News

May 31, 2023

Vojtěch Drbohlav: Everything can be automated, but at what cost?

No one wants to obtain results late, inaccurate or, in the worst case no results. However, this can happen in manual work,” says Vojtěch Drbohlav, Managing Director of Beckman Coulter Czech Republic.

No one wants to obtain results late, inaccurate or, in the worst case no results. However, this can happen in manual work,” says Vojtěch Drbohlav, Managing Director of Beckman Coulter Czech Republic.

Up to 75 percent of errors happen in just the pre-analytical phase of patient sample analysis. This can be avoided using automation. In an interview with The Slovak Spectator, Vojtěch Drbohlav speaks about the current trend in laboratory automation, the challenges, the future, as well as the various solutions and innovations Beckman Coulter offers in this regard.

In general, what problems do laboratories face?

The low unemployment rate, which results in a lack of qualified personnel, although on the whole that is a problem not only in healthcare itself. There arises a problem related to managing a process, be it in the laboratory or production, where you want to standardise the process as much as possible.

What does that mean?

When a sample arrives to a laboratory, you want it to be processed exactly the same way every time. Historically, different people focused on different roles, becoming experts. However, since there is a turnover of people, or just daily and night shifts, at the moment patients arrive you want their samples to be processed in the same way every time. So if a sample needs diluting, we want a laboratory worker to always dilute it the same way in the same ratio. This is standardisation.

In this context, the fact that laboratories are under pressure to perform, to deliver the sample on time, naturally arises. When a patient arrives in a critical health condition and needs to have laboratory parameters quickly determined, the pressure is on to get the result within an hour, if not faster. It follows that the laboratory needs to have the sample under constant control, from collection, through transport to a lab, analysis and the processing of the result and we need a clear idea of where is each sample in the laboratory.

When you think about how the process usually works today – the sample comes in, you put it in the rack, but as they are processed in batches, you need to wait for more, resulting in delays, a sample can be forgotten on the table, etc. To process urgent samples, laboratories usually duplicate equipment to process them indifferent workstreams. This increases costs, requirements for qualified staff and maintenance of equipment.

Where does automation come in?

The aim of automation is that when you receive a sample and you put it in the device one by one, it performs all the pre-analytical, analytical and post-analytical steps – checks whether the sample is correctly labelled, whether it has been centrifuged or not, how much material it contains. Then the sample is sentto analysers based on smart routing to avoid delays. After all the tests, auto-validation software tells you whether the sample meets the criteria you have set. Automation is important, it helps laboratories with the main issues – a lack of qualified personnel and standardisation. It also allows for total process control and consequently speeds up the process, because the samples are not processed in batches, but based on priority.

You mentioned that the lack of qualified personnel is a problem. With automation, is there not the risk that insufficiently trained people will enter the process?

I don’t think so. We live in a highly regulated industry and we have clearly set standards for the qualification of healthcare professionals. On the contrary, automation helps optimise human resources, saving time for workers to focus either on problems that arise in the laboratory setting or on self-education. When you have someone who is highly skilled, you don’t want them to spend time moving racks between devices, searching for tubes etc. You want them to focus on making sure that the results are properly evaluated, that if there are results that don’t match the set criteria, they focus on these problems and implement countermeasures.

Are there different types of automation?

The process is divided into three parts: the pre-analytical, analytical, and post-analytical part. Historically, companies have been concerned with automating the analysis itself. In the 1970s and 1980s most tests were performed using a pipette, plates or test tubes. Then automation of the analysis and tests gradually began. Now our focus has switched to pre-analytical and post-analytical parts because we think this is the future. According to data, 60 percent of manual steps occur in the pre-analytical part. Currently, this is the stage in which the greatest number of errors occur.

Such as?

The sample may have a labelling error, damaged tube, insufficient volume, wrong storage, various clots may appear when the sample is poorly prepared. Up to 75 percent of errors occur in this part.

What about the post-analytical part?

You want samples to either be stored in cold storage or sorted for further minor tests. At this stage, when you need to verify a test, you are able to use automation to find the appropriate tube in a few seconds and send it back for analysis. Without post-analytical automation, after analysis samples are usually sortedinto racks according to days, their positions marked and stored in a refrigerator. This meant further manual steps, the lab staff could spend several minutes looking for the appropriate sample. From a patient’s point of view, they want the result to be on time, so post-analytical automation can help to reduce the searching time and avoid possible errors.

So the trend is to eliminate as many steps as possible?

Exactly. As the analysis itself has already been automated, we are now looking at the remaining two parts, as well as working with clinical data, auto-validation, remote access. Imagine you have two labs with the same management. With remote access, you are able to see the status of instruments, initiate calibration, troubleshoot or validate the results in one lab that were measured in another lab. Health care will, of course, follow the path of sharing results between laboratories and hospitals.

Are there processes that cannot be automated?

Good question. In the end, I think it’s possible to automate everything, it’s just a matter of cost. But if you want everything automated, the cost will be enormous and that makes no sense. In every process, you balance between cost and performance. It makes sense to automate 80 percent of steps, which brings the greatest benefit, and frees up resources to focus on special tasks representing the remaining 20 percent of tasks. Laboratories are one of the hospital departments able to generate a profit. Of course, that means pressure to maximize efficiency. But as a responsible manager not only in a laboratory, you think about what you want to automate and what you don’t, so that the cost is optimal.

What are the requirements for automation in terms of healthcare?

I think the question is more the demands and expectations of laboratories. Not all are ready for automation, some are even afraid. Instead of making things easier, they think it will bring more work for them. During the implementation of automation, there is a great demand on the laboratory to set standards, processes, to change the way they function and work. But that may be difficult for people with established long-term habits. There are laboratories that prefer to work with individual analysers.

Does this not mean there is a risk they will be left behind?

It does. As a laboratory must be prepared in terms of IT and software skills, we also encounter age conservatism within the lab staff. Of course, software solutions are interesting, but if you have older lab staff, working with a computer may pose a problem for many. This may be one of the barriers that make theselaboratories afraid of automation.

That almost sounds like the fears accompanying artificial intelligence, that people will lose their jobs. But the counterargument is that yes, certain jobs will be replaced, but at the same time other jobs will open up, because people will have to learn completely different skills so that the technology can be used at all.

That’s right. We are not there yet in terms of artificial intelligence, but the software of some devices is already at the level that it can learn by itself, evaluate based on the collected data, and then give people information to make decisions. Here, according to the legislation, AI cannot make decisions for us. It justgives us suggestions on how to proceed and that’s the right way. This allows people to educate themselves, to gain qualification and make the right decisions, and not spend time with unnecessary manual work such as sorting and transferring samples between facilities.

The uninteresting steps.

Exactly, because you don’t attract young people to laboratory work by telling them that they will be doing manual steps, when a young generation can use various tools with, for example, artificial intelligence.

Is AI the future of this field?

Both AI and machine learning. As I said, there are AI-like parameters in some of the software solutions already offered, whether it’s learning to evaluate blood cells or patient prioritisation in emergency department. But it should not make decisions for us, but offer solutions.

More than two years ago, you introduced the DxA 5000 system, which included as many as 24 innovations. What was that about?

The system allows a laboratory to automate the pre-analytical part. Our idea is that upon entering the device, eight parameters of the sample are detected – for example, the colour of the lid, the amount of material, its weight. Then, the most optimal path to the result is traced for the sample. The decision on whether the sample is an urgent one, also known as a stat and which has priority, or a routine one is made. Then the device sends samples for centrifugation, done in four minutes. The next step is sending samples to individual analysers on a conveyor with four tracks – two in each direction. Moreover, there is no queueing. The conveyor only sends as many samples as any given analyser is capable of processing. This allows prioritisation if more urgent samples come. After that, the analysis itself takes place, or the measurement of all parameters on the relevant devices. What follows after that is storage in a refrigerator. It’s aimed at large laboratories.

What do you mean by large laboratories?

A large laboratory is based in a university hospital or big laboratory chains, a medium one in a district hospital in towns such as Znojmo, Czech Republic, or Ružomberok, north Slovakia. A medium lab mostly had a standalone analyser and automation was difficult for them due to space constrains. For this reason, we introduced the DxA 5000 FIT line. The footprint is significantly smaller, but it’s suitable for medium or smaller laboratories that have been largely dependent on using individual analysers. There are three laboratories in Moravia where these smaller systems have been installed.

Are there any DxA 5000 or DxA 5000 FIT systems in Slovakia?

We are in contact with laboratories, and I trust that next year we will install at least one system in Slovakia.

What do the DxA 5000 innovations mean for workers?

In the past for example, workers manually transferred samples, carrying them into the centrifuge. With a manual centrifuge, it requires a certain number of tubes in it, which means you have to wait for them to accumulate. Thus, there is a delay. Our line is able to weigh and evaluate the volume of material,automatically balance itself and rotate even a single sample.

This year you are presenting the DxI 9000 system, which contains a unique camera system within the device. Why is it unique?

The camera system offers total process control during analysis. For example, it allows the detection of the amount of material in a sample in order to perform a sufficient number of tests from a given tube, the detection of its quality, and more. On the one hand, it also increases the reliability of the processand the result’s correctness; and on the other it’s also able to minimise the manual actions of the operator and reduce daily maintenance to zero. It’s always ready for measurement next morning.

Before the camera system, what did you use?

Sensors. They used to check whether, for example, a test tube reached the place it should, the given place; but they could not tell us whether the contents of the test tube were correctly aspirated, for example. This is where cameras allow full control.

Do you think other companies in the field of automation are taking on the idea?

I hope that we are a leader and that other companies will follow suit. We have to keep in mind that we need a good competitive environment, and we want to deliver the best possible results to patients.

According to you, what would be the most fundamental innovation in automation?

An advanced software solution that allows to link results from samples so that a doctor has an overview of what stage the sample is in, without, for example, having to call the laboratory. Imagine you have a patient in the operating room and you can glance at a display regarding how long it will take to get the result. There are parameters determined only during the operation. You will have everything “live”. That would be a big deal in terms of health care.

Is this something that is possible in our lifetimes?

I think these solutions will develop at a great pace.