Wisplinghoff Laboratories is one of Europe’s largest diagnostic labs, processing approximately 12,000 specimens by conventional microbiology each day. The department operates 24/7 to assess swabs, blood cultures and urine, among other sample types, for more than 5,000 private practitioners, 120 hospitals, health departments, and other health care institutions. With that sort of volume, it comes as no surprise that the lab has machines do a lot of the manual tasks involved. Professor Dr Hilmar Wisplinghoff oversees the microbiology at the establishment and he says much of the efficiency gained through automation is in plate streaking and reading.
In microbiology, the lab uses a variety of instruments from different manufacturers to automate manual processes. As part of its microbiology automation concept, Wisplinghoff Laboratories uses Beckman Coulter’s workstation automation solution – which includes the DxM Autoplak, APAS Independence and DxM Trio – to achieve efficiencies in workflow.
Among these, the APAS Independence has made the most difference. “APAS has the highest impact with regard to workflow,” he says. “But the combination allows for almost total automation, covering streaking, reading, bacterial identification and antimicrobial susceptibility testing.” This means the only real manual step carried out in between each process is transporting the samples from machine to machine. The flexibility of workstation automation allows the lab to prioritise the most significant manual bottlenecks from the pre-analytical to analytical stage leading to improved productivity and accuracy.
Integrating automation into a laboratory is an arduous process, as setting up the machines means not only finding a space for them, but also detailed workflow planning and programming of the equipment, as well as connecting it with the existing laboratory information system to keep a record of processes. All told, Wisplinghoff says, depending on the type of automation, the process may take anywhere from a few days to several months before a system can run near independently.
Total automation isn’t the only valid approach, however, and due to the high cost of equipment and the fact that not every laboratory can spare the time and resources necessary, he says some laboratories may be better suited for workstation automation. “It takes very little planning and the actual implementation can be done in a matter of days,” he says. “Workstation automation replaces a specific process or task, meaning it usually comes preconfigured and there’s no extensive teaching necessary to use it”.
Another thing Wisplinghoff says had been important in the laboratory’s planning process is that automation needn’t take place in all in one go, and if you can only implement one machine now, there will hopefully be the opportunity to expand in the future. However, in doing so, it is important to minimise future complexity as far as possible – for example, by using standalone machines from one manufacturer or distributor, and having a common middleware with little to no requirements for interfacing.
APAS is a good example of this philosophy, as it will accept any type of plate regardless of the manufacturer, the type of plate, the specimen type, the streaking and the labelling. While the mechanical part obviously plays a huge role in the reliability of the various machines, the impact in the laboratory is also highly dependent on the software, especially when automating more complex tasks such as AST or plate reading.
The next steps
For those laboratories that have the resources to push the envelope even further, the next step is likely to be a more advanced AI system. While automation per se increases consistency, traceability and decreases turn-around time, AI can also help to automate more complex tasks. Even with the AI currently built into the APAS Independence, the possibilities for process automation can be extended far beyond the task-automation currently seen in conventional microbiology laboratories. But Wisplinghoff says this is just the beginning and more sophisticated algorithms will be coming in the near future.