This is a longer version of an article published in Contrast, the quarterly magazine published by NHS EMRAD on 24 September 2020 available at https://emrad.nhs.uk/news-alias/670-autumn-2020 “Does this mean a "robot" reads them [mammograms]? Not sure I like that idea much.” When TaoHealth researchers asked women in the East Midlands earlier this year how they felt about using artificial intelligence in breast screening, this was, not surprisingly, a common response. Very few of us really know what artificial intelligence is and the part it can play in healthcare if it proves to be effective and safe. As one of the participants in a recent focus group said; “A lot of my understanding is from movies where robots take over the world.” And yet it is increasingly embedded in our everyday lives in the search engines like Google, virtual assistants like Siri, entertainment platforms like Netflix and email platforms like Microsoft Outlook that we use constantly. So, what is artificial intelligence? The UK government use the definition “Technologies with the ability to perform tasks that would otherwise require human intelligence, such as visual perception, speech recognition, and language translation… and learn and adapt to new experiences”. In plain English, you could think of artificial intelligence (AI) as computers and robots understanding patterns, pictures, speech and language. They can learn from their understanding and make decisions. The BBC has a short video that is a helpful start https://www.bbc.co.uk/news/av/technology-34224406/what-is-artificial-intelligence NHS EMRAD is working with UK-based AI companies to test how their technology can be used safely in the NHS breast screening programme. As part of the evaluation of this project we are talking to women of all ages to understand what they think about using AI. What worries many of us about AI is the idea that machines will start making decisions without human input, the machines will ‘take over’ and potentially put many of us out of a job. There are two ways in which we, as a society, are working to ensure that does not happen, specifically in breast screening but this applies in many situations where AI is being used. A lot of work is being done at the moment by governments, academics and regulators to establish how AI will be developed, tested and regulated in the real world in ways that make sure it is trust-worthy, understandable, un-biased, protects our privacy, doesn’t exclude anyone and allows us to have control over who and what we share our data with. This project is working closely with the UK government and regulators to shape the approach that will be taken to regulating AI in healthcare to make sure it is ethical as well as safe and effective. We know that health professionals are in short supply in the UK and across the world and radiology as a specialty is possibly more affected that some other professions. Introducing AI to support professionals with the human retaining ownership of the decision is broadly acceptable to the women we talked to. This would free up professionals to do other work, like seeing women referred by the GPs with breast lumps or other symptoms they worried about, and speed up the time women have to wait for these appointments, where the wait can be agonising. As one woman I talked to put it: “If it can free up more human time to then spend more time on the complex cases or actually just to be able to process things more quickly, I think that's got to be a positive. To get the confidence of the woman, [there] has there been a human involvement, I think it’s a partner in medicine not a replacement.” The full evaluation report for this project will be published early in 2021. TaoHealth would like to thank all of the women who took part in the survey and focus groups during 2019 and 2020. Their contributions have been hugely valuable in informing the project and future application of AI in breast screening already.
|