How many participants do you need for medical device usability testing?

The short answer

Formative evaluation: five to eight participants per user group, per round, over several rounds. The aim is to find problems, and small rounds repeated find more than one large round.

Summative (validation) testing: at least 15 participants per distinct user group. That is the minimum in the FDA's human factors guidance for medical devices, and the MHRA's human factors guidance takes the same approach. IEC 62366-1, the usability engineering standard, sets out the process but does not fix a number.

So the total for validation depends on how many distinct user groups the device has. That is where most of the work is.

How to define your user groups

A user group is a set of people whose characteristics could change how they use the device, and so which use errors they might make. Typical splits:

• Patients and healthcare professionals
• Lay carers, such as a parent injecting a child
• Adolescents and adults, where both will use the device themselves
• First-time users and people experienced with a similar device
• People whose condition affects dexterity, vision or cognition, where that is part of the intended population

An auto-injector intended for adult patients, adolescent patients and carers is three groups. At 15 each, that is 45 validation participants before anyone has been recruited twice by mistake.

Merging groups to save money is the most common shortcut and the riskiest. If a regulator thinks two groups could use the device differently, a combined sample of 15 does not show either one is safe. Grouping decisions should come out of your use-related risk analysis and be written down with the reasoning.

Formative testing: smaller, earlier, repeated

Formative evaluation runs throughout development, from early concepts to near-final prototypes. It can use interviews, expert review and simulated-use sessions, and it is where the design actually improves.

Five to eight people per group per round is usually enough to surface the main use problems. Fix them, then test again. Two or three rounds like that, with the critical tasks in every one, are what make summative testing a confirmation rather than a surprise.

We ran formative design validation on five 3D-printed prototype variants of an insulin device with people living with diabetes. You can read about it in the insulin device case study.

Summative validation: what the 15 have to show

Summative testing is not a satisfaction survey. It shows that representative users can complete the critical tasks, the ones where a use error could cause harm, on a production-equivalent device, with the real instructions and labelling, in a realistic setting.

• Representative participants. People with the condition, or the clinicians who will use the device, recruited to match the intended users. Not healthy volunteers acting the part.
• Realistic training. If users will get a five-minute demonstration from a nurse, that is what participants get, with a realistic gap before they use the device.
• Every use error and close call recorded, with a root-cause interview afterwards to understand why it happened.
• A simulated environment close enough to real use: a kitchen table, a ward, a noisy clinic.

Recruiting 15 people per group with a specific condition, in person, at the right point in their treatment, is the part that most often slips a programme. Plan it early.

What it costs

A formative round of eight in-person sessions on a regulated product comes to £15,250 to £17,500 in our cost calculator, including recruitment, participant payments and three days on site.

Summative validation is scoped once the user groups and critical tasks are known, because the number of groups drives everything. Complex validation work starts from £16,000. See design validation.

Questions device teams ask us

Can we reuse formative participants in summative testing? Avoid it. People who have seen earlier versions are no longer representative of first-time users.

Is 15 enough for a big user population? It is the minimum per group, not a target. Higher-risk devices or very varied groups may justify more.

Do we need to test in the UK? Regulators expect participants who represent the people who will use the device in that market. Healthcare systems, training and prescribing differ between countries, and those differences can change how a device gets used.

Can testing be remote? Formative interviews and some software-only devices, yes. Simulated-use testing of a physical device almost always needs to be in person.