Home

Cover Story

Table of Contents

E-Newsletter

Article Archive

Editorial Calendar

Datebook

Writers' Guidelines

Orgs/Links

Opinion Polls

Reprints

Forum

Search

December 18, 2006

Biometric Identification — You See It in the Movies. Will You See It in the Hospital?
By Thomas G. Dolan
Radiology Today
Vol. 7 No. 25 P. 14

As handling security and privacy issues evolves, the technology will expand, but clinical care presents some unique challenges.

Biometrics, which may be described as the capturing of unique body characteristics for the purpose of both identification and access, “is not a new technology but has been around for two to three decades,” says Eric Stotz, founder of the San Pedro, Calif.-based Omni Identity, Inc., a distributor of biometric products. “Most of the applications had been in high-security government and military applications. Only in the past 10 years, especially the past five years, has the price come down to make it feasible in other fields, such as healthcare, so biometrics is currently getting a lot of traction.”

Fingerprinting, which has been around for more than 100 years as a means of identification, remains the most common form of biometrics. Stotz says that five years ago, the cost of biometric fingerprinting technology was $300 to $400 per user; today it’s dropped to $150 to $200 per user—still pretty expensive when compared with the magnetic stripe card or password. Yet Stotz says the cost of this technology will continue to fall as it becomes more of a commodity and is mass produced.

There are other forms of biometrics, Stotz says, some of which may not be suitable for a healthcare environment. These include eye or face recognition, which may be too intrusive, and hand geometry, which may be more suitable in a bank or point-of-sale situation as a way to access a cash box, or as a substitute for punching a time clock. There is even gait recognition, which can capture the way a person walks.

Healthcare Challenges
Even though fingerprints are more suitable to healthcare, some problems associated with their care remain. For instance, doctors and nurses, whose fingers become naturally worn, may apply lotion, which in turn interferes with the reading, as does their fingers getting colder from the lotion. Similar problems can result from latex gloves.

There are other forms of biometrics at different stages of development and/or implementation that may address the problems associated with fingerprints. They include, Stotz says, voice recognition, keyboard stroke recognition, and light recognition, which identifies users by how light passes through their skin. There is also a new technology that recognizes the network of veins in the palm of the hand.

Not only are there multiple modes of biometrics, they are also used for a variety of applications, sometimes overlapping, sometimes different from each other. The technology may help doctors and nurses gain quicker access to records or it may more effectively deny others this same access. It can lock out unauthorized personnel from both physical spaces as well as data areas on their computers. It may help patients access their records or prescribed medication. Biometrics can also protect patients from receiving the wrong medication or undergoing surgery intended for someone else.

The question is, with all the different types of biometrics—some old—others just arriving, as well as its varied applications, does biometrics really work, is it practical, and is there a real need?

According to industry experts, the technology basically works, it’s easier to implement than cards and passwords, and it’s much more accurate, which enables it to better protect patient identity.

However, it’s worth noting that there are drawbacks to its use in healthcare that are unrelated to the viability of the technology itself.

Paper Challenge
One big problem, says consultant C. Maxine Most, principal and founder of Acuity Market Intelligence, “has to do with healthcare information technology. It doesn’t exist. Anybody familiar with healthcare knows that recordkeeping is abysmal. There is no centralized management for records. Where I live in Boulder, [Colo.,] when I go to the hospital, I still have to bring a physical piece of paper, and that’s xeroxed if I go to the lab or another department. It enters different computers which are not integrated. You talk about the right to privacy. You see a PC with a bunch of files piled up around it. It’s kind of a joke. The amount of paper is ridiculous, especially when compared to industries such as financial services or insurance. This is one of the reasons why healthcare costs are so out of control.”

Another problem, Most says, is the expertise gap between the biometric manufacturers and system integrators. All too often, Most maintains, the manufacturers may have a worthwhile technology, but its inability to integrate with other systems and the manufacturer’s unfamiliarity with the healthcare industry makes it of little use. Conversely, the system integrators understand integration and the industry but don’t have the specific knowledge of biometrics and how it fits within healthcare.

These factors, in turn, relate to a third problem: the acceptance of biometrics by healthcare users. “In healthcare, people are interested in taking care of patients, not in using technology,” Most says. “A doctor who has resistance to new technologies may not want to be sticking his finger in a slot every time he wants to get something done. Biometrics is highly dependent upon users and how they use it. Unless healthcare personnel really understand and see the correlation between the technology and helping patients, they are not going to embrace it. If they don’t embrace biometrics, it won’t work properly.”

Most says many people in healthcare have a general awareness of biometrics, if only because of their knowledge of its use in other fields. But they are skeptical of its use in healthcare—a skepticism well-justified, Most says, because of the industry’s experience with so many other technologies. “The industry has pushed technology rather than provide more creative, integrated solutions to help them take care of patients,” she says.

However, Most believes biometrics is ultimately not only much more accurate than current identification tools but more convenient for healthcare staff. Cards can be lost or stolen, and passwords can be forgotten or learned by someone else. Many institutions require new passwords or personal identification numbers every 60 to 90 days, which irritates many people. Also, many healthcare situations require quick access to information and personnel can ill afford to spend time typing in passwords or otherwise being bogged down.

“I have faith in biometrics and believe this will ultimately be a large-scale solution,” Most says. When asked whether the variety of different biometric modalities may spawn even more lack of integration, she says forms have different uses and all can have their place in a healthcare environment.

Keystrokes & Veins
For instance, she mentions the keystroke option introduced by iMagic Software, Inc. “They actually link and integrate the password with the pattern of the keystroke,” Most says. “The technology monitors you and keeps track of your muscle memory until it has a good template. So the password doesn’t have to be changed. And the person doesn’t have to change the way he [or she] is doing things. He [or she] doesn’t even have to know that his [or her] stroke is being monitored.” Most says this is one form of biometrics that addresses the user’s resistance to change or reluctance to adapt.

Vein recognition in the palm is another biometric solution designed to overcome any resistance to change. Hiroko Naito, manager of business development for Fujitsu Computer Products of America, says the new product’s method of capture “is infrared. The sensor captures not point-to-point but the mathematical pattern of the veins.” Naito explains that because this device does not require direct contact, virtually anyone can use it. Doctors and nurses don’t have to worry about inserting their finger in a slot, and lotion or latex gloves will not present any interference.

Moreover, Naito says, “the keyboard in the nurse’s station has been found to be one of the most contagious devices in the hospital.” Merely wafting the palm over the keyboard avoids at least that particular contact with the keyboard, as the motion unlocks the computer without the need of a card or a password.

Vein pattern recognition has been around for approximately five years, but it’s even newer in terms of application. “We are a pioneer of this technology and introduced it into the Japanese market about 21/2 years ago,” Naito says. “You must keep in mind that no biometric is 100% proven from a medical standpoint. Fingerprinting has gained a market confidence since it’s been around for over 100 years. So how do we build that same confidence? We have to take the same path. The only way we can prove this biometric method works is to collect significant samples in Japan, Europe, and the U.S., which is what we are doing. We now have over 150,000 templates.”

Naito adds that the first big success was with 21 Japanese financial institutions. It has since been integrated into the administrative process, though not the clinical recordkeeping, at the Tokyo Medical University Hospital, and is currently undergoing its first trial in a U.S. hospital. Because the technology is new, it’s more expensive than other biometrics, but Naito says customers understand this.

Gaining a Fingerhold
While these new forms of biometrics search for their place in the healthcare marketplace, the fingerprint scanning method has found a home in many venues. Based in upstate New York, Ultra-Scan Corporation uses fingerprint scanning to help hospitals improve patient safety, according to president John K. Schneider, PhD.

“Prior to biometrics, when a patient came to a hospital, he was identified on the basis of demographics: name, address, phone number, and so on,” Schneider says. “But when you look at the studies from Health and Human Services, the error rates based on demographics are as high as 30%. The problem of patient identification is epidemic throughout the U.S. President Bush, in presenting his current Health Care Reform Act, has said that one of the major necessities is developing a universal patient identification format.”

There are basically two fundamental sets of errors, Schneider explains. The first results from duplicate records. This occurs when a patient is entered into a hospital’s system but is not identified with records that have been entered on previous visits. This duplication often creates a splitting of clinical information. Record No. 1 may show that the patient is allergic to a particular antibiotic, while record No. 2 does not contain that data.

While duplicate records are the most common error and can result in serious mistakes, the second, less common kind—the overlay record—is more dangerous and more apt to result in drug overdoses or operations on the wrong patient. An overlay record occurs when two different individuals who may have the same name or similar demographics are assumed to be the same.

Schneider maintains that Ultra-Scan’s TouchLink, in conjunction with two other simple tests, reduces the error rate to nearly zero. The enrollment process takes approximately one minute. Both of the patient’s index fingers are imaged, a digital photo is taken, and insurance cards are scanned. The entire database is scanned so nicknames, aliases, or fraudulent attempts are caught. On return visits, the patient’s finger is scanned, he or she provides an easily remembered number such as their birth date, and a digital photo is displayed for the healthcare worker to provide a human check between the photo and face. “Put these three factors together and you reduce that 30% error rate to almost zero,” Schneider says.

For a medium-sized acute facility with 200 to 250 beds, the system is fully operative within two to three weeks at a one-time cost of $150,000 to $250,000.

Although this form of fingerprint scanning technology was introduced only roughly two years ago, Schneider says, “acceptance has been tremendous. We start with patient records, but once the hospital sees how fast is the return on investment, it immediately wants to extend it to access control, time and attendance, and other applications.”

— Thomas G. Dolan is a medical/business writer based in the Pacific Northwest.