How are Algorithms Transforming Medical Diagnostics?
The use of algorithms is prevalent within modern life, from social media feeds being tailored to your preferences, to smart motorway traffic control. Within healthcare, advanced algorithms are used during pharmaceutical discovery, within medical device operating systems and in diagnostics. Many are referring to technology of this nature as ‘Artificial intelligence (AI) ’, however, it is vital to remember that true ‘AI’ does not yet exist, and its current form is simply the use of pattern matching and machine learning to exhibit intelligence.
Information Technology company IBM have developed a supercomputer known as ‘Watson’, which algorithms arrive at a diagnostic suggestion and assign a probable success rate to them. The MD Anderson Cancer Center was one of the first leading clinics to start using Watson. At the clinic, Watson aided the physicians by coming up with a list of options for treatment and assigning a confidence value to them, from very low to very high. Researchers at the center evaluated Watson’s success rate and found it to be very efficient. This demonstrates how powerful and successful algorithms can be in medical diagnostics, but there is surely still a long way to go before systems like Watson can be used to diagnose autonomously?
The UK Government has pledged to deploy ‘AI’ in an attempt to transform the NHS and prevent over 20,000 cancer related deaths a year by 2033. The technology will attempt to spot trends at the early stages of cancer by using medical records along with information about patients’ genetics and habits, which will then be cross referenced with national data. 2033 may seem a long way off, but with growing strain on healthcare services, change is imperative going forward.
Individuals will likely not want to put a diagnosis of a potentially life-threatening illness in the hands of an algorithmic system, ask yourself the question: would you trust a machine to make a diagnosis for yourself or a loved one? That is why we put our trust in a qualified practitioner to make a diagnosis, we trust human intelligence to be able to extrapolate, assess and question the information provided; therefore we view the decision they arrive at as trustworthy and reliable.
Watson, the aforementioned supercomputer, was able to improve with each case as the physicians would value each suggestion that Watson came up with. Evaluation of this nature is an ideal way to keep the Watson improving and could lead to a day where it is used independently to diagnose cancer, which could pave the way for its use in clinical settings across the world.
The use of algorithms in healthcare can result in more efficient health services worldwide by easing the burden on overworked healthcare professionals. With continuous strain being added on public healthcare services as a result of a growing worldwide aging population, this technology is evidently well worth continuing to develop and trialing in smaller territories.
Despite these findings, algorithms in healthcare will always be shadowed by one major doubt; will patients ever trust and feel completely satisfied with the reliability of the results it arrives at, and will patients and machines still require the support of a qualified professional?
Clearly, we are still some time away from allowing machines complete control over our medical diagnostics, however the rapid development of this technology opens numerous channels of opportunity as we enter the future of healthcare.
Procorre are fortunate to be working with some very innovative companies operating at the forefront of AI and algorithmic diagnostics and are constantly seeking new opportunities in this field The potential for improving a patient’s health is monumental, but very challenging – the future holds a great number of possibilities!
A Blog by Paul Fletcher-Dyer