News

If you read our previous article ‘Government funds £20 million into breakthrough AI projects’, you would have seen that we briefly touched upon the AI projects the researchers hope to pursue and eventually introduce into the Healthcare industry. We recently came across an article written by Fortune 100 Data Scientist Eugenio Zuccarelli on ‘Towards Data Science’ talking about the three AI trends that are hoped to be introduced to ‘revolutionise healthcare’.

Not only has AI in Healthcare been invested in recently by the government but, was also a major investment back in 2016 more than any other sector by the global economy - being called ‘the hottest AI category for deals’. Eugenio states, ‘this explosive growth is due to a wide variety of causes, stating with an ever-increasing adoption of big-data solutions and the need for technological solutions to adapt healthcare to crises such as the COVID-19 pandemic’. ‘AI’s penetration in healthcare, a sector accounting $2,487.7 billion in 2019 in the U.S. alone, will have a significant impact in one of the key aspects for every country, the human wellbeing.’

Three trends have particularly stood out and defined themselves in the Healthcare space.

#1 Electric Health Records (EHR) In order for AI to be introduced into healthcare and to be beneficial, with the current systems and processes, there needs to be a foundation for it to work from, which is where the electric health records (EHRs) come in. The Chartered Society of Physiotherapy explains that ‘EHRs are any digital document or system that contains information on an individual’s health and care. This could be online, on an internal network, or on a device.’ The records can include a wide variety of data and information such as past medical history, diagnosis, immunizations, progress notes, surgical history and much more.

Eugenio goes on to explain further that, ‘there are currently many standards for record-keeping, but the most famous one, FHIR (Fast Healthcare Interoperability Resources), is becoming the leading protocol used by companies of the like of Google and Apple. FHIR uses a modern suite of APIs, HTTP-based RESTful protocols, HTML and CSS for UI integration and allows to use JSON, XML or RDF for data representation.

‘One of its goals is to facilitate legacy healthcare systems to communicate with each other to easily provide information to medical providers and individuals. This is allowed on a wide variety of devices from computers to tablets and cell phones, and, more importantly, allows third-party developers to provide medical applications which can be easily integrated into existing systems.’

Through collating such a large amount of data, companies will be able to apply Machine Learning processes in order to extract insights from the data to provide to the medical providers with all of the key information. ‘An example, strictly related to electronic health records, is shown by the use of Natural Language Processing algorithms to extract information from unstructured text. For instance, by running an NLP algorithm, we can extract clinical characteristics and diagnoses from the doctor’s notes and store them in a structured format. These can be a list of diagnoses related to a specific patient, or a series of procedures to undertake.’

#2 Diagnosis Prediction One of the main ways that AI will look to transform Healthcare is by using Machine Learning to analyse a patient’s data to cross-reference features and look to predict a disease before it presents itself. An example of how this can be used to detect illnesses was presented by IBM, on their ‘AI Models Predict Breast Cancer with Radiologist-level Accuracy’ report where they stated, ‘our team was able to create a unique and novel algorithm that is – to our knowledge – potentially the first to incorporate both mammograms and comprehensive electronic health record data for the prediction of breast cancer. Built on deep learning models, our team was able to train this system to achieve an accuracy comparable to radiologists, as defined by the American benchmark for screening digital mammography.’

Other companies such as Google DeepMind have partnered with the likes of Moorfields Eye Hospital to develop and introduce AI systems to detect over 50 different sight-threatening eye diseases – more information can be found on this here. Through introducing these breakthrough innovations, a whole new way of efficient working in Healthcare is hoped for, which in turn will increase productivity and decrease turn-a-round times, as well as expenditure.

#3 Telemedicine Due to the NHS being overwhelmed because of the recent pandemic, the economy has had to quickly find alternative processes to relieve the pressure, but to still be efficient enough to rely on in the absence of many Doctor’s full attention and time.
On an article published by Wired they report that the NHS are currently trialling an AI-bot on an app created by company Babylon, which is intended to reduce the pressures of the 111 non-emergency service. “Babylon’s AI technology can process billions of symptom combinations much faster and more accurately than the human brain.”

Through the AI chatbot speaking to individuals on the app and receiving a set of information, it makes it easier and quicker for the bot to determine what the potential issue or disease would be through suggested symptoms. ‘Staff from Babylon have said that they believe the NHS-approved app will help the organisation save money as the process of interacting with the bot takes around 12 messages and is quicker than speaking to a human on the phone.’

In conclusion, the different AI technologies mentioned above are and will need to be tested and accepted by healthcare professionals in order to solely rely on in the future. But as discussed, due to recent COVID-19 pandemic and its aftermath a lot more research, investment and testing is underway to make this a significant positive impact in all of the way’s healthcare can utilise these innovative technologies.

https://towardsdatascience.com/3-ai-trendsthat-will-revolutionise-healthcare-da4198dbb31d https://www.csp.org.uk/professional-clinical/digital-physiotherapy/electronic-health-records https://www.ibm.com/blogs/research/2019/06/ai-models-radiologist-level-accuracy/ https://deepmind.com/impact#real_world_impact https://www.wired.co.uk/article/babylon-nhs-chatbot-app