Resources
Medical Device Due Diligence Checklist: The Technical and Clinical Questions
By Lloyd Marks · Published
Medical device due diligence asks whether the technical and clinical story behind an investment or acquisition holds up under examination. Financial, legal, regulatory and intellectual-property diligence each have their own specialists. The technical and clinical review sits underneath them. It tests whether the device does what is claimed, whether the clinical need is real, whether the evidence supports the claims made for it, and where the assumptions in the plan are most fragile.
This article sets out what that review should examine, where uncertainty usually sits, and what an independent review can and cannot conclude. It is general information, not investment, securities, legal or regulatory advice, and nothing in it predicts the outcome of any particular investment.
What technical and clinical diligence is for
A prospective investor or acquirer is usually handed a polished account: a presentation, a summary of clinical results, a regulatory timeline and a market estimate. Each element was assembled by people who believe in the company. That is not a criticism; it is the reason an outside reader is useful. The job of a technical and clinical review is to trace each important claim back to its underlying evidence and to say how well it is supported.
It helps to agree at the outset which decision the review supports (a first investment, a follow-on round, a license or an acquisition) and which claims matter most to that decision. A follow-on investor may care most about whether the next development milestone is achievable. An acquirer may care most about whether the product fits its clinical and sales channels.
The device itself
Start with what physically exists and what it has been shown to do.
- What it does, and how
A clear explanation of the mechanism, in terms a clinician and an engineer would both accept. A mechanism that cannot be explained plainly is an early warning.
- Technical maturity
Concept, bench prototype, frozen design or production units? Projections frequently assume a later stage than the evidence supports.
- Verification data
Bench, simulation and animal testing: what was measured, against what specification, on how many units, and by whom.
- Usability
Who will use the device, in what environment, and which tasks could cause harm if done wrong. FDA's human factors guidance frames these questions around intended users, uses and use environments.
- Manufacturing and cost
Whether the device can be made reliably at the volume and cost the plan assumes, and how the prototype differs from what would actually be built.
- Software and connectivity
Where the device depends on software: what the software does, how it is updated, and how it interacts with other equipment.
Clinical need: the 'solution looking for a problem' test
An impressive technology can still be a solution looking for a problem. The review should establish who has the clinical problem, how it is managed today, how often it arises, and whether the people who would buy the device consider it a problem worth paying to solve. Direct conversations with practicing clinicians and with purchasing decision-makers in the target setting are often more informative than a market report, because they show whether the claimed benefit matters to the people who would have to change their practice.
Reading the clinical evidence
Clinical results are where a review most often finds a gap between what the data show and how they are described. The questions are straightforward, even when answering them takes care:
- Design. Was there a comparison group? Were patients randomized? Was the primary endpoint defined before the data were collected?
- Population and setting. Were the patients and sites representative of the intended market, or a selected group at an expert center?
- Endpoints. Does the primary endpoint measure what patients, clinicians and payers care about, or a surrogate that may not translate?
- Size and precision. How many patients were studied, and how wide is the uncertainty around the result? Small studies can produce large apparent effects.
- Sponsorship and analysis. Who ran and analyzed the study, and were all the predefined outcomes reported?
- Portrayal. Do the investor materials describe the results the way the study report does? Watch for subgroup findings presented as main results, relative effects without absolute numbers, and pilot results presented as if they were confirmatory.
Frameworks such as GRADE give a structured vocabulary for this: the certainty of evidence is assessed outcome by outcome, considering risk of bias, imprecision, inconsistency, indirectness and publication bias. A review should say, for each claim that matters, how much confidence the evidence supports.
Regulatory status: a fact to verify and an assumption to test
Regulatory status is usually presented as a milestone. In diligence it is something to check and, for a device not yet on the market, a plan to question.
- Verify what has been granted. FDA's public 510(k) database can be searched by applicant, device name and product code. Confirm that the cleared device, its stated intended use and the product being sold or planned are the same thing.
- Understand what a clearance means. A 510(k) clearance reflects an FDA finding of substantial equivalence to a legally marketed predicate device. On its own it is not evidence that the device improves outcomes compared with the alternatives.
- Check the postmarket record. FDA's recall database covers device recalls classified since November 2002, and the MAUDE database holds adverse-event reports from mandatory and voluntary reporters. These sources can be incomplete, and a count of reports is not a rate. They prompt questions; they do not settle them.
- Test the planned pathway. For a device not yet on the market, the plan assumes a pathway, such as 510(k), De Novo or premarket approval, and a timeline. Whether a predicate exists, whether clinical data will be needed and whether FDA has given feedback through a pre-submission are questions for the company and for qualified regulatory counsel.
A technical reviewer can point out where regulatory assumptions look fragile. Conclusions about regulatory strategy belong to qualified regulatory counsel, and decisions belong to FDA.
Intellectual property: technical questions for counsel
Patent counts are easy to put on a slide and easy to misread. Issued patents, pending applications and provisional applications are different things, and patent families and current ownership are different again. A U.S. patent grants the right to exclude others from making, using, offering for sale or selling the invention. It does not grant the right to make or sell the product, which can still be blocked by someone else's patents.
A technical reviewer can support counsel by explaining how the device works, comparing it with patent claims in plain technical terms, and identifying earlier devices that appear similar. Validity, infringement, freedom to operate and ownership are legal questions for patent counsel, and a technical review should not be read as answering them.
Competition, direct and indirect
- Direct competitors
Devices that address the same problem in a similar way, whether on the market or in development.
- Indirect competitors
Other devices, drugs, procedures or workflow changes that address the same clinical problem differently.
- The current standard of care
What clinicians do now. It has the advantages of being familiar, already purchased and already trained on.
- The next generation
What established companies may launch while this device is still in development. A comparison with today's market can be out of date by launch.
Adoption, workflow and payment
Adoption depends on more than clinical merit. The review should ask who uses the device, who decides to buy it, how it fits the existing procedure and equipment, how much training it needs, and whether its form factor suits the setting. On payment, it should ask whether the device fits existing coding and coverage or depends on new coverage. Medicare national and local coverage determinations are public in CMS's Medicare Coverage Database, a reasonable starting point for asking how comparable devices and procedures are covered. Detailed reimbursement analysis is a specialty of its own.
Where assumptions are most uncertain
Some assumptions in a device plan are more uncertain than they look. A review should identify which of these the plan depends on and what evidence would reduce the uncertainty.
- Market size
Top-down estimates multiply a large population by an assumed share and price. The more useful question is how many specific clinicians or facilities would buy, at what price, and why. Check that the arithmetic and its inputs are internally consistent.
- Effect size
Results from early, small or single-site studies can shrink in larger and more representative ones. Ask what the plan assumes the confirmatory result will be, and what happens if it is smaller.
- Regulatory timing
Timelines often assume the most favorable pathway and no requests for additional data.
- Speed of adoption
Clinical practice usually changes gradually, and purchasing review adds time. A plan that assumes rapid uptake should explain why.
- Cost to manufacture
Prototype costs say little about production costs, and low-volume production costs say little about costs at scale.
- Coverage and payment
An assumed coverage decision or payment level is a major dependency if the device adds cost to care.
What an independent review can and cannot conclude
An independent technical and clinical review can:
- explain how the device works and how mature it is;
- assess whether the clinical need is real, and for whom;
- judge how well the technical and clinical evidence supports the claims made;
- identify where the plan's assumptions are weakest and what evidence would test them;
- frame specific questions for regulatory counsel, patent counsel and financial advisors.
It cannot:
- predict investment returns, valuation or commercial success;
- provide securities, investment, legal or tax advice;
- determine patent validity, infringement or freedom to operate;
- determine the regulatory pathway or predict an FDA decision;
- verify information it was not given, or substitute for testing that has not been done.
A review is a reasoned opinion about the evidence at a point in time. Its value is in making uncertainty visible before money is committed, not in removing it.
I have done this kind of diligence myself: Datascope (now Mindray) sent me to evaluate a startup's automated external defibrillator for in-hospital use, and after reviewing the data and studying the device I advised against the acquisition. For how I write up a review, see my sample assessment of a robotic blood-drawing device.
For acquirers: fit as well as merit
Corporate acquirers have additional questions. Does the device fit the existing sales channel and the clinicians the company already serves? Does it complement or compete with current products? Can the acquirer's manufacturing and quality systems absorb it? Could the company's own development team build something comparable sooner or at lower cost? A technology can have real merit and still be a poor fit. See medical technology assessment for corporate development.
Starting a review
A first conversation can usually establish fit without confidential information: the device category, the stage of development, the decision being made and its timing. Detailed technical and clinical materials are best exchanged only after the parties have agreed confidentiality terms in writing. Materials that make a review useful typically include technical descriptions and test reports, clinical study protocols and reports, regulatory correspondence, the patent list with the status of each item, the competitive analysis, and the assumptions behind the commercial plan. For investors, see due diligence for medical technology investors.
Discuss a due diligence review
Describe the device category, the decision you are weighing and your timeline. Please leave confidential details out of a first inquiry.
Questions and answers
Is a technical and clinical review the same as investment advice?
No. A technical and clinical review examines the device, the clinical need and the evidence, and identifies where assumptions are uncertain. It does not recommend whether to invest, value a company or advise on securities. Those decisions remain with the investor and their financial and legal advisors.
Can a review tell us whether the company's patents are valid or whether the product infringes?
No. Validity, infringement, freedom to operate and ownership are legal questions for patent counsel. A technical reviewer can help counsel understand how the device works and how it compares technically with the patents and earlier devices.
Does an FDA 510(k) clearance show that a device works better than the alternatives?
Not by itself. A 510(k) clearance reflects an FDA finding that the device is substantially equivalent to a legally marketed predicate. Whether the device improves outcomes compared with alternatives is a separate question that depends on the clinical evidence.
What if the company will not share its underlying clinical or test data?
Then the review can assess only what it is given and should say so. The absence of underlying data is itself a finding, and it limits how much confidence the summary results can support.
How is diligence for an acquirer different from diligence for an investor?
Both examine the technology and its evidence. An acquirer also has to judge fit: with its sales channel, its existing products, its manufacturing and quality systems, and the alternative of developing something internally.
Sources
- HTA Glossary (INAHTA/HTAi et al.): definition of health technology assessment — Supports describing HTA as a systematic, explicit-methods comparison of a technology's intended and unintended consequences with existing alternatives.
- GRADE Working Group: overview of the GRADE approach — Supports the certainty-of-evidence domains: risk of bias, imprecision, inconsistency, indirectness, publication bias.
- FDA guidance: Applying Human Factors and Usability Engineering to Medical Devices — Supports framing usability around intended users, uses and use environments, and reducing use errors and resulting harm.
- FDA: Premarket Notification 510(k) — Supports the substantial-equivalence standard and the definition of a predicate device.
- FDA: 510(k) Premarket Notification database — Supports that cleared 510(k)s can be searched by applicant, device name, product code and decision date.
- FDA: Medical Device Recalls database — Supports that the database covers recalls classified since November 2002, with caveats on posting dates and unclassified actions.
- FDA: MAUDE (Manufacturer and User Facility Device Experience) database — Supports that MAUDE holds mandatory and voluntary reports and may not include all reports (exemptions, variances).
- FDA: Classify Your Medical Device — Supports risk-based classes I-III, classification by intended use/indications, the product classification database, and 513(g) requests.
- USPTO: Patent essentials — Supports that a patent is a right to exclude others, not a right to make, use or sell the invention.
- CMS: Medicare Coverage Database — Supports that national (NCD) and local (LCD) Medicare coverage documents are publicly searchable.