Most medical device projects do not fail because the idea is wrong. They stall because important decisions are made too late: the intended use is vague, risks are found after the design is frozen, or the evidence needed for approval was never planned.
Here is how we approach medical device development, from first principle to first patient, and what to get right at each stage.
1. Define: the problem, the physics and the risk
Before any CAD, write down:
- Intended use and users. Who uses it, on which patients, in which setting? This shapes everything from usability to regulatory class.
- The physics. What has to be true for this to work? A quick feasibility calculation or simulation can save months.
- The biggest risks. Technical, clinical and commercial. Start a risk file early; standards such as ISO 14971 describe how.
The output is a short set of design inputs: measurable requirements you can later verify.
2. Design: architecture, CAD and simulation
Good architecture separates what is novel from what can be bought. Decide early which subsystems you will develop yourself and which you will source.
- Simulate before you build. Finite-element models of stress, heat and acoustics find failure points on screen rather than in the lab.
- Design for test. Plan how each requirement will be verified while you design, not afterwards.
- Think about standards now. Electrical safety (IEC 60601-1 and its particular standards), software (IEC 62304) and usability (IEC 62366-1) all influence the design. For diagnostic ultrasound, for example, IEC 60601-2-37 sets specific requirements.
3. Build: prototypes, rigs and firmware
Iterate quickly, but deliberately:
- Works-like and looks-like prototypes answer different questions; do not try to answer both at once too early.
- Custom test rigs make measurements repeatable and save enormous time over a programme.
- Keep records. Even early prototypes benefit from version control and simple design records, which become valuable when formal design controls start.
4. Validate: bench, preclinical and clinical evidence
Two words are often confused:
- Verification asks: did we build it right? (Does it meet the design inputs?)
- Validation asks: did we build the right thing? (Does it meet user needs in real use?)
Evidence typically builds from bench testing, through preclinical studies, to a clinical investigation where needed, designed according to ISO 14155. A well-designed study answers the regulator’s questions and the clinician’s questions at the same time. Plan it early, because it often sets the timeline for the whole programme.
5. Scale: documentation, handover and team
To reach the market you need a quality management system (commonly ISO 13485), a technical file for your regulatory route (for example the EU MDR or the US FDA), and a design that can be manufactured consistently. Just as important is the team and knowledge to carry the product forward.
What saves the most time
From our experience across ultrasound, optical sensing and drug delivery programmes:
- A clear intended use written down in the first weeks.
- Risk thinking from day one, not after the design freeze.
- Simulation and good test rigs before expensive prototypes.
- Planning the evidence: knowing what studies you will need, and designing for them.
- Senior technical leadership that connects engineering decisions to the clinical and commercial goals.
If you are planning a device programme, or one has stalled, we can help with technical leadership, engineering design and the path to clinical studies. Start a conversation.
