Artificial Intelligence (AI) permeates every single industry vertical. In Health, there are abundant opportunities to introduce AI to both clinical and administrative workflows. Health faces significant macro-economic challenges with budgets cut and staff shortages persistent; it’s clear that technology could be a powerful enabler for health providers to continue to deliver the safety and quality standards of care we all expect.
Fujitsu research with the Australasian Institute of Digital Health (AIDH) found that 82% of survey respondents believe that AI will be an important part of healthcare and that it can be used for good. The benefits of AI have the potential to be realised across multiple aspects of workforce satisfaction and patient care.
Administrative tasks can be automated, freeing up highly skilled people to focus on higher-value activities such as directly caring for patients, training junior clinicians and preventing medical errors. AI can assist with allocative efficiency ensuring that the most effective use of resources is made in a complex environment that has historically experienced high levels of waste and poor productivity.
A data-driven, AI approach that relies on historical performance data rather than human estimation and approximation can improve the probability that resources will be scheduled when they are most likely to be needed. There is a correlation between improved productivity and improved governance, so it is plausible to plan for improved achievement of environmental, social and governance (ESG) objectives for healthcare providers.
One of the most complex and urgent activities in the hospital Emergency Department is triage. AI is well suited to support clinicians with predictive, informed triage to increase the likelihood that patients with most pressing or life threatening needs are treated first. For example, AI sepsis prediction models are currently being successfully trialled across the Australian public health system.
Patient care is undergoing a technology revolution. Based on the Fujitsu and Australasian Institute of Digital Health (AIDH) research, one in every five clinicians are already using AI to support diagnosis. These clinicians using AI to augment and assist their practice should set and achieve higher standards of patient care.
Health service management is still in the process of implementing Responsible AI – carefully and judiciously selecting the right policies, ethics, privacy and expertise to meet their enterprise needs. Many of them are lagging behind the workforce especially where the AI is used in clinical decision support.
One way to focus the workforce on Responsible AI adoption is wearable technology for use by patients either in the hospital or at home where it might monitor chronic conditions and predict an acute exacerbation thereby avoiding an ambulance and an admission. This use case has the potential to bring a stream of diagnostic data to clinicians with patient consent, participation and ownership. In the aged care setting, wearables and remote monitoring devices may assist with fall detection and ensure wandering patients can be found quickly and safely.
Before the Covid-19 pandemic, the development of vaccines and medicines took many years. AI, along with other technologies, enabled new vaccines to be developed faster than ever before saving millions of lives and greatly minimising the number of people that required hospitalisation or acute care. We can expect AI to continue to be used in the discovery, design and development of new medicines and vaccines that improve public health. Fujitsu’s Quantum Computing as a Service technology (Digital Annealer) combined with AI is already being used in drug discovery by research and development units, Universities and Clinical Research Organisations globally.
Beyond the benefits we see today, AI has the potential to positively impact healthcare in many ways. We can expect greater personalisation of healthcare as drugs can be customised to meet our specific needs. Known as Precision Medicine, as we learn more about the genetic and environmental factors that influence health, we are able to develop medicines faster than ever before and tailor genomic therapies to individuals.
In order to achieve AI’s full potential, there are hurdles to be overcome. Fujitsu’s research with Australasian Institute of Digital Health (AIDH) found that while some organisations have AI policies, experts to advise on AI ethics and governance, and education plans to enhance AI skills in place, just 10% have all three of these critical elements.
State agencies have differing views on the use of AI during this rapidly changing adoption period. Queensland and New South Wales seem to be fostering the use of AI in clinical settings while it is currently discouraged in Victoria until firmer guidance is available. The Federal Government has allocated $57M to research how AI can be used to diagnose and manage a variety of illnesses and conditions including multiple sclerosis, cancer and cardiac issues.
Building trust in AI remains an important challenge to overcome. There is still scepticism about whether AI can be trusted and its impact on workforces. However, this can be managed by working alongside clinicians and healthcare professionals to understand where the pain points are in their work and what opportunities there are to use AI and other assistive technologies to support them.
The introduction of new technology cannot be successfully achieved without appropriate consultation and education. The journey starts with a plan for where AI can deliver benefits today. Many concepts about AI adoption will be tested and the majority are likely to yield no measurable benefit – only a few will succeed. Once a project has been identified, put governance plans in place to ensure AI is used ethically, safely and responsibly to achieve those benefits. A critical element of those plans is having a plan to develop staff skills and digital literacy so they can better understand and build trust in the technology.
AI can be used to deliver significant health benefits to the workforce and to the population. While there may be understandable reticence to use AI in clinical settings, it can be a powerful tool to streamline administration, aid in triage and enable clinicians to spend more time directly caring for patients. Successful deployment of AI depends on having a well defined scope, appropriate governance in place and improving the digital literacy of healthcare teams.
