Safe and Secure

This past July, Pfizer and FDA made a distressing announcement: Counterfeit Viagra (silendafil) had been discovered in California retail pharmacies. The fakes weren't perfect. The tablets were sharper-edged than the real thing, and the distinctive blue was a bit off; the trained eye could spot typographical glitches on the packaging. But perfect or not, they were apparently good enough: More than a year after the notorious counterfeiting attacks on Lipitor (atorvastatin) and Procrit (epoetin alpha), and the establishment of an FDA task force to fight counterfeiting, a prescription drug supply system that was already on the alert had again been penetrated. (See "Counterfeits: The Cost of Combat," Pharm Exec, November 2003.)

Safe+Secure Pharmaceutical Supply Chain

The sobering incident came on the heels of an interrelated set of developments that in their own way threatened the security of the drug supply chain:

• congressional action to open US borders to drugs from other countries

• an explosion of internet pharmacy purchases

• Interpol warnings about terrorist involvement in counterfeiting

• an increasing number of offshore manufacturing facilities, making the supply chain longer and more vulnerable.

Alone, these issues would be worrisome. Together they make it clear that pharma desperately needs a way to control the integrity of its products.

The good news is that highly effective technology for tracking and accounting for medicines is near at hand, in the form of an electronic product code (EPC) information management system, using radio frequency identification (RFID) technology. In essence, an EPC can give a unique individual numerical identity to every single unit of medicine-a process called mass serialization. RFID, meanwhile, automates the reading and tracking of these numbers and adds a layer of technological security. Fully implemented, an EPC/RFID pedigree system should enable authorized users to automatically identify and account for each unit of authentic medicine in real time as it enters and moves through the distribution system. It could identify the current location of all suspect products in the event of a recall, track disposal of damaged and out-of-date product, and allow law enforcement full and accurate supply chain visibility if terrorists were to launch an attack using tainted medicine. And it should do all this with minimal increases in staffing and packaging costs.

The US drug system has embraced EPC/RFID and is betting most of its "anticounterfeiting chips" on the expectation that a fully implemented system will be in place and operating (at the unit packaging level) by the year 2007 or earlier. The state of Florida already requires pedigrees for some products. Today's compliance challenge will grow exponentially when the Florida law expands to all drugs and all wholesalers in July 2006.

This article explores some of the hurdles and questions that currently exist. It also identifies key steps that must be taken to move forward and examines approaches that might be useful if full implementation cannot be achieved in this time frame.

Whats the Timeline?

An Internet of Things

RFID is not new. During World War II the Allies used it to help fighter pilots distinguish their planes from enemy aircraft. The systems used then were expensive and bulky. Today's RFID chips are about the size of the period at the end of this sentence and may ultimately cost just a few cents to make and apply.

The revolution in the size and cost of chips has been married with, and stimulated by, a consortium of businesses and academic institutions. In 1999, the Uniform Code Council (UCC), several leading international companies, and the Massachusetts Institute of Technology formed the MIT Auto-ID Center, under the leadership of Kevin Ashton and Sanjay Sarma. The center was founded on a vision of an "internet of things," in which all products would be tagged with inexpensive chips that linked them to the global business information infrastructure.

For three years, the center and its affiliates worked to conceptualize a numbering approach, an information management system, and a governance/ privacy structure that could bring its vision to life. On November 1, 2003, the effort moved from concept to initial implementation with the creation of a new international standards body, EPCglobal. The corporation is nonprofit, jointly owned by UCC and the other main international bar-code administration organization, EAN International.

The shift of EPC/RFID from its research phase to initial implementation coincided with the emergence of drug counterfeiting as a major issue in the United States and the resulting struggle to implement drug pedigrees. It's no surprise that a special connection has grown between pharma and the EPC/RFID movement. EPC provides a nearly perfect solution for many of pharma's problems with counterfeiting, diversion, and the like, while pharmaceuticals offer a somewhat price-insensitive application for launching EPC/RFID systems.

Electronic Pedigree

EPC/RFID quickly became the anticounterfeiting focus of FDA, the pharma distribution industry, many manufacturers, and even state regulators. The system holds transformational promise to bring safety and security benefits, as well as logistical efficiency savings to the industry. A recent study by Cap Gemini Ernst & Young estimated that the five-year net present value of full adoption in the US drug system would result in a $2.3 billion benefit to manufacturers, $8.2 billion to hospitals, and $660 million to distributors. And in its February 2004 "Combating Counterfeit Drugs" report, FDA said: "RFID tagging of products by manufacturers, wholesalers, and retailers appears to be the most promising approach to reliable product tracking...[it should] provide cost-reducing benefits in areas such as inventory control, while also...mak[ing] the copying of medications either extremely difficult or unprofitable." The report goes on to call for "mass serialization for most packages of pharmaceuticals; and...use of RFID technology...by all manufacturers, all whole- salers, all chain drug stores, all hospitals, and most small retailers [by December 2007]." (See "What's the Timeline?")

The pharma industry is making a serious effort to move EPC/RFID from the drawing board into practice. FDA, the National Association of Boards of Pharmacy (NABP), the National Association of Chain Drug Stores (NACDS), Johnson & Johnson, Pfizer, and many others in the industry, as well as Wal-Mart and the Department of Defense, actively back the use of EPC/RFID.

The Healthcare Distribution Management Association (HDMA) is at the forefront of this effort. In November 2003, HDMA published the white paper "Protecting Safety and Improving Efficiencies in the Health- care Supply Chain: Using Electronic Product Codes," which described the proposed system and background technology and sketched the potential benefits of the system. Around the same time, it convened the Product Safety Task Force (PSTF), an industrywide coalition "to assess technological approaches to combating the existence of counterfeit activity in the healthcare industry, specifically in pharmaceuticals." Early in the process, PSTF identified EPC/RFID as its central focus. Several PSTF members have published studies, comments, and position papers supporting the use of EPC/RFID.

There are several preliminary pharma-based EPC/RFID initiatives under way. Possibly the most developed is the Wal-Mart "C2 mandate" that required its suppliers of Schedule II drugs to begin EPC/RFID tagging individual units by March 2004. Although the initiative has received less public attention than the retailer's top 100 supplier case/pallet tagging requirement, it is essentially a small but functional prescription drug electronic tracking system, designed to facilitate inventory control and Drug Enforcement Agency (DEA) audit requirements. Wal-Mart's application is a microcosm of the prescription drug marketplace, involving only one customer and a limited number of products.

Resources

Another visible initiative is the Accenture Jump Start program. In the summer of 2004, 11 industry participants (five pharma companies, two distributors, two chain pharmacies, and two industry associations) began moving a set of EPC/RFID tagged, commercially packaged medicines from manufacturer to pharmacy, using the tags to track their movement. This test program will be "challenged" with simulated counterfeits, recalls, and other incidents requiring intercompany exchange of information. Accenture plans to present the results in the fall, and several more pharma manufacturers and other companies are in line to join the next phase of the project.

On July 27, 2004, Cap Gemini and SupplyScape announced the opening of the first pharma RFID center of excellence focused on anticounterfeiting. The center was established to allow pharma companies to pilot their own drug pedigree authentication programs.

In addition to HDMA and the other groups mentioned, two organizations are central to efforts to implement EPC/RFID in the pharma industry. On July 1, 2004, EPCglobal held an organizational meeting of its Health Care Action Group, which will work on business structures and practices, including the implementation of electronic pedigree. And MIT Auto-ID Labs has created the Healthcare Research Initiative (HRI), to research various technical aspects of EPC/RFID in healthcare-for example, the effect of RF energy on product stability and RF collision in the hospital environment. Both organizations are open to industry membership.

The Practicalities of Progress

So the stage is set. Now, the primary questions are: How quickly can a full prescription medicine supply chain system be implemented? What barriers stand in the way? How can they realistically be overcome? And, what interim steps are available if the full implementation schedule slips? Unfortunately the answers to these tough questions are not entirely clear.

On July 13, 2004, PSTF released "The Healthcare Indus- try's Business Requirements for Electronic Track and Trace (An Anticounterfeiting Strategy)." The paper was intended to summarize the business issues to be resolved in implementing electronic track-and-trace and identify the next steps industry needs to take. The list includes at least 26 major issues and more than 60 sub-issues grouped into four main categories: technology, implementation, business processes, and governance. The following sections outline just a few of the critical issues that must be resolved.

Mass serialization: the naming of products. The core of the EPC system is the serialization/naming of products. The PSTF paper calls it "the most critical component of the track-and-trace system." "

Although existing EPC standards provide a basic protocol (a 96-bit code that identifies the company, product, and serial number), there is no agreement within the industry on exactly what information should be used in the "product" field.

Many argue for using FDA National Drug Code (NDC) numbers, which already play various roles in prescribing and administering drugs and would thus be easy to integrate with current practices. Privacy advocates, however, point out that if NDCs are used, third parties might be able to determine what drugs an individual was taking by reading data from bottles carried by the patient. Privacy concerns of this sort have led some to propose labeling products with a random number that does not directly reveal identity of the product but can be looked up in an associated secure database. That approach would be compatible with the overall conceptual design of EPC/RFID, but it could put greater demands on the system's information architecture and lead to a more complex and expensive data management system. Other suggestions -such as securing the information on the chip-would necessitate new security standards for EPC devices and would significantly increase the cost of both chips and readers.

It is not clear exactly how and when the naming standard will be developed. If the industry can agree quickly, the issue may not present much of a hurdle. On the other hand, if FDA has to step in and adopt a standard by regulation, with the required notice and comment periods and possible litigation, the process could take years.

Database architecture: pedigree representation. Another basic issue concerns how the track-and-trace database will be constructed. Four approaches are under consideration:

• a central database, managed by a third party, that contains pointers to the current location and past history of every product

• " company-by-company databases maintained by manufacturers or distributors and containing the current location and past history of all products produced or distributed by that company

• a distributed, or "daisy chain" (one-up, one-down), database in which each company (manufacturers and distributors) records who it received the product from and to whom it was sent

• an electronic pedigree document that accompanies each product, recording its life history as it travels through the supply chain.

Each approach has different economic, legal, and regulatory implications. The central-database approach requires the selection of a single entity to manage a shared service for all products-EPCglobal and VeriSign are the obvious choices. Access to the data under this approach would have to be strictly controlled, because no company would want to share its information with competitors.

The company-by-company approach would distribute the data and eliminate some concerns about an overly ambitious "database in the sky." But it would provide a few supply chain members with a view of the full supply chain and the ability to leverage it into near-real-time control. Clearly this would generate resistance. It could also extend companies' legal responsibilities into the supply chain, for example, by inserting them into the processing of product shipment and receiving up and down the chain, not to mention making them a center for counterfeiting investigations within the distribution chain.

In the daisy-chain model, each company maintains "from whom, to whom" records for each product. NACDS and others support this approach, and it is the system called for under the recent FDA food antiterrorism legislation. One weakness, however, is that if any link in the chain does not maintain full and accurate records, the pedigree breaks down. Also, both Florida and NABP model legislation require full pedigree to precede delivery of the product as medicines move through the distribution chain, so this approach could require changes in the law.

The electronic pedigree document model is a practical adaptation of the current state and federal approaches in an electronic setting. In this approach, manufacturers select the product identity approach that best meets their business needs and provide their customers with a digitally signed document containing the product's EPC number. As the product moves through the supply chain, each transaction is appended to the document and a new digital signature is added to the pedigree as certification of the authenticity of the information. Some of the technological and operational processes required to implement this system are already in place-such as Advanced Ship Notices (ASNs)-possibly making adoption easier. However, the full upstream history of custody, including transaction dates and companies, is available to each participant in the chain, as it is today in Florida. Even though the pedigree does not include pricing or discounting information, companies may resist having this much information available.

Although there are advantages and disadvantages to each approach, the point is that this basic dimension of any future electronic pedigree system has yet to be specified, and implementation depends on agreement or regulatory specification of a single approach.

Operational realities. Several business processes and other pragmatic realities will have to be ironed out to make an EPC/RFID system work. These issues could become a major consideration in the timely adoption of the system.

One example is how the system handles exceptions and inconsistencies. Exceptions will result from a variety of simple errors such as read errors, damaged tags, and equipment failures. How will these eventualities be handled, and who bears the risk of loss and the obligation to correct? "

Cost, of course, is an issue-not just the expense of obtaining tags and applying them to products, but the cost of readers, database systems, and the integration of all the pieces of the system. Some manufacturers currently advocate using bar codes, arguing that the cost of RFID tags and readers is too high and the technology is not ready for the quality standards required for pharma. The bar-code approach is indeed cheaper-by an order of magnitude-and would require only limited upgrades to work with current systems. But bar codes require "line of sight" to read numbers from the package. This would ultimately mean higher labor costs. There would also be additional costs to upgrade to RFID later on. Also, if criminals were able to create or steal a set of correct EPC product codes, it would be easier to print them on bar codes than to source or create encoded RFID chips.

One size may not fit all in the adoption of an electronic pedigree solution. Different companies may adopt different product identity approaches. Bar-code, RFID, and varying approaches to serialization may be used. Furthermore, a manufacturer cannot be guaranteed that a distributor or pharmacy will have the technology in place to electronically read the identity from an RFID chip, so there may have to be redundant marking-bar coding, for instance-for some members of the supply chain.

Some believe that for the pedigree system to be truly comprehensive, the electronic system will need to interoperate with paper processing, so entities that have not implemented an electronic process can both send and receive products. This would require a means for printing, e-mailing, and scanning pedigrees.

Another practical detail: RFID tags can operate at several different frequencies: 13.56 megahertz, 915 megahertz, or 2.3-2.4 gigahertz. The different frequencies will have different prices, functions, and forms. More important, the physics of operating at a given frequency offer certain advantages. For instance, 13.56 tags can be read through liquids and metals but are often limited in range. The higher-frequency tags can have greater read ranges-feet rather than inches.

And then there is the issue of the stability of the pharmaceutical products under prolonged exposure to radio frequency waves. Although the frequencies at which the tags and readers communicate are commonly used in home and industrial environments, the power level is an order of magnitude higher. Fortunately, researchers at MIT have embarked on an industry-sponsored initiative to verify the common belief that RF exposure will not affect the stability of drugs.

The full range of 60-plus issues identified by PSTF is well beyond the scope of this article. The point is that a great deal of work remains to be done in a relatively short time. And the decisions companies must make involve significant investments. An upcoming HDMA Healthcare Foundation study, based on case studies developed by AT Kearney, puts system integration costs alone for a large manufacturer at $10-$16 million and at $3-$16 million for a large distributor, not including hardware, data-processing software, and operating expenses.

The promise on the one hand and the uncertainties and investments required on the other appear to be prompting companies to ask how they can shorten the implementation path, accelerate the benefits, and hedge the uncertainties.

How to Move Forward

It is not surprising that there are substantial uncertainties. The key question is not whether a system will be implemented, but when? And what can be done in the meantime?

A recent AT Kearney presentation on the HDMA study called for the industry to agree on an overall rollout schedule. It outlined an approach that would set a priority schedule for full EPC/RFID implementation with different categories of drugs, beginning with products highly susceptible to counterfeiting (as defined by the NABP and Florida rules), controlled substances, and high-priced/high-charge-back drugs. It also suggested an industry schedule for action. The first item on the list is for HDMA, NACDS, and PhRMA to set a timetable. The presentation suggested that recommendations for technology standards, data standards, and data sharing should all be completed before the end of 2004.

At a recent international conference on pharmaceutical counterfeiting, HDMA's new president, John Gray, reemphasized the organization's commitment to driving adoption of the system. Other industry leaders reemphasized their commitment as well. Nonetheless, there were also repeated references to the possibility that complete implementation will take longer than expected. The following suggestions may help advance different parts of the effort:

Product identity standards. The pharma industry should adopt the EPC standard and extend it to address both privacy and cost concerns. Manufacturers should be able to use either a serialization scheme that includes the NDC product code or omits it. Furthermore, a bar-code standard for representing EPCs either as a redundant or sole marking should be defined.

Electronic pedigree document format. The industry should leverage the legal framework for paper documents established by Florida and the NABP to define the required data fields that need to be shared among companies. It should also adopt the electronic data exchange standards developed by UCC to create a global electronic pedigree document format.

Security infrastructure. The industry should establish a root digital certificate authority for securing electronic pedigree documents. A similar structure was established by DEA for controlled-substance regulations. The root needs to be designed so that it can extend to secure global transactions.

HDMA appears to be driving as hard as possible to build consensus in the industry; EPCglobal Healthcare Action Group and MIT Auto ID Labs Health- care Research Initiative are ready to respond. Their coordinated efforts with users and technology vendors should rapidly create the required technology standards.

There are two other areas in which cooperation should be enhanced: FDA and state pharmacy boards. FDA should work closely with leading states and NABP to develop a national electronic pedigree regulatory framework. A safe and secure supply chain requires both mass serialization and electronic pedigrees. To date, serialization has been an FDA-driven initiative and pedigrees a state-driven initiative. A common regulatory framework for both serialization and pedigrees that starts with manufacturers is needed to achieve a national solution. The industry should help drive this and stimulate state adoption.

HDMA, PhRMA, and NACDS. These organizations need to establish the common business practices to handle exceptions and set consistent best practices. The intercompany processes for resolving damaged tags and data errors need to be efficient and result in fair business practices.

Although many advocates of EPC/RFID do not want to consider investments in implementation projects without a full definition of the standards and complete industry adoption, they must recognize that the gradual integration of the technology does create a progressively safer supply chain. Unless the industry acts, its runs the risk of having deadlines come and go without substantially improving security. Some immediate benefits can be gained by serialization and basic pedigree verification:

EPC serialization. FDA's "Combating Counterfeit Drugs" called mass serialization "the single most powerful tool for securing the supply chain." Even before a full electronic pedigree system is developed, EPC numbers encoded on RFID tags attached to medicines could be verified by pharmacies against manufacturer lists to authenticate product to a high degree of certainty. And this approach would be completely compatible with later full-scale implementation.

Basic pedigree verification. A perfect implementation may require that all items in a case and all cases are read upon each movement of the product. But again, the industry will benefit and patient safety will improve with even a partial implementation. For example, if pedigrees are recorded at a case level and members of the supply chain are provided with information about the serial numbers inside the case, there would be significant improvement in the ability to trace individual products-perhaps enough of an improvement to cut the time required to triangulate on counterfeiters from months to weeks.

As FDA said so clearly in the "Combating Counterfeit Drugs" report, "There is no 'magic bullet.'" The rapidly emerging EPC/RFID system holds extraordinary promise for making electronic pedigree a reality-bringing transparency, security, and efficiency to the pharmaceutical distribution system. It will take the focused efforts of the entire industry to move to the point of full adoption, and it will require honest flexibility to take some of the steps along the path. But it will be worth it in the long run.