Artificial Intelligence (AI) is increasing in scale, power and applicability. It is transforming businesses in every sector, including healthcare. Drug development, condition diagnosis, disease prediction and patient management are all being disrupted by AI. In this article – Healthcare Embraces Artificial Intelligence featured in HefmA pulse Magazine January/February edition, we highlight current AI developments and applications within the healthcare setting, their benefits and potential concerns that could arise around the use of AI in healthcare.

Artificial Intelligence (AI) is a computer science discipline that seeks to create intelligent machines that can replicate critical human mental faculties. Key applications include speech recognition, language translation, visual perception, learning, reasoning, inference, strategising, planning, decision-making, and intuition. There are several underlying disciplines within the field of AI which include big data, data mining, rules-based (expert) systems, neural networks, fuzzy logic, machine learning (ML), deep learning (DL), cognitive computing, natural language processing (NLP), robotics and recognition of images, speech and video.

AI in the broadest sense, i.e. the ability to replicate the entire functioning of a human brain, may be some way off. However, AI in a narrow sense is already embedded in many the systems, devices and digital services we use. For example, digital assistants such as Siri, Cortana and Amazon Echo all use AI to recognise speech patterns and run a growing range of processes to answer queries, search for information and learn or needs and behaviours. Within healthcare, the range of current and potential applications is growing – as outlined below.

Diagnosis and prediction

DeepMind, the UK-based Google owned AI research and applications company, has launched DeepMind Health; a subsidiary focused on applying AI to a vast range of healthcare problems. In a five-year deal with the UK’s National Health Service, the company has been given access to patient data to develop and deploy its Streams healthcare app. The app will provide doctors and nurses with cell phone alerts about a patient’s condition. Applications include spotting people at risk of kidney problems, detecting blood poisoning and coordinating patient treatment.

IBM Watson has developed IBM Watson health, devoted to improving healthcare with cognitive computing. The massive processing power of technologies such as these can revolutionise myriad aspects of healthcare; delivery, diagnosis, patient experience.

Many modern healthcare providers have huge amounts of data on each patient, most of which is unstructured. While it may take a person weeks to go through reams of patient data, an AI, such as IBM Watson Health, can ‘read’ 200 pages of information in 3 seconds. Using powerful processing technology, links can be made between disparate sets of data, old trauma, a previous illness, and health and lifestyle information can be interpreted and extrapolated to form an accurate diagnosis in seconds.

Equally, using such processing techniques, an AI can extrapolate existing healthcare problems and life style choices into the future. A healthcare AI may be able to warn you about potential problems such as type 2 diabetes, cancer, high blood pressure, and high cholesterol – years before they would normally become apparent.

Drug Development and Clinical Trials

With the cost of genetic testing falling exponentially, and the increase of stratified medicine tailored to our individual genetic make-up, drug development can be assisted by AI. Genetic data can be analysed, extrapolated and combined with the chemical composition of drugs to create individualised medicine. Equally, AI can benefit clinical trials: a mobile phone app called AiCure records people taking their medication, identifying the patient and drug, and deploying sophisticated features such as facial recognition to ensure it is not being tricked. The adherence data is available in real time to organisations conducting clinical trials, ensuring for the first time that they are based on genuine hard data.

Healthcare Management 

AI is already in place in ‘smart cities’ – locality infrastructures designed to inform management decisions on everything from street lighting, to traffic control and policing. Smart city networks capture massive amounts of information about the population and its patterns of behaviour using a range of sensors and data via diverse devices ranging from street cameras and building management systems through to mobile phones and wearable technology. This Internet of Things (IoT) generates huge volumes of data, which is transmitted and shared via cloud computing and interpreted using AI processing capabilities.

As part of the smart city concept, we will see the emergence of smart hospitals, with the medical equipment, beds and physical fabric of the hospital all providing data about patients, professionals, the building infrastructure, and logistics flows of people, goods and health information through the hospital. Which patients have been left unattended for an excessive period? What services are being used the most? How did the infrastructure of the emergency room cope? Where are resources laying idle? Which department is under, or even, over staffed? How can resources and personnel be deployed more effectively? Using AI, trends can be monitored in real time and predictions made as to where and when to deploy people and physical resources to maximum benefit. So, for example, if an AI ‘notices’ that the emergency room sees the most heart attacks on Friday afternoons, extra emergency room doctors and cardiologists can be deployed ahead of time.

Cost and Time Efficiency

Across the healthcare sector as a whole, AI could bring massive efficiency increases. Artificial Intelligence chatbots can be trained by and learn from specialists across a range of typical patient and professional interactions. Once in use, the chatbots would deliver advice with the same expertise as a general practitioner and only require human input when previously unseen situations arise. By using patient medical data drawn from multiple sources and combining it with personal data, lifestyle, interests and hobbies – an AI chatbot can answer questions with more precision and depth than a human ever could. This can free up a doctor’s time to build relationships with patients and focus on the broader aspects of their health and wellbeing.

The NHS has partnered with Babylon Health, a digital healthcare app, to provide its AI-powered chatbot to over 1.2 million patients in North Central London. Patients type in symptoms, the app asks questions to determine severity and advise whether to seek medical assistance, visit a pharmacy, or manage the issue at home. The process typically involves around 12 interactions and takes around 1.5 minutes – less than the time of the average call to the NHS 111 non-emergency helpline.

Using AI means diagnosis can be almost instantaneous and comprehensively researched and stratified medicine can ensure the drugs prescribed for a patient are of genuine benefit.

Ethics, Trust and Compliance

There is extraordinary and almost limitless potential for AI to change the nature of healthcare delivery, outcomes and patient wellbeing. Equally it raises new ethical questions and concerns for patients. For example, the predictive power of AI could warn you of heart disease problems that may become apparent many years down the line. For some, this could cause stress if they have not consented to such advanced predictions; it may not seem of any importance to their life at this moment. Special attention must be paid to the human side of healthcare in these cases; a process of informed consent, facilitated by a medical counsellor, may be necessary to explain the potential insights an AI may deliver.

Knowing When

For some an AI chatbot may present an unnatural and untrustworthy visage to a patient. For example, an older person, less well versed in digital technologies might find it difficult to use and be unforthcoming in giving it sensitive information about their health problems. This is not just a user service experience problem but also represents an ethical concern; if a patient comes to harm because the doctor’s surgery has not provided a service that patients feel comfortable using, then where does the responsibility lie? Equally, an AI chatbot can spend unlimited time listening to the patient’s concerns, will act consistently on every occasion, will never get tired, is available 24/7/365 and may be the perfect solution for out of hours’ enquiries, housebound individuals, patients in remote places or those with anxiety about talking to a human. What is critical for healthcare providers is knowing when and where to deploy these AI based healthcare solutions to maximise benefit, minimise harm and create a sustainable model going forward.