Jens Koegler, healthcare industry director, EMEA, VMware
We can have data without information but cannot have information without data. Nowhere is this more aptly demonstrated than in healthcare, which has experienced a tidal wave of data at every stage from patent records to complex surgery. It is a development that has simultaneously highlighted that humans need help with the demands of healthcare today and that artificial intelligence (AI) is ideally placed to answer the call. It’s a topic we discussed recently on a webinar with Craig Rhodes, EMEA industry lead AI for healthcare and life sciences from our partner, NVIDIA.
We ‘AIn’t seen nothing yet
The use of AI in healthcare was growing before the pandemic but, like many technologies, has been rapidly accelerated as a result of the last couple of years. But the way the world is heading means we ‘ain’t seen nothing yet. Globally there are more patients, who are living longer and have more complex conditions as a result. According to McKinsey, by 2050, one in four people in Europe and North America will be over the age of 65. At the same time the industry is facing a staffing crisis, huge pressures – both time and money – on drug development and an ever increasing volume of scans, images and modeling data that requires phenomenal computing power. This is set in a backdrop of complex procurement, hospitals that already contain hundreds of applications and a challenging regulatory environment where technology moves far more quickly than the people trying to manage it.
The sector is facing a perfect storm of challenges, which it is looking for AI to solve. Little surprise that the AI in the healthcare market is projected to reach $194.4 billion by 2030. But the users in the hospitals – doctors, nurses, caregivers and support staff – have enough challenges already without the integration and application of AI being one of them. So, where is it making a difference and what can other healthcare providers learn from those ‘doing AI right’?
How AI is being delivered in healthcare
In the webinar, Craig talks about the growth in graphics processing units (GPUs) embedded into individual machines that are capable of running algorithms – putting edge computing power in devices. This means clinicians can make detailed determinations around images or scans. For example, The Netherlands Cancer Institute (NKI), one of the world’s top-rated cancer research and treatment centers, is using the NVIDIA AI Enterprise software suite to test AI workloads on higher-precision 3D cancer scans. NKI’s AI model was previously trained on lower-resolution images. But with a higher memory capacity, researchers can now use high-resolution images for training, helping clinicians better target the size and location of a tumor every time a patient receives treatment.
This is a great example of using AI to better make use of the data that already exists but this can be applied to all areas of healthcare. In genomics, for example, analysis of a whole genome has gone from weeks to realistic ambitions to be able to do it in under 30 minutes very soon. It means that datasets can be compiled in close to real-time to understand specifically what is happening with a patient and how to treat them. Another excellent example is from the University Hospital ESSEN, which sees more than 300,000 patients a year benefit from its pioneering approach to giving all medical staff, access to all applications and data at their fingertips at any time and anywhere to provide the best and most appropriate care for every patient. It has a methodology founded in AI.
Scalable, secure and resilient
One of the reasons AI is becoming so pervasive in healthcare is the network of partners all aligned to collaboratively embrace problem-solving to the mutual end of improving the care and experience for patients. Whether it’s in digital biology, point of care, or in drug discovery or at GSK to emerging start-ups the speed and scale of adoption is rapid and is making a difference on an almost daily basis. Craig talks of an AI centre at King’s College London that includes 10 NHS Trusts working on neurology, cardiology, and oncology. But this is not a bunch of technologists that are getting in a room creating an algorithm. It’s being driven by the medical team to look at informing the clinical pathway – disrupting it in the best possible way, ensuring there is no encroachment on the valuable time of the clinical team, and improving patient flow. Getting data to clinicians at the right point in time and giving them the ability to look at the data in the best possible way is where AI is making a major difference.
One of the biggest challenges with using data in healthcare is governance and security and this is something federated learning is addressing. This is where algorithms are built in each location looking at, for example, breast cancer, and centrally stored and consolidated to create a supermodel that then is delivered back into the hospital. It allows for much greater variance in subject, sample size and therefore accuracy. This is only possible because of AI but the healthcare industry needs to avoid working in black box environments – we don’t want to create these amazing AI tools if nobody understands how they’re delivering the results. We need it to be open but as part of that, we need to ensure it complies with the governance of that data.
Maximise the immense value that AI can bring
Indeed, the only way the sector can move forward and maximise the immense value that AI can bring is to ensure the tools and solutions that are being used are able to run anywhere and be built on infrastructure to make it scalable, secure and resilient. The potential is almost limitless but when looking at large-scale healthcare problems and challenges, you need the right computation. This means the right infrastructure, a platform that enables organisations to start small and scale and the right partners and people alongside you.
Our mission is to use the knowledge of tomorrow for the data we have today to better understand it and deliver improved results for patients. This is impossible for us to do alone. But we’ve seen the art of the possible now and we’re only going to see more and more hospitals turning to AI to help humans in healthcare.
If you would like more information, please get in touch with us at email@example.com.