Major issues faced by LIS users and providers will be influenced by technologies that few people understand. Laboratory managers should learn as much as they can about systems critical to the operation of the lab.

By Gary F. Braley

The health care environment is placing continued stress on suppliers and users of all types of information systems. Vendors respond with increasingly capable and complex systems; users react favorably to these capabilities but may not be able to provide the additional resources required to install and maintain such complex products.

The technology of the modern laboratory information system (LIS) has come a long way since the punched card and magnetic tape versions of the mid-1960s. Many feel, however, that the law of diminishing returns is catching up with us, as time saved in lab operations due to increased LIS capability is offset by increased resources required for ancillary tasks, including maintenance, training, installation and upgrades. While we demand a high level of sophistication in the products we purchase -- whether LIS or PC network -- we do not show the same level of sophistication in the way we manage the purchase and operation with these systems.

This article discusses the competing pressures and technology advances faced by those who must deal with what is arguably the most sophisticated technology employed by health care providers today. Since it is impossible to divorce the LIS from information technology in general, a variety of computer topics are discussed so the reader can understand the factors that will influence information systems of all types, including the LIS of the 21st century.

There is no question the Internet will affect virtually every aspect of information processing, communication and storage in the next few years. At one extreme, major computer systems will be accessed online or will, in fact, be converted to Internet applications. At the other extreme, users will employ numerous devices to access information on the Net.

The new paradigm will be characterized by variety -- no one method or approach will dominate. Users will communicate wirelessly with specialized devices as PCs provide routine services. Individuals will communicate with each other and with businesses at the same time as business systems communicate directly over the Internet. Peer-to-peer (P2P) networking will change the way we think about our computer.

P2P used to mean simple, in-office networks connected for file sharing with no main "server." The Internet has dramatically expanded this concept to the point where underutilized computers around the world can be set up to share resources.

And the Internet itself will change dramatically as high-speed, high-bandwidth infrastructure becomes widespread.

In conjunction with the rapid ramp-up in Internet communications capabilities, few IT concepts have caught on faster than the application service provider (ASP). However, the ASP is actually an old concept with a new justification.

Thirty years ago users were intimidated by the size and complexity of any computer system, including an LIS. Meditech was a leader in the development of remote processing that allowed users access to LIS capabilities without maintenance of an onsite system. Other systems in health care and many business service bureaus operated this way in the early years.

As the size of computers decreased, maintenance problems were reduced and users became more comfortable with onsite systems. Users also expected more powerful systems with significantly greater data flow between user devices (terminals and instruments) and the central processor. Telecommunication technology in the 1980s was incapable of delivering such service at a price users could afford.

What changed to revive the ASP approach? Why is a shared service once again in vogue? Users are no longer intimidated by the hardware, but they are increasingly concerned with software issues. Although some vendors have expressed concern about the ability of the Oracle database to serve the needs of a clinical laboratory -- primarily due to performance questions — such a database will gradually be incorporated into many systems.

However, nowadays, database administrators are difficult to find anywhere and unaffordable by many health care providers. The ASP solution has been proposed to address this issue. ASP providers promise to run your system and take care of the database.

Other issues are also pushing the movement. Software testing and upgrades might be handled more effectively by the vendor’s staff. Backup and other maintenance tasks could be performed by specialists, not the user’s third-shift operator who may not have sufficient experience to handle such a critical task.

Before looking at these forms of "justification" in detail, we should note one additional potential benefit. It has been suggested that having an off-site system would make it easier to throw out a "bad" system. Anyone who has ever installed an LIS realizes how little effort goes into installation of the processor compared to numerous hours devoted to vendor selection, report design, process improvement, file builds, user training, terminal installation, data conversion and interface development.

Lori Opsal, director of Acute Care Systems at the 11-hospital Allina Health System, with facilities in Minnesota and Wisconsin, justified an ASP in one instance but could not do so in the laboratory. She noted, "ASPs will likely take off first for small hospitals in markets where information specialists are hard to find. I also think some applications lend themselves better to transaction-based charging than others."

While the fact that an ASP would shift costs from capital to operating was attractive to Allina Health System, Opsal said that lifetime costs for a full LIS would be significantly higher than the traditional approach.

Occasionally, vendors will use the term "standard system" when talking about their ASP. Every major LIS vendor has at some time tried to sell a low- cost system by providing standard data files for the user, including standard test files, report formats, labels, etc. The user had limited ability to modify these files at the low price quoted. While the future of information systems may rely on more standardized procedures, it has not been an approach that was well received in the past. A true price comparison should include functionality in the ASP model that is equivalent to that which the user would prefer from a traditional system -- more than likely one that does not include standardized tables and formats.

The prospective client should think about some important questions:

• Where will you save money? Programmers cost money no matter where they live or work. They won’t be a free service if provided by the vendor. Will remote hardware result in substantial savings? How much support is required for your current hardware? Aren’t most costs related to application setup, training and maintenance?

• What "infrastructure" problems/guarantees/costs will there be? Redundant telecommunications facilities with guaranteed performance will not come cheap. Introducing third-party communication services into the mix could lead to "finger pointing" when problems occur.

Information technologies that work in other fields do not always translate well or quickly into the laboratory. The benefits are not proven and the total costs are uncertain. It’s difficult to perform a cost-benefit analysis when neither the costs nor the benefits are known.

Some vendors are still trying to convert to client/server, which has been a grueling process with huge barriers along the way. Any technology conversion in the LIS field as complex as switching to an ASP model could be equally traumatic. Some would say it’s just a matter of moving the machine from one location to another but, in fact, the entire service approach by the vendor must change and this will not be easy.

Information services may some day convert to a "utility" model -- one where the service is piped in much like electricity and natural gas, but that’s not likely to happen soon. Unquestionably, a large number of lab services will be handled online, including test ordering, results reporting, training, etc. Whether or when the LIS will move to an online ASP model remains to be seen.

Two powerful forces are pulling in opposite directions. LIS software is becoming increasingly complex while personnel resources to install and maintain the systems are dwindling.

Complexity results from not only "feature bloat" based on the vendors’ desire to offer the most complete product, but on the long-standing desire by vendors to be the best and offer every possible combination of capabilities to their customers. While vendors and clients alike are increasingly reluctant to customize software because of support issues, thousands of customizations can creep into systems when tables are built to meet the desires of users inside and outside the department.

A third factor involves the increasing complexity of the client organization. An early LIS provided basic capability to a single moderate-sized hospital or reference lab. Now, a dozen hospitals of all sizes may be involved, along with numerous clinics and a reference lab. Replacing a single facility system might have been performed in nine to 12 months in the past, but now a complex provider system requires two years or longer. Combining that lengthy installation with a time-consuming selection and negotiation process means a system conversion can require four years or more. Unfortunately, replacement is often triggered by a vendor discontinuing service on the current system and the quality of that service could decrease dramatically over time.

Additionally, providers’ needs will likely change during those four years, requiring continuous redesign, reorganization and additional costs.

One possible solution is more standardization in the user base, whether it’s standardization throughout the numerous labs of a single user organization or standardization in laboratory testing in general. It would be hard to imagine another industry where vendors attempted to produce high-quality products for hundreds of major clients and give everyone their own custom product or service. Such customization drives up costs and introduces delays at every step of the process. This is not a reference to customizing code since vendors and users generally avoid doing that. It is customization through tables and options that is now causing the problem.

One of the most difficult but necessary steps to evaluating vendors is to judge where they will be in 10 years. The process of judging the likely future of a vendor involves a complex analysis of history, the organization, its technology and current clients.

While new companies with leading-edge technology often appear attractive, they all face one hurdle -- expanding from the proverbial "mom-and-pop" style to a well-structured business. Many LIS vendors today are not much bigger or more important than they were 10 years ago, often because they do not know how or do not want to make that transition. Unfortunately, the perpetually small company will find it harder to compete in an increasingly complex field with a dwindling number of small clients.

It’s possible that the leading vendors will be those who have addressed the "usability" issue mentioned at the beginning of this article. They will provide implementation, training and support systems that use the latest technology to minimize labor and time required in all aspects of LIS installation and operation.

Interfacing among the numerous systems involved in a typical provider organization requires communication and decision-making skills that few organizations possess. They have not developed processes to cope with the problems of interface design; managers are not trained in the required techniques and most people turn pale when programmers begin to discuss interfacing protocols.

At the same time, most health care organizations select modules from different vendors. Often, these vendors are competitors. Few vendors have the resources to attack all the interfacing problems they face so they are forced to pick and choose. A vendor that is preparing to discontinue a product or that has only a limited number of units in the field is not likely to put much energy into a complex interface. It is important for users to pick vendors based on long-term relationships and deal with them as partners. A vendor that feels cheated in negotiations or other dealings will not be as likely to help with an interface, particularly if the new product was selected in competition with one of their own.

While speed and capacity go up and price goes down, other aspects of PC technology have remained constant for many years. A major change is occurring, though, in how desktop computers connect to all types of peripherals. In the past, you generally had a choice of serial port, parallel port or the Small Computer Systems Interface (SCSI). USB (Universal Serial Bus), a vastly superior technology for low-speed devices, has been around for several years but until Apple in 1998 began shipping computers with no other interface, the technology languished on the shelf. Most PCs today offer USB ports and a high percentage of printers, scanners and digital cameras provide such a connection.

At the high-speed end, the SCSI interface has always been difficult to work with but it was the "only game in town" for many years. Now there are two new competing technologies and SCSI is not likely to survive much longer. One of these is generally referred to as IEEE-1394 (the industry standard code), iLink (Sony) or Firewire (Apple). PCs are beginning to include IEEE-1394 and virtually all digital cameras advertise Firewire connections. However, a group of companies led by Intel would like to see USB-2 (a faster USB product) used, so it is likely that IEEE-1394 and USB will co-exist for the next few years.

USB and IEEE-1394 both offer the benefits of being hot swappable (you do not have to shut down to install them), the possibility of connecting dozens of devices to a PC in one "daisy chain" and commonality among platforms so the same printer or scanner will work on any type of computer.

While these are not strictly speaking LIS issues, most lab systems include dozens of PCs so the technology of these devices cannot be ignored, particularly since many interfaces such as barcode wands and instruments use PC connections.

For many years, P2P networks were considered the poor cousins of client/server technology. A few PCs connected in a small office with no central server or controlling machine constituted a typical P2P network. These networks are now spreading across the Internet and changing dramatically as they do. The controversial music-sharing service Napster is the best known example, but more important applications are under development.

Since most PCs are idle most of the time (usually all night and frequently during the day), it should be possible to link PCs in a way that others could tap this valuable resource. Why not let your computer work for someone else when you’re not using it, provided, of course, you are compensated and strict security is in place for your protection? Many companies, including Datasynapse, Entropia, Popular Power and United Devices, are attempting to construct such systems while Intel has formed an industry consortium to develop standards and protocols for such applications.

Numerous core technologies and specific products will come into play in the next few years. Although not developed for the health care market, they will impact many industries -- and health care will be among them.

- Voice dictation. This technology may be coming into its own. Previous systems required users to pause between words so that software could determine the beginning and end of each word.

Modern software can distinguish words without pauses and provide reasonably high rates of recognition -- rates that improve with practice. The user gets better and the software "learns" as well. The one caveat is that users must speak clearly and not utter extraneous sounds. The environment can also be critical. While it might not be difficult to speak to your computer in the privacy of your home, voice dictation in a busy laboratory or emergency room could be a different matter given the sensitivity of most systems to background noise.

Pathologists and radiologists have long contemplated this application and there have been a number of successes in the past. In the future, the driving force may not be that the technology is available, but that the alternative -- standard dictation and transcription -- is becoming increasingly expensive and time-consuming.

- Digital imaging. Digital images, whether in radiology, pathology or point-and-shoot consumer cameras, offer several advantages over traditional "analog" methods. These new images consist of absolutely precise patterns of 1’s and 0’s that can be copied and recopied with no loss of quality. In addition, they require less space for storage and can be transmitted around the world at high speed and low cost.

The primary limitations of digital imaging are cost and resolution. Pixels in an image are analogous to grains in film. Photos taken with high-speed film have graininess that is obvious upon enlargement, just the same as the pixels (dots) in a low-resolution digital image show up when enlarged. Dramatic gains in resolution in the last two years have resulted in digital cameras with more-than-adequate resolution and prices that are plummeting. Considering the significant savings in the cost of film and processing, digital photography can no longer be ignored due to cost. Medical applications require a certain image resolution based on the detail that must be observed. As that level becomes attainable at reasonable cost, there will be no reason to continue film-based systems.

- Medical record automation. It is likely the medical record will continue to be fragmented for a long time and in only isolated cases will we see the type of integrated format we would envision as best serving the patient and physician. Whatever the future in this important area, the modern LIS will be required to respond to a variety of new interfacing requirements to supply accurate, timely information for these systems.

- Hand-held and wireless services. Vendors have struggled for years to find the right combination of size, price, functionality, user interface and connectivity to crack the hand-held market. Palm computing leads in the small, general-purpose device category, but there’s a lot of room for others in this wide-open space. Many people predict that wireless connections will become pervasive in a few years. No doubt this will be an important trend, but old architectures are difficult to supplant and wired devices of some type will be around for a long time.

- XML, WML and other new languages. As the Internet begins to dominate virtually every aspect of IT, a host of new technologies are being introduced. Wireless Markup Language (WML), Web Clipping (Palm Pilot Web format) and eXtensible Markup Language (XML) go beyond HTML to greatly extend the reach and power of the Internet. XML, in particular, will significantly alter the way organizations exchange information, including health records.

These advances generally complicate the life of Web developers since they are primarily ways of offering multiple displays of the same information. This means a Web site might have to be coded in several different ways if it is to be accessible by individuals using a cell phone, PDA and traditional computer.

- Free services. Fifteen years go, cost was a major impediment to online communication. Compuserve, the earliest popular information service, charged from $6 to $15 per hour depending on time of day and modem speed. Twelve hundred bits per second (bps) was considered high speed compared to the current dial up rates of 56 thousand bps. Most early users could only afford a 300 bps connection. And these costs did not include long distance phone charges that were often required or extra charges for special services such as database research.

Because digital communication today is so inexpensive, many services are being provided free. The hope is that they will be supported by advertising, much like commercial television. These free services include disk storage and file sharing, group calendars, meeting planners, personal health records, e-mail and Internet access. When costs-per-user drop so low, it eventually becomes a burden to collect pennies per transaction from each user and it is better to develop another cost model. Advertising and other marketing-related plans are thought to be the light at the end of the tunnel.

- Convergence. Information processing and communication technologies are evolving rapidly and in the process they tend to merge with other technologies to produce cell phones that can access Web pages, PCs that show movies, telephone calls on the Internet, and LIS and other health care capabilities provided as online (ASP) services. This "convergence" makes it increasingly difficult for purchasers of information systems to decide what they really need. This trend represents the latest complication in the fuzzy areas we saw in the past where a scheduling system might be purchased from a radiology system provider or as part of the hospital information system; the transcription package might be standalone or part of the anatomic pathology system.

At the same time, we see specialized "appliances" developing to solve the ease-of-use problem. One way to make a device -- whether PC, cell phone or toaster -- easier to use is to limit its functionality. A Web access screen on the refrigerator or microwave could be used only to retrieve recipes, a device that has already been demonstrated. The competition between convergence/complexity and specialization typified by Web appliances is likely to continue far into the future.

- PC hardware, operating systems. A high percentage of the gains from desktop computers occurred many years ago when the first word processors and spreadsheets were introduced. Few managers or users today would argue that the latest software upgrade -- whether it’s an operating system or application -- was worth the cost. This is one reason many users still run Windows 3.1 and obsolete versions of their applications.

Apple also has just released an entirely new operating system, OS-X, based on Unix as its standard desktop OS. This software combined with the fact that many professional Apple desktop computers will contain multiple processors could offer significant performance and stability gains.

Many users of current PC technology do so primarily because they see no other choice. That may change now that Linux, a Unix-based PC operating system that’s been around for many years, is capturing an increasing share of the market. Although it’s been a favorite of many Internet service providers, PC companies were discouraged from pre-installing it on their machines because Windows was so dominate. Today, many companies offer a choice of Windows or Linux. Later this year, Microsoft is likely to release the next version of Windows -- the first consumer version to be free of the DOS legacy. This product, code named "Whistler," named Windows XP, will represent a bigger change than the Windows 95 and 98 products did so it may not be easy for users to convert; many will choose to stay with the product that came with their computer.

On the hardware side a new firm, Transmeta, began selling a CPU, named Crusoe, aimed at the portable market where low power consumption is essential. Features of the most powerful processors are often disabled when used in portables to save power and reduce heat. A new player entering the field at a time when Intel is fighting with AMD for bragging rights about CPU power and laboring to introduce its next generation technology could be disruptive. The Intel/HP processor development effort to replace the x86 architecture that began in 1994 (which has suffered repeated delays) is finally bearing fruit in the form of the Itanium processor. Intel announced this new chip would be targeted at servers and high-end workstations, not the general desktop as once thought.

Many experts, including Gordon Moore, founder of Intel and creator of "Moore’s Law" concerning the doubling of transistors on a chip every 18 months, agree that there are basic physical limitations as to how small circuits can be made using silicon technology. Fortunately, molecular electronics may save the day. Scientists Jim Tour at Rice University and Mark Reed at Yale University say they believe individual molecules will act as switches in electronic devices much as transistors do today. The concept originated in a 1970s paper by IBM researchers but not until Tour and Reed began serious study in 1991 did anyone foresee the possibilities. Today, major corporations and universities are investing heavily in research into this field that could result in computers thousands of times faster than current devices manufactured at a tiny fraction of current costs.

Meanwhile, the Internet is praised for its power and criticized for being too slow. Billions of dollars are being poured into the infrastructure -- cables and switches that control the flow of information among millions of computers. Two projects referred to as Internet 2 (I2) and Next Generation Internet (NGI) are addressing current limitations. I2, driven by university research and education requirements, and NGI, a federal program announced by President Clinton in 1996, will look at all the limitations of the Internet and assist in developing ways to overcome those limitations. These are not quick fixes, but the way the Internet works 10 years from now, and specifically how it might function as a remote computing service such as an ASP, will in great part be due to these efforts.

Major issues faced in the future by LIS users and providers will be influenced by numerous technologies that few people understand today. It’s imperative that laboratory professionals learn as much as they can about these subjects if they are going to be masters -- not servants -- of the systems that are critical to the operation of the laboratory. Because many of the future issues revolve around usability and not functionality, users have a real opportunity to get involved and determine the direction of the LIS of the 21st century.