Service of Medical Informatics
University Hospitals of Geneva
Identification management is a wide-ranging topic. It is an everyday experience at stores with barcode-based product identification or electromagnetic measures against theft. But that is only part of the story. Electronic chain production is part of the usual processes in large consumer goods industries and there is growing interest in applying identification management techniques to the healthcare sector as a whole. This is strongly emphasised by the involvement of GS1 (formerly EAN International) in the healthcare domain, and the increasing leveraging efforts of EPCglobal to apply these techniques, including radiofrequency carriers (RFID tags), to healthcare. However, uniform identification management is not used or not available in the majority of the healthcare sector. Here we investigate what the challenges are.
Increased supply chain efficiency Healthcare is at the crossroads of countless supply industries – pharmaceuticals, medical devices and disposable products, but also numerous other suppliers, such as food and groceries, beds, clothes, computers and office equipment. Hospitals are small cities, with stock management and internal markets embedded in the global market, yet they have low efficiency.
Authentication Identification of genuine products is a growing concern, with the increasing problem of counterfeit products in the healthcare sector. This can have a dramatic effect on patient safety, such as the mishandling of infectious diseases.
Identification Applying identification to all products and actors (citizens as patients or care providers) is a major need and challenge, with strong political, cultural and ethical challenges, including privacy, which is not yet solved in any manner.
Prevention of medical errors Since the US Institute of Medicines 1999 report on medical errors, safety has received considerable interest. Better processes, with identified actors, are one of the promising ways to address these concerns. For example, encoding of the drug unit dose or unit of use package to enable automated verification and ensure the right dose for the right patient at the right time is only possible if identification is fulfilled.
Tracking and tracing The use of an interoperable framework of identification schemes allowing identification of actors, objects, locations and actions in a temporal net of events is of numerous benefits: building the cascade of events leading to an unwanted or unexpected incident, improving billing, understanding flows and processes, leveraging clinical research, boosting epidemiology by facilitating the detection of causal links, to name but a few. Track and trace is a very important goal that is reachable only if global identification is achieved.
Healthcare is a complex sector, and it is more complex than what is usually the case in consumer goods production. There have been many unsuccessful attempts from industries to address this issue, mostly using information technologies.
Industrial suppliers There are countless industrial sectors involved in healthcare, from goods production to high technology and food supply. The health sector is a transforming production industry, though often considered as an end-user industry. It is a convergence sector, at the crossroads of numerous production chains. The healthcare sector should rely on object identification made during the manufacturing process and already carried by objects, which implies that there must be an informational integrity and interoperability of all these converging chains.
Coverage Most objects carry some kind of identification. Numerous drug packaging systems have EAN barcodes. However, this encoding does not cover all objects; many of them do not carry any encoding, or have incomplete, nonstandard, unreadable or unreliable encoding.
Granularity In many cases, such as drugs or disposables, encoding is at the packaging level instead of being on each item. This level is enough to manage the supply chain, but it is not sufficient for end-to-end traceability. Drugs and needles, for example, should carry codes at the level of each single item used for patient care.
Serialisation Usually encoding is devoted to production management. The encoding carries company identification, object category, sometimes batch number and, rarely, further information, such as expiration date. With few exceptions, such as pacemakers, it never includes unique item serialisations.
Durability One of the important characteristics that distinguishes a screw used in healthcare from a screw used in a car is the durability in time. For many reasons, including legal aspects, objects used in the healthcare sector have a durability of around 80 years (in industrialised countries). This duration is, however, increasing for scientific research, case-based databases and genomic aspects. The identification schemes used in healthcare must therefore support persistence in time, which means that once an ID has been used it can never been used again.
Pedigree and traceability US federal laws implement a certified chain of custody, or pedigree, for pharmaceuticals, identifying all steps between the raw production and the final end-user. Such a “pedigree” is in place in Europe with the General Food Law, which entered into force on 1 January 2005. The BSE/”mad cow disease” crisis, which made such political decision possible and also made it economically necessary, shows the feasibility of these approaches at very large scales, even on low-cost items.
The ability to identify every actor, including citizens, is an important challenge. It is a very difficult objective to achieve, but it is required to facilitate the continuity of care and the interoperability of patient data, to decrease errors and to consolidate administrative procedures, including billing. A unique patient and professional identification is required. The realisation of such a unique, global and shared identification leads to important political, legal, ethical and privacy concerns. The trend is to have insurance cards, which provide some unique identification linked to the insurance coverage.
Healthcare as a knowledge industry
Healthcare is a knowledge industry, and medical knowledge is increasing at an incredible rate, as illustrated by the more than 623,000 new references added in 2006 to the Medline database.
Moreover, this is only a small part of the complete information flow in healthcare. There is a strong need for an interoperable semantic framework that can be linked to items. Numerous classifications, nomenclatures and ontologies are used in the healthcare sector, with the highest difficulties being in having converging meanings and coherent representations. To date, nothing indicates that a unique representation is feasible, or desirable. Therefore, in addition to identification, all elements should be linked to one, and often several, representations. Providing an overview of these representations is beyond the scope of this article, but some examples are given to illustrate the point:
- The Logical Observation Identifier Names and Codes (LOINC(®)) database facilitates the exchange of results, such as blood haemoglobin, serum potassium or vital signs, of clinical care and of outcomes management.
- The Systematized Nomenclature of Medicine core terminology (SNOMED CT) provides a common language enabling a consistent way of indexing, storing, retrieving and aggregating clinical data.
- The WHO family of international classifications, the International Classification of Diseases (ICD), the International Classification of Functioning, Disability and Health (ICF) and the International Classification of Health Interventions (ICHI).
- The nursing classifications – observation, diagnosis and intervention.
- And many others, such as Medical Subject Headings (MeSH) and Online Mendelian Inheritance in Man (OMIM).
In addition to having semantic use for the entities used, they should be seen within an overall process model. This can be a patient record, for example. Again, there are several options possible, such as open EHR, the HL7 reference information model or the CEN 13606 prestandard.
The way all these aspects are “carried” or “tagged” on the object is an important technological challenge. A complete discussion of the carriers is unfortunately beyond the scope of this article. There are many possible carriers, such as barcode, data-matrix, RFID tag, smartcard, each subdivided into several types. For example, RFID tags have three different frequency ranges: 125 kHz, 13.56 GHz and 900 mHz. Each range shows specific advantages and disadvantages. Barcodes are commonly used identification carriers, and the EAN.UCC encoding is very frequent. Beside carriers, one cannot miss speaking about readers. Specific readers exist for each type of carrier, and few readers can read several types of carriers. Readers able to read all type of carriers, such as all RFID frequencies, are still uncommon. This is a major problem in healthcare, where care providers, especially nurses, would have to use different readers.
Entity identification and tracking is a major challenge in healthcare. Correct identification of all entities involved, including actors, will increase supply chain efficiency, reduce medical errors, enforce clinical and epidemiological research, and leverage management understanding of processes involved in healthcare. However, there are many issues to be solved. Technical issues, such as serialisation of all items in the supply chain, taking into account the cost of such procedures, are still unsolved; uniform reliable carriers and universal readers are missing. There are political and ethical issues, especially around the identification of actors, citizens and care providers. There are also important issues around semantic and knowledge representation. This is an open field requiring progress in order to reach a sufficient level of function to cover the needs.