Medical Device Daily Washington Editor

WASHINGTON – The second day of the annual meeting of the Healthcare Information Management and Systems Society (HIMSS; Washington) included a presentation that served as a business case study for data tagging of drugs dispensed from hospital pharmacies.

William Churchill, director of pharmacy services at Brigham and Women's Hospital (BWH; Boston), offered a firsthand account of both the upsides and downsides of deploying barcodes and radio frequency identification (RFID) tags to track drugs.

Churchill said that his initial interest in barcodes was for use in tracking intravenous (IV) medications because errors in this mode at BWH accounted for as many as six of 10 severe adverse events associated with medication errors.

He said a key driver for his interest was that any hospital that dispenses more than 6 million doses of drugs in various forms each year – even with an accuracy rate of 99.1% – triggers “more than 44,000 errors per year.” These numbers, generated by a study in which Churchill participated, addressed the function of a hospital with more than 700 beds, a volume matched by the bed count at BWH (the study, in the Journal on Quality and Patient Safety, does not name the hospital, but all the authors work at BWH).

In other industries, Churchill observed, “we would slap our hands together” in glee for an error rate of less than 1%, but such numbers are cold comfort to those patients who fall into the 0.9% crack.

The ideal, he said, is a system that ties together all the components of drug delivery within a hospital, including a real-time digital link between the pharmacy and a smart IV pump that would require little more from the nursing staff than placing a vial into the pump.

“If you've ever seen a nurse trying to program an IV pump, it can be frightening,” Churchill said, given ongoing multiple distractions and all the data that he/she must enter into the pump's computer. Automation of such devices “should lead to more efficient use of the nurse's time,” including dealing more directly with a patient's needs, he said.

While barcode formats and their ability to hold larger quantities of data have grown over the past few years, Churchill said that the current FDA regs are “just the starting point.” The agency calls for little more than the NDC number and has not issued a regulation that fixes the format of barcodes.

Churchill said he felt that “the linear barcode requirement is short-sighted” and should require information on expiry and lot number, which could be included in a barcode “relatively easily.” Lot numbers are required for many biologics, including blood, but not for many conventional drugs.

CDRH stated in a document earlier this year that medical devices are not included in barcode regulations because of a “lack a standard and unique identifying system comparable to the National Drug Code system.” The FDA “is reconsidering whether some form of unique device identification (UDI) is warranted for medical devices,” but the document does not indicate whether action is impending.

Churchill recommended that barcodes should include dose form data in addition to NDC numbers because a number of medications come in multiple dose forms.

BWH opted for a barcode format known as PDF-417, one of the so-called 2-D barcode formats, because of problems encountered with traditional linear barcode. One of the difficulties engendered by this older format is that most barcode readers will not pick up the code if it is mounted horizontally on a vial of small diameter because readers are not designed to deal with the resulting wrap-around effect.

PDF-417 – a type of data matrix code – “is widely used in a lot of industries in which data exchange is important,” according to Churchill. BWH's use of this format has paid dividends despite skepticism on the part of some of Churchill's peers at other facilities, and he acknowledged that this code type requires 2-D scanners costing up to $1,000, with some models topping that figure substantially.

The next step up is image-capture technology – units that are simply imagers – capable of capturing all currently available code types but more expensive yet than 2-D scanners.

The payoff is that data matrix codes offer greater scanning accuracy than conventional linear bar codes and can still function even when as much as 60% of the code face is damaged. Churchill said that standard printers can print these codes with relatively inexpensive applications.

And Churchill encouraged consult with nursing staff when looking at such systems, saying that “the greatest field engineer in the world is a nurse in need,” and that hospital managers should avail themselves of that savvy.

Overall, Churchill said BWH has cut medication dispensing errors by 85%, with the greatest benefit coming in dose form error reductions. He said BWH's dose strength errors fell by more than 70%, and medication type errors fell by almost 60%. The system has also intercepted almost 2,100 doses of expired drug, a number that “FDA needs to see,” he said.

As for return on investment, Churchill offered fairly promising numbers.

In the first five years of BWH's barcode drug program, the cost ran to $2.3 million but saved the hospital $5.5 million, a net savings of more than $3 million. Of the cost, the hospital coughed up $1.4 million in the first two years and the running cost of operating the system tallies to just over $340,000. BWH broke even on the system within a year of “go-live,” in the third year.

Hospitals may find that RFID technology offers benefits not offered by bar codes, but the expense is substantially greater, and Churchill said that forward-looking institutions may want to selectively deploy both. However, he acknowledged various hang-ups with RFID that were unknown to him going in.

“The thing that really surprised me is that RFID is impacted by product and package composition,” he said, including even viscous liquid drugs interposed between the RFID chip and the reader.

“I do not see RFID as a 100% solution,” he opined, but potentially very useful in keeping track of those very expensive, biologic-based drugs and higher-risk drugs. However, RFID tags can help a hospital keep track of various devices that occasionally end up in the hospital basement to the amazement of all, when recovered.

RFID tags can cost $5 apiece at the lower end, a substantial increase in the penny-apiece cost of high-end barcodes.

RFID readers were described in his slide show as running between $2,000 and $5,000, whereas a top-of-the-line barcode reader peaks at $2,500.

But barcodes cannot be used to track the locations of surgical patients, an important point for friends and family during those long hours in waiting rooms. Barcodes also cannot tell supervisors where their charges are at any given time, a feature Churchill appreciates given his responsibility for a staff of several hundred in the hospital pharmacy department.

Churchill warned: “no one should leave here thinking that [either of these technologies] will be a quick-and-dirty implementation.” Some establishments will put in barcode systems now and wait to see what RFID has to offer, while others will take the plunge and go after a system that tracks everything from top to bottom.

In either case, the overall perspective is the same. To rationalize the expense, he said, “you have to have faith that your average length of stay and your medical errors will go down.”

That comment emphasized a key point: Given the lag between up-front deployment costs and accrued benefits, a key executive must champion such programs in hospitals with razor-thin bottom lines.