The Eyes Have It

Problems associated with patient identification at the point of service include inefficiencies in data entry, privacy breaches and medical identity theft, such as the fraudulent use of insurance cards.

More seriously, errors in transcribing can prevent the right patient’s record from being used in caring for that patient. The use of biometrics can improve the accuracy and integrity of care and should be considered as options to automate identification.

Retiring “Non-Smart” Cards

The Medical Group Management Association advocates for standardized patient identification cards through its Project SwipeIT. They estimate that leveraging decades-old technology, such as magnetic stripe and barcode readers, should save approximately $1 billion by reducing repetitive data entry and transcription errors. But the savings won’t come easy, as implementing these process improvements requires cooperation from a latticework of payers, providers and technology vendors.

Automated identification technologies such as bar coding, magnetic striping and smart cards all depend on the systematic encoding and representation of key identifiers onto labels or cards that must be printed or produced and carried around until they are needed. The Workgroup for Electronic Data Interchange (WEDI) has published standards that outline essential information to be contained in a standard card, as well as data structures to govern the encoding and decoding of machine-readable information. Specifically, the WEDI Health ID Card Implementation Guide calls for two-dimensional symbols or a magnetic stripe to carry essential identification information.

The use of smart cards is a clear improvement over current manual processes, but additional benefits can be achieved by using more secure Auto-ID technologies. Biometrics eliminate layers of representational apparatus such as labels, tags or cards, and rely on unique, person-specific characteristics that cannot be forgotten or fraudulently used.

Fingerprints

The use of biometrics has been common for some time with certain healthcare applications, in particular, those which address vulnerabilities associated with password-driven user authentication. For example, Cardinal Health has been shipping fingerprint readers as a standard component of their Pyxis MedStation automated dispensing cabinets since 2004. These readers enable more secure authentication as the user manually enters a user ID and then swipes a finger over the reader as a password substitute.

Early usage of fingerprint readers has been subject to technical limitations that hinder successful scans. One limitation relates to the inability to scan “excluded” people, such as those with worn-down ridges on finger pads, or individuals with certain genetic profiles. Other limitations are caused by particular usage factors such as wearing gloves, using cleaning fluids and general user difficulty in scanning.

Such exceptions cause inconsistent processes and end-user applications have introduced features to deal with these exceptions. For example, Omnicell Inc., another vendor of automated dispensing cabinets, allows a “fingerprint-capable” user to serve as a witness for an excluded user to sign in with a password. Also, newer fingerprint readers have introduced improved hardware, firmware and decoder software, which more effectively deal with imperfections in the reading process.

The Eyes Have It

While fingerprint readers continue to gain momentum in use, some providers are adopting more advanced technologies such as palm vein recognition and iris identification. Instead of smiling and saying “cheese,” patients at the Urban Health Plan, a community health center in Bronx, N.Y., look into the camera to prove their identity. This community health center is one of the first in the U.S. to use advanced iris identification technology, which employs a camera that performs a near-instantaneous scan of the iris pattern to uniquely identify each person. Iris ID is the only commonly-used biometric that requires no physical contact between the reader and the individual. As such, iris scanning doesn’t require device cleaning between reads, a benefit from an infection control standpoint.

Palm vein technology uses near-infrared light that “illuminates” the veins in a person’s palm to record an image that may be uniquely associated with that individual. Palm vein readers don’t require contact with the sensor, but do necessitate cleaning the plate onto which the palm is placed for reading.

Urban Health Plan is using technology provided by Eye Controls LLC, Chantilly, Va., as a risk management tool. For clinical safety, the iris scan facilitates a right-patient/right-record match, which is more foolproof than validating against a photograph. “Our practice exposes us to potential for misidentification, as we may have 1,000 patients named Rodriguez,” explains Samuel De Leon, M.D., chief medical officer for Urban Health Plan. Positive matching is especially critical when indexing patient information across a broader spectrum of care, such as a regional health information organization.

Financial risk is mitigated through reducing insurance card fraud. The practice is reimbursed according to fixed-payment guidelines and so additional fraudulent visits result in uncompensated care. The health center further recognizes the ability of iris scanning to reduce its exposure to legal/medical risk by preventing a patient identification error from becoming a treatment error.

Patient flow is improved as iris scanning immediately brings up patients’ data in their electronic health records, negating the need to choose them from a pick list of patients with the same last names.

Iris identification is not yet commonly-accepted due to patient unfamiliarity with the technology. “In order to allay patient fears, we explain that the reader is a camera, although admittedly, a complex camera. This makes the technology far more approachable than if we called it ‘iris recognition,’” explains De Leon.

Evolving Toward Choice

The availability of newer techniques such as iris and palm vein scanning allows providers choice when selecting identification technology. Biometrics, like other Auto-ID technologies, simply facilitate a means of inputting information. The complex ecosystem of cooperation required between payers, providers and vendors benefits from the standardized data structure. However, the choice of equipment used to input information can be more flexible as future needs evolve.

For example, in order to set the stage for pharmacy and point-of-care medication scanning, the FDA rule that went into effect in 2006 requires that pharmaceutical manufacturers encode the National Drug Code on the lowest level of packaging shipped. This rule enforced the use of linear bar codes largely to accommodate the abundance of already-installed laser scanners that are incapable of reading more advanced technologies, such as two-dimensional symbologies.

“Despite the FDA’s linear ruling, the industry has moved to adopt Aztec 2-D symbols for patient wristbands and Data Matrix for medications packaged in the pharmacy. Hospitals are increasingly purchasing camera-based imagers in order to be prepared for the future,” according to Jamie Kelly, co-founder of The unSUMMIT for Bedside Bar coding. Three years later, the FDA is considering more flexibility in the choice of Auto-ID technologies to be used for medication e-pedigree,

This recent industry lesson underlines the wisdom of allowing providers to choose the technology that works best for them in ensuring positive patient identification. For many, biometrics readers will be the right choice as hospitals seek to eliminate the costs of encoding data onto labels or cards and offer patients greater convenience.

Mike Wisz is a healthcare IT consultant and member of the HIMSS Auto-ID taskforce. Contact him at [email protected].