AI wound-analysis apps: Check intended purpose, image quality and storage
Storing a wound image differs from providing treatment recommendations. Assess intended purpose, correct patient attribution and the complete image path.
Révision éditoriale :
Le texte de cet article est disponible en anglais dans cette version du site.
Contexte et sources
Software managing photographs differs from software producing information for medical decisions. MDCG guidance separates medical-device qualification from classification. The label wound app alone establishes neither class IIa nor suitability for every setting. [1]
An identifiable wound image can be health data through its account, case number or context. Skin features alone do not automatically make it special-category biometric processing; the GDPR's specific conditions apply. [2]
Security must cover capture, local copies, transmission, analysis and recipients, using appropriate technical and organisational measures. [3]
Portée et limites
No product's clinical accuracy or treatment effectiveness is certified here. CE marking does not replace checking the actual intended use. Written consent is not automatically the sole possible basis for every necessary clinical photograph.
En termes simples
Fictional example: A team uses authorised test material to check accidental gallery copies and incorrect patient attribution before a limited clinical deployment.
Recommandations pratiques
Our recommendation: Ask which functions store, measure or recommend, and review the corresponding documents. Test patient identification, poor-quality images and record export. Define when a result must be rejected or referred for clinical assessment.
Review local copies, personal photo backups, access and retention. Explain the purpose and minimise the image area. Provide a correction route; numerical precision on screen is not proof of a correct clinical result.