decided that piggybacking RFID technology for
blood transfusions on a bar-coding system for medication dispensing would
resolve both problems found in the study. (The research is purely investigational;
Georgetown University Hospital doesn’t have a bar-coded medication system.)
Dr. Sandler reasoned that if the nurses used the PDA program to dispense
medications several times a day, they’d maintain their competency in using
the program. In the envisioned system, blood products (platelets, plasma,
and red blood cells) would pop up on the PDA screen just like an antibiotic
or other medication, he says. Patients would wear a dual bar-code/RFID
wristband. A dual scanner would automatically read the RFID tag if the
bar code failed, which would save the nurse time and aggravation. Also,
unlike bar codes, RFID tags can be read through bedclothes without disturbing
a sleeping patient.
Using a bar-coded medication dispensing system as the platform for RFID
technology makes sense given that the FDA has promulgated a rule to move
pharmacies in the direction of bar coding to improve patient safety, Dr.
Sandler says. Cost was another consideration: You can print a bar code
for about 10 cents, but an RFID tag costs about $1 to $1.50, says Dr.
Sandler, which makes the technology far too expensive now to use in lieu
of bar codes in a pharmaceutical dispensing system.
Georgetown’s current three-way comparative study in the outpatient oncology-hematology
unit is a first step in examining the role of RFID tags as an adjunct
to bar codes to ensure the right patient gets the right blood. For the
study, Georgetown is partnering with Precision Dynamics Corp., San Fernando,
Calif., which is providing "smart" patient identification wristbands that
include human-readable print, a bar code, and RFID tag. Precision Dynamics
is also providing RFID tags placed on the crossmatch label generated by
the hospital blood bank.
The RFID tags used in the study contain 2,048 bits (256 characters) of
semiconductor memory (similar to a computer’s memory) and a small antenna,
embedded in the wristband, says Charles Feldman, PhD, a Harvard researcher
in cardiovascular medicine and science advisor to Precision Dynamics.
The RFID scanner’s computer-driven electromagnetic field can read and
write to the RFID chip. Dr. Feldman likens this process to a radio transmitter
and a radio receiver operating at a low frequency of 13.56 megahertz (by
contrast, cell phones operate at about 900 megahertz). The low radio frequency
means that an RFID scanner can’t read an adjacent patient’s wristband
RFID, which is an important safety feature, Dr. Sandler says.
In the study, patients wear the dual bar-coded/RFID wristbands. Before
transfusing the outpatient at the bedside, the nurse scans his or her
bar-coded ID badge to create an electronic signature. (The researchers
skipped the added expense of putting an RFID tag on the ID badge, but
the tag would have allowed the employee to scan the badge upside down,
through a pocket, even in a purse, Dr. Sandler says.)
The nurses perform the hospital-required two-nurse visual inspection
of the patient’s identification and the blood product unit, which becomes
the procedure of record. With the patient’s informed consent, the nurse
uses the PDA to scan the bar-coded wristband and the American Red Cross
whole blood number on the front of the blood unit-and the bar-coded crossmatch
label that the hospital blood bank places on the back of the unit. The
nurse then uses the PDA to scan the RFID tag on the wristband, the RFID
tag on the crossmatch label on the back of the unit, and the bar-coded
Red Cross whole blood number on the front of the unit.
The nurse documents the time required to perform each of the three patient
identification approaches and comments about the procedures.
The study shows that a second nurse can do the visual check in less time
than it takes to use the bar codes—but not the RFID tags, Dr. Sandler
says. Yet using the PDA to scan either bar codes or RFID would eliminate
the need to have two nurses perform the visual identification inspection,
which saves nursing time. The researchers have measured the interval of
time for the nurse to retrieve a second nurse to do the double check,
which Dr. Sandler says ranges from a few seconds up to 20 minutes. The
bar code or RFID process also offers a more precise way of making sure
the right patient receives the right transfusion, he notes.
The study has encountered technical glitches, such as printers that don’t
work and the PDA’s software being programmed so that if the nurses punched
one wrong button they couldn’t go back and change it, says Barbara Arnett,
RN, an oncology nurse participating in the study. Though the nurses believe
the RFID works "beautifully when there are absolutely no problems with
the system," she says, they prefer the visual inspection using a second
nurse until the glitches can be worked out.
Even when a bar-coding or RFID system is in place, it’s not fail-safe,
Dr. Sandler says. For example, a patient could get the wrong wristband,
though even then he would receive the right blood since his blood sample
would generate the crossmatch.
But, says Dr. Sandler, hospitals can have errors that RFID tags can’t
prevent. A few years ago, two pediatric patients receiving intravenous
transfusions were playing checkers in a playroom. The children changed
places at the table without repositioning their IV poles. A nurse came
in to hang blood on one of the children. The nurse had a second nurse
read the intended transfusion recipient’s wristband and check the information
on the blood container, as required. The nurse then began to hang the
blood on the IV pole positioned closest to the identified transfusion
recipient, which belonged to the other child. Fortunately, Dr. Sandler
says, the child’s mother entered the room just in time to alert the nurse
to the error before the wrong child received the blood.
Hospitals can use RFID technology to set up gatekeeping functions where
an alarm would go off if a nurse brought the wrong blood product or drug,
for example, into a patient’s room or the operating room, Precision Dynamics’
Dr. Feldman says. "Any gatekeeping function would probably operate in
a higher frequency range than the current wristbands, most likely 915
megahertz," he adds.
For now, the cost of RFID technology stands in the way of using it widely
in blood transfusion safety or other applications. "To equip a hospital
with a bar-code or RFID system for transfusions would cost on the order
of a million dollars in an initial outlay," Dr. Sandler says. But Georgetown’s
next study planned for this summer will address how one could minimize
the ongoing expense of such a system by using RFID chips that have been
part of the supply chain from the start. Dr. Sandler also notes that if
the Department of Defense, Wal-Mart, and other potential high-volume users
move forward with RFID technology, as planned, its cost should drop.
In its study this summer, Georgetown will partner with the American Red
Cross blood center in Baltimore and Precision Dynamics. Precision has
produced 10 multi-write RFID chips that contain "prompts" for gathering
information for the collection and manufacture of blood in accordance
with good manufacturing practices.
If the small experiment shows there is a "strong place" for RFID technology
in blood collection and manufacturing, a larger study would look at the
cost and benefits of the technology and its impact on workflow, says Irwin
Thall, Precision Dynamics’ health care RFID manager. "We can’t quantify
cost saving in the study planned for this summer because we’re just doing
10 units to see how the technology fits and what information we would
want to write on the chips, and how it’s used and where it’s needed."
Could use of RFID technology in blood collection, manufacturing, distribution,
and transfusion become the standard in blood banking? Yes, Thall says,
and how soon would depend on how quickly the research gets published.
"Once you have the technology of safer, secure special identification
on blood test tubes and for all lab testing—and the infrastructure
to manage data and do patient identification comparisons in real time—one
would see the impact on workflow and the bottom line," Thall says. Nurses,
for example, would have more time for patient care. "And the system would
greatly reduce patient errors."
Karen Lusky is a writer in Brentwood, Tenn. |
|
|
