Sharing lessons learned from an RFID deployment at Boston's Beth Israel Deaconess Medical Center, Dr. John Halamka, CIO of CareGroup and Harvard Medical School, discusses how RFID can help organizations achieve key goals and what the impact may be on the IT infrastructure.

When the world's most famous wizard, Harry Potter, wants to discreetly explore the Hogwarts School, he consults the Marauder's Map, a magical device that shows every classroom, office, and secret passageway at the school. The map also shows the identity and location of everyone wandering its grounds. So Harry knows when the dreaded Professor Snape is busy in his office. And he can tell when the caretaker, Peeves, is just around the corner, requiring Harry to take evasive action.

While the Marauder's Map may seem like a schoolboy fantasy, it's not so farfetched. From a technology standpoint, a crude version is feasible today, thanks to the evolution of radio frequency identification (RFID) technology, which makes it possible to track people and objects with increasing accuracy. In fact, many organizations are exploring innovative and strategic uses of RFID, beyond the well-known applications of tracking inventory and reuniting lost pets with their owners.

From the gas pump to the summer Olympics to the ER

Contact-less payment systems, such as the Exxon Mobil Speedpass, enhance consumer convenience and speed revenue collection by letting consumers make credit card purchases with the wave of a card or token. Dow Chemical is involved in an initiative that combines RFID and GPS to enable real-time tracking of railway cars carrying hazardous materials. In preparation for the 2008 Summer Olympics, Beijing is rolling out an RFID ticketing and fare collection system for its transport network of subways, buses and taxis, which handle 10 million journeys a day.

In the health field, a Florida-based center for Alzheimer's care has enlisted 200 volunteers who will be implanted with a passive RFID chip (such as the one I have in my arm) that will help identify a person and provide medical history data if an individual arrives at an emergency room in a confused or unresponsive state. In the near future, we expect to see RFID sensors that can check a diabetic's glucose level, eliminating painful and frequent finger pricks.

At Beth Israel Deaconess Medical Center, which is part of CareGroup, we decided in 2006 that the technology was sufficiently mature to explore whether RFID could enhance our mission-critical processes in a cost-effective way. Knowing that the two basic forms of RFID—passive and active—offer different value propositions, we decided to test both. (See Related Article, "Passive vs. Active vs. Barcodes.")

A mission-critical task: patient identification

To explore passive RFID, we focused on the challenge of positively identifying patients. Accurate electronic identification helps ensure a patient is given the right medications, samples sent to the lab are correctly labeled, and the patient is not accidentally subjected to the wrong procedure. It also lays the groundwork for electronic administration of medications.

Before committing to a solution, however, we weighed the benefits of RFID against those of barcodes. We determined that bar-coded wristbands were adequate for identifying adult patients, but passive RFID was a better choice for neonatal infants. For example, when a bar-coded wristband is curled around a baby's tiny wrist, it's difficult to read and may need to be flattened out or repositioned. In contrast, a passive RFID wristband can be scanned without touching the infant. This means less disruption for a fragile newborn and less likelihood that slender feeding tubes and monitoring wires will be disturbed. Passive tags are also used to label and track mother's milk, which is stored in containers. A software application and RFID scanner ensure each infant receives the right milk and creates a record for feedings.

Making better use of equipment and staff time

For testing active RFID, we chose a different application: keeping track of medical equipment such as electrocardiogram machines, bedside monitors, and IV infusion pumps. These costly devices are mobile, so it's easy to lose track of them in a large facility. In addition, some staff members hoard scarce equipment to be sure it's available when they need it. As result, $500,000 worth of hardware goes missing in a typical year, forcing the hospital to carry extra inventory rather than risk patient safety. In addition, doctors and nurses were spending an average of 20 minutes a day looking for misplaced equipment, at a cost of up to $100 an hour.

The active RFID application we are now implementing addresses all these issues. An active RFID tag is attached to a medical device. Every few seconds, the tag emits a signal that identifies the device to any readers within a 10-meter radius. By measuring relative signal strength, multiple readers can determine where the device is located. In turn, this data is passed to an application, which is available from desktops throughout the hospital and allows staff members to see where the nearest piece of equipment is located so it can be quickly retrieved. Preliminary results suggest this new application will reduce the average equipment search to five minutes.

Assessing the impact on infrastructure

With an estimated 500 of the planned 5,000 devices now tagged, Beth Israel Deaconess continues to roll out the equipment tracking solution. However, a number of clear lessons have already emerged from our passive and active RFID deployments regarding the impact of RFID on IT infrastructure and related costs.

Leveraging Wi-Fi: In implementing active RFID, we lowered our solution cost by leveraging the Wi-Fi network already in place at the hospital, rather than using a proprietary frequency. A proprietary system can provide more accurate location, to the level of about a square meter. However, it would have required us to spend an additional $2 million in network infrastructure. The tradeoff we accepted is that we can only tag items that are large enough for staff members to easily spot within the current 10-meter accuracy. As the economics and accuracy of active RFID improve, we will be able to track items that are smaller and less expensive.

Cutting edge vs. safe harbor: In deploying PanGo Networks active RFID tags and Cisco lightweight wireless access points, we chose Cisco network components that incorporated cutting-edge features over Cisco "safe harbor" code that maximizes reliability and stability. Our subsequent experience suggested that cutting edge was not stable enough for a clinical setting. Now, in 2007, geolocation features are part of the "safe harbor" release, providing both advanced features and stability.

Tags and access points: In purchasing active RFID tags, we found it's important to purchase tags that do not associate with a specific access point. Access points downgrade their speed to support the most distant or weak associated device, which can degrade performance as the tags move around.

Radio frequency interference: All new devices being introduced into the hospital require clearance from Clinical Engineering to ensure their frequency does not conflict with existing devices. We followed this standard procedure for our RFID projects.

Battery life: At the time we deployed our solution, the batteries in active tags had to be replaced every six months. Newer tags have a battery life of two years, making maintenance easier.

Security: Passive tags are encoded with a patient's medical record number but no patient data. Due to other safeguards, it is unlikely this information could be used to compromise a patient. No data is stored on active tags, so security is not an issue.

Privacy: Employees were adamant about not wanting active RFID used on employee badges, which would have created the potential to track people's movements without their knowledge. We adopted bar-coded badges instead, which require the employee to swipe or wave their badge near a reader for their location to be identified.