Integration of artificial intelligence into health technology assessment: optimizing clinical analysis and synthesis of evidence on effectiveness and safety

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Clinical assessment is one of the most labor-intensive components of Health Technology Assessment, as it requires systematic literature retrieval, critical appraisal, and synthesis of scientific evidence. Recent advances in artificial intelligence have created new opportunities to optimize these processes; however, comparative analyses of artificial intelligence-based tools supporting the clinical component of Health Technology Assessment remain limited. The aim of the study was to systematize current evidence on the application of artificial intelligence in the clinical component of Health Technology Assessment and to identify its advantages, limitations, and prospects for evidence synthesis. The study was based on a structured review of scientific publications using descriptive, comparative, and content analysis methods. The findings demonstrated that contemporary artificial intelligence-based tools support different stages of evidence synthesis, including literature retrieval, publication screening, structured data extraction, risk-of-bias assessment, and evidence synthesis. According to their functional purpose, the identified tools were grouped into two principal categories: specialized platforms developed to automate specific stages of clinical assessment and large language models capable of supporting a broad range of analytical tasks. Comparative analysis revealed substantial differences in their functional capabilities, degree of automation, and potential applicability to Health Technology Assessment. None of the identified artificial intelligence-based tools currently supports the complete clinical assessment workflow, indicating that they should be regarded as complementary rather than interchangeable technologies. For Ukraine, the integration of artificial intelligence-based tools into national and hospital-based Health Technology Assessment offers promising opportunities to improve the efficiency of evidence synthesis, optimize the use of expert resources, and reduce the time required to prepare the clinical component of Health Technology Assessment reports. The implementation of artificial intelligence-based tools should be accompanied by mandatory expert validation of the obtained results. Further integration of artificial intelligence into Health Technology Assessment practice requires the development of methodological guidance together with clear principles governing the application and validation of artificial intelligence-based tools.