Design Validation

Design validation is the process of confirming that a medical device, as a finished product, meets user needs and its intended uses under actual or simulated use conditions. It is one of the last stages of design controls, followed by design transfer to production, and must be completed before the device is released to customers. It determines whether the device that was designed and manufactured actually works as intended for its target users in its target environment. Design validation is required by ISO 13485 (clause 7.3.7), under the EU Medical Device Regulation (MDR), and by FDA: since February 2, 2026, FDA's Quality Management System Regulation (21 CFR 820) incorporates ISO 13485 by reference, replacing the former US design validation section, 21 CFR 820.30(g).

A device can meet every technical specification (verified through design verification) and still fail in practice if it is confusing to operate, uncomfortable for patients, or unsuitable for the clinical environment. Design validation bridges the gap between engineering specifications and real world performance. Regulatory agencies give particular weight to validation evidence because it represents the closest approximation to how the device will actually be used. Inadequate validation has been implicated in high profile device recalls where products that passed bench testing failed in clinical use. Validation is also essential for regulatory submissions; without adequate validation data, a premarket application is unlikely to receive clearance or approval.

Design validation testing is conducted on production equivalent or production units, not on prototypes, because the goal is to evaluate the device as it will actually be manufactured and distributed. The validation plan specifies the test methods, acceptance criteria, sample sizes, and statistical rationale. It draws its requirements from the design inputs, which capture user needs and intended uses.

Clinical validation may involve clinical investigations (clinical trials) that evaluate safety and performance in the target patient population. Simulated use validation involves representative users (clinicians, patients, or both) performing representative tasks in environments that simulate actual use conditions. Human factors validation testing, described in FDA's human factors guidance and in the usability engineering standard IEC 62366-1, assesses whether representative users can safely and effectively perform the critical tasks, those that could cause serious harm if performed incorrectly or not at all. FDA's guidance is nonbinding; it generally recommends at least 15 participants from each distinct user population, and it asks for a qualitative analysis of the root cause of every use error rather than a numerical pass rate.

The validation protocol must account for the full range of intended users (including those at the extremes of the user population, such as elderly patients, users with impaired vision, or emergency department staff under time pressure), the full range of use environments (including home use, ambulatory settings, and intensive care units as applicable), and the full range of use conditions (including foreseeable misuse scenarios).

Results are documented in a validation report that includes test protocols, raw data, statistical analyses, deviations from the protocol, and conclusions about whether each acceptance criterion was met. If validation testing reveals deficiencies, the design must be modified and re validated before the device can proceed to market. Any modifications trigger an assessment of whether previous verification activities need to be repeated.

Design validation is a mandatory regulatory requirement for all medical devices subject to design controls. It should be conducted after design verification is complete and the device is manufactured using production processes. It is also required whenever significant design changes are made to an already marketed device, to confirm that the changes do not adversely affect safety or performance. Simulated use testing is not a substitute for clinical evidence in higher risk device categories, where clinical investigations may also be required as part of validation.

A frequent error is conducting validation too early, using prototypes that do not represent the production device, which means the validation does not reflect what patients and clinicians will actually receive. Another mistake is designing validation studies with insufficient sample sizes or with users who do not represent the actual target population (such as using trained engineers instead of naïve end users). Confusing validation with verification remains common; a bench test that confirms a thermometer reads within plus or minus 0.1°C is verification, while a study showing that nurses can correctly read the display and interpret results in a clinical setting is validation. Failing to include foreseeable misuse scenarios in validation planning can result in devices that perform well under ideal conditions but fail when users interact with them in unexpected but predictable ways. Finally, treating validation as a pass/fail gate without learning from the process, ignoring near misses or use difficulties that did not rise to the level of critical errors, misses opportunities to improve the device.

A company developing a home use blood glucose monitor conducts design validation with 90 diabetic patients representing the intended user population, including 30 patients over age 65 and 15 with visual impairment. Participants perform blood glucose testing using the production device in their homes over a two week period, without any assistance from the manufacturer. The validation protocol specifies that 95% of self tested values must fall within plus or minus 15% of laboratory reference values, and that every use error on a critical task (one that could lead to incorrect insulin dosing) will be analyzed for its root cause and residual risk rather than judged against a numerical error rate. Results show 96.8% of values within specification, meeting the accuracy criterion. Use errors on critical tasks occurred in 0.9% of attempts; each was analyzed for its root cause, and the manufacturer documented why the residual risk was acceptable before the results supported the regulatory submission.