Resources
Medical Device Technology Assessment: What It Should Cover
By Lloyd Marks · Published
A medical device technology assessment is a structured review of whether an invention deserves the next investment of money, time or credibility, and of what has to be learned before that investment is made. Done well, it is not a verdict on whether an idea is clever. It is an organized account of the clinical problem, the alternatives, the strength of the evidence, the questions a prototype or study must answer, and the regulatory and commercial questions that remain open.
This article sets out what an assessment should cover and where its limits lie, for inventors, university technology-transfer offices and corporate development teams.
Start with the decision the assessment must inform
Every useful assessment begins by naming the decision it supports. "Is this a good invention?" is too broad to answer. "Should we spend on a working prototype before approaching licensees?", "Is this faculty invention a priority for the office's patent budget this year?" and "Should our product team take a second meeting with this inventor?" are decisions with a cost, a deadline and an alternative.
Naming the decision sets the depth of the review, identifies the assumptions that matter most, and makes the conclusion useful: the assessment either supports the decision, argues against it, or shows that it should wait for specific evidence.
The underlying economics are simple. Patenting, developing, licensing and commercializing a medical invention is expensive, so whoever pays for the next step needs a reasoned view of two separate things: whether the technology can do what it is meant to do, and whether anyone will adopt it if it does.
Clinical need: is there a problem worth solving?
A common way for a medical invention to fail is to work as designed while solving a problem no one is willing to change their practice for: a solution looking for a problem. An assessment should treat clinical need as a claim to be tested, not a premise. In concrete terms it asks:
- Which patients have the problem, in what setting, and how often does it arise?
- What happens today when it arises, and what does that cost the patient, the clinician and the institution?
- Who notices the problem? A problem clinicians tolerate quietly is harder to sell than one that causes delays, complaints or adverse events.
- Who would decide to adopt a new approach, and who would pay for it? Those are often different people.
The most direct way to answer these questions is to ask the people involved. Interviews with potential users and buyers, such as the clinicians who would use the device, the nurses or technicians who would set it up, and the purchasing or value-analysis staff who would approve it, often show that the real problem is adjacent to the one the inventor set out to solve.
Alternatives, including doing nothing
A new device competes with everything a clinician could do instead.
- Existing devices
Products already on the market for the same use, including older devices that are cheaper, familiar and already stocked. A new design has to be better at something the buyer values, not only at something the engineer values.
- Other forms of care
Drugs, procedures, imaging strategies or referral pathways that address the same clinical problem without a new device.
- Workflow changes
A checklist, a protocol change or better training sometimes solves the problem well enough that a device is not worth its cost and training burden.
- Doing nothing
If the problem is rare, self-limiting or tolerated, the status quo is a real competitor and should be treated as one.
- Work in progress elsewhere
Issued patents, published patent applications, conference abstracts and registered clinical studies show where others are heading. Reviewing existing patents and applications belongs in this map of alternatives. It is a technical review, not a patentability or freedom-to-operate opinion, which is a question for patent counsel.
Evidence quality: what is known, and how well
A review of the medical literature establishes what is already known about the clinical problem, the alternatives and any earlier attempts at a similar solution. The harder question is how much confidence the evidence deserves.
Health technology assessment, as the field defines it, uses explicit and systematic methods and judges value by comparing a technology's intended and unintended consequences with those of existing alternatives. The GRADE approach, widely used in evidence review, rates the certainty of evidence outcome by outcome and considers risk of bias, imprecision, inconsistency, indirectness and publication bias. Those domains make a practical checklist even for an early-stage device:
- Risk of bias. Were the data collected and analyzed by people who stand to benefit from a favorable result, without blinding or a predefined analysis?
- Imprecision. Does the result rest on a handful of cases, where a few different outcomes would change the conclusion?
- Inconsistency. Do bench results, animal work and early human use point the same way?
- Indirectness. Was the device tested in the population, setting and hands it is intended for? Does the endpoint measure something patients and clinicians care about, or a surrogate?
- Publication bias. Are the favorable studies the only ones visible?
Often the honest finding is that direct evidence does not yet exist. Stating exactly which evidence is missing, and what question it would answer, is one of an assessment's most useful outputs.
Prototype and testing questions
Decide whether a working prototype is necessary
A prototype is expensive, so the first question is what it would answer. Some questions, such as whether the mechanism works at all, need hardware. Others, such as whether clinicians want the function, can often be answered with drawings, a simple model or structured interviews.
Define what the prototype must demonstrate
State the specific performance, size, safety or usability question in advance, along with the result that would change the plan. A prototype built to impress tends to answer no question in particular.
Consider the intended users and use environment early
FDA's human factors guidance frames usability around the intended users, uses and use environments, and around reducing use errors and the harm they cause. For an early concept the practical questions are who will handle the device, under what conditions, with what training, and which steps could go wrong in ways that matter.
Decide whether a clinical study is advisable now
A clinical study is justified when the remaining uncertainty can only be resolved in patients and the answer would change the decision. If bench work, simulation or user research could resolve it more cheaply, that work comes first.
If a study is warranted, design it around the decision
The design questions are the population, the comparator, the primary endpoint, the number of participants needed to detect a meaningful difference, and who will run and analyze the study. In the United States, clinical investigation of a device that is not yet cleared or approved generally requires an investigational plan approved by an institutional review board, and a significant-risk device study also requires an investigational device exemption approved by FDA. How those rules apply to a particular study is a question for qualified regulatory counsel.
Regulatory pathway questions to answer with qualified regulatory counsel
An assessment should surface the regulatory questions early, because the likely pathway affects the evidence required, the time to market and the cost. It should not answer them in place of qualified regulatory counsel, and only FDA makes classification and marketing decisions. Its useful contribution is a precise list of questions:
- What exactly is the intended use and indication? FDA classification depends on intended use and indications for use and is risk-based, so a small change in the claimed use can change the analysis.
- Does an established device type and product code fit? FDA's product classification database identifies classification regulations and product codes for established device types.
- Is there a legally marketed predicate? A 510(k) must show substantial equivalence to a predicate: the same intended use, and either the same technological characteristics or differences that do not raise new questions of safety and effectiveness.
- If there is no predicate, could De Novo classification apply? FDA describes De Novo as the pathway for classifying novel devices for which general controls, or general and special controls, provide reasonable assurance of safety and effectiveness, but for which no legally marketed predicate exists.
- Could the device be Class III? Premarket approval applies to the highest-risk devices and requires sufficient valid scientific evidence that the device is safe and effective, typically including clinical investigations.
- Would early feedback from FDA be worth seeking? FDA's Q-Submission program provides ways to request written feedback or a meeting before a marketing submission or clinical study, and a 513(g) request asks FDA for a formal classification determination.
Any answer to these questions in an assessment is a working assumption to be confirmed, not a conclusion.
Adoption: form factor, workflow and the people who decide
An invention can be sound and still fail because of how it is packaged. Size, weight, the number of steps needed to set it up, how it connects to existing equipment, whether it is reusable or disposable and how it is cleaned can decide adoption more than the core idea does. Sometimes the form factor has to change before a good invention becomes a product clinicians will use.
An assessment should also identify the people whose opinion carries weight in the clinical field: the specialists who set practice, publish, train others and sit on purchasing committees. Their early reaction is evidence about need and adoption, and relationships with them can matter later for study design and acceptance. Their input should be reported candidly, including when it is unfavorable.
Payment belongs in the same discussion. A device that adds cost to a procedure without a clear route to coverage, or a clear saving elsewhere, will meet resistance however well it works. An assessment can frame those questions; detailed reimbursement strategy is usually a separate specialty.
What a useful assessment hands back
The output should be organized around the decision named at the start. In practice that means:
- the decision and the question the assessment was asked to answer;
- the scope, including what was reviewed and what was not;
- the evidence considered, with its sources and a candid view of its quality;
- the assumptions the conclusions depend on, stated so they can be checked;
- the findings, separating what is established from what is judgment;
- the open questions, ranked by how much they affect the decision, with the least expensive way to answer each;
- the date, because evidence, competitors and regulatory context change.
The approach described here expands the assessment methodology Medical Invention Consulting has published since 2020: interviewing potential users or buyers of the technology, reviewing existing patents and patent applications, reviewing the medical literature, determining whether a working prototype is necessary, determining whether a clinical trial is advisable, designing clinical trials, and establishing relationships with opinion leaders in the field where the invention will be used. It combines engineering, clinical and business perspectives because the questions that decide an invention's future often sit where those disciplines meet.
For a worked example, read my sample assessment of a robotic blood-drawing device.
The limits of an assessment
An honest assessment says what it cannot do.
- It cannot predict commercial success. It can identify the conditions success would require and how well supported each one is. Markets, competitors, payment and timing all change.
- It is not a legal opinion. Reviewing patents and applications informs the technical picture. Patentability, infringement, freedom to operate and ownership are questions for patent counsel, and a patent is a right to exclude others rather than a right to make or sell the invention.
- It is not a regulatory determination. It can frame pathway questions. Classification and marketing authorization are FDA decisions, and regulatory strategy belongs with qualified regulatory counsel.
- It is not clinical validation. Reviewing evidence and designing a study is different from having run one. Only data from appropriate testing shows how a device performs.
- It depends on what it is given. Conclusions are only as complete as the information, access and time available, and should say where those were limited.
- It is dated. Its conclusions apply at the time it was written.
How the questions shift with who is asking
- Inventors
The decision is usually whether to spend more of your own money or time, and on what. The most valuable output is often a short list of the least expensive tests that would tell you whether to continue. See consulting for medical inventors.
- Universities
A technology-transfer office has to decide whether to protect, develop or license a faculty invention within a limited budget. An assessment helps compare inventions on clinical need, differentiation and evidence, and anticipates what a prospective licensee will ask first. See assessment for university technology transfer.
- Corporations
Device companies receive outside submissions and internal proposals and need to judge them consistently. The questions widen to fit with the existing portfolio, sales channel and manufacturing, and to whether the idea is a solution looking for a problem. See assessment for corporate development.
Discuss a technology assessment
If you are weighing a decision about a medical device invention, describe the decision and the stage you have reached. Please leave confidential technical details out of a first inquiry.
Questions and answers
How is a technology assessment different from a patent search?
A patent search looks for prior patents and publications, usually to inform a patentability question for patent counsel. A technology assessment uses that information as one input among several. It also examines clinical need, alternatives, evidence quality, prototype and testing questions, and adoption. It does not give a legal opinion on patentability, infringement or freedom to operate.
Does an invention need a working prototype before it can be assessed?
No. Many important questions, including whether the clinical need is real, what the alternatives are and what evidence exists, can be examined before anything is built. Part of the assessment is deciding whether a prototype is needed at all, and if so, what specific question it should answer.
Can a technology assessment tell me which FDA pathway my device will follow?
It can identify the questions that determine the pathway, such as the intended use, whether a predicate exists and the device's likely risk class, and flag where the answers are uncertain. The pathway itself should be confirmed with qualified regulatory counsel, and classification and marketing decisions are made by FDA.
When is the right time to commission an assessment?
Before a significant spending decision: filing or continuing patent protection, building a prototype, starting a clinical study, approaching licensees or taking an outside submission further. An assessment is most useful while its findings can still change the plan.
What should I prepare before a first discussion?
A short, nonconfidential description of the clinical problem, the stage of development and the decision you are facing is enough to start. Detailed technical information should be shared only after confidentiality terms have been agreed in writing.
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: Classify Your Medical Device — Supports risk-based classes I-III, classification by intended use/indications, the product classification database, and 513(g) requests.
- FDA: Premarket Notification 510(k) — Supports the substantial-equivalence standard and the definition of a predicate device.
- FDA: De Novo Classification Request — Supports the description of De Novo as the classification pathway for novel devices without a legally marketed predicate.
- FDA: Premarket Approval (PMA) — Supports PMA as the review for Class III devices and its 'sufficient valid scientific evidence' standard.
- FDA: Investigational Device Exemption (IDE) — Supports the IRB-approved investigational plan requirement and FDA IDE approval for significant-risk device studies.
- 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 guidance: Requests for Feedback and Meetings for Medical Device Submissions (Q-Submission Program) — Supports the statement that submitters can request written feedback or meetings with FDA before IDE, De Novo, 510(k) and PMA submissions.
- USPTO: Patent essentials — Supports that a patent is a right to exclude others, not a right to make, use or sell the invention.