Balancing optimisation, flexibility and robustness in laboratory workflow design. (FutureLabsLive)
Donald Peter Fraser- May 16, 2023
- 4 min read
Updated: Oct 18, 2023
The scientific community faced unprecedented challenges during the COVID-19 pandemic.
We didusn't quite know what we were facing for a long time... and once we did know - uncertainty persisted. Our existing systems of diagnostics, logistics, healthcare, informatics and policy just weren't ready, and resulted in hugely disruptive changes to the way we live and work.
The ongoing threat to health from COVID-19 has been successfully mitigated through vaccine programs, therapeutics, and overhauls of our health care and diagnostic capabilities. Three years since the start of the pandemic, close to 7 million deaths from COVID-19 have been reported to the World Health Organisation. Now, public health organisations seek to reflect on the learnings of the pandemic - and give ourselves a sturdier framework to meet Pathogen X.
An observation of existing diagnostic capabilities we made at Omix while working in the diagnostics sector were that there were two schools of thought when it came to lab workflows:
Many labs employed off-the-shelf solutions. Such systems were highly specialised and optimised for particular tasks and gave advantages when it came to consistency, efficiency, supplier guarantees and regulatory strategies. Typically, these systems were relatively inflexible, would tie a lab into a single supplier and were difficult or impossible to re-configure.
On the other hand, there were labs that adapted open platform systems. These were often research grade instruments - and hence were extremely precise and versatile, but not designed for heavy industrial usage. These systems were compatible with standardised plasticware and reagents from any supplier; and they tended to require significant up-skilling of operators, and bespoke qualification, validation and maintenance plans.
Similar observations apply to Laboratory Information Management Systems (LIMS). Historically, vendors would produce a rigid piece of software that would attempt to be a one-size-fits-all solution. But just like a one-size-fits-all shirt, these systems would fit many badly. Every lab is different, and hence if a lab required a piece of software to support their custom workflow, they would often be forced to develop their own, and take on the risks associated with software development. More recently, LIMS vendors have started selling configurable-off-the-shelf solutions, and such software can be customised to better fit a lab's workflows. LIMS solutions are expensive with a notoriously long lead time to implementation. While the configuration of these systems can be handled by the end-user, or by the vendor, both approaches carry risks and benefits.
Open systems allow for greater flexibility in system design, and in late 2021 we were tasked with developing an entirely bespoke SARS-CoV-2 RT-qPCR genotyping service to manage the Omicron wave.
Pinning down a precise set of requirements to allow us to develop the optimal solution however proved a challenge. There was uncertainty over procurement, suppliers, assay targets and equipment. Meanwhile the window for delivery closed dramatically from 6 months to one month. With limited resources we set to work at pace to build a suitable, flexible, system.
What we found was that there was a balance and a cost when it came to flexibility.
With a certain set of requirements - flexibility in the solution is not needed, but as certainty decreases, the requirement for flexibility increases. But flexibility usually comes at a cost to optimisation: it is often less efficient than specialised systems, and is harder to automate as it requires more human intervention. This is most easily illustrated by considering the lab-based extremes:
a sample-to-answer machine gives a fully optimised, automated and consistent service;
a set of hand-pipettes and a lab book gives maximum flexibility, but control is fully dependent on the operator.
The pressures and requirements of this project crystallised into a fully in-house developed workflow solution that was reagent, plasticware, equipment and assay agnostic.
Our team built the liquid handler scripts, optimised the assay performance and created the workforce model. The entire process was tied together by a python-based LIMS, ReflX.
Three key stages of the workflow gave the opportunity for controlled decision-making. The result was a traceable and compliant diagnostic system that incorporated several classes of flexibility:
Product: any Taqman PCR assay, with any plasticware
Routing: any 96-head liquid handler, any PCR instrument with correct filters
Operational: re-working and scheduling parallel or sequential activities
Volume: scalable data and lab architecture
By the time of its retirement, ReflX had processed almost half a million genotyping tests and tracked the emergence of the Omicron variant and its sub-lineages. The flexible system allowed the lab to pivot and respond to emerging threats and challenges with little to no requirement for re-configuration.
The success of the project was down to the holistic development approach taken by our team:
Early engagement of stakeholders to secure buy-in
Synergy with all aspects of the organisation
Investment in people and culture
Parallel development of lab and data architecture
Designing a flexible system architecture with boundaries
Labs must consider their options strategically when designing and implementing new workflows as there are risks and benefits to both off-the-shelf and open architecture solutions.
New solutions are expensive. Change is resource intensive. Every organisation wants to maximise their return-on-investment and ensure business continuity. This could be done in a couple of ways:
Maximise usage through optimisation of throughput, efficiency and consistency.
Increase solution lifetime by building in flexibility.
Each application is different, and there is no single 'right' answer. Service flexibility, whether it's a pathology lab or a manufacturing production line, helps manage strategic uncertainty by allowing diversification. But such flexibility cannot be at the expense of quality. How labs achieve that flexibility could be with rapidly re-configurable solutions made up of standardised building blocks, or solutions with flexibility built-in. We have implemented flexible solutions in both ways, and have found that they both have their pros and cons.
Comments