Years ago, I stood before the committee that ultimately awarded us an EIC Accelerator grant and argued that our product was at TRL 6, approaching TRL 7.
I was not lying. I presented what I believed was solid evidence. The claim was scrutinised and accepted by experienced evaluators. We even had written confirmation from the hospital where our test had already been used in specific circumstances. And we are incredibly proud of the lives saved as a consequence.
And yet, with the benefit of hindsight, the product was far less mature than I then cared to admit.
The technical evidence was real. The use cases were real. But neither answered the question that mattered for that evaluation: was that diagnostic ready to make the difficult journey from a promising test to something that could be used, authorised, adopted and sustained in routine care?
Recently, I was reminded of this while participating in an innovation evaluation exercise for a private foundation. Several teams confidently described themselves as TRL 5, 6 or 7. In one sense, they were right. Their technologies had reached meaningful technical milestones.
In another, more consequential sense, most were much earlier than they realised.
That is why I have come to think that TRL may be the most dangerous number in diagnostics.
To be clear, I don't claim that the framework is useless. It is a useful description of technical maturity. The danger begins when a single technical-readiness number is allowed to stand in for the maturity of an entire diagnostic programme.
A useful answer to a narrower question
TRL was developed to answer a focused and useful question: how mature is a particular technology for a defined application?
A technology that has worked only on a bench is not equivalent to one that has been demonstrated in a relevant environment, integrated into a system or proven in operation. NASA’s TRL definitions make this clear: the scale concerns the maturity of a particular technology, component or system in its relevant application and operational environment.1
But the question becomes seriously incomplete when it is imported into the diagnostics world without changing what “ready” actually means.
A diagnostic is not simply a technology moving towards use. It is a regulated product or service entering a clinical, commercial and institutional system. In Europe, for example, its performance evaluation must support the manufacturer’s stated intended purpose and address scientific validity, analytical performance and clinical performance.2 Beyond regulatory requirements, it must also fit the operational and economic realities of the laboratories, clinicians, purchasers and health systems expected to use it.
Those are not administrative tasks to be completed after the technology is mature. They shape what must be developed, tested and funded from the beginning, and in my experience considering a technology ready without addressing these constraints is, to be blunt, not useful at all.
This is why my own TRL claim could be technically defensible and still be misleading about the programme as a whole. We had evidence that the test worked. We had evidence of use in a defined clinical circumstance. What we did not yet have was a diagnostic that was ready, in the fuller sense, to enter routine care.
The distance concealed by a single number
There is a vast distance between a research-level diagnostic that has been tested in an operational context and a final assay that can be responsibly offered to the market.
The first may have been developed close to an excellent clinical team, with access to samples, patients and clinicians willing to use it in a defined circumstance, and technologically savvy teams. It may generate very promising results, and that is real progress.
But it is still a research construct.
The second is a defined product or service. Its reagents and components are specified. Its supply is secured under appropriate contracts. Its manufacture and performance are governed by a quality system. Its intended use is defensible and specifically validated. Its result is shown to be reliable beyond the people who first developed and operated it.
It must work independently of the original team, in the hands of other operators, within the realities of laboratories, regulation, procurement and clinical practice.
The distance between these two states is measured in years and millions of Euros/Dollars/etc., to the surprise of TTOs, entrepreneurs and investors alike. Have you ever heard that hindsight is 20/20?!
Yet a research-level test can feel uncomfortably close to the finished product when both are described through a single TRL number. A team may have a technically credible TRL 6 or 7 claim and still be far from possessing a diagnostic ready for regulated, routine and commercially sustainable use. Even recently I was evaluating one project where the specific diagnostics research tool was used in a specific department for a subset of patients, with the assay conducted in the research centre associated with the hospital, by the graduate student that developed it as part of her thesis. For that team, if feels like "they made it to the clinical use", so the TRL is obviously high.
But that number will describe a real technical achievement while concealing the work that will determine whether that achievement ever becomes a product.
The supposedly non-technical problems that change the technology
The problem is not that TRL is useless. It is that, in diagnostics, the technology is conditioned from the beginning by the realities of regulation, supply, operation and market access.
A researcher may choose an enzyme because it gives the best analytical performance in the laboratory. Later, the team discovers that it cannot be obtained under the supply and quality arrangements needed for an OEM product. Or that it cannot tolerate the logistics of the markets in which the diagnostic is meant to be used. Keeping a product at “room temperature” through a Spanish or Saudi summer is not an abstract supply-chain question. It has a cost: refrigerated transport and temperature monitoring during transit is costly and will add far more than most people imagined to the final price of the product.
The answer may be to replace the enzyme. But a replacement changes the assay, will likely alter linear range, sensitivity, specificity or reproducibility. It will require new specifications and a return to technical validation.
The same is true when a critical supplier becomes unavailable in a target geography because of export controls or sanctions. Or when a component or logistics makes the eventual test too expensive for the laboratory, payer or health system expected to buy it.
Each of these looks, at first, like a commercial, regulatory or operational problem. In practice, each can become a technical problem again.
By the time a diagnostic reaches the market, it may bear little resemblance to the research assay from which it began. The reagents may be different. The format may be different. Sometimes the underlying technology is different. I have seen assays move from NGS to qPCR, or from mass spectrometry to ELISA, not because the original science was wrong, but because the original assay was incompatible with the conditions of real-world use.
This is what a single technical-readiness number obscures. The question is not simply whether the technology works. It is whether the technology can remain coherent while it becomes a regulated, manufacturable, affordable and usable diagnostic.
When the number produces the wrong investment decision
Used on its own, a TRL number can lead an investor, grant agency or founder to choose the wrong type and scale of support.
A programme described as TRL 6 or 7 may attract expectations of rapid commercialisation: a modest bridge round, a short grant, or early pressure to generate sales. Yet the diagnostic may still be years from market and require substantial work in assay redevelopment, clinical validation, quality systems, regulatory preparation, manufacturing, supply-chain qualification and market access. And I am not even considering clinical validation, but let's leave that to a different article.
The result is far more than simply a funding gap. It is a mismatch between the reality of the programme and the capital placed behind it.
Too little money is committed and it is committed for too short a period. Milestones are set around the wrong achievements. The balance between non-dilutive and dilutive funding is designed around an imagined route to market rather than the route the product will actually have to travel.
By the time the gap between aspiration and reality becomes visible, the company may have spent scarce capital proving the wrong thing, while the work that would make the diagnostic investable and adoptable remains unfunded.
This matters particularly in funding systems that use TRL as a decision shorthand. The EIC Accelerator, for example, frames grant support around innovation activities from TRL 6 to TRL 8.3 That may be entirely appropriate for a technology whose maturity maps cleanly onto its path to deployment. For a diagnostic, however, the same headline number can mask radically different clinical, regulatory, commercial and operational positions.
I don't think that a funder or investor should ignore technical readiness. But I believe that technical readiness, which frequently better describes a technological proof of concept achievement, should not be permitted to imply readiness for the investment being proposed.
Building on TRL, rather than discarding it
This is what led me, through accumulated personal experience - also known as mistakes - and through repeated encounters in grant evaluation, due diligence for VC investors and conversations with scientists and entrepreneurs, to develop Diagnostics Readiness Levels, or DRL.
DRL does not discard TRL. Technical readiness remains a necessary part of the picture. But it places TRL within a broader assessment of the conditions that determine whether a diagnostic can actually reach and remain in use.
At each stage, it asks questions across five parallel streams:
- scientific and technical;
- clinical;
- regulatory and quality;
- commercial and market; and
- strategic and systemic.
The purpose is not to produce a more impressive number. It is to reveal where a programme is already strong, where it remains exposed, and what work must be funded before the next technical milestone can be mistaken for market readiness.
The framework currently consists of three connected tools. A white paper sets out its logic. A practical, canvas-like tool helps entrepreneurs and innovators ask the right questions early; it is inspired by the spirit of the Business Model Canvas, but is designed around diagnostic development rather than business-model design. A simple Excel-based evaluation tool is intended for technology-transfer offices, grant panels and due-diligence exercises.
The framework is beginning to be tested by TTOs and investors in different countries. It remains a work in progress, not a validated universal score.
Its central discipline is simple: a diagnostic is only as ready as its least-ready condition for reaching the patients.
If you work with diagnostics, I would invite you to test it against your own reality. Did it reveal something you had not yet considered? Did it prove useless? At this stage, both answers are valuable. I am interested in what the framework helps people see, and where it fails. Please do get in touch with comments, criticisms, suggestions for improvement. My favourite AI was wonderful in making a dedicated site in short order (the DRL site), where you can find all relevant information and developments.
Sources for factual context
NASA, Technology Readiness Levels; NASA, Technology Readiness Levels Demystified.↩︎
Regulation (EU) 2017/746 on in vitro diagnostic medical devices, Annex XIII: performance evaluation and performance studies.↩︎
European Innovation Council, EIC Accelerator and guidance for applicants. Funding terms and eligibility criteria can change; verify these immediately before publication.↩︎