Qualitative Evaluation of an AI Tool for Opportunistic Detection of Vertebral Fractures in the National Health Service
Author(s)
Catriona Inverarity, PhD1, Minjie Gao, PhD2, Jiyoun Chang, PhD2, Manjot Brar, BSc3, Ruphinder Kaur, MSc4, Lauren Gatting, PhD3, Juan I. Baeza, PhD5, Anna Barnes, PhD1.
1King's Technology Evaluation Centre (KiTEC), King's College London, London, United Kingdom, 2Department of Public Services Management & Organisation, King's College London, London, United Kingdom, 3King's College London, London, United Kingdom, 4Clinical Engineering, Guy's and St. Thomas' NHS Foundation Trust, London, United Kingdom, 5King's Business School, King's College London, London, United Kingdom.
1King's Technology Evaluation Centre (KiTEC), King's College London, London, United Kingdom, 2Department of Public Services Management & Organisation, King's College London, London, United Kingdom, 3King's College London, London, United Kingdom, 4Clinical Engineering, Guy's and St. Thomas' NHS Foundation Trust, London, United Kingdom, 5King's Business School, King's College London, London, United Kingdom.
OBJECTIVES: This study sought to assess the impact of an AI tool for automated incidental finding of vertebral fractures from CT scans, exploring the impact of the AI tool on staff, clinical workflows and patient experience, and identifying any challenges and limitations in adoption of the AI technology and the factors that contributed to or resulted from these.
METHODS: Semi-structured interviews were carried out via Microsoft Teams with both patients and healthcare professionals in different roles. These were transcribed in full, and explored themes of perceptions of AI in clinical practice, workflow integration and operational challenges, and patient trust and engagement with AI-assisted diagnoses. Thematic analysis followed an inductive coding approach using NVivo software.
RESULTS: Five key themes were identified: workforce and workflow impact, practical implications of adopting AI, perceived impact on patient care, issues of AI adoption and integration, and future potential of AI. Interview data, supported by pathway and process mapping, revealed that while the AI tool improved opportunistic vertebral fracture detection and identification of at-risk patients, it also introduced new complexities, such as increased service pressures, administrative burdens, and technical limitations. Notably, instead of reducing workload, introducing AI shifted and expanded responsibilities for radiologists, nurses and managers. Cost of AI and increased staffing costs are borne in radiology and fracture liaison clinics, and any benefits through health resource savings are recouped in other settings (i.e. social care, inpatient hospital care), raising important questions of how AI tools like this should be costed, charged and resourced.
CONCLUSIONS: The integration of AI tools like this into healthcare presents significant opportunities but also new challenges. While AI has improved fracture detection and patient identification, it has also created new service pressures, required additional workforce adjustments, and highlighted the need for support navigating information governance and integration challenges specific to AI technologies.
METHODS: Semi-structured interviews were carried out via Microsoft Teams with both patients and healthcare professionals in different roles. These were transcribed in full, and explored themes of perceptions of AI in clinical practice, workflow integration and operational challenges, and patient trust and engagement with AI-assisted diagnoses. Thematic analysis followed an inductive coding approach using NVivo software.
RESULTS: Five key themes were identified: workforce and workflow impact, practical implications of adopting AI, perceived impact on patient care, issues of AI adoption and integration, and future potential of AI. Interview data, supported by pathway and process mapping, revealed that while the AI tool improved opportunistic vertebral fracture detection and identification of at-risk patients, it also introduced new complexities, such as increased service pressures, administrative burdens, and technical limitations. Notably, instead of reducing workload, introducing AI shifted and expanded responsibilities for radiologists, nurses and managers. Cost of AI and increased staffing costs are borne in radiology and fracture liaison clinics, and any benefits through health resource savings are recouped in other settings (i.e. social care, inpatient hospital care), raising important questions of how AI tools like this should be costed, charged and resourced.
CONCLUSIONS: The integration of AI tools like this into healthcare presents significant opportunities but also new challenges. While AI has improved fracture detection and patient identification, it has also created new service pressures, required additional workforce adjustments, and highlighted the need for support navigating information governance and integration challenges specific to AI technologies.
Conference/Value in Health Info
2025-11, ISPOR Europe 2025, Glasgow, Scotland
Value in Health, Volume 28, Issue S2
Code
OP19
Topic
Health Technology Assessment, Organizational Practices, Real World Data & Information Systems
Topic Subcategory
Academic & Educational
Disease
Musculoskeletal Disorders (Arthritis, Bone Disorders, Osteoporosis, Other Musculoskeletal), No Additional Disease & Conditions/Specialized Treatment Areas