Meeting FDA's Bar Coding Rule
Three professionals from major drug companies discuss the agencyï¿½s proposal to reduce dispensing errors.
Well before FDA proposed linear bar codes for all drugs, the industry was abuzz with concern and excitement over the prospect of unit-of-use bar codes. Some packaging professionals wondered how they could fit codes onto labels already crowded with other required information, while others were delighted the agency finally offered a plan to curb medical errors. And now that FDA’s March 14 proposal has settled in, real solutions to drug bar coding are emerging.
For this year’s pharmaceutical packaging roundtable, PMP News has brought together three professionals who are leading their firms into the bar coding era: D. Bruce Cohen, director of packaging technology for GlaxoSmithKline North America; Rich Hollander, senior director, packaging services, for Pfizer Inc.; and Joe Mase, senior marketing manager for Baxter Healthcare. In this discussion with PMP News editor Daphne Allen, these professionals share their firms’ reactions to FDA’s proposed rule.
Was the proposal a surprise?
Mase: FDA did an excellent job of surveying both the healthcare industry and the manufacturing industry to learn their needs and concerns. The agency collected data, listened to opinions, and devised a workable solution.
Cohen: The PhRMA bar code team met with FDA on several occasions, and we showed them what happens with bar codes, how the bar codes are used, what formats we could use, what was currently on the market, and what we believe would be the easiest way to implement a system. We were very happy that FDA listened to us—most everything we asked for is in the proposed rule.
What do you think about the proposal to require just the National Drug Code (NDC)?
|“There is certainly a difference between trying to provide all this data on a blister versus on a unit-of-use label for a vial or a syringe.”|
D. Bruce Cohen
Director of packaging technology for
GlaxoSmithKline North America
Cohen: Hospital groups were pushing for just the NDC because it would give them about 90% of what they needed to prevent medical dispensing errors in the hospital. The NDC would give them product identification at the bedside using any kind of scanning system and would be relatively easy for manufacturers to implement.
Mase: From our perspective, FDA did the easiest and probably the smartest thing by proposing to mandate the linear bar code with the NDC. But what I think will happen is that large group purchasing organizations will ask for additional information above and beyond the NDC. They have already stated that they want lot number and expiration date to be bar coded as well. There will be some advantage for manufacturers such as Baxter and Pfizer that have included a lot number and an expiration date. By using an expiration date, you can ensure that the patient never receives an expired product. With the lot number, you can ensure the patient never sees a recalled product.
Hollander: One concern is the use of the NDC versus the Universal Product Code (UPC) for over-the-counter products used in hospitals. There are some implications that need to be reviewed and digested relative to what that means for OTC products that do not generally have an NDC bar code anywhere on their packages.
Do manufacturers of OTC drugs need to decide for themselves whether their products are included in this rule?
Hollander: The agency has an approach for prescription drugs—all drugs need to comply with this requirement whether or not they are labeled for institutional use, as it is not always clear which drugs will end up in a hospital. But the OTC side of the argument is a little different. For OTC products, FDA says you need to apply this rule to any product that is likely or intended to go to a hospital. Most of us would say that means if the product is labeled for institutional use, but the rule does not say that.
Cohen: There certainly are some OTC products that manufacturers may not intend to distribute in a hospital because they are in bottles of 30. But a repackager or a hospital may choose to use those products and order them for use in a hospital. That presents another problem—who puts the code on there? If it is a repackager, what NDC does it use? How does the repackager distribute it? In most cases, repackagers will buy bulk product and pack them up into unit-dose containers for sale to clinics, nursing homes, and hospitals. I have seen a repackager put its own NDC on something, and I have seen a repackager put the manufacturer’s NDC on it, which raises other issues, because the repackager does not own that number.
Will the rule encourage unit-dose packaging and bar coding?
|“We in the packaging community got excited about being able to make an impact, and we all worked long and hard to get to where we got.”|
Senior director, packaging services, for Pfizer Inc.
Mase: FDA requires that you bar code down to your lowest packaging level. So at Baxter, we bar code down to our individual unit of use, which would be an individual IV solution. If you do not place individual tablets in an individual package, you do not have to bar code each dose. You have to bar code down to the lowest level of use or lowest level of packaging, which is the bulk container. So you are not required to go back and redesign packaging, which would be cost-prohibitive for some small manufacturers. But what the regulation will do is create competition between those products that do have unit-dose packaging and those that do not.
Cohen: The wording in the proposed rule was “unit of use,” which can be described as unit dose, but certainly unit of use is the lowest form of the pack that you can distribute to a patient. In some cases, it may be an aerosol container that holds 60 doses of a respiratory product. And that would be the smallest form of that product. So that pack, even though it is 60 doses, would be what the patient actually receives in the hospital, and it would have a bar code on it. It may even have a UPC code on it because it is a trade pack that is regularly available via retail sale, but it can be scanned by a system in the hospital.
Hollander: This proposed rule by itself is not going to do anything to either encourage or discourage the use of unit-dose packages. In the end, it is how hospitals develop their own systems and infrastructure to support unit-dose blisters and the bar codes that are on them. Hospitals are going to reach a point where these types of bar coded packages are essential for them to dispense product based on systems and practices they have developed. They will tell industry, “We want blisters, and if you want to sell product to us it is going to be in a blister.”
Cohen: There are some products in blister packs that were not originally designed to be separated and were not designed to be distributed in a hospital. They were designed to be dispensed intact to a patient. Those configurations will either have to be modified to make them large enough to support bar coding and to allow them to be separated for a hospital, or the company will decide to make a hospital unit-dose pack where you can provide all the required information for the hospital.
Pfizer has demonstrated that a standard hospital unit-dose blister can bear NDC, expiration date, and lot code. How is that being received, Rich?
Hollander: We did not do anything that I would consider “rocket science.” We simply demonstrated we could do what we thought the right thing to do was and used existing technology to do it. We were just first. I do know of at least two other companies that are doing exactly what we have done. These three companies provide a significant percentage of the hospital unit-dose blisters that enter hospitals. So if you take that, plus what Baxter is doing with IV bags, there is going to be a lot of packages very shortly in the hospitals with all three data elements on them.
Cohen: There is certainly a difference between trying to provide all this data on a blister versus on a unit-of-use label for a vial or a syringe. Rich makes light of Pfizer’s efforts, and he deserves a little more credit. I understand what Baxter has had to do to get to that point. And it is not as easy as some would like it to be. It takes concerted effort among a number of folks and plant sites. And the blister is probably one of the easier packages on which to do this because you are often already printing lidding in-line. The process is relatively easy compared with that of vaccine vials, which are relatively small with virtually no space for anything, and now we are going to be asked to encode an NDC. It is going to be difficult to print them at the speed that those lines run.
FDA is suggesting the use of a linear code, rather than a two-dimensional code. Does that present another challenge?
|“The regulation will create competition between those products that do have unit-dose packaging and those that do not.”|
Senior marketing manager for Baxter Healthcare
From the vaccine standpoint it does. The vaccine manufacturers’ group, VISI, has been working with the Centers for Disease Control, and they certainly would like to go to a 2D code. They have, in fact, petitioned the Uniform Code Council Inc. (UCC) to look at Data Matrix as a standard code. And UCC is, in fact, investigating that and at some point we believe they will adopt the Data Matrix. We can use it on some very small labels.
What about data structure?
It was pretty important that FDA choose a bar code standard. The retail industry has a very specific standard—the UPC. Until this point the hospital industry could have chosen the Health Industry Bar Code standard. But pharmaceutical manufacturers use the UCC standard, which FDA has proposed to be mandated.
Pfizer has chosen to use the data structure it has put forward for all three data elements based on working closely with UCC/EAN standards in existence and presently used in other areas of our industry. We have used standard application identifiers and published this widely so that the people developing their systems will be able to know how to interpret the data that we are giving them. In other words, we need to develop a standard data structure for the NDC that adequately represents the unit-dose package level [Global Trade Identification Number] as well as for those who want to encode lot number and expiration date. Computer systems working with data from bar code readers will therefore be able to know how to manage that data once they receive it. Otherwise the whole thing goes bust. FDA is not in a position to develop those standards for us; it wants industry to establish them.
Has anyone explored new in-line printing technologies?
We print directly on the IV drug container. To print a bar code, you typically need a white background with a dark bar printed on top of that background for a bar code to work. So you need a reflective surface and you need an absorbing surface to get the contrast between the bars and spaces for a bar code to read. IV drug manufacturers in the United States use a hot-stamp printing process, which requires a metal-engraved printing plate and a hot-stamp ribbon to transfer the image to the containers. But it really does not have the resolution needed to get a consistently readable bar code. So we developed a new high-resolution technology called enlightened HRBC (High-Resolution Bar Coding) that will allow us to print directly on the container at the time of manufacturing. We are able to get a high-resolution bar code on our bag, with lot numbers and expiration dates.
For printing blister packs, flexographic printing is probably as good as you are going to get in terms of print quality. There are other types of printing that are used for blister packaging, such as laser printing. I understand that these printers are now capable of printing RSS composite codes, however, I have not seen any data to demonstrate print quality. We are embarking on evaluating them for some sites outside of the U.S. For other packages, like vial labels and plastic bottle labels, which is where we are now beginning to focus, we are investigating steered-beam lasers.
What could vendors do better?
Help us understand and use the ANSI bar code grading system for print quality as an in-process control tool. Pharmaceutical companies do not have a lot of experience printing bar codes. Typically, they come to us already preprinted on labeling, as they are not variable from lot to lot. It has been the vendor’s responsibility to ensure that we receive printed material with an ANSI grade of C or better. Now that we are becoming printers, we are responsible for that ourselves. What we are finding is that the tools that are available to help us inspect for print quality are not up to the standards that the pharmaceutical industry requires in terms of algorithmic robustness. In other words, can we validate them?
Assuming that you can have the software validated to a CGMP standard, it would be nice to be able to verify print quality in-line. First, the verifier could tell us that the overall quality is C or better. Second, it could provide us information that will help us better monitor printing. Verification looks at nine different attributes of the print quality, and some of those we cannot control because they are driven by the type of substrate, the ink, and the print mat or the ink carrier. Those things are fixed during qualification. However, there are things we can do, like adjust pressure or clean the roller, that can positively affect print quality. Reviewing these attributes as close to the process as possible will enable us to adjust the process to avoid printing lower-quality bar codes. No one has a system available that allows in-line bar code verification according to ANSI standards. There’s an opportunity for someone to develop such a system.
Verification checks so many different bar code attributes, and it is a slow process when compared to bar code scanning. We scan each bar code that we print to make sure it is readable, and then we do periodic quality checks with a verifier where we take a sample off the manufacturing line to make sure that we are still maintaining a high level of print quality during the whole manufacturing process.
So adding in-line verification can slow down a line?
Boy, that’s an understatement.
It takes seconds for a verifier to give you a read back, which is too long for manufacturing line speeds in many cases. To make sure you maintain print quality, you develop an off-line plan with statistically valid sampling. That is where we are going.
But what if you want to be able to monitor trends in your printing process so that you can react long before you have a problem? It gets prohibitively expensive to stop the line, fix the problem, and reinspect the blisters that might be of dubious quality. We definitely need in-line verification. Off-line verification is time consuming, resource intensive, and slow.
Will bar coding give you a competitive advantage?
We first started our evaluation about a year before the NCCMERP paper came out. Our CEO at the time, Bill Steere, asked the question, why do we not have bar codes on our hospital blister packs? He had heard that hospitals were interested in having bar codes to prevent dispensing errors. Management said it wasn’t concerned about cost, just doing the right thing. We in the packaging community got excited about being able to make an impact, and we all worked long and hard to get to where we got. But cost never came up. Is it a competitive advantage? Time will tell. The market is going to dictate it. We think it’s the right thing to do and hope hospitals will develop systems to take advantage of it.