June 2026—Instrumentation for all laboratory testing sites—large and small—was the roundtable topic on April 21 when representatives of Roche, Beckman Coulter, and Siemens Healthineers met online with Moira Larsen, MD, MBA, of MedStar, and Joe Baker of Baylor Scott & White. CAP TODAY publisher Bob McGonnagle led, with questions about, among other things, AI, testing consolidation, what laboratories expect of vendors, and what laboratories are willing to contribute to the partnership. “We’re happy to help figure out what the bugs are,” Baker says of new instrumentation.
CAP TODAY’s guide to chemistry and immunoassay analyzers for low-volume laboratories and for use at the point of care begins on page 47. Next month: analyzers for mid- and high-volume laboratories.
In last year’s roundtable, we talked about the importance of the core lab and its automation and instruments across the scale of large health care systems, as well as the increasing importance of automation in smaller settings. Many of you on the industry side said the lines are blurring between analyzers and automation systems and middleware, LIS, EHR, and many other controlling pieces of software. Jeff Shockley, what has changed, if anything, in the last year or so?
Jeff Shockley, senior technical product manager, automation and digital solutions, Roche Diagnostics: We’re seeing many of the same trends. The degree of integration we want to do and are expected to do has increased. Everything around cybersecurity has ramped up. Artificial intelligence is probably the biggest change. We were starting to integrate it into products, and customers were asking about it. Over the past 12 months, customers now want to know what our strategy is with AI and how it will be used in our products. I went to the HIMSS meeting this year and it seems like AI was a buzzword at every booth, whereas the year before it was maybe 50 percent.
Nick Culshaw, do you agree that the AI tsunami is alive and well in in vitro diagnostics?
Nick Culshaw, MD, vice president and managing director, clinical chemistry and immunoassay business, Beckman Coulter: Yes. I try to think about AI as a tool to help solve problems. We’re only scratching the surface of what it will do for laboratory operations, clinical medicine, patient education, and delivery of care. It’s undoubtedly exciting. But it’s most important now to focus on what customers’ top issues are and what our role is in addressing them in the most efficient and effective way. Do we address it through partnerships, which is one way Beckman Coulter has gone about it. Do we address it through building our own products—we’re thinking deeply about how we do that.
Jeff Shockley (Roche): I like thinking of AI as a tool because we’re seeing the same customer issues we’ve seen in past years—workforce shortages, lab consolidations. Now artificial intelligence is an additional tool we can use to address those.
Martu Richards, I hear many companies speaking about the use of AI to improve and understand laboratory workflow. Are you hearing that at Siemens Healthineers?
Martu Richards, marketing director for Atellica instrument acquisition and diagnostics, Siemens Healthineers: Certainly. We can’t escape it. Before, patients would use WebMD to look up everything medical. Now, because of Microsoft Copilot, ChatGPT, physicians are being inundated to order patient-requested testing and are probably doing more consultation. We’re starting to have conversations about that and how it impacts us and the laboratory. How can we equip physicians and pathologists to have more time for their patients who are asking more questions today? We’re using machine data to keep our instruments running and in better shape. We’re also looking at its impact on design and with our technology and how we can automate more processes to free up laboratory time.
Moira Larsen, what are your thoughts on AI as you listen to your colleagues on the instrumentation side?
Moira Larsen, MD, MBA, physician executive director, MedStar Medical Group Pathology, MedStar Health: The AI question is challenging to answer because the definition has been changing and developing. What one person calls AI, someone else doesn’t. The systems on our analyzers that assess the tests that have been requested, how tests are queued, and the most efficient way to perform tests, that’s what we call AI. The analyzer looking at a specimen and warning of an HIL interference, that’s the type of decision support we’re now calling AI. So much of what is today branded as AI in the laboratory is what we have always considered just part of the tools that help us do our job and do it correctly. It’s getting quite confusing.
It would be valuable to have early warning systems about when an instrument will need service or preventive maintenance. We have systems that our vendors monitor from afar, but there needs to be a smarter way to check the instrument to keep it from going down. Everything needs a redundant system. You can’t afford to have something go down that will compromise your operations and patient care, yet it still happens. My concerns are that these machines still go down, sometimes with double downtime, and we weren’t warned it was coming and don’t know why it takes so long to fix.
Joe Baker, you’ve gone through a big Epic Beaker consolidation. How are things on your core laboratory side? Has the implementation gone smoothly?
Joe Baker, senior vice president of diagnostic services, Baylor Scott & White Health: Our Epic Beaker transition was very successful. There was a lot of pain in the beginning as our teams worked through issues, and we’re now at the point of optimization. I’m grateful to our talented LIS and IS teams for shepherding us through this vital project. They understood the pain points on each side and helped us navigate through the challenges of such a large-scale implementation. Our success was largely due to the full team effort within and outside of laboratory.
With an organization as large as ours, our challenge early on was having so many new tickets while we’re still trying to resolve some of the complicated items. We are 18 months post-go-live. We were in optimization mode within nine months, which was three to six months earlier than we anticipated. We still have difficult decisions around where we house certain things like autoverification rules—is it in the middleware or within Beaker? That has a downstream impact when we look at closed-loop communication for critical values.
Regarding how we define AI, we have set up a process within our organization for anything truly AI-based, and these requests must be reviewed through that pathway. It is important to distinguish when something is algorithmic or part of the support model versus if it is a large language model. We are mindful of how we use the term because that determines the evaluation pathway.
Nick Culshaw, what are your thoughts about the AI separate track that systems are trying to implement?
Dr. Culshaw (Beckman Coulter): You have to acknowledge a few of the realities. Laboratory data is deeply fragmented. You can’t deploy meaningful AI solutions without first solving the data architecture problem, and that includes cybersecurity. Lab leaders should rightly be skeptical of AI that overpromises. We have to come back to solving the problems that are top of mind, most important, and probably largely operational. Are the people who are coming to you with solutions solving those problems? Can they solve them with the architecture you have?
Near-term value will likely be in the operational piece—turnaround time optimization, predictive maintenance, test utilization, analytics—but also in being able to measure ROI. You’re going to have to invest, which means answering to a CFO.
Finally, there is the training piece. We talked last year about shortages in the workforce but also a less experienced workforce. We will have to upskill operators in this new technology and in how to use AI, as well as in the basics of traditional laboratory medicine. That’s no mean feat; it’s extra workload on the hospital and for laboratory leaders. But it’s vital because the people who are most excited about using AI in the workplace are also the most knowledgeable and most engaged and best understand how the tools can be used.
Consolidation is increasing among laboratories, health care systems, and manufacturers. The market share of top in vitro diagnostic companies, like the three represented here, is enormous around the globe. Moira Larsen, how are you experiencing consolidation?
Dr. Larsen (MedStar): We have long worked as one another’s internal reference labs, trying not to replicate unneeded services. That has stood us in good stead with instrument and reagent issues because we’ve been able to rapidly pivot to consolidate testing to conserve reagents when needed, always with an eye to the clinical need. We have found areas where consolidation makes sense, such as standard histology and microbiology. It’s more challenging to get consolidation with the day-to-day chemistries, hematologies, and immunoassays because of the volume of work and pressure of wanting those results. You have to choose the more esoteric tests that allow for more grace in terms of turnaround time for it to make sense to consolidate.
We don’t always realize the turnaround time that’s required for some tests in the clinical world. For example, drug levels. We’ve learned of an increasing clinical demand for methotrexate level turnaround times. For certain diseases they use a very high dose, and in order to continue dosing to cure the patient or control the tumor, they need a turnaround time of the methotrexate level in an hour or less to plan the rest of the dosing. We had consolidated this testing because we didn’t know this clinical need had arisen.
Joe Baker, have you had similar experiences with assays in your system?
Chemistry is a great example. At some of our smaller sites we’re limited by the number of different analytes we can perform on certain equipment. We’re finding some of our smaller facilities want to get more in the oncology or cardiac space. They need access to tests they may run only 20 to 30 times a year, but when they need it, they need it quickly to decide whether to admit a patient. We’re working through those challenges now.
Martu Richards, I’m struck by how much better laboratories are in dealing with their medical staffs and the decisions made on a system basis, a hospital basis. Do you see improvement in communications with your customer base and, as a result, in the pointed questions they can direct to you as their supplier?
Martu Richards (Siemens Healthineers): I would say yes. We were talking about how the definition of AI can be different for different people; I think that applies to analyzer consolidation and configuration. From a manufacturer perspective, it’s trying not to be a one-size-fits-all. For Siemens Healthineers, it’s traditional for us to integrate chemistry and immunoassay. Most manufacturers have an integrated platform, so what’s the next level? It’s how we integrate automation functions available on an automation track, like sorting, archiving, decapping. Now how do you integrate these functions into your standalone instruments, because you don’t have the space or funding for a larger track-based solution?
Take it a step further. What can a software do that’s more integrated and more intelligent? You keep building on it. How do we deploy intelligent sampling, optimizing how we get results faster based on your sample work order? We listen to our customers and laboratory partners and ask where we can help them in the design of new instrumentation to meet their needs. The next level of instrumentation will have to go beyond testing and offer more automation and software functions and other things we’re probably not thinking about today. Having everything consolidated and integrated to the best of our ability to impact laboratory workflow and help laboratory staff will put the manufacturers that solve those issues in a better position now and in the future.
Nick Culshaw, that means labs are going to expect more from their instrument and system vendors because the brainpower and global reach are there.
Dr. Culshaw (Beckman Coulter): It’s tempting to think about consolidation as something that simplifies, whereas it often increases complexity. If you’re going to provide a solution to an entire network, you have to supply the most complex, esoteric part of the network, as well as outreach labs, while meeting the needs of every lab in between. That is crucial in your product and portfolio management strategy. You have to ensure you are not neglecting the ultra-high-volume or ultra-low-volume segment, have platforms that fully scale up and down, introduce the right levels of automation, and have commutable results across all of those because patients move around networks more and more.
Our customers have a desire for consolidation at the industry level. Beckman Coulter is owned by Danaher, and at last year’s ADLM meeting, Danaher had a diagnostics booth for the first time. Many people were confused about who Danaher was and didn’t know all the operating companies existed under one umbrella. You could see lightbulbs going off around the opportunities for customers, not just on the procurement side but also the research and collaboration sides. You will continue to see consolidation and companies with diverse portfolios looking to grow strategic relationships with their customers as they themselves consolidate and have to tackle their own complexity. Both trends can be positive for the industry and laboratories.
Jeff Shockley, I want to ask you a question about the changing definition of in vitro diagnostics. Companies make acquisitions, some of which are beyond traditional IVDs, and the Danaher decision to buy Masimo—a provider of pulse oximetry and other patient monitoring solutions, largely for acute care settings—is a good example. It’s a diagnostic test but not in a way we normally think of a pure IVD. What will this new definition of diagnostics mean for the traditional laboratory customer you all serve?
Jeff Shockley (Roche): It puts an onus on us, the major traditional IVD vendors, to expand our portfolios and capabilities. When you think about medical algorithms, whether or not they’re using AI, they don’t fit the traditional sense of chemistry tests and others. The boundary is becoming blurred.
Laboratory consolidation drives the need for solutions to be scalable and reliable and introduce a degree of automation, and not the traditional track automation. We have to continue to innovate, not with just the new technologies like AI and machine learning but with taking our traditional sense of equipment and evolving it to deliver higher testing capacities, increase automation within the solution, whether in software or hardware, and find ways of doing more with less. As labs consolidate and sites go to a core laboratory and hub-and-spoke model, space is a limitation. We have to find ways to put higher throughput and higher capacity into a smaller footprint. We have to evolve that beyond the IT and digital technology.
Joe Baker, you have a huge geographical footprint. What is the expanse of territory you cover?
Joe Baker (Baylor Scott): We cover roughly 30,000 square miles in Texas. It equates to about the size of Virginia. We go from the Dallas–Fort Worth area through Waco and Austin and to College Station and beyond. We are the largest nonprofit in our state.
I agree with Jeff that space is a concern. When we think about equipment and partnerships with vendors, we have shared goals. But we want to make sure the partnership can meet the breadth of differences we have, because we have small micro hospitals to midsize hospitals to large flagships, as well as clinic laboratories and others. If we’re working within a specific area, whether chemistry or hematology, we want to be aligned across the organization with that vendor. Having several scalable options is important to us, as are the intangibles. We are willing to be an early adopter to trial new things. Whereas before we wanted vendors to come with the finished product and with all the bugs worked out, now we’re happy to help figure out what the bugs are.
Moira Larsen, you have a smaller footprint but with great complexity. Tell us about MedStar and its laboratory system.
Dr. Larsen (MedStar): MedStar Health is in Maryland and in Washington, DC. We are a fully integrated care delivery network, so in addition to seven community hospitals in Maryland and two large hospitals in DC, we have multiple ambulatory care centers, multispecialty clinics, physician offices, rehab centers, urgent cares, et cetera. Most of the non-hospital-based are CLIA waived, so they have a lineup of instruments on counters and reagents that can be kept at room temperature.
Our hospitals are great; we have two giant instruments and our smaller hospitals may use only 10 or 15 percent of the capacity of those instruments, but they need the menu and the redundancy. Back in the days when my parents would buy a car, they would choose their options and four weeks later a car would appear with what they wanted. It would be cool if we could build a chemistry analyzer and say, “This is going into an oncology clinic; this is the suite of tests I need. How small a box can you put that suite in?” or “This endocrine or fertility clinic needs this suite of tests.” As medicine moves into outpatient centers and out of hospitals, it would be beneficial for those sites to get what they need, where they need it, and when they need it, instead of having to accommodate an instrument they’ll use at only five percent capacity. Doing it in a modular way would be helpful, because we’ve had instances where we wanted to change the analyzer we’re using to serve the ED but the space isn’t big enough for the smallest X analyzer.
Jeff Shockley (Roche): That’s an interesting challenge because sometimes scaling up is easier than scaling down. There’s a balance of point-of-care solutions that need to deliver laboratory-grade results at the point of care. We see that too, as well as the need for small ambulatory clinics and others to have these solutions. But I can’t put a large automated analyzer system that requires water hookup and all of that; it needs to be more of a benchtop-type solution.
It is interesting to me how we keep moving tests and machines from one place to another. We consolidate in a core lab and then, as several of you have indicated, it has to move back close to the patient and specialist physician for their needs. It is becoming a matter of consumer choice in many ways, isn’t it?
Jeff Shockley (Roche): Yes, and even including home care, looking at ways to get samples directly from patients, especially in rural communities where access to a hospital is limited.
Nick Culshaw, how are you looking at the relationship and balance between core laboratory and point of care?
Dr. Culshaw (Beckman Coulter): Danaher’s main point-of-care franchises are Cepheid for molecular and Radiometer for blood gas. Beckman Coulter is mainly focused on the core laboratory. That breadth is important to us, and we see an increasing demand for menus at the point of care. Crucial to point of care is staff training and ensuring the reliability of results, and then the commutability of results to the core lab, because patients often have repeated tests. Confidence in point-of-care results is key.
Martu Richards, when we reflect on the relationship of instrument manufacturers and the laboratory, with all that goes with it—IT, automation systems, middleware, the EHR—how can it be deepened and improved?
Martu Richards (Siemens Healthineers): First, I love what Joe Baker said about his laboratory getting involved early. Many laboratory partners would not have volunteered for that in the past but are open to it now. Having laboratory partners participate in the conversation early in development helps a product come to market. It’s of utmost importance to listen to what’s happening in the lab and bring that into your organization as you look to improve your platform.
The relationship can be deepened by listening more intently and understanding their operations more intimately so it informs what we do, whether on the software or instrumentation side, with the development of new assays, how we deal with automated functions, and even how AI gets deployed within the diagnostic industry in a way that is meaningful to laboratories.
Jeff Shockley, same question.
Jeff Shockley (Roche): I agree with Martu. The need for deep, strategic partnerships will increase. The more we share with one another, the more we can address the challenges labs and hospitals face. We take voice of customer seriously, but sometimes we can only know what we’re told. I learned early in my career the strength of partnerships comes to fruition when we come together to collectively address challenges, whether with an implementation or instrument. We can work together for a better solution, one that puts us in a better place than we were headed. We can take it to the next level when the relationship is treated as a true partnership and not a vendor-supplier type of relationship.
Nick Culshaw, how is Danaher going about the need to get closer to the customer earlier in the decision-making?
Dr. Culshaw (Beckman Coulter): Within the Danaher Business System we have a number of tools around customer insights, one of which is “Problem to Portfolio.” The goal is to understand the high-value clinical and/or operational needs and then do rigorous iterative cycles of voice of customer to understand what the product requirements are and how to deliver that in a finished product or portfolio.
It’s crucial to understand the segments of customers—large hospital systems, small hospitals, academic institutions, private institutions—and address that in multiple ways. You want to talk to happy customers and less happy customers, key experts and those who are on the ground day-to-day, so you get insight across the board. Follow that up rigorously into your design briefs, business case, quality system, and new product development.
Joe Baker, have you ever invited a major vendor and partner to shadow you for three or four days in Dallas?
Joe Baker (Baylor Scott): We did when we did Lean several years ago and were making sure we aligned appropriately with where opportunities for efficiencies existed. Where we haven’t done this well is with a strategic discussion about the future. We do a good job operationally and contractually and have conversations with our vendors about the clinical side. But I don’t think we’ve done a good job, speaking for myself, at blending strategic and clinical so we’re all walking the same pathway and saying, What do you all have coming down the pipeline? Do either of us have solutions that can benefit the partnership? Even thinking beyond pathology and our laboratory by including key leaders in other service lines like the ED or cardiology, thus connecting the dots a little better. Everything is open and can be considered.
Moira Larsen, how are you connecting the dots? What role do some of your partners play in walking in your shoes for two or three days?
Dr. Larsen (MedStar): We’re working at connecting the dots. We meet regularly with vendors to review instrumentation performance across all nine hospitals, where and what the problems are, establish parts depots closer to our sites, as well as what our clinical colleagues are telling us—which service lines are growing and what’s coming down the pike—and what the vendor has to support it.
One challenge is that while we can track instrumentation issues and failures across our nine locations, we have no idea how our experience of instrument use and performance compares with that of the rest of the laboratory world. If hospital A has a light source that dies every three days, we don’t know if that’s a one-off or if 30 to 50 percent of customers are experiencing it because it’s inherent in the engineering and the company is working on it versus it being an error in the way our staff is handling the machine. I would like to get perspective as part of an open dialogue.
For example, many years ago it took us years to figure out why our chemistry instruments were crashing each time we had Renal Week. Even though the patients were getting dialyzed, there’s something about renal patient serum that impacted instrumentation. Once we figured that out with the vendor, we set up preventive maintenance prior to and after Renal Week, and it cleared everything up. Those in-depth conversations about quality that delve into whether it is us, our patient population, or an engineering issue would help deepen relationships.