Singapore radiographers' perceptions and expectations of artificial intelligence: a qualitative study

• Autor(a) principal: Ng, Chloe Theresia
• Data de Publicação: 2022
• Outros Autores: Roslan, Sri Nur Aidah, Chng, Yi Hong, Choong, Denise Ai Wen, Chong, Ai Jia Letty, Tay, Yi Xiang, Lança, Luís, Chua, Eric Chern-Pin
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.21/14973
• topic: Artificial intelligence; Radiography; Radiographers; Focus group discussion; Singapore

Introduction: With the emergence of artificial intelligence (AI) in medical imaging, radiographers are likely to be at the forefront of this technological advancement. Studies have therefore been conducted recently to understand radiographers’ opinions on AI adoption. This study extends that work by using a qualitative approach to further explore radiographers’ knowledge, perceptions, and expectations of AI. Method: Six online focus groups were conducted with 22 radiographers from the three public healthcare clusters in Singapore. They were purposively sampled, and participants were recruited from a broad demographic background with varying years of working experience and designations. The focus group sessions were transcribed verbatim and thematic analysis was performed on their responses. Results: Participants demonstrated limited knowledge of AI. Their perceptions of AI were mixed, recognising its benefits in increasing efficiency and improving patient care, but also aware of its limitations in accuracy and bias. On how patients may perceive AI, participants felt that patients would accept AI if they felt it improves their care but may reject it once they lose trust in it. Expectations-wise, participants envisioned several applications in pre-, peri‑, and post-procedural workflows including order vetting, patient positioning, language translation, and artifact removal. On radiographers’ role and career opportunities, some participants see an opportunity for radiographers to specialise in AI, becoming involved in algorithm development and its clinical implementation. Discussion: Our findings suggest that widespread implementation of AI would require limited knowledge amongst radiographers and current AI limitations to be addressed. While radiographers are positively anticipating the integration of AI into their practices, they should also become actively involved in the development of AI tools such that those they envisioned. This would help align the optimal use of AI tools and radiographer role changes. Patients’ acceptance and reactions to AI also warrant further research.