Circadian

View Original

Artificial intelligence and the future of healthcare

In recent years, artificial intelligence (AI) has taken the centre stage in the field of medical technology worldwide. Many industry leaders envisage AI as having the potential to revolutionise healthcare delivery amidst a backdrop of ever-important challenges in tomorrow’s healthcare systems. Through enabling system efficiency, clinical effectiveness and patient satisfaction, AI is vetted to be a powerful tool in appropriating the needs of ageing populations, an increasing burden of chronic disease and growing resource pressures that will be commonplace.

 Artificial intelligence is defined at its most basic as computer systems that are capable of learning, reasoning and decision-making. Such mimicry of human intelligence behaviours is achieved through analysis of vast amounts of data, or so-called ‘big data’; all requiring the application of logic, statistics and biological principles. Some academics quite excitedly refer to AI and its implementation at large as fueling our transition to a ‘fourth industrial revolution’.

 The positive impact of AI developments could extend across many dimensions of healthcare. Within the realms of clinical care, AI shows strong potential in aiding diagnosis of disease with this being viewed as a frontier for AI-driven healthcare progress. Its potential usage as a diagnostic or screening tool pertains mostly to the fields of medical imaging and pathology, where a multitude of patient medical scans and tissue samples already pre-exist to ‘train’ AI systems. If this potential is realised, AI could reduce time and cost spent for radiologists and pathologists to achieve the same result, yielding greater time for improved decision making and for more human analysis for complex cases. Additionally, several studies suggest that AI may be inherently better at detecting skin cancer, pneumonia and eye diseases than the trained professional. In medical research, AI is now able to comfortably analyse large varied datasets the data-rich field brings to advance medical knowledge and discovery; for example, combining genetic and medical patient data with data existing medical research studies to identify therapeutic targets for disease. On a service provision level, AI is already being used via ‘chat-bot’ style interfaces to offer generic healthcare advice at home and even personalised health assessments; this may in theory alleviate unnecessary GP appointments as well as potentially instilling a sense of independence to one’s own healthcare.

 A common question, especially directed towards AI’s diagnostic abilities is “will doctors ever be replaced by AI?” Industry leaders stress that this will be highly unlikely in the future with the primary goal of AI to augment and the human doctor’s intelligence rather than to replace it through freeing up time for personalised patient care, being akin to a human-AI symbiosis. Furthermore, it has been said that AI is limited, at least currently, in that it cannot replace the human attributes of a doctor such as empathy and compassion, and is not able interpret patient data in the social context.

 As with most technologies when applied to healthcare, a complex and dynamic system, AI is likely to pose challenges in its implementation. Due to the large scale of patient data usage, ethical issues of data privacy and ownership may be significant challenges to overcome before even examining the issues of the security of such data, where stores of sensitive information are prone to being breached maliciously. 

 Analysts warn that AI in healthcare is still in the ‘hype phase’ and is ‘still in its infancy’ within modern healthcare systems: few AI is used in routine medical practice today apart from the select institution; and therefore it may be difficult to predict its effects and challenges that may arise when introducing such ‘intelligent’ systems.

 As the NHS launches its first ever AI lab, the prospect of an AI-driven UK healthcare system is perhaps only further being realised.