Patient and practitioner perceptions around use of artificial intelligence within the English NHS diabetic eye screening programme
Charlotte Wahlich, Lakshmi Chandrasekaran, Umar Chaudhry, Kathryn Willis, Ryan Chambers, Louis Bolter, John Anderson, Royce Shakespeare, Abraham Olvera‐Barrios, Jiri Fajtl, Roshan A. Welikala, Sarah Barman, Catherine Egan, Adnan Tufail, Christopher G. Owen, Alicja R. Rudnicka
Abstract
AIMS: Automated retinal image analysis using Artificial Intelligence (AI) can detect diabetic retinopathy as accurately as human graders, but it is not yet licensed in the NHS Diabetic Eye Screening Programme (DESP) in England. This study aims to assess perceptions of People Living with Diabetes (PLD) and Healthcare Practitioners (HCP) towards AI's introduction in DESP. METHODS: Two online surveys were co-developed with PLD and HCP from a diverse DESP in North East London. Surveys were validated through interviews across three centres and distributed via DESP centres, charities, and the British Association of Retinal Screeners. A coding framework was used to analyse free-text responses. RESULTS: 387 (24%) PLD and 98 (37%) HCP provided comments. Themes included trust, workforce impact, the patient-practitioner relationship, AI implementation challenges, and inequalities. Both groups agreed AI in DESP was inevitable, would improve efficiency, and save costs. Concerns included job losses, data security, and AI decision safety. A common misconception was that AI would directly affect patient interactions, though it only processes retinal images. CONCLUSIONS: Limited understanding of AI was a barrier to acceptance. Educating diverse PLD groups and HCP about AI's accuracy and reliability is crucial to building trust and facilitating its integration into screening practices.