IN PERSON: EXPANDED ONLINE VERSION
 |
Helmut Koehler, PhD, is executive managing director at Olympus Life Science Europa GmbH (Hamburg, Germany) and Olympus Life Science Research Europa GmbH (Munich, Germany). Since joining Olympus in 2002, he is responsible for the European life science business at Olympus, which includes diagnostics and microscopy, and all Europe-based global R&D and manufacturing activities in diagnostics, including lab automation, reagents for clinical chemistry, and immunochemistry. |
Demands for instrument connectivity to information management systems and for higher throughput continue to be heard from laboratories that IVD manufacturers serve. Meanwhile, lab staff and lab funding are always in short supply—perhaps even more so as the general economic prospect darkens. Instrument developers have to respond to the challenge of providing new equipment that can turn more tests more quickly and less labor-intensively while remaining affordable. Of course, holding cost down cannot be at the expense of profitability.
To learn how instrument manufacturers are coping with demands for innovative and productive yet economical laboratory systems, IVD Technology editor Richard Park spoke with Helmut Koehler, PhD, executive managing director of Olympus Life Science Europa GmbH and Olympus Life Science Research Europa GmbH (Munich). In this interview, Koehler talks about finding ways to make analyzers more versatile, balancing evolutionary and revolutionary changes in technology, and the importance of giving customers what they ask for.
IVD Technology: What do you see as the biggest technological advances in IVD instrumentation over the past few years? And what have been the latest trends in technology development?
Helmut Koehler: I think there are several. Definitely, there was a lot of development on the electronics side, leading to more combination systems which, in the end, reduced the instrument’s cabling or wiring and provided more intelligent programming and control of the components in the instrument.
I would also mention graphical user interfaces. These in our day profit from developments that are very beneficial, making them easier to use, more self-explanatory. Graphical user interfaces are now much easier to develop and to use than they were before.
A general trend has been to make instruments smaller, more compact. Basically, you reduce size but at the same time also adjust performance toward more speed or throughput. Or, when it comes to handling reagents, reducing those to smaller volumes than was possible before.
I think these are probably the most common trends I see at the moment.
Also, clearly, what I see in research diagnostics today is modularity. The advances I just described make it possible to compose instruments from modules that are easy to attach or detach. Then, definitely, trackability or connectivity, so that you can directly connect instruments to any tracks or conveyer systems.
Last but not least—and I would say very important for serving laboratories of different sizes—is the effort to be scalable in the instrument design approach. That means having instruments based on a common technology or reagent application which can be configured in small to large versions by using a group of, say, 10 compatible components or technologies and the same reagent application.
What clinical laboratory market factors are driving these technology improvements and trends?
It’s consolidation of laboratories as such, or labs getting bigger. But also, in a laboratory network, you have to serve laboratories ranging from the very small to the very big and, of course, ideally by using the same reagents or applications. The effect is a consolidation of disciplines. In the past, you had a clear separation among laboratories. But now, everything’s together and is getting integrated into a work flow so that the throughput and turnaround time in the laboratory are optimized.
When you speak of disciplines in the lab, do you mean different areas such as clinical chemistry or coagulation or hematology?
Yes. Even in, say, immunochemistry, where in the past infectious disease testing was separate from thyroid facility testing, this is now coming together.
So, now that there’s this greater consolidation, there are greater demands for one machine to serve for all these various disciplines?
Yes. And I mean also optimizing the laboratory as such, where you now have separate rooms or task areas that have not been optimized or built to support optimized work flow. It is more and more the case that labs are constructed in a completely different way than before, to allow for workstations and instrumentation being arranged in the fashion that is most efficient for handling samples in the laboratory.
Development Challenges
What are the primary financial, regulatory, or other obstacles that IVD manufacturers encounter when developing instruments and trying to do business in this area?
From my point of view, they’re definitely on the financial side, because it’s an investment, and return of investment is of course important. Also, you do have to make sure the instrument supports system projects. Issues surround the life cycle management of products already developed and brought to the market and the decision between evolution and revolution when considering how to go forward.
It’s like in the car industry. Usually, carmakers generate a platform only occasionally and in between do not issue completely new developments but basically just upgrades or updates. Only after a certain period of time has passed do they develop another new platform which involves, of course, the investment of much more money.
So, I think it’s just life cycle management and the balance between evolution and revolution. Your equipment has to remain backward and forward compatible, for example. In terms of chemistry applications and relations, everybody tries not to revolutionize the technology used to such an extent that it is no longer compatible with your installed instrument base.You basically have to compromise.
What are your thoughts with respect to the current global financial crisis—specifically, what sort of challenges might it pose for IVD manufacturers?
First of all, I don’t think that the crisis we are now seeing will stop before the healthcare sector generally is affected. The economy worldwide seems to be going into a recession. In Europe, a number of states definitely now are in some kind of recession. And Japan is going into this.
So, that means that, after a certain delay—I would say maybe one or two years—healthcare spending will be impacted. I think it would be fanciful to expect that the recessions will not lead to pressure on the healthcare system to cut down costs.
This fact will lead, especially in the routine laboratory business, to an intensification of a lot of things that we already see happening today. Number one is that commercial laboratories—the private laboratories—are very careful about investing. Where in the past year, because of very good profits, there was much more willingness to invest, now this is slowing down.
I believe that one result of this will be that instruments are used even longer than they were in the past. A chemistry instrument, for example, that had a normal lifetime of something like five years, it now would be used for seven years, or even more.
And that comes back to what I said earlier. When you are an IVD company with a huge installed base in your market, you have to look at this situation from a certain perspective. You have to go forward carefully into your next generations, because you should never lose the connection between your installed base and your future instrument platform.
Therefore, having backward and forward compatibility is very important in order to maintain profitability and return on investment.
Laboratory instrumentation and automation are very costly, of course. How will manufacturers deal with the potential lack of funds to invest in next-generation machines and, at the same time, work with customers affected by financial constraints who may be unable to buy another machine?
We are already seeing different business models coming up. Of course, there is still a certain market that I would call simply a cash state, in which instrumentation is bought. But more and more now, especially in Europe, instrumentation is being purchased as part of what we call managed-service solutions, or turnkey solutions. Here, basically, hospitals or healthcare providers try to contract for a complete solution, including services, that the IVD provider has to deliver.
And that, of course, is linked to much longer contract terms. A normal contract of the pay-to-reportables or pay-to-tests type, depending on the country, I would say is usually something between two and five years. With these managed-service contracts we in Europe are now getting into, we are talking about 15, 20 years of contract.
This is basically the compensation for what the service provider has to invest. Of course, when you are not only responsible for putting an instrument in and then providing reagents but also for several imponderabilities in the total project, you carefully have to monitor the whole laboratory. You have to provide all third-party consumables and whatever that entails. Sometimes you have to manage the personnel; then, of course, your primary investment is very high. But, as a benefit, those contract periods are generally getting longer and longer, which is of course good for you.
So, it’s an interesting trend based on the fact that some customers are less willing to be burdened with management decision making for the lab, and they just toss that out to somebody who’s doing the whole service at a fixed price. But, at the same time, they are willing to contract for a longer period of years.
When developing instruments, what model have IVD companies tended to follow lately? Do they generally try to come up with completely new instruments, or are most building on current offerings by developing upgrades, modular additions, and so on?
Definitely the second for two reasons. Number one is the return on investment, or the investment period you do.But I think also, especially in the routine laboratory business, today’s instrumentation offers reliability quite often based on proven components.
Look at chemistry: nowadays, you find a lot of components that were basically developed for instruments generations before. They are reused and reimplemented because they have been proven to be very robust. A lot of manufacturers will use old components even with new instruments, in order to provide very high performance reliability from the beginning. A good mean time between failures not only helps to save costs on the manufacturing side but even more so helps to win much better acceptance from customers.
As I said, the revolutionary approach usually carries manifold risks. The manufacturer is no longer comfortably working with its older-generation technology, and it is also taking a risk in that its new instrument’s reliability and robustness may suffer as a result.
Therefore, especially for the routine work at any rate, IVD companies quite often use proven components in their machines. Of course, you often can find new electronic parts in newly designed components. But a close look in those machines will reveal a lot of parts that have been used in them, sometimes for 50 years.
Do you expect this successful modularity model to continue? Is it primarily driven by financial factors or by customers?
Both, I think. From a manufacturer’s point of view, it is very economical. But on the other side, the company can provide more flexible and customized solutions, which customers want. The two things go hand in hand. Sometimes you have to sell to customer expectations, when modularity is something customers consider technically necessary. Then, of course, modularity can be used as a marketing tool; you can sell it to the customer as a benefit of the instrument.
The problem is that there’s still a way to go. The modularity idea is that, at the end of the day, you would have a base module from which a very high end analyzer could be built with the addition of other modules. That would be the ideal case. However, when you look into today’s instruments from several companies you don’t actually find this. Instead, they are fundamentally different machines—sometimes even with different technologies—that are sold as modules, or used as modules.
But, for me, real modularity would mean having a base module as a foundation for an integrated superanalyzer made up of multiple modules.
Connectivity
What role has connectivity played in instrumentation development?
Increasingly in recent years, of course, and especially for the market-leading companies, it has been essential for instrumentation to be connectable - connectable to equipment tracks of whatever type—even third-party track solutions—and also connectable to stand-alone automation solutions. The latter are also quite important: you can do certain pre- and postanalytical things even on separate workstations that are not completely consolidated or linked into a track system. That is additionally a very useful and efficient approach to lead automation. I think it is clearly significant and essential.
On the other side, there are some drawbacks from those developments that are still not overcome, involving tracks, for example. There’s the problem that, if the track is down, usually the lead is in big trouble. Also, some instruments linked to the track do not provide a backup solution that, for example, could load samples in a conventional way.
So, there are a lot of modern modules that are nicely linked and directly connected to tracks, but, if the track is down, don’t allow any loading of samples. That is something that needs to be much better addressed in the future by technical means.
How have the rapid advancements in information technology (IT) affected laboratory instrument development? And not just IT, but also patient-data collection in hospitals generally.
Of course, you have on one end the laboratory information systems (LIS), and on the other the hospital information systems (HIS) that you have to link to. The proper interfacing of results is mandatory.
But in the meantime, while those systems were being developed, there were very important developments in the area of middleware. Middleware is software components that are easier to interface with the various instrument systems being used in the laboratory that are taking over an essential part of work-flow management. In the end, the LIS and HIS are basically more responsible for patient data.
Middleware systems are extremely important. For today’s automation solutions, no track or even stand-alone automation is possible without fast middleware solutions.
How do IVD manufacturers like Olympus coordinate efforts between their lab automation development team and middleware software engineers, whether in-house or outsourced?
This is one of the problems. You have the choice to do things like that by yourself, but then you have to bring the connectivity-enabled instrument into a market where, usually, a lot of third-party products from dedicated companies are established. And, unfortunately, those middleware providers are not the same around the globe. In Europe, providers even vary by country, and between Europe and the United States there are different providers.
So you have only these two choices: either you describe your interfaces to these middleware systems very clearly and incorporate with various middleware providers by territory or by country, or you have to do it by yourself.
At Olympus we decided to do it the first way ultimately. We started out the second way and found it difficult pushing our products into the market because of all the established middleware providers that were very highly accepted by the laboratories and also by laboratory instrument providers.
The POC Factor
How has point-of-care (POC) testing affected instrumentation development?
From my point of view, not really. The fully scalable concept is still missing., Everybody’s dream, of course, would be to have a technology that could be used in a point-of-care instrument and at the same time in a routine analyzer for the high-throughput laboratory. Unfortunately, that is not available. The main reason is that there still is a big cost difference between POC technologies and lab technologies, because the prices that can be achieved with POC are much higher compared with those for routine testing.
Again, thinking in terms of chemistry, a high-throughput lab in chemistry usually has very strict limits in regard to cost, and some of the POC devices that are very sophisticated are of course much more expensive than that.
But, to be very clear, if there would be a technology that could fulfill the requirements of both types of testing, this would revolutionize instrumentation development.
Do you think the concerns expressed by some that POC testing would take market share away from the lab instrumentation area are justified?
I don’t think so. Of course, in the hospitals there may be certain movements toward more bedside POC testing, but owing to the centralization of laboratories—even in hospitals—certain functions that are very clearly established are not being eaten up by POC systems.
I think there is still a clear separation between what you do at the point of care and what you do in the routine labs. But, as I said, as technologies become available where the distinctions, from the cost and scalability perspectives, are more fluid, then this may break down.
Both connectivity and POC testing fall into the realm of making testing operations faster, smaller, sometimes less expensive. What other similar factors in this area have emerged or promise to emerge?
Scalability is important, so I would emphasize having modules or technologies that are adaptable in terms of size so that, basically, you can meet the expectations and needs of very small, medium-sized, and very high end laboratories with the same technology.
By the way, I would question that connectivity is making things faster. I think it is a market misperception that, for example, track systems accelerate operations. They don’t yet, though customers believe in that. What certain systems available today do, of course, is automate the process, and this is what the user likes so much.
But what they don’t do is decrease the turnaround time of a sample in the laboratory. For example, running reflex testing on a track system is a more cumbersome option if stand-alone workstations with technicians to serve them are available.
I would say a track may be counterproductive in terms of optimizing turnaround time, but customer needs are sometimes different, and they believe in certain things. I think that we in the industry should not try to educate the customer. The customer that wants a track and believes in it should be served with a track if it’s willing to pay for that.
Other customers may be served with other available solutions that are more in the task-targeted automation area. These are basically automation islands for certain tasks in the pre- and postanalytical processes. This is what private laboratories that are seeing more samples and are more keen to generate better turnaround times are commonly using instead of tracks. But at hospitals, tracks are very much in demand nowadays.
To return to the question, I think scalability is one thing. Another thing is that quality control and checks in instruments are getting more and more important.
I mean, it sounds stupid, but positive identification of things that have been done is sometimes not a simple matter. Technologies are still not there that would deliver 100% assurance that, for example, a certain pipetting step was performed.
Of course, physically the step was done, but confirmation of whether reagents or samples were delivered to the test is not that simple. So, in this area of quality control—those checks of steps performed and finished—providing a warning to the customer that there may have been a long pipetting, or air pipetting or whatever, is something to look for.
It is becoming more and more important of course that the control steps are automated so that mistakes in laboratories can be reduced.
Molecular Diagnostics and Instrumentation
How have developments in molecular diagnostics affected laboratory instrumentation development?
Here again I would say not very much, because polymerase chain reaction (PCR) technology is still quite distinctive and, I think, more complicated in terms of the steps required for testing, the duration of tests, and so on. Of course, it has affected the pre- and postanalytical parts of the process such that automation instruments are now being designed to integrate with and serve PCR machines or PCR tests. This is very important and is being done.
But regarding the technology as such or the solutions used, I don’t see that they have influenced very much the routine instrumentation technology. They are on a completely other level. Chemistry test durations are a few minutes nowadays, and with immunochemistry, you’re talking 15, 20 minutes. PCR tests are much, much longer and involve much more testing, so that means two probes.
But studies have projected that the molecular diagnostics field will grow quite rapidly in the next few years. Given that, do you think that IVD manufacturers like Olympus will have to take a closer look at molecular diagnostics and develop instruments and automated machines for that discipline?
Yes, of course. I think so. None of the IVD manufacturers can ignore that.
But the question of what you can do with molecular diagnostics remains. The majority of PCR testing still is done in infectious disease. Other applications have been expected and may be coming, but they are not much available today.
So, that means the market for molecular diagnostics is still relatively small compared with clinical chemistry or immunochemistry. But you are perfectly right. I think the most important challenge in molecular diagnostics is to find technologies that make the whole process easier to automate and that make it cheaper.
Don’t forget that reagents used in PCR are extremely expensive. That’s also because of the volumes that are used. The miniaturization of PCR processes is definitely something that would be very beneficial for both customers and the IVD industry.
Customer Input
How do IVD manufacturers find out what their customers—laboratories, laboratorians, physicians, hospitals, and other end-users—need the most from their IVD instruments?
By talking to the customer and doing surveys or focus groups. I’ll give you an example of something we went through; we did some quite interesting learning.
You know, with our OLA 2500, we have, I think, one of the best stand-alone lab automation solutions available. Of course, when we developed that we were very proud of it and felt that it was a much better solution than tracks that laboratories were using. We were saying among ourselves, “Let’s argue against tracks, because our technology is in certain respects much superior, and we can argue about the benefits that it brings to the customer.”
But, as I said, there are customers who strongly believe in tracks and who want them. To those customers you should not say such things, as we found out. You should accept that this is what the customer wants, and that you have to provide solutions that enable your instrumentation to be connected to existing tracks—or even tracks that you can provide to this type of customer—even though you may have other solutions in your portfolio that you would prefer to sell.
So, I would say to listen to the market and hear what the customer needs. It’s very important. Sometimes you will not be able to educate the market in the way you might think you can.
Requirements differ among laboratories of different types and sizes. What sort of challenges does it pose to IVD manufacturers to deal with a variety of needs and demands?
Basically, you need scalability, not only on the analyzer side but also on the automation side. We have to offer a system that covers a range of laboratories and solutions that will grow in line with customer needs and visions. That is sometimes tricky, because tracks, especially, are, from the hospital’s point of view, very expensive.
So, that means you need a lab of a certain size in order to justify a track. But of course smaller laboratories have automation needs that must be met. It is getting more and more important that, on the lab automation side, a manufacturer is able to provide tools and solutions that even a small or medium-sized laboratory will find attractive.
Coming Trends and Challenges
What new trends can we expect to see this year and in the near-term future in the area of laboratory instrumentation development?
I think that the most important thing everybody is working on at the moment—and there are already instrument launches in clinical chemistry, for example—is throughput. The task is to provide modules that are capable of throughput of 2000-plus tests per hour. That is a very tricky thing, technically speaking, because you have to reduce cycle times to a very few seconds in order to have systems that really can provide 2000 tests per hour, or more.
Then, of course, there are things like autoloading of reagents, applying also to immunochemistry. Most analyzers nowadays still require that you place reagent bottles in a reagent compartment manually. To automate that, so that the user has to perform fewer interventions on the instrument, is something that is more and more a concern in the design of modern instruments.
Then there is contact-free mixing, another of the trends in chemistry. The standard has been invasive mixing technology, but that is being replaced by noninvasive approaches.
On the lead automation side, I would say that the objective is, of course, higher-throughput track systems. Achieving greater throughput with track systems also is not an easy thing. Some providers are claiming that they have technology that theoretically can process 2000 chips per hour, but this is usually not the case.
At the end of the day, of course, if a track system could performs tests at a speed of 2000 or even 3000 chips per hour that would provide laboratories with a much faster turnaround time. But most track systems today are more in the below-1000-chips-per-hour range—600 or 650, something like that.
How do you expect IVD manufacturers to address the challenges?
The demand for higher throughput puts a lot of pressure on the area of reagent applications, for example. Even though you have less time to prepare and to mix, you still have to guarantee that your assays are performing well.
Instrument manufacturers need to miniaturize their reagent volumes, because the speed at which the rotors in analyzers have to turn is such that they have to be of reduced weight. Smaller assay volumes are very much in demand in order to allow lowering the weight of the rotor so that it can turn faster.
Also, smaller cuvettes are necessary in order to fit more of them in a cuvette ring. If you want to accommodate a throughput of 2000 tests per hour, you need a lot of cuvettes in the ring.
So, that goes hand in hand with the miniaturization of assay volumes. Basically, it’s getting to be a standard now that you have to be able to do presized-sample pipetting in the range of 1 µl, or even below.
That’s on the chemistry side. Immunochemistry still is suffering from the fact that the heterogeneous immunoassay leads to complex instruments. Just hardware-wise, because of the various steps necessary in the separation processes, immunoassay analyzers are more expensive. And, of course, assay times are longer so there is much lower throughput in immunochemistry.
I think this is possible to overcome only if another technology is developed. Everybody would love to have homogeneous immunoassay technology that can be used for high-standard assays.
Suppose you are able to do a fourth-generation TSH at an extraordinarily high sensitivity level, with the homogeneous immunoassay technology. So far, nobody has found anything like that, but a lot of attempts have been made in order to go in that direction.
There’s been a lot of talk about emerging IVD markets, especially in Asia, Latin America, and Eastern Europe. What are your thoughts about the potential growth of lab instrumentation sales in those areas?
The interesting thing to me is that we are so very active in China and Russia already. It is surprising how fast those emerging markets adopt the available technologies and instrumentation.
I mean, usually you associate emerging markets with small laboratories too and only small analyzers. And those labs must be less automated and well organized. That is to a certain extent true. However, very readily we see the development of labs in these new markets that, by size and by their equipment and standard, are quickly coming very close to what you find in Western countries.
So, one should not underestimate those economies, especially Russia and China, who specifically are very keen to change their standards in a much shorter time then we did in the West. They want to do things in two or three years where we spend maybe 10, 20, or 30 years to make that progress.
Do you have any final comments, perhaps on a topic that we may not have covered?
Two important things might be mentioned.
Especially in the routine segments, cost is becoming very important. Since you cannot expect to get more money for better technology—a new-generation instrument or assays—the rule becomes that your next-generation analyzer technology or reagent must be cheaper from a manufacturing standpoint. This is the only way manufacturers will be able to cope in a market where the price is very much under pressure and they will not be able to ask customers to pay more.
And the other thing is something I mentioned before: while once you offered instruments and reagents, and then you offered systems that brought hardware, software, and reagent applications together, nowadays more manufacturers are offering complete solutions. They provide not only the analyzer and the software and reagents but the whole surrounding connectivity and lead automation environment, and go even further to include other services.
I think it’s a very important trend in the industry. In the near future, differentiation among instrumentation suppliers, especially in the routine business, will be less by technology—all IVD manufacturers will be comparatively equal in their technical offering—than by whether they can provide more intelligent and more customer-specific solutions.
