CEO SUMMARY: Since 1994, thoughtful lab executives have wrestled with the concept of total automation for their laboratory. Despite concerted marketing efforts by some of the world’s most successful diagnostics manufacturers, only a handful of total laboratory automation (TLA) sites are operating today in the United States. In operation, the economics of TLA failed to deliver on its promise of lower costs. As the curtain falls on this first-generation TLA in 1999, maybe now is the time to share some untold stories about TLA’s pioneering efforts. Despite its limited success, TLA did spur development of a worthwhile successor: modular automated workcells. Here is THE DARK REPORT’S follow-up on this important story.
Story Update From 1997
PROBABLY THE GREATEST untold story in the laboratory industry today is the failure of total laboratory automation (TLA) to deliver on its promise of higher productivity, lower cost, and improved quality.
Since 1995, THE DARK REPORT has counseled that most laboratories should take a “go slow” approach towards total laboratory automation. Our reasons were clear and we felt the arguments were compelling.
The case against total laboratory automation was best presented in our classic An Industrial Engineer Looks At Laboratory Automation And Robotics. It was published in the February 17, 1997 issue of THE DARK REPORT (back issues available complementary to clients and subscribers upon request.)
Authored by Mark Smythe, an industrial engineer with four decades of experience in manufacturing, distribution, and clinical laboratory operations, it made three key points about TLA.
First, Smythe pointed out that the typical clinical laboratory in the United States, by the fundamental design of its organization and production throughput, does not have the intrinsic potential to benefit from total automation.
For comparison, he noted that a car manufacturing plant will operate at full production volume for three shifts (24 hours per day), six and seven days per week. A clinical laboratory operates at maximum volume only one shift per day (8 hours) and only five days per week. Thus, the opportunity to amortize the cost of automated equipment over a huge production run is severely limited for laboratories.
Second, Smythe also noted that total automation tends to “freeze”? the production process of the laboratory. Once expensive automated equipment is installed, the laboratory finds it difficult to alter processes and workflows which might generate incre- mental productivity and cost improve-ments. In that respect, a TLA installation “handcuffs”? the laboratory to its expensive equipment, for better or for worse.
Smythe’s third point was most telling. “First, whatever automation equipment is chosen for a laboratory, the manufacturer should specify an expected return on investment (ROI),” wrote Smythe. “Calculations to arrive at this number should be clearly understood. Both the laboratory buyer and the vendor should be prepared to work together to achieve that ROI.”?
ROI Data Never Made Public
At the time he wrote the article, Smythe noted that public information about the ROI of existing TLA installations was notably absent. “I eagerly scan the clinical laboratory press for detailed financial documenta- tion as to how laboratory automation has reduced costs, improved productivity, and delivered a market return on investment to those few laboratories which have pioneered the installation of such technology. Such documentation is not forthcoming…”
Smythe’s assessment? “It would be a reasonable conclusion that neither the automation vendor nor the laboratory customer is totally satisfied with the performance of their laboratory automation systems to date.”
Written in 1997, Smythe’s words stand just as true in 1999. During the past two years, no vendor or laboratory user has been brave enough to publish a rigorous ROI analysis of an operational TLA laboratory installation.
Because every successful TLA installation represents $2 to $4 million in sales to a vendor, it is reasonable to assume that laboratories with operational TLA systems would be quick to make the good news public. The lack of such public information is one good clue to the economic disaster known as TLA.
It is instructive to review the list of total laboratory automation projects which became operational in recent years. SmithKline Beecham Clinical Laboratories (SBCL) was probably first to get a TLA site into operation.
In the early 1990s, SBCL chose the King of Prussia laboratory to be its first TLA site. Several years of engineering effort and expense were required to get TLA into full operation at the King of Prussia laboratory.
Since TLA become fully operational at King of Prussia, SBCL has not published any detailed numbers on the costs required to develop and install the equipment. Nor has any rigorous analysis been published which measures the specific productivity enhancements against those costs.
Several articles appeared in clinical lab publications about SBCL’s King of Prussia total laboratory automation package. Although complimentary about the TLA equipment in operation, these articles were silent on the subject of return on investment and cost effectiveness.
Some Operational TLASitesIn The U.S.
Only a limited number of laboratory sites have installed a total laboratory automation system. Here are some examples:
1. SmithKline Beecham Clinical Labs: King of Prussia laboratory.
2. Quest Diagnostics Incorporated: St. Louis laboratory.
3. Quest Diagnostics Incorporated: Denver laboratory.
4. Beth Israel Medical Center: Central hospital lab, New York City.
5. Mt Sinai Medical Center: Central hospital lab, New York City.
6. South Bend Medical Foundation: System core laboratory, South Bend.
7. AutoLab & Columbia/HCA: Regional core laboratory, Atlanta.
Diagnostics vendors with instruments in the King of Prussia laboratory tell a disappointing story. When SBCL originally decided to automate its first laboratory, it decided to engineer its own equipment. Over a period of as long as five years, they say SBCL spent as much as $18 million on the total laboratory automation project!
If this is true, then SBCL’s TLA project at King of Prussia is a financial disaster, an economic write-off. The fact that SBCL has decided not to automate its other high-volume regional laboratories is strong evidence that it does not believe the current state of TLA technology is a good investment of capital.
Next to install TLA was Quest Diagnostics Incorporated. During the time when it was still called MetPath, a decision was made to install TLA at three high volume laboratories: St. Louis, Denver, and Detroit.
MetPath had an advantage over other clinical laboratories. Its corporate parent was Corning Incorporated, a world class manufacturer. Corning’s experienced industrial engineers would aid MetPath in the design and build phases of their TLA.
In 1995, the St. Louis TLA installation became operational even as installation of TLA at the Denver lab progressed. When the economic performance of the St. Louis TLA project was assessed, MetPath allowed work at Denver to be completed, but pulled the plug on plans to automate its Detroit laboratory.
In the years since 1995, not only has Quest Diagnostics refused to bring TLA to any other of its laboratories, but during 1998 it downsized the St. Louis facility. Those actions, taken after careful financial analysis of the TLA installations in St. Louis and Denver, would indicate that Quest finds the return on investment for total laboratory automation fails to justify even its existing installations.
Among the three blood brothers, only Laboratory Corporation of America did not invest in TLA. For various reasons, LabCorp never pioneered a TLA site in the 1994-95 period. Since that time, there has been nothing to change LabCorp’s financial assessment of other operational TLA installations.
One of the national reference laboratories decided to be an early TLA pioneer. Many laboratory observers know that just a few years ago, Mayo Medical Laboratories actually signed a contract with one TLA vendor and began installation at its core laboratory in Rochester, Minnesota.
TLA Vendor Ejected
Executives at Mayo quickly recognized that the TLA project was not going well and would not deliver the promised results in a cost-effective manner. Mayo ejected the TLA vendor and ceased work on the project.
Moving to the hospital laboratory segment of the industry, results are not much different. New York City provides a perfect example. Beth Israel Medical Center and Mt. Sinai Medical Center, both located in Manhattan, decided to be early adopters of total laboratory automation.
Beth Israel’s total laboratory automation project became operational in the fall of 1997. Mt. Sinai’s was turned on in 1998. Adminstrators at both sites acknowledge that their systems run at less than 20% of potential capacity. Both hospitals had declared that their excess capacity would be absorbed by increased specimen volume from two sources: test referrals from other hospitals in the region and lab outreach programs to physician offices.
More than one year after TLA start-up, neither hospital has generated any substantial increase in specimen volume. Accordingly, the return on investment for these TLA projects must be dismal.
There is an interesting anecdote which further illustrates the naivete of the TLA vendors and their laboratory customers. When THE DARK REPORT toured one of the New York hospital TLA sites, it was told a remarkable fact. Only after the TLA system became operational at this particular laboratory, was it discovered that 25% of the specimen volume that was supposed to go onto the automated line would not fit.
Unlike specimens coming into a commercial laboratory, hospital test collections generate a wide range of variety in tubes and containers which cannot be accommodated by this particular TLA equipment.
The TLA vendor, having no prior hospital laboratory experience to draw upon, had not anticipated this situation. Consequently, there were even fewer specimens available to run on this TLA line than projected, further eroding the original economic basis of the project.
Why Truth About TLA Was Never Revealed
Did you ever wonder why lab industry rumors contained more truth about total laboratory automation (TLA) than most of the laboratory publications?
After all, the arrival of TLA in 1994 and 1995 was a major industry development. If the technology worked, it was expected that those labs with TLA would have a significant competitive advantage over those that didn’t. So why didn’t lab publication get the story right? Why didn’t they seem to get the story at all?
One critical factor is advertising. Instrument vendors buy thousands of dollars of advertising in lab publications each year. Any lab publication which printed a story telling the truth about the problems with the pioneering TLA installations stood to lose large amounts of money from upset advertisers.
So lab publications simply ignored the real facts about TLA. That is why rumors proved to be a more reliable source of accurate information than most TLA stories which appeared in the clinical laboratory press.
Learning From Pioneers
When South Bend Medical Foundation turned on their TLA installation in 1997, they had the benefit of learning from some of the pioneering laboratories mentioned above. It is reported that their TLA installation operates acceptably. But, as in all other cases, neither the vendor nor the laboratory have volunteered to publish detailed information about the ROI and productivity performance of the TLA project.
In fact, during 1998, a former employee with the vendor’s TLA team told THE DARK REPORT several interesting facts about the South Bend project. First, because it was the vendor’s first TLA site, the equipment was priced at discount, near cost. Since this was a pilot project, that would not be unusual. But the vendor also wrote-off expenses approaching $600-$800,000 for software programming and engineering adjustments.
As a result, the vendor did not charge South Bend Medical Foundation a price which accurately reflected the true manufacturing and installation cost of the TLA project. Yet, even with this discounted price structure, neither the vendor nor the laboratory customer have made public a detailed ROI and productivity analysis of the project.
There is one remaining player in the total laboratory automation marketplace that continues to be active. It is AutoLab Systems, a division of MDS, Inc. of Canada.
AutoLab’s approach to marketing TLA is unique. When several years of sales effort generated no buyers in the U.S., AutoLab developed a strategy of partnering. AutoLab retains ownership of the equipment and shares in the profits and/or losses of the laboratory venture.
In partnership with Columbia/HCA, it constructed a totally-automated laboratory in Atlanta that became operational late last year. The partnership in Atlanta is called Integrated Regional Laboratories (IRL). AutoLab and Columbia are proceeding to build a similar laboratory in Miami.
AutoLab’s approach overcomes the specific problems, noted by Mark Smythe that “the manufacturer should specify an expected return on investment (ROI).”? If AutoLab owns the equipment, it doesn’t have to promise an ROI. Instead, it must deliver financially through its own hands-on management of the TLA-equipped laboratory.
Too Soon For Evaluation
Since the Atlanta laboratory only became operational in October of last year, it is too soon to expect a detailed report on whether it has met acceptable ROI targets and delivered productivity and cost gains. But, like the other TLA projects mentioned earlier, should neither party publish detailed financial and productivity data in the near future, it can be assumed that the project was not completely successful by those measures.
On the other hand, if AutoLab is able to enter into several new partnerships with additional hospital labs during the coming 24 months, that may be evidence that the performance of TLA in Atlanta proved to be adequate.
Overall, however, this review of existing TLA sites is discouraging. During the past four years, the news has not been good. THE DARK REPORT believes that total laboratory automation will not expand its installed base much in the coming years. There are a variety of reasons.
First, automation is most justified where high volumes of specimens are handled daily. In the clinical laboratory industry, the highest specimen volumes are found at laboratories operated by the three blood brothers.
SBCL installed TLA into one labora- tory, then stopped. Quest Diagnostics put it into two regional laboratories and stopped. Its Teterboro facility is one of the largest in the world, yet no TLA project has been announced. LabCorp’s facility in Burlington is also one of the world’s largest, but no TLA is contemplated there.
Given the first-hand experience of the national laboratories in working with TLA, the fact that they refuse to install TLA at additional laboratory sites means that this existing generation of TLA is uneconomical and does not deliver the productivity and cost enhancements once expected of it.
Third, the diagnostics vendors learned from their TLA mistakes. All vendors are now engineering modular instrument systems and workcell clusters. These are groups of instruments which logically tie together. Upcoming generations of modular systems will harvest a cost-effective return on productivity, without incurring the seven figure expense of TLA. In other words, labs may get a lot of bang for a smaller investment buck.
Fourth, evolving diagnostics technology will erode the dominance of central, or core laboratories. Miniaturization will encourage near patient and point- of-care testing.
More specifically, breakthrough technologies, such as Luminex Corporation’s multiplex bioassay system (see TDR, December 21, 1998) will permit more testing to occur outside the central laboratory.
Taken collectively, the facts indicate that TLA cannot deliver the improvement promised by TLA vendors. These are the reasons why THE DARK REPORT declares the current generation of TLA to be dead.
For TLA to be feasible in the future, diagnostics companies must engineer a higher level of performance at a much lower cost. Such TLA equipment has to deliver an unquestioned level of productivity gains and cost savings before it will be accepted by laboratory executives.
The reason is simple. The first gener- ation of TLA badly burned the pioneers who installed it into their laboratories. An entire lab industry will be justifiably skeptical of any TLA product which does not demonstrate unquestioned capability.
In the meantime, look toward modular automated systems and emerging diagnostics technology as the most likely sources of cost improvement and productivity enhancement for today’s laboratory organizations.
Smythe’s Value Analysis Tools
“In my work with clinical laboratories, I am always surprised at their reluctance to seek out and use proven industrial techniques,”? said Mark Smythe, Principal of Management Mentors.
“The techniques of Value Analysis and Deliberate Methods Change,”? he continued, “when applied in a clinical laboratory, can improve processes while slashing costs by a factor of 15% to 40%. These techniques do not require staff layoffs and actually increase morale and productivity of the med techs. They can be rapidly implemented, often in less than 30 days.
“That is certainly a better way to reduce costs than investing in expensive equipment or laying off large numbers of loyal med techs,”? Smythe observed. “Manufacturers have used these ‘secrets’ for years. Maybe one day laboratorians will adopt them as well.”?