Opinion: How do we do no harm with AI in health care?

Artificial intelligence (AI) is poised to play a significant role in the delivery of health care for all Australians, but if we are to realise its full potential, safely, we need an urgent plan.

AI has the possibility to not only enhance many of our existing services, such as the interpretation of imaging and pathology, but also to open the door to new opportunities. Research is already underway for instance using AI to improve treatment of rare brain tumours¹. AI can also be used to support and educate consumers in self-care and prevention and to optimise hospital processes.

It may come as a surprise to many then that Australia lags behind other nations in investing in the development of AI for health care. The US leads with industry investment in AI for health care estimated to exceed US$6 billion in 2021, driving potential savings of US$150 billion by 2026. In the UK, the Artificial Intelligence Laboratory (AI Lab), established by the NHS, is part of a GBP 1 billion investment in AI. Former UK health minister and renowned surgeon Professor Lord Ara Darzi’s review² of the NHS identified productivity improvement from smart automation of £12.5 billion a year, 9.9% of the NHS England budget. Specifically, the review found that AI enhanced breast cancer³ image interpretation could reduce the time UK radiologists spend reviewing images by 20% (890,000 hours annually).

While Australia is ideally positioned to become a global powerhouse in the real-world application of AI, we have only now achieved a clear plan for how this will be done — safely, ethically and sustainably. Through a consortium of universities, healthcare bodies and industry, the AI in Healthcare Roadmap⁴ has been developed.

The AI in Healthcare Roadmap shows the path that Australia should take to embrace the opportunities that AI brings. The roadmap identifies the current gaps in Australia’s capability to translate AI into effective and safe clinical services and provides guidance on key issues such as workforce, industry capability, implementation, regulation and cybersecurity. It also acknowledges the extensive work already undertaken nationally and internationally and builds on this work.

The Australian Alliance for Artificial Intelligence in Healthcare (AAAiH) developed the roadmap through extensive community consultation and a national survey.

The roadmap provides 24 recommendations across eight priority areas. The highest community priority identified was for healthcare AI to be safe for patients and developed and used ethically. AI privacy and data security was a major concern — emphasised further by recent findings of serious problems with privacy and inconsistent privacy practices in mHealth apps⁵. Respondents also identified the need for genuine whole-of-nation leadership in the healthcare AI space as well as robust governance of the sector.

Gaps in our workforce capability to build and use healthcare AI were clearly identified as was the need for consumers to be fully engaged in shaping the healthcare AI agenda. Respondents also rated the gaps in our capability to adopt AI into practice and the need to enhance local industry capability as issues needing clear attention.

AAAiH has been working intensively since it was founded in late 2018 to influence and guide the implementation of AI in health care in Australia. It brings together almost 100 national and international partners and stakeholders in academia, government, consumer, clinical, industry organisations and peak bodies to translate artificial intelligence (AI) technologies into real-world health services.

We are aiming to create an AI-enabled healthcare system capable of delivering personalised healthcare safely, ethically and sustainably supported by a vibrant AI industry sector that creates jobs and exports to the world, alongside an AI-aware workforce and AI-savvy consumers.

To make this a reality, Australia needs to commit to funded research and development. The recently announced federal government grants to establish AI and Digital Capability Centres are the best opportunity yet for achieving this, added to CSIRO Data61’s statement that health care is one of the three highest priority opportunities for the nation.

While the development of new AI technologies is an exciting and necessary prospect, how these technologies exist safely and ethically within the healthcare system is equally important. We need funded research for the provision of expert guidance, as well as a healthcare and IT workforce that is sufficiently skilled with long-term as well as accelerated training programs. Ultimately, to maximise the benefits and opportunities of an AI-enabled healthcare system, nothing less than a whole-of-government and whole-of-nation approach will do.

In a world of smartphones and self-driving cars, it would be easy to think that AI is already upon us. There is, however, still much to do to ensure that all Australians benefit from the use of AI in health care, and we do no harm.

*Professor Enrico Coiera is Director, Centre for Health Informatics, Australian Institute of Health Innovation; Director, NHMRC Centre of Research Excellence in Digital Health; Founder, Australian Alliance for Artificial Intelligence in Healthcare; Professor in Medical Informatics, Macquarie University. Coiera trained in medicine and holds a computer science PhD in Artificial Intelligence (AI). He has a research background in both industry and academia and a strong international research reputation for his work on decision support and communication processes in biomedicine. Coiera was elected Foundation Fellow and first President of the Australian College of Health Informatics, is a foundation member of the International Academy of Health Sciences Informatics and an International Fellow of the American College of Medical Informatics. He is an Australian representative on the OECD International Global Partnership on Artificial Intelligence, AI and Pandemic Response Working Group. Coiera and the Centre for Health Informatics will celebrate 21 years of cutting-edge research on 11 July 2022 with a live event at Macquarie University.

Top image credit: Professor Enrico Coiera. Image: Supplied.

^References

^{1 Liu S, Shah Z, Sav A, Russo C, Berkovsky S, Qian Y et al. Isocitrate dehydrogenase (IDH) status prediction in histopathology images of gliomas using deep learning. Scientific Reports. 2020 May 7;10:1-11. 7733. https://doi.org/10.1038/s41598-020-64588-y}

^{2 Darzi A. Better health and care for all: A 10-point plan for the 2020s. The Lord Darzi review of health and care. Final report. 2018 June 15 London: Institute for Policy Research 2018. https://www.ippr.org/research/ publications/better-health-and-care-for-all}

^{3 Topol, E. The Topol Review: An independent report on behalf of the Secretary of State for Health and Social Care February 2019 Health Education England, 2019. https://topol.hee.nhs.uk/}

^{4 Coiera E, Waldie J. A Roadmap for Artificial Intelligence in Healthcare for Australia, 1 December 21, Sydney: Australian Alliance for Artificial Intelligence in Healthcare, ISBN: 978-1-74138-486-4. https://aihealthalliance.org/2021/12/01/a-roadmap-for-ai-in-healthcare-for-australia/}

^{5 Tangari G, Ikram M, Ijaz K, Kaafar MA, Berkovsky S. Mobile health and privacy: cross sectional study. The BMJ. 2021;373:1-12. n1248. https://doi.org/10.1136/bmj.n1248}