In November 2019, Sydney played host to the first Australian AIMed conference. Dr. Anthony Chang, AIMed’s founder, hosted the vibrant event, which attracted 140 delegates to the two-day meeting of workshops, panels, a Shark Tank, and presentations. Held in partnership with the Australian Alliance for AI in Healthcare (AAAiH), the agenda was packed with learning opportunities through its workshops, panels and sessions. It featured user case studies and heard from clinicians who are using AI today to benefit patient safety and improve clinician efficiency.

A common theme for discussion was the critical need to focus less on technological innovation, which is happening apace in areas such as industry, and to focus more on clinical implementation or ‘last mile’ challenges for AI. Health service delivery problems require different approaches to the problems of design AI, such as demonstrating technical performance e.g. diagnostic accuracy. Solving implementation challenges in health services is strongly dependent on understanding and exploiting human to human interactions, and human to machine partnerships.

Australia is a nation with many unique attributes, which makes it an ideal place for the development of clinical and consumer AI systems, and conference delegates also spent much time exploring these natural advantages. The nation is home to one of the most genetically diverse populations on earth. Two centuries of immigration have seen the original indigenous population joined by people from the Pacific, South East Asia, China, the Indian subcontinent and Europe. This diversity makes the nation an ideal place to both develop and validate genetically based algorithms, in contrast to nations with more homogenous populations.

Australia is also home to one of the world’s highest performing health systems, delivering outstanding clinical outcomes for the population in a very effective manner. The widespread digitisation of hospital records, and nearly complete digitisation of primary care records for over two decades, means that the nation has a great supply of the ‘new oil’ – data. Interestingly, because of the great differences in the role of private insurance health system compared to countries like the US, clinical documentation burdens are significantly lower. It is thus likely that clinical data will be of a higher quality here than in other nations. With a reputation for being an early adopter of technology, all these benefits make Australia an ideal global testbed for clinical AI.

Conference delegates also spent much time exploring the challenges of transforming Australia’s health system into a smart and AI-enabled one. These challenges are shared by most other nations and include developing a workforce that is sufficiently skilled, not just in building and evaluating AI, but also in using it safely and effectively in routine care.

These challenges – workforce and safety, quality and ethics– are two of the four core streams of the AAAiH, which ran a lively and interactive panel on nationwide challenges. AAAiH is an internationally unique collaboration with over 90 member organisations, and over 20 individuals associated
with the AI Alliance presented in one way or another at the conference. The AAAiH brings together members from industry, academia, health services, consumers and government and aims to catalyse the creation of national policy and infrastructure needed to adopt

AI routinely in the health system. The conference closed with reflections from Dr. Chang on the next steps forward in the development of healthcare AI, both globally, and in Australia. With so much still to do, planning is already underway for AIMed Australia 2020.

KEY TAKEAWAYS

  • AI should focus on solving problems first and not be focussed on the technology (design AI).
  • Solving implementation challenges in health services is dependent on understanding and exploiting human to human interactions and human to machine partnerships.
  • Australia is a globally unique nation exceptionally well suited for early stage development of AI for healthcare.
  • With one of the most genetically diverse populations on earth, a high performing health system, and widespread digitisation of clinical records, the nation is well placed to support development, evaluation and implementation of clinical AI.
  • The Australian Alliance for
    AI in Healthcare (AAAiH) is a globally unique whole of nation collaboration aiming to fast track developments needed to see effective use of AI, including workforce training and expansion, and national frameworks for governance of safety,
    quality and ethics.