Prototyping Medical Devices: Stages, Design Controls and Costs

Prototyping medical devices follows the same physical stages as any hardware — looks-like, works-like, verification unit, pilot build — but each stage also has to produce documented evidence, because the regulator eventually reviews how the device was developed, not only what it does. Teams that treat documentation as a final task rebuild prototypes they already paid for. Teams that start recording user needs, requirements and risk analysis in the first month usually reach submission with the prototypes they already have.

Splitting looks-like from works-like early is what keeps a program affordable. Combining them too soon produces one expensive prototype that answers neither question cleanly and is painful to change.

In the US the framework is FDA design controls under 21 CFR 820.30, now converging with ISO 13485 through the Quality Management System Regulation. Anything contacting the patient adds biocompatibility evaluation under ISO 10993, and any device with software adds IEC 62304 lifecycle expectations.

A printed resin housing is fine for a usability study and unacceptable for a patient-contacting verification unit. Choose the production polymer, its grade and its supplier documentation before verification testing, because switching material afterwards can invalidate biocompatibility, sterilization and mechanical test results. The same applies to sterilization method: gamma, EO and autoclave each impose constraints on materials and packaging, and picking one late forces redesign.

Use the prototype cost calculator for a first estimate of the physical build, and add regulatory, testing and quality-system effort on top — for a Class II device, that overhead frequently exceeds the engineering budget.

Early medical prototypes exist to answer questions. At some point they stop being experiments and start being evidence: the unit used for a usability study, a bench test, or a biocompatibility sample has to be built from documented materials, by a documented process, from a controlled drawing. Teams that keep building shop prototypes past that point end up repeating the work under design control, which is the single most common source of schedule slip in early-stage device programs.

The practical rule is simple: the first build intended to support a claim is the first build that needs a build record, a bill of materials at a fixed revision, and a record of every deviation.

Costs assume a small-volume US build of a Class II electromechanical device and exclude testing fees. Sterile, implantable or software-heavy devices move well above these ranges.

The physical work is similar; the evidence requirement is not. Medical prototypes have to be traceable to documented user needs and design inputs, tied to a risk analysis, and built from materials and processes you can justify later. That documentation trail, not the hardware, is what usually determines the schedule.

Not for the earliest exploratory models, but design controls apply once you begin the design that will become the product, and a formal quality system is required before manufacturing for clinical or commercial use. Most teams stand up a lightweight system early and mature it, rather than retrofitting one under submission pressure.

Yes, and it is the standard approach for looks-like models, fixtures and early works-like builds. Printed parts are generally not acceptable for verification or clinical units unless the printing process and material are themselves qualified for the application, including biocompatibility and sterilization.

Plan for at least three or four: one or more looks-like rounds for usability, one or two works-like rounds for function, and a design-locked round for verification. Programs that budget for a single prototype almost always overrun, because the first build's job is to reveal what the requirements missed.

The difference between a bench model and a build that supports a submission is traceability. Auditors ask which drawing revision a unit was built to, which lot of material went into it, who built it and what deviated. Capture that from the first functional prototype and the verification build becomes an administrative step; capture it late and you rebuild units purely to generate paperwork.

Serialize every unit from the first functional prototype onward, even the ones you expect to destroy. When a test result later looks wrong, the serial number is what lets you tell a build problem from a design problem — and that distinction is usually worth weeks of schedule.

A medical technology pitch is judged on evidence, not enthusiasm. Clinical reviewers want to know which workflow the device changes and what happens to patient outcomes and staff time when it does. Investors want to know the regulatory pathway, the cost of getting through it, and whether reimbursement exists or has to be created. Both audiences discount claims that a prototype cannot support, so the prototype you bring matters as much as the deck.

Bring the roughest artifact that proves the hardest claim. A works-like rig that demonstrates the core measurement is worth more than a beautiful non-functional model, because the measurement is the risk. Keep the looks-like model for the market conversation and be explicit about which one you are holding.

Work with LA NPDT: if you are moving from here to execution, start with our product development consulting or talk to us about end-to-end product development.