Patient Perspectives on AI-Powered Medical Robots in Breast and Prostate Cancer Care: Qualitative Study

TL;DR

This study explores how patients with breast and prostate cancer feel about using AI-powered robots in their care, finding that they are cautiously open if the technology is safe, reliable, and complements human care.

Contribution

The study provides novel insights into patient perspectives on AI in cancer care, emphasizing the importance of transparency, education, and maintaining human involvement.

Findings

Patients are open to AI-assisted care if it improves accuracy, efficiency, and reduces wait times.

Key barriers include concerns about reliability, loss of human interaction, and technical failures.

Facilitators include timely results, education, and ensuring AI complements human expertise.

Abstract

Nearly 250,000 cancer cases are diagnosed annually in Canada, with breast and prostate cancer representing 25% and 22% of new cases, respectively. Artificial intelligence (AI) applications can potentially improve the accuracy, efficiency, and timeliness of cancer care, including screening, diagnostic imaging, and early treatment. However, patient acceptability of AI in cancer care remains underexplored.

This study aimed to understand the feelings and perceptions of and acceptability to patients with breast and prostate cancer regarding the inclusion of AI-powered medical robots for cancer screening, diagnosis, and early treatment and to identify barriers and facilitators to implementation.

In this qualitative study, semistructured interviews were conducted with 15 patients with breast (n=6) or prostate (n=9) cancer. Participants (mean [SD] age 67 [12] y; range 41‐88 y) were recruited from Sunnybrook Health Sciences Centre between May and November 2022. Each participant completed 2 semistructured interviews, each approximately 1 hour in length, conducted by telephone or Zoom by a research assistant. The first interview explored initial reactions and expectations regarding AI-assisted medical technologies, allowing us to tailor questions for the second interview to better understand practical means of introducing AI into care, while further exploring salient concepts. Data were analyzed using a conventional content analysis approach. Two research assistants independently and inductively coded transcripts, iteratively generating and refining a codebook. Data saturation was assessed after 10 interviews and confirmed through analysis of all 15 participants.

Three overarching categories were developed: (1) individual beliefs, understanding, and attitudes; (2) integration of AI into care; and (3) health structure, systems, and processes. Participants reported openness toward AI-assisted medical robots but emphasized the importance of reduced wait times, demonstrated safety and reliability, and patient-centered care. Patients indicated that with appropriate education and transparent communication, they would be willing to accept AI-assisted care due to its enhanced accuracy and efficiency. Key barriers included concerns about reliability, potential loss of human interaction, and inadequate mitigation strategies for technical failures. Facilitators included timely results, improved efficiency, accessible educational resources, and assurance that AI would complement rather than replace human expertise.

Participants demonstrated cautious willingness to accept AI-powered medical robots in cancer care if positioned as complementary to, rather than substitutes for, human-provided care. These findings underscore the need for maintaining the presence of health care providers during AI-assisted procedures, providing clear and accessible education, and ensuring transparent communication about safety and reliability. Personalizing patient education and offering multiple modes of information delivery may foster confidence and improve acceptability. While findings are exploratory and reflect the perspectives of a small, predominantly urban sample, they provide actionable insights into patient concerns and priorities that may inform future research and guide early implementation strategies in integrating AI into cancer care pathways.

Figures