Transforming care delivery with AI

Artificial intelligence represents a game-changer in the field of diagnostics and will be a critical tool in the move to value-based care.

Sign Up

Get the New Statesman's Morning Call email.

In recent years artificial intelligence has become part of our everyday lives. Intelligent computer algorithms are utilised in internet search engines, smart phones, financial transactions and self-driving cars. In 2017, the machine learning-based computer programme “AlphaGo” defeated the world champion of “Go”, the strategic board game, once again demonstrating the potential of AI1.

Meanwhile, in the healthcare sector, and specifically in the field of diagnostics, a fundamental transformation is on the horizon. According to a recent poll by The Economist, more than 50 per cent of global healthcare leaders expect an expanding role of AI in monitoring
and diagnosis2.

In the United Kingdom, an ageing population, shortage of radiologists, strain on pathology, as well as economic austerity has underpinned a difficult workforce shortage. Demand for CT and MRI scans has grown by more than 30 per cent between 2013 and 2016, while growth in the number of radiologists over the same period staggered at eight per cent. Additionally, the number of pathologists reporting that they worked well beyond a 39-hour week is also increasing, which is directly linked to adverse events4.

With the above in mind, and these 12-hour plus workdays becoming more common3, what role could AI play in managing and standardising workload, while mitigating against the risk of reading fatigue and clinician burnout?

We at Siemens Healthineers believe that AI is poised to play an increasingly important role in medical imaging and pathology. We believe the role of the radiologist and pathologist will be augmented by these new tools, and we believe these tools will improve the patient experience while also helping to meet the rising demand for diagnostic imaging and faster turnaround times in the lab. Digitisation will make way for more efficient and data-driven decision making.

At the core of our approach to healthcare digitisation is a drive to not only cut costs but to impact real and positive change on both the working lives of clinicians as well as patient outcomes, and to always ensure patient safety and privacy is kept paramount. It is no secret that reading and reporting on medical images or pathology results can be labour intensive and repetitive at times. With that in mind, we set ourselves the goal of giving clinicians an accurate, reliable aid to help minimise repetitive tasks such as manual labelling and measurement of organs, to give them more time that could be spent reporting on more severe cases.

Enter the AI RAD companion; after being “trained” on more than three million images, the AI RAD companion is a cloud-based software that will automatically analyse CT Chest scans to report on metrics such as aortic diameter, heart calcification, emphysema staging, and lung nodule size and position. This is all done within minutes of a patient’s scan being completed. While the AI RAD companion is a classic example of AI in action within healthcare, the applications of machine learning within healthcare can sometimes be subtle and often reach beyond the limits of one hospital department or discipline.

An example of such an application is our AI Pathway companion, another AI-driven software that Siemens Healthineers is developing with a key focus on the management of patients through their journey along an often complicated, painful and life-altering clinical pathway.

The Pathway Companion currently focuses on some of the main killers of our time; namely cancers of the prostate, breast and lung, as well as heart disease. It connects to data sources and extracts insights using machine learning and natural language processing from pathology and radiology results, patient records and notes to locate and analyse the most relevant data and facilitate a comprehensive clinical picture of the patient .

Imagine you’re a patient with suspected prostate cancer, and in a single appointment your oncologist can see your latest blood results as well as a history of those results over the last few years. They can also see the results of your latest biopsy and radiological exams, and lastly, they see how all this data would fit into the clinical pathway, and what the recommended next step would be for you as an individual. Data from three or four different sources and/or departments in one place, with a recommendation for action to be considered by your doctor – or doctors in a multidisciplinary team setting.

Value-based care – leveraging AI to deliver data-driven precision medicine

The digitisation of diagnostic results and shortening the time required by clinicians to interpret what those results mean is a key output of artificial intelligence; but what it really delivers is value-based care, where the patient is at the centre of diagnostic decision making and success is determined by patient outcomes, not just operational efficiency.

While certain technical metrics such as specificity, sensitivity, and reliability are crucial to the acceptance of any AI solution into clinical use, equally important is how it fits within the clinical workflow and the value it adds to the clinical decision-making process. Far from replacing any clinicians, an AI system is only worth as much as the value it produces for clinicians.

Artificial intelligence is not the answer to every issue being faced in the NHS, nor will it be a perfect and infallible solution. However, the potential for AI to introduce highly practical and life-changing implications for both patients and clinicians is high enough that it must be given a real chance to prove its worth in our NHS.

1. Mozur, Paul (May 23, 2017) “Google’s AlphaGo Defeats Chinese Go Master in Win for A.I.”

2. The Economist (2017) The future of healthcare – Realities or science fiction?

3. The Royal College of Radiologists (2017): UK workforce census 2016 report.

4. https://thepathologist.com/outside-the-lab/all-in-a-days-work

5. https://www.siemens-healthineers.com/en-uk/press-room/press-features/lun...

6. https://www.siemens-healthineers.com/press-room/press-releases/pr-201811...

By Matt Gibson, head of sales and digital services – UK&I, and Hasan Jouni, business development – digital services UK&I, at Siemens Healthineers.