Automating Your Clinical Laboratory

For laboratories handling high volumes of samples, every minute of time and effort is precious. That’s why, for the last 150 years,¹ laboratory professionals have looked for ways to automate their most tricky, tedious, or time-consuming tasks. From simple liquid-filtering devices to fully autonomous workflows capable of artificial intelligence (AI)-driven learning and decision-making,² automation has made inroads into almost every part of the clinical laboratory. But for labs new to the technology, what’s the best way to get started?

Thinking ahead

The first step in any automation process is understanding the lab’s existing workflows and where challenges arise. Are there particular stations or instruments that are busier than others? Are there specific tasks that create bottlenecks or increase the risk of error when performed manually? Often, these “trouble zones” are the areas that might benefit most from automation—or need the most planning and care during the automation process.

Not all workflows can be fully automated, but for those that can, labs must consider whether to opt for partial (modular) or total laboratory automation (TLA). Most laboratories already feature some degree of automation,³ typically in the analytical phase of testing, but increasing numbers are adopting TLA systems that incorporate pre- and post-analytical phase automation.⁴ These systems may include sample transportation, sorting, preparation, storage, disposal, and more, using technologies from simple conveyor belts⁵ to advanced robots⁶ and AI-based software.⁷ TLA offers numerous advantages: long-term cost savings, less need for staff to move around the lab, lower sample volumes, increased standardization and traceability, reduced potential for human error, and more.⁸ But it comes with challenges that may make it inappropriate or impossible for some labs: high initial costs, space and infrastructure requirements, potentially higher maintenance and consumables expenses, potential equipment-based bottlenecks, increased risks of downtime or software errors, and loss or disruption of manual skills.

“Automation is a cornerstone of my work,” says Giuseppe Lippi, a professor of clinical biochemistry, dean of the University of Verona’s Faculty of Medicine, and director of the clinical chemistry and hematology laboratories at the University Hospital of Verona. “We receive almost 10,000 tubes per day, all of which are automatically checked in on arrival, sorted, aliquoted (if necessary), and then—in some cases—transferred directly from the preanalytical modules to the analytical modules (clinical chemistry and immunochemistry). All of these manually laborious procedures would require additional staff, which can instead be saved for more qualified and empowering activities.” But even large labs are not immune to the challenges of TLA. “I decided to remove hemostasis and hematology from the TLA to avoid bottlenecks (too many tubes managed by the same system) because the samples are different and require specific handling and, above all, because the testing is completely different and has different requirements.”

Choosing technology

After determining their path to automation, clinical labs must choose from a wide variety of vendors and technologies, weighing factors such as the total number of samples they handle in a day, the range of specimens and tests they handle, the features and capabilities of each instrument or system, and practical considerations such as cost, size, performance, and maintenance requirements. For example, a comparison of three systems for automated bacteriology—all of which required twice-yearly maintenance—revealed similar mean times between failures, but significantly different time requirements for each preventive maintenance cycle (although all systems allowed continuous work during maintenance). In the event of failure, all three systems offered multiple backup options and semi- or fully manual operation modes, although some offered more backup automation than others.⁹

A group of researchers at Harvard Medical School has developed a process for selecting automation technology in the clinical lab.¹⁰ Their recommended steps:

1. Form a decision-making committee that includes members of each group involved in the process.
2. Perform workflow analyses to determine existing strengths and weaknesses.
3. Draft a request for proposal to vendors; this should include the lab’s key needs or expectations from the automated system.
4. Conduct site visits to client laboratories to gain a better understanding of each vendor’s systems.
5. Trial the top candidates in your laboratory before making a final decision.

Low-cost automation

For small or resource-limited labs, even partial automation can seem out of reach—but that’s not necessarily the case. Increasingly, open-source software, minimalist hardware, and 3D printing are making advanced technologies accessible to even the smallest labs.

Researchers at Germany’s Albstadt-Sigmaringen University used an affordable robot arm in place of specialized laboratory automation equipment.¹¹ Although the robot itself is only capable of simple automation, combining it with free scripting tool AutoIT allowed it to not only operate without manual control, but also interface with third-party lab software. Specific end effectors (the arm’s “hands”) were 3D-printed for lab tasks. Although the system’s speed, throughput, and functionality were limited, the laboratory nonetheless saved time because staff could focus on higher-level work while the automated system tackled simple tasks like pipetting and autosampler operation.

Such stripped-down solutions are popular approaches to democratizing automation. The Minimum Viable Option (MVO) automation platform, developed by Brian Iglehart, exemplifies these principles. Using low-cost Raspberry Pi boards and accessories with open-source software and 3D-printed parts, Iglehart created a system that costs only a few hundred dollars and can be adapted for multiple assay types.¹² The MVO has caveats; it’s not as accurate or precise as a commercial system designed for laboratory use and accommodating data security requirements may present a barrier. Nonetheless, Iglehart intends to continue improving the device for use in resource-limited settings. And he’s not the only one; from 3D-printable liquid-handling robots¹³ to free, universal, modifiable software coding packages,¹⁴ accessibility remains a focus for laboratory automation technology.

To automate . . . or not?

Although Lippi is an automation advocate, he cautions clinical lab professionals not to lose their manual skills. “In October 2023, [my hospital] fell victim to a cyberattack that led to a complete shutdown of the laboratory information system and forced us to do everything manually. It was a nightmare because most of the staff had never worked without automation before, and those who did were no longer used to doing so.” Nevertheless, he remains enthusiastic about the benefits of automating. “It’s about more standardized procedures—and therefore better quality—from sample receipt to test result transmission, lower sample volumes needed, lower overall costs, fewer staff involved in low-skilled activities, and greater personnel satisfaction.”

References:

1. Olsen K. The first 110 years of laboratory automation: technologies, applications, and the creative scientist. J Lab Autom. 2012;17(6):469–480. doi:10.1177/2211068212455631.
2. Xie Y et al. Toward autonomous laboratories: convergence of artificial intelligence and experimental automation. Prog Mater Sci. 2023;132:101043. doi:10.1016/j.pmatsci.2022.101043.
3. Zaninotto M, Plebani M. The “hospital central laboratory”: automation, integration and clinical usefulness. Clin Chem Lab Med. 2010;48(7):911–917. doi:10.1515/CCLM.2010.192.
4. Bakan E et al. Automation in the clinical laboratory: integration of several analytical and intralaboratory pre- and post-analytical systems. Turkish J Biochem. 2017;42(1):1–13. doi:10.1515/tjb-2016-0234.
5. Niazi MA. Automated and On Track: What is Total Laboratory Automation? Control Automation. December 5, 2022. https://control.com/technical-articles/automated-and-on-track-what-is-total-laboratory-automation.
6. Miller JA. Robots Get Ready to Roam in Clinical Labs. Clinical Laboratory News. October 1, 2020. https://www.myadlm.org/cln/articles/2020/october/robots-get-ready-to-roam-in-clinical-labs.
7. Cope S. M095 AI infrastructure for the digitization and automation of clinical laboratories. Clin Chim Acta. 2022;530:S243. doi:10.1016/j.cca.2022.04.192.
8. Lippi G, Da Rin G. Advantages and limitations of total laboratory automation: a personal overview. Clin Chem Lab Med. 2019;57(6):802–811. doi:10.1515/cclm-2018-1323.
9. Croxatto A et al. Laboratory automation in clinical bacteriology: what system to choose? Clin Microbiol Infect. 2016;22(3):217–235. doi:10.1016/j.cmi.2015.09.030.
10. Melanson SEF et al. Selecting automation for the clinical chemistry laboratory. Arch Pathol Lab Med. 2007;131(7):1063–1069. doi:10.5858/2007-131-1063-SAFTCC.
11. Rupp N et al. Establishment of low-cost laboratory automation processes using AutoIt and 4-axis robots. SLAS Technol. 2022;27(5):312–318. doi:10.1016/j.slast.2022.07.001.
12. Iglehart B. MVO automation platform: addressing unmet needs in clinical laboratories with microcontrollers, 3D printing, and open-source hardware/software. SLAS Technol. 2018;23(5):423–431. doi:10.1177/2472630318773693.
13. Barthels F et al. FINDUS: An open-source 3D printable liquid-handling workstation for laboratory automation in life sciences. SLAS Technol. 2020;25(2):190–199. doi:10.1177/2472630319877374.
14. Johnson JL et al. PLACE: an open-source Python package for laboratory automation, control, and experimentation. J Lab Autom. 2015;20(1):10–16. doi:10.1177/2211068214553022.