This focus on adding maximal user-friendliness to a suite of already powerful applications is good news for laboratories of all sizes because it reduces training time and keeps employees happy, but the friendliness of middleware is not limited to its own designated functions. In cases where interaction with the laboratory information system (LIS), hospital ordering system, or off-site database may be confusing and slow, for example, middleware can become the user’s smart friend, acting as a translator that takes simple, natural-language input from the user and turns it into the arcane, complex commands that an older (and less friendly) information system requires.

Enhancing Connectivity

The ability to serve as an interface between systems is at the core of middleware’s functions. Initially, middleware was brought into many laboratories to allow individual instruments to communicate with the LIS or with automation systems. Later, middleware also acted as the translator to make the disparate protocols (such as XML, HL7, and ASTM) used by hospital and laboratory systems work seamlessly together.

Since then, middleware interfaces have been continuously improved, in addition to being extended wherever they are needed. Middleware now connects devices and systems both within the hospital (including central data repositories, reporting systems, specimen-tracking systems, point-of-care devices, bedside identification devices, autoverification systems, and purchasing/inventory systems) and outside it (such as external disease registries, public-health surveillance systems, reference laboratories, and telepathology systems).

Beginning in 2007, small to medium-sized laboratories became major buyers of middleware to interface with the electronic medical record (EMR) and with physicians’ office-management systems. This activity was driven by the rapid adoption of the EMR by referring physicians, making EMR connectivity necessary for laboratories that intend to compete for outreach business generated by those offices. Now, and through 2009, many middleware sales will be based on extending connectivity to microbiology systems and blood-banking systems, where implementation activity seems particularly heavy.

Vendor Capabilities

While the number of installations that a middleware vendor has in place can vary widely, both large and small companies can offer advantages to the laboratory. In this issue’s survey of middleware providers, for example, the installation figures vary by a factor of 50 between the smallest and largest; nonetheless, each company may be providing the ideal software solutions to its clients. Small companies often concentrate on highly challenging installations requiring a great deal of custom programming, while large companies are likely to give customers the benefit of their extensive experience and active user communities.

Experience over time is also an important indicator of general customer satisfaction with a vendor’s middleware offerings, since this is a highly competitive field. A vendor unable to meet its clients’ expectations is unlikely to remain in business for long in this environment, so longevity can be counted as a positive sign. Of course, software companies, by their nature, will always be newer than equipment and supply companies, but middleware has been needed long enough to let many vendors accumulate several years of experience in providing it. Survey respondents, for example, have spent from 5 to 17 years fine-tuning their products since their first installations went live, so they are unlikely to encounter operational problems that are new to them.

This does not mean that a middleware vendor can simply release a product and stop ongoing research and development. Information systems evolve so rapidly that the tasks that middleware is expected to perform change constantly, along with the circumstances under which those tasks must be completed. Each newly released laboratory instrument or LIS update calls for new responses from middleware, so vendors must adapt with great flexibility and speed. They release upgrades on a recurring schedule as accelerated as that seen in any other information-related industry, and they supply single-purpose updates (such as patches and new interfaces) as needed between upgrades. As one might expect in this highly fluid setting, all survey respondents report the recent release of upgrades, some of which are as little as a month old.

Predicting the cost of a middleware installation can be challenging because each vendor has its own methods for determining the pricing structure. Fortunately, vendors can do much of the work involved in providing a cost estimate, so long as the laboratory can state its needs clearly. Middleware companies have a reputation for flexibility, so it may be possible to negotiate favorable package pricing; because middleware competes with LIS upgrades in the laboratory’s information-system budget, however, middleware vendors will almost always price their products at a point lower than that of a LIS upgrade that serves the same purpose.

Most middleware pricing involves an initial licensing fee or a flat rate per interface, but some companies also require annual maintenance fees. Monthly charges based on transaction volume (of orders received or results distributed, for example) may also be assessed. Regardless of the structure involved, middleware is consistently described as a high-value purchase with a rapid return on investment, not only because it can reduce the need for expensive LIS upgrades, but because it can add capabilities that improve the efficiency and productivity of laboratory staff.

Of course, that staff will not get the highest level of benefit from a middleware installation unless the new capabilities are thoroughly understood. Vendors compete today based on the ease with which their middleware can be used by novices, making user-friendliness a major selling point. This is greatly to the advantage of laboratories where personnel may have a wide range of computer experience, ranging from next to none to many years of high-level use.

Given the staff shortages that many laboratories face, easy-to-use middleware can be very helpful by making it possible to hire computer novices. An older LIS can be quite opaque to the untrained new employee, for example, but a friendly middleware interface can act as an interpreter, letting intuitive interaction with the system replace a memorized series of commands and getting the new user up to speed quickly.

Naturally, the middleware company must be prepared to provide solid training for new users and to see that good technical support is there if questions or problems arise. All survey respondents provide online training, and some supplement this with on-site training during installation and scheduled classroom training at regular intervals. Documentation is usually supplied, in addition to traditional technical support. Some vendors help users help one another (and themselves) by archiving past questions and maintaining searchable databases of problems that users can access online.

Ongoing Innovation

Because middleware exists to fill gaps, it must change as those gaps shift. While some gaps narrow as adoption of data standards widens and as older information systems and instruments are replaced, other gulfs open as new applications move through the implementation cycle from wished-for capability to cutting-edge ability to everyday necessity. Most laboratories face pressures to spend less, handle more work, and use scarce staff time efficiently, so today’s middleware is designed to support those efforts. Other needs may arise more suddenly as the requirements of regulatory agencies, accrediting bodies, payors, and referral sources change. This is the arena in which middleware truly shines, since it can be plugged in quickly (and usually inexpensively) to meet unanticipated needs.

While lawmakers and regulators will probably always concoct some surprises for health care, some of the demands that will be placed on the laboratory tomorrow can be predicted. Middleware vendors are already positioning themselves to meet future demands for enhanced connectivity, for fewer proprietary components (and more open systems), and for secure access from any point. Dashboards are already available to help users keep all of the activities that need their attention on a single screen; for example, on-screen alerts that bottlenecks in workflow, or results outside preset ranges, are being seen can let managers know, in real time, that an instrument is failing or a technologist is overloaded. Instant warnings of this kind are likely to become much more common as productivity pressures increase, since a single snag in the laboratory’s production process can have ripple effects that waste both time and money.

Increasingly, middleware will be used to meet reporting demands from external entities automatically, thus saving compilation time and increasing accuracy. Public-health reporting is already done automatically by some middleware. Integrated disease management is likely to increase, and middleware can reduce the associated administrative/clerical burden through automatic generation of the necessary results and reports.

Bookmark our site
To track developments in middleware, bookmark our website.

Middleware also handles decision support, and this particular sphere is likely to expand greatly over the near term. The probability of regulatory and accreditation inspections being conducted with little to no warning continues to grow, and middleware can be a great help in maintaining preparedness for these events. For example, middleware can log and integrate quality-control information for all instruments and processes in real time, all the time, so the laboratory will always have its data ready for review.

Over the long term, middleware can become more than the individual user’s friend. By maximizing efficiency, reducing costs, and keeping track of even the most complex activities in real time, middleware has shown that it can be the good friend of the laboratory, referring office, and health system as well.


Kris Kyes is technical editor of CLP.