Assessment and implementation of AI-enabled digital health technologies in hospitals: a review and qualitative document analysis of HTA reports and guidance documents.

Authors

Abstract

Artificial intelligence-enabled digital health technologies (AI-enabled DHTs), encompassing computing platforms, connectivity, software and sensors that incorporate AI or machine learning components for healthcare applications, are increasingly used in hospitals, yet practical methodological frameworks to guide investment decisions remain unclear. This publication, based on a comprehensive Health Technology Assessment (HTA) report, aims to provide hospital managers with structured approaches for assessing the added value of AI-enabled DHTs in clinical settings. We conducted a review with qualitative document analysis: (1) a focused hand search of 51 HTA agency websites through May 2024, supplemented by results from the ASSESS DHT project; (2) narrative comparison and thematic analysis of HTA methodology documents; (3) standardised extraction from HTA reports and (4) synthesis into practical recommendations to inform hospital procurement decisions. We identified five methods guidance documents and 30 HTA reports from 13 HTA agencies. The frameworks varied in scope, with most addressing DHTs and AI broadly, while some focused specifically on AI. Each framework highlighted AI-specific assessment domains beyond standard HTA criteria. The majority of HTA reports (27/30) focused on diagnostic and screening applications, particularly in radiology and internal medicine. In line with the study objective of providing structured guidance for hospital decision-makers, our synthesis shows that conventional HTA methods suffice for many domains, but AI-enabled DHTs require additional technical, ethical and organisational considerations. The European Network for HTA Core Model, supplemented with AI-specific guidance, provides a practical evaluation toolkit for hospitals.

Tags

Topics