Table 5.

Analytic validation in practice.

Documentation you can expect

Description of analytical validation studies conducted according to the requirements of Good Clinical Practice (GCP). This description can be in any one or more of the following forms: • Internal documentation • Regulatory submission (510 k) • White paper • Published journal article In the documentation, the evidence for every algorithmic output in their system: • Description of the output metric • Overview of how the metric was calculated, including specific details where possible • Which reference standard was used as the comparator to validate the metric • Results from a direct comparison between calculated metric and reference standard, including statistical analysis methods • Description of the human subjects population and experimental conditions and protocol used in the aforementioned direct comparison testing If this validation testing was undertaken as part of a clinical trial with human subjects, then the Institutional Review Boards (IRBs) or Ethics Committees (ECs) documentation should also be provided.

Clinical users’ questions answered by analytical validation

Can an algorithm acceptably measure, detect, or predict the presence or absence of a phenotype or clinical condition when that algorithm is applied to sample-level data captured by a verified sensor in accordance with a specific data collection protocol in a particular population?