Middleware—computer software that connects software components or applications—allows multiple processes running on one or more machines to interact across a network. For clinical labs, middleware is used to connect laboratory information systems (LIS) with the analyzers that do the actual lab testing. It’s an interface between the two.
But middleware has gotten more complex, and its functions have expanded vastly. In fact, middleware is a term that is difficult for most in the health care industry to define, and some, like Gregory Vail, president of Data Innovations Inc, South Burlington, Vt, are trying to redefine it because the very term can be misleading.
Curt Johnson, vice president of sales and marketing at Orchard Software, Carmel, Ind, says the difficulty in defining middleware is that it performs so many (or so few) functions, depending on the users’ needs, that a single definition really cannot apply. He says “middleware” is a broad and sometimes vague term for software attached to the LIS to accomplish tasks that the LIS cannot do, or cannot do as well as middleware. And it can be outreach software in front of the LIS, or it can sit between systems or exist at the back end.
A typical use is for middleware to sit between a LIS and analyzers in labs to enhance analyzers’ capabilities and take advantage of the features that a LIS offers.
Middleware as a Messenger
According to Steward Macis, director of products and support services for Antek HealthWare LLC, Reisterstown, Md, the basic definition of middleware can be stated as a way to provide messaging services allowing different applications to communicate or the systematic tying together of disparate applications. From the lab perspective, middleware often serves as an intermediary between the analyzer and the LIS or from the LIS to a third-party software solution such as an HIS, an EMR, or a PMS. In addition to converting data for disparate systems, middleware often applies advanced rules to help facilitate functionality that may not be available in the LIS.
“It is certainly a good solution for an aging data-management system,” Macis says. “As an example, a hospital may have an outdated LIS and incorporate middleware to improve functionality. This will allow it to stay relevant by addressing certain issues for a few years but will not achieve the functionality and productivity a comprehensive and full-featured LIS can provide. In the coming years, the need for middleware will grow as the adoption rate of EMRs will increase. Middleware will provide the necessary technology of directly interfacing point-of-care testing.”
He adds that in larger laboratory facilities such as hospitals or large community clinics, middleware is often used to add more sophisticated functionality to a legacy LIS with features such as auto processing or advanced rules technology. In the physician’s office laboratory (POL) environment, middleware can be used to capture and manage laboratory data from clinical analyzers and exchange demographics, orders, and lab results to an EMR or a HIS. Other uses of middleware in the POL environment involve interfacing billing information with the HIS, EMR, and PMS. In the smallest environments where waived testing is allowed, middleware is used to connect point-of-care testing (POCT) with various health care management systems.
Role in Automation
Macis notes that middleware is designed to automate processes that allow for less user intervention. In the larger hospital facilities, most training is concentrated on configuring the middleware to best fit the current workflow. Depending on the complexity of the workflow that is in use, training could take as little as a few days to as much as several weeks. However, validating a comprehensive rules set may take several months. For POLs and waived testing environments, the same rule applies: The amount of training involved depends on how the middleware is being used. The users who actually work with the software can range from laboratory managers and technicians to various health care professionals as well as information technology (IT) staff at the facility.
Today’s middleware even incorporates some features that LIS may (or may not) include. For example, if a particular lab’s LIS doesn’t feature inspection preparedness, there is middleware that does. Most middleware is modular and can be made as sophisticated as a lab needs it to be, or can afford.
Vail says the term “middleware” came from the interface software used between LIS and instruments, in the days when a lot of protocol conversion was being done. Middleware puts data into plain, consistent formats.
But now, he says, things are more sophisticated. Data Innovations’ products, for instance, are complete workflow managers, used for preanalytic, analytic, and postanalytic stages. Today’s middleware can connect EMRs, LISs, instruments, automation, and more to make labs as efficient as possible. And although any software that connects other software has become lumped into the “middleware” term, Vail says he’s trying to redefine it as software in the “middle” of lab operations.
Middleware’s usage is not limited to connecting LIS and analyzers, agrees Anne Tate, senior product manager for IT products at Sysmex America Inc, Mundelein, Ill. It is also a bridge to and between automation systems, according to Tate.
“Automation systems today are increasingly sophisticated and more complex. New and updated technologies such as sorters, return lines, and other elements make it challenging for the LIS vendors to keep pace. At Sysmex our customers also want a very high level of integration between equipment, workflow, and software. The pressure of a decreased workforce in the laboratory is very much driving the need for this kind of middleware,” Tate notes.
Getting ROI from Middleware
Data Innovations’ Vail says the uses of middleware depend on an individual lab’s needs. Some uses of middleware are so basic that virtually no training is required to operate it, while with other uses, such as autoverification, training may be required to learn the concepts of writing rules and to get them added to and tested in the system.
“To allow customers to get the best ROI from our products, we offer classroom training, both in our facilities and on-site, along with e-learning modules. We have put a lot of effort into ramping up our training offerings,” Vail says.
He notes that middleware users run the gamut from physicians’ labs with only a couple of instruments to large reference laboratories, such as Data Innovations’ customers Quest Diagnostics and the Mayo Clinic. Because of its FDA 510(k) clearance, Data Innovations is able to offer its middleware functionality to blood banking and screening—not areas where middleware has traditionally been used before, he says. Data Innovations partners with most of the IVDs that have molecular arms, and its middleware is utilized as instrumentation traditionally used in academia. Flow cytometry is also a fast-growing area, he adds.
Orchard’s Johnson notes that many LIS systems have been around a long time and don’t necessarily take advantage of newer software. That’s why separate middleware software packages are available to enhance LISs and analyzers, he says.
And he adds that the newer middleware is meant to enhance analyzers, not LISs. Such middleware is sometimes developed by the analyzer manufacturers themselves and sometimes by other companies. Analyzer buyers and users are also a part of the development process because they tell vendors what they need, he says. Large diagnostic companies developing middleware specifically for the analyzers they make and sell is more appropriate, he says. “That type of middleware should be developed by analyzer companies, although that’s not always the case,” Johnson says.
Services that surround middleware are in increasing demand by labs, Vail says, including guidance in effective implementation. Autoverification is a much sought-after outcome of middleware implementation, but it takes some time to put in place. There are other areas where return on investment can be more rapidly realized, however. For instance, middleware can manage refrigerated samples and save medical techs time in locating particular samples—and the process can be automated very quickly.
“To address the time to implement auto verification, we’re developing base sets of rules via our consulting process that address requirements that are similar for many labs so that we can more easily help them save time and money,” he says.
But a major area of concern, and one that middleware can help significantly to alleviate, is inspection preparedness. The US government mandates several organizations, notably the College of American Pathologists, to conduct lab instrument inspections with 1-week notices, Tate says.
According to Vail, inspections are made to affirm lab accreditation by ensuring that labs are adhering to good practices. Data Innovations software can schedule maintenance, automate troubleshooting, and track both scheduled and unscheduled events to make sure that labs are inspection-ready at all times. Schedules can be customized and work something like the Outlook calendar, prompting inspection and maintenance at user-defined times.
Leslie Dakarian, MT (ASCP), senior manager, Global Informatics Marketing, Siemens Healthcare Diagnostics, Deerfield, Ill, says that for accreditation, labs need to ensure they access up-to-date documents and procedure manuals. With Web-based applications, Siemens’ middleware can utilize the Internet to send software version updates or provide access to current package inserts, as well as update user guides and other information stored in its middleware.
Driving quick, electronic updates to a central location with prompt removal of out-of-date information raises the quality of service provided. It also can be used to help meet quality or patient safety initiatives for accreditation purposed with features like stat alerts or required commenting, which ensures that appropriate documentation procedures are followed. Patient identification check software is another use of Diagnostics IT products that can help ensure accuracy in correct collection and labeling of specimens and samples.
Role in QC
For quality control management, middleware can streamline data management, consolidate date into a single location, and offer options that allow labs to identify QC failure or instruments issues earlier. It can also store 2 years’ worth of records for inspections.
Eric Olson, vice president, Informatics & eBusiness, Siemens Healthcare Diagnostics, says that middleware also can provide operator training and keep track of which techs have taken which qualifying courses for operating certain systems. “Some middleware products also offer audit trails that enable a lab to identify who did what when, which is extremely valuable from an accreditation perspective,” he says.
And some middleware can also drive consistency throughout a hospital’s network. She says that browser-based systems enable hospitals to compare and analyze QC and other data across hospital departments or sites, to do such things as meet standards for turnaround time. Database export features offer easy access to data for data mining, which can determine what is causing delays, Dakarian says.
Macis says that with middleware, “You should validate the software in the same manner you validate other software applications. Extensive testing and validation is recommended to be certain that the software will perform reliably in reference to the workflow that is designed by each practice. Testing the software in every scenario that may pertain to the practice workflow is highly recommended.
“Inspections occur on the practice itself, and not on the software,” he continues. “Since the software is a part of the patient care process, the middleware will have to demonstrate that it performs its designed duties as specified in the practice’s standard operation procedures. In order to stay prepared for inspections, it is recommended that a comprehensive audit be performed routinely to validate that the middleware is performing optimally.”
What’s the future of middleware, and why can’t analyzers operate without it?
Tate says that more and more manual tasks will be off-loaded to technology. This will mean that more intelligence will be built into laboratory analyzers and equipment, and that in turn will drive an increase in the need for more sophisticated software, including middleware systems. “At Sysmex we have seen a tremendous demand for middleware, not only to support a single laboratory but labs want systems that manage an array of analyzers: automation systems and single analyzers, across multiple sites and locations. This makes network connectivity increasingly important,” he says.
“At Sysmex we are collaborating more closely with our customers’ IT departments and LIS vendors than ever before,” Tate continues. “When that happens, the lines begin to blur between equipment and information systems, and that’s when real operational savings can be achieved and the best information becomes available to the clinician for higher-quality patient care.”
One such customer, Laurie Emanuel, lead hematology tech for the University of Texas Medical Branch (UTMB), Galveston, says the benefit of Sysmex’s WAM hematology information system to her lab is that “it’s the birthplace of virtual hematology of the future, because it makes streamlining and managing workflow possible.” The Sysmex WAM manages the data flow for orders and results between the CellaVision digital hematology cell imaging device.
Sysmex WAM consolidates the CellaVision automated differential results with the hemogram instrument results and can apply integrated reporting logic across these two data sets. A physician within UTMB can view results on the CellaVision workstation remotely and view the combined Sysmex instrumentation and digital cell manual differential data without having to look at two different data sources.
Future Directions in Middleware
Orchard is actually eliminating the need for middleware in some of its new LIS offerings by building middleware features into them. Users of these LISs will not need middleware. With old LIS systems and middleware, though, and the billions of dollars invested in the systems and the mountains of data they store, replacing them with the newer middleware-inclusive LISs could be a difficult proposition. For labs starting fresh, though, the middleware-inclusive LISs are an exciting advance.
If a LIS has the functions the analyzer needs, he explains, labs do not need the middleware software. Typically, a legacy LIS requires middleware with up-to-date systems. But with up-to-date LISs, labs may not need the middleware.
For some labs, it’s helpful to imagine middleware as a Band-Aid, Johnson says; for them, cost, lost data, and time are factors, and middleware may be the answer for them to enhance legacy LIS systems.
Siemens’ Olson says there was a time when LISs and analyzers connected directly. As analyzers became more detailed and added more features and higher levels of functionality, middleware provided an opportunity to support the new analyzer functionality and connect them to LISs.
“But we do see integration coming back. Lots of software programs are being developed that combine middleware features. Siemens makes both LIS and analyzers, so we’re integrating middleware into those systems more tightly and seamlessly,” Olson says, adding that he thought integration would take place by segment, from high-end labs on down, since nearly all labs have LIS but not all have middleware.
For more on middleware and other LIS news, bookmark our website.
And middleware can help labs deal with staffing shortages, Dakarian adds. “We offer scalable solutions. With less funding, middleware fits in nicely to fill the LIS-analyzer gap. And it’s expandable, so a lab can use it with only one instrument or dozens of instruments. It runs the gamut in size and affords opportunities for growth in volume and applications,” she says.
What’s next? Jeff Shockley, group marketing manager for IT solutions for Roche Diagnostics, Indianapolis, says that data-mining enhancements are being sought by more labs, which want to be able to get into analysis or statistical reporting, and tighter integration with analyzers will continue to be a trend.
And consolidation of control, where the customer has a central command screen from which to be able to access information from all analyzers, is another trend that’s getting lots of demand from customers, Shockley says.
Gary Tufel is a contributing writer for CLP.