Acceptability of artificial intelligence in breast screening: focus groups with the screening-eligible population in England
Lauren Gatting, Syeda Ahmed, Priscilla Meccheri, Rumana Newlands, Angie A. Kehagia, Jo Waller
Abstract
Introduction Preliminary studies of artificial intelligence (AI) tools developed to support breast screening demonstrate the potential to reduce radiologist burden and improve cancer detection which could lead to improved breast cancer outcomes. This study explores the public acceptability of the use of AI in breast screening from the perspective of screening-eligible women in England. Methods 64 women in England, aged 50–70 years (eligible for breast screening) and 45–49 years (approaching eligibility), participated in 12 focus groups—8 online and 4 in person. Specific scenarios in which AI may be used in the mammogram reading process were presented. Data were analysed using a reflexive thematic analysis. Results Four themes described public perceptions of AI in breast screening found in this study: (1) Things going wrong and being missed summarises a predominant and pervasive concern about an AI tool being used in breast screening; (2) Speed of change and loss of control captures a positive association of AI with technological advances held by the women but also feelings of things being out of their control, and that they were being left behind and in the dark; (3) The importance of humans reports concern around the possibility that AI excludes humans and renders them redundant and (4) Desire for thorough research, staggered implementation and double-checking of scans included insistence that any AI be thoroughly trialled, tested and not solely relied on when initially implemented. Conclusions It will be essential that future decision-making and communication about AI implementation in breast screening (and, likely, in healthcare more widely) address concerns surrounding (1) the fallibility of AI, (2) lack of inclusion, control and transparency in relation to healthcare and technology decisions and (3) humans being left redundant and unneeded, while building on women’s hopes for the technology.