Research and validation

Research scenarios include cohort comparisons, repeated measurements and program outcome reviews.

Access to a feature does not imply independent clinical validation of every model.

Specify samples, metrics, windows, missing-data handling and quality criteria before analysis.

Separate correlation, prediction and causal effects. Report uncertainty.

Research partnerships, data scope and publishing terms are agreed separately.

Ethics and privacy requirements remain applicable.