The smart lab: The shift to more digitization is picking up speed

They have probably never been in the spotlight as much as during the pandemic: laboratories. In Germany alone, around 73 million COVID-19-tests have been evaluated since the beginning of 2020. And even away from Corona, laboratory physicians have a lot to do – blood, urine, aspirates and much more have to be evaluated and documented every day. Together with transporting the samples to the analysis site, this results in an enormous amount of work, just in terms of organization.

A classic laboratory workflow involves the arrival of samples and their testing, as samples must meet certain requirements to be meaningfully tested. Next, they are given a laboratory number and sampled so that the readings can then be analyzed. Particularly monotonous in this regard are repetitive tasks that do not actually require humans, but can be done by machines. To relieve the burden on trained personnel whose skills could be put to better use elsewhere, a Danish company has developed a robot with associated transport boxes. Intelligent boxes ensure the traceability of samples so that they arrive at the laboratory already checked for their usefulness. A robot then takes care of the rest of the sorting process. Anders Kjergaard Madsen, technical director at the engineering company LT Automation, explains in an interview for whom such a system is worthwhile and what work it can actually do. He is convinced that the digitalization of laboratories will continue to advance: "The bottom line is that automating such workflows is worthwhile, especially in the long term, because you no longer have to rely on human personnel to keep the process running."

Before a laboratory device like a robot can be used, however, it must first be produced. But what do manufacturers actually want in terms of the product development process and what is a "Lab 4.0"? The WILD Group is active in the development and production of laboratory equipment – CTO Wolfgang Warum knows what important features modern laboratory equipment must have. He says that laboratories were in great demand during the pandemic and so there was an enormous boom for laboratory automation: "The topics of productivity and automation became particularly important here. Processing measurements in large quantities and in a short time is in high demand in laboratories right now." This creates the perfect use case for robots, as they automate monotonous tasks and thus speed them up.

The use of artificial intelligence (AI) is also no longer a rarity in medicine today: it is particularly popular for evaluating data, as it can recognize correlations or evaluate laboratory data more quickly in case of doubt. The practical thing about AI is that it not only recognizes patterns that have already been stored, but can also learn new things independently. The problem with using artificial intelligence or machine learning is that it is rarely reproducible. If an AI is to be compared with other applications, there is usually not enough data available to make a comparison. This is particularly unfortunate in the biomolecular field: the more the transparency and reproducibility of AI systems could be increased, the more "the trust in such systems [...] grows - for example, by people who could use such systems in clinical practice," says Prof. David B. Blumenthal. He is conducting research on making artificial intelligence reproducible and transparent with the application AIMe.

Robots in the lab, smart transport boxes – the world is becoming more digital and so are laboratories. At this year's MEDICA, a lot will revolve around the modern laboratory – find out in the trend theme of laboratory medicine how we keep track in times of viral variants, long covid and vaccinations and how we keep track of pathogens worldwide. In which forums and conferences experts will give you a first-hand overview of the state of the art and the future of laboratory medicine, and much more about the topic, you can already find out in the trend topic laboratory technology.

But regardless of whether it is directly an AI for data evaluation or "only" a robot that sorts the samples: The digital transformation has long since begun in laboratories. Such systems may not yet be able to perform the sampling itself, but only to evaluate the data obtained from it – but even at the current state of affairs, they are already a great help when it comes to relieving laboratory staff of time-consuming work. To what extent they will be able to take over important tasks in sampling and diagnostics themselves in a few years' time remains to be seen.