Can Artificial Intelligence deliver the support that our medical professionals urgently need?
The long road to clinical excellence
When we find ourselves in need of medical aid and put our health and our future in the hands of others, it’s reassuring to reflect on just how well prepared they are for the task. Doctors in Britain may train for up to sixteen years before qualifying. That includes five years for their degree, or six years if that degree allows them to take time out to work on an additional relevant subject, followed by a two-year postgraduate foundation course and then between three years and eight years of specialist training. The skills acquired over that period are priceless, and the intuition and good judgement we find in experienced clinicians can’t ever be undervalued. But how close are we to a point where it can be ably supported through artificial intelligence? In some key areas, we may already have reached that point. A study in medical journal Nature has suggested that in the diagnosis of breast cancer from the evidence of a mammogram, an AI programme may be more accurate than a human practitioner. Some results even showed an AI diagnosis as more accurate than two doctors working together. So who needs doctors anyway?
Artificial intelligence, real healthcare benefits
A team including medical professionals and researchers from Google Health designed and trained a computer model to analyse X-ray images taken from almost 29,000 women. When it came to analysing mammograms, the program outperformed experienced radiologists, even when the doctors had the advantage of a patient history to work with. So are we working towards a doctor-free future? Far from it. This AI program may have performed well in one narrowly defined task, but even there it required human design and improvement to function. What it does potentially offer is a back-up service which, managed correctly, can give our doctors the support they urgently need. X-ray analysis is a crucial element of the diagnostic process, and it takes time that practitioners may not have. In Britain alone there’s an estimated shortage of more than 1,000 radiologists. It’s been suggested that AI could help us move away from the current practice of dual reading of mammograms by two doctors. If we can work towards a process that delivers an AI reading followed by a human review without adversely affecting patient care, the time saved could be game-changing.
Technology and human ingenuity; the perfect marriage
Conversis Project Manager Jack Garner sees clear advantages:
We often hear medical professionals talk about the relentless nature of a disease. Viruses and cell mutations are constantly looking for ways to outwit our defences, because they have nothing else to do. If we can harness that same relentless quality in artificial intelligence it will surely deliver major benefits.
Jack also sees interesting parallels with the way the translation industry has incorporated AI:
As language service providers we use cutting edge software to develop translation memory that can save time and money in repeat projects. We also offer a machine translation service that can produce an effective first draft of text in a new language. In both cases, though, the process is managed by suitably skilled people. The final judgement on a project is always made by a qualified professional. We marry technology with human ingenuity every day, notably in support of the Life Sciences sector, and I find it fascinating to see clinical professionals walking a similar path.
Who needs doctors? We all do. The real question is, what support do doctors need, and can AI help give it to them?
Conversis: helping you speak to the world
To find out more, contact us using the form below.
Please use the form to get in touch
or alternatively give us a call on: +44 (0)1869 255820