From LIMS to AI: Modern laboratory medicine
Anthony Jackson-Crawford, Principal clinical biochemist, Doncaster Royal Infirmary
Digital systems in laboratory medicine are often underused despite their central role in service delivery. The day focused on three key themes: getting the most out of LIMS and wider integrated systems, sharing digital solutions to common laboratory challenges, and the inevitable discussions on the growing role of AI. I left with greater confidence – albeit still without the time - to start thinking about how these approaches could be applied in my own laboratory.
Maximising LIMS and system integration
A consistent theme throughout the day was that the true value of laboratory data lies in how well systems are integrated. The laboratory’s main output is information, not tests, yet fragmentation between LIMS, GP systems and hospital platforms continue to limit its impact. Standardisation (such as SNOMED and structured reporting) and emerging interoperability frameworks like FHIR were highlighted as key enablers. A key message was that the main barrier is no longer technology, but organisational and patient trust in how data are shared and used.
Digital Solutions for Everyday Laboratory Challenges
Several talks demonstrated how digital interventions can significantly improve workflow and efficiency. Examples included rebuilding paperless order comms from the ground up and embedding rule-based decision support directly into requesting and reporting pathways. Practical strategies such as internal prompts, workflow redesign and reworking “trapping” rules were used to demonstrate how laboratory scientists can take control of the LIMS rather than allowing the LIMS behemoth to dictate practice. The emphasis was on being intentional by reviewing processes from first principles and focusing on high-impact, high-volume areas.
From DIY Automation to AI in Practice
Others highlighted how innovation does not always require large-scale investment. Local “DIY” solutions, ranging from Excel tools to Python scripts, demonstrated clear efficiency gains when applied to specific problems such as minimum retesting intervals, user acceptance testing or administrative tasks. Alongside this, discussions around AI focused on validating output, governance and risk management. While AI has significant potential, a key takeaway was that meaningful progress can already be achieved through smaller, well-designed informatics solutions with the right skills and mindset in place.
Conclusion
As the first meeting of its kind, organised in direct response to member demand, this felt like a valuable opportunity to reflect on how we can operate safely and effectively within an increasingly digital laboratory landscape. I’d welcome future events as this area continues to grow.