Artificial Intelligence (aka AI) is on everyone’s lips right now, but what is it and how might it affect the world of healthcare? This is a short introduction to the subject for those working in healthcare leadership.
AI is a pretty broad term, but where it gets interesting is something called machine learning. This means that computers are using their ‘intelligence’ to process huge amounts of data to teach themselves to get better and better at a given task. This is in contrast to computer programmes that simply follow the same rule over and over, without learning anything.
Even better results can be achieved by so-called neural networks, a more sophisticated version of machine learning that emulates our logical processes. It’s the intelligence that’s the most human-like.
A match for humans?
How good are these new machine brains? Well. clever enough to match humans at diagnosing certain skin cancers and doing diabetic retinography, for example. That’s because currently, AI is particularly good in the area of visual and speech recognition.
AI needs lots of data to work properly, essentially to train itself to a point where it can spot an anomaly, do a comparison, or react to a human interaction, all in a constantly-changing environment.
This need for data to learn from means that medical records become invaluable, so issues of patient confidentiality and data integrity need to be resolved until its full potential can be experienced more widely.
Currently, AI is being applied most commonly to medical imaging and diagnostics, as well as patient data insights and risk analysis.
Opening up faster and more convenient routes to medical care is another AI application. This is typically via an app-based system that can do online consultations and then direct the patient towards the best source of further care, based on the interaction between the app’s chatbot and patient.
Similarly, heath-orientated apps using AI tools on handheld devices like phones and watches can be a way of monitoring personal ‘live’ health data. It is easy to imagine a future where signals from those personal monitors can be an alert to a health crisis.
As Sebastian Thrun, Professor of Computer Science at Stanford University put it in the New Yorker: “Our cell phones would analyze shifting speech patterns to diagnose Alzheimer’s. A steering wheel would pick up incipient Parkinson’s through small hesitations and tremors.”
For those needing ongoing medical care, AI might power a virtual nurse to support people by monitoring their condition and treatment. Are they taking their medicines, for example?
On the administration front, computers using machine learning are capable of combing huge and complex healthcare records and treatment plans to highlight regular mistakes that would normally remain hidden. Or to suggest a more intelligent way to manage patient flow within a healthcare system.
Work in progress
The potential for improving healthcare via AI is clear and getting clearer every day. But there is still a lot of work to be done on how to turn a vision of the future into an everyday reality. Issues like overcoming regulatory obstacles, answering ethical concerns, reassuring the public, and integrating this new technology into existing systems, all will have to be navigated before AI is a success.
Balancing risks and regulation
“Leaders need to make decisions in areas that may be outside their immediate comfort zone and scope of knowledge, balancing the risks and regulatory implications of the use of AI with the incredible potential rewards for improving patient care. Walking this line is very demanding for healthcare leaders, but the value of AI for transformational change is potentially enormous”.
This is the first of a series about the issues of interest to those working in Healthcare leadership. If you have suggestions on topics you might want us to cover, get in touch with your suggestions.
Disruption is a reality for all multinational companies, and Japan is no exception. Hidden cultur...
As Brexit begins to impact one of the UK’s most internationally-connected sectors, Alex Acland an...