ProPath Deploys Sybase RFID Technology to Streamline Laboratory -- ADTmag

ProPath Deploys Sybase RFID Technology to Streamline Laboratory

By John K. Waters
February 23, 2005

Processes don't get much more hands-on than the work carried out in a specialized branch of the clinical laboratory services industry known as anatomic pathology. In most AP labs, the procedures for processing the hundreds of bits of skin, gallbladder, breast lumps, and other tissue specimens that flow into the facilities every day are carried out by specially trained technologists. Each specimen must be described, sectioned, dehydrated, and embedded into small blocks of paraffin that are sliced for slides. As they are traditionally carried out, manually, these procedures are both time-consuming and vulnerable to human error.

But one of the oldest AP labs in the country has set out to streamline the biopsy biz with an inventive implementation of RFID technology.

Founded in 1966, Dallas-based ProPath Laboratory is a 24-hour operation staffed by 30 pathologists and a group of veteran lab techs. Although it is thought of as a local operation, ProPath receives more than 1,200 specimens every day from around the country.

"We're not a small operation, but AP is a small part of the industry," says Krista Crews, ProPath's executive director, "and it's one that hasn't had the attention of [software] development companies and doesn't have the luxury of some of the automation and instrumentation that you see in a clinical laboratory. We are trying to take some of the technologies that are available in other industries and use it to automate parts of a very manual process."

When most of us think of lab work, what comes to mind are those blood and urine tests for things such as cholesterol and glucose levels that our doctors use to convince us to lay off the cheeseburgers and Krispy Kremes. These are all... well... renuable substances that can be replenished if something were to go awry with a specimen. But an AP lab specializes in preparing specimens from an often finite amount of material. "You can always draw more blood or provide more urine," she says. "But that suspicious mole the clinician sent in just might be the only one."

Moreover, modern blood and urine tests are highly automated, says ProPath's IT director Mel Lively. "Blood and urine samples going to a clinical lab are going to be put into machines that test them and generate numbers," Lively explains. "It's a highly automated process right through to the results. But in an AP lab, it's mostly about preparing specimens that are reviewed by a doctor."

To make their processes both more efficient and more secure, ProPath is now tagging each specimen as it enters the lab with a tiny RFID chip. The chips identify the tissues, announce the specimens' movements through the laboratory, and provide information about which processes need to be done by querying ProPath's anatomic information system (AIS) database.

The lab has utilized bar-coding for this purpose, but because the specimens are embedded in paraffin, this is not an especially good implementation of that technology, Crews says. Labels can slide off, peel off, or otherwise not survive processing. And even hand-lettering the labels on the plastic cassettes that hold the paraffin blocks doesn't work very well, because they're often unreadable by they time they're through the process. Unlike bar codes, RFID tags can be embedded in the paraffin blocks.

But switching to RFID tags was only part of the solution, Lively says. After slicing, dicing, and analyzing the lab's procedures for slicing, dicing, and analyzing specimens, the members of ProPath's four-person development team realized that they would need a middleware layer to manage the large volumes of fine-grained data generated by the new system. Traditional AP databases are simply not designed to handle the kind of information ProPath wanted to collect.

"Our diagnoses are textual," he says. "There's a narrative about what is seen in the specimens. We don't just give them a 140 for cholesterol. These text-based diagnoses don't lend themselves to easy sorting and database-type functions. The traditional AP reporting system isn't designed for capturing processes."

After much research, and a fairly limited vendor search, ProPath settled on iAnywhere Solution's RFID Anywhere middleware platform for this purpose. iAnywhere is a subsidiary of Dublin, CA-based enterprise infrastructure and mobile software company Sybase, with which ProPath already had a good relationship.

iAnywhere is a relative newcomer to the budding RFID middleware market. Its current RFID offering is based on technology from XcelleNet, the maker of device management software for mobile databases and wireless devices that Sybase acquired last May and rolled into iAnywhere. The product includes connectors and controllers designed to handle the interfaces with commonly used RFID hardware components, such as readers and printers, freeing developers and integrators to focus on business logic instead of the code-level complexities of evolving devices and standards.

RFID Anywhere was built around a .NET-based service-oriented architecture, explains David Barrack, senior director of solution services at iAnywhere. "We won't be using traditional client-server architecture for RFID data," he says. "What we'll have instead will be very close to a peer-to-peer network, except that one of the RFID Anywhere nodes will serve as a kind of a master, listening for events that come from the other nodes." The company chose .NET rather than Java, Barrack adds, because Java app servers can't handle the large volumes of "bursty data" that an RFID network generates.

The ProPath RFID project is still something of a work in progress. ProPath was an early adopter of the RFID Anywhere. Crews and Lively expect to have the new RFID system at ProPath fully implemented in March. Crews expects it to yield a 15 to 20 percent reduction in overall processing times. The system will also improve accountability, she says, because it will be tracking who did what and when.

"The hard part was getting into the lab, understanding the processes, and writing down all of the steps that a specimen goes through in the lab," Crews says. "Some were obvious, but others were anything but. Histo-techs, the people who prepare the specimens, are trained to know what to do for gallbladder versus stomach versus liver versus lungs. How do you convert that intelligence into an automated process? We found ourselves building a database that doesn't exist anywhere. You can't just go out and buy a database on histology protocols."