By World Healthcare Journal-
Artificial Intelligence or AI has recently received plenty of limelight in the healthcare sphere. Deep learning AI has developed new compounds for vaccines, drones have been used to revolutionise medical logistics, and governments are realising the potential of AI in the healthcare sector with the UK government announcing a £250m investment into developing a state of the art AI lab for the NHS.
And now, a scientific review published in The Lancet has found that deep-learning AI may even be as effective at diagnosing health conditions as expert human clinicians. However, more research is needed to discern if AI really can be as effective as human diagnosticians.
Researchers from University Hospitals Birmingham NHS Foundation Trust carried out what is believed to be the first-ever systematic review of AI technology in the diagnostic sector, and published their findings in the Lancet Digital Health Journal.
Despite finding that expert AI diagnostic systems can match human success rates, and in a few cases surpass human efforts, Professor Alastair Denniston, an author of the study, has urged for more stringent standards of research and study to improve further assessment of AI within healthcare.
"We reviewed over 20,500 articles, but less than 1 per cent of these were sufficiently robust in their design and reporting that independent reviewers had high confidence in their claims," says Professor Denniston.
"What's more, only 25 studies validated the AI models externally, using medical images from a different population, and just 14 studies actually compared the performance of AI and health professionals using the same test sample."
However, even though large percentages of the original articles examined were insufficient for the purposes of the study, the findings of the systematic review show that AI has ‘enormous potential’ for improving the precision and speed of diagnosis.
The analysis of the data from the 14 studies showed that, when operating at best, deep learning AI systems correctly detected disease in 87% of cases. The ability to accurately rule out patients who did not have disease was similar - 93% for the machine algorithms, compared to 91% for doctors.
However, the clinicians in these scenarios who had to make the decisions were not provided additional patient information which they would have access to in the real world – which would aid diagnosis.
"We found deep learning could indeed detect diseases ranging from cancers to eye diseases as accurately as health professionals,” says Professor Denniston.
"But it is important to note AI did not substantially out-perform human diagnosis."
Dr Xiaoxuan Liu, the lead author of the study, addressed notions that AI systems are overtaking human capability, saying; “There are a lot of headlines about AI outperforming humans, but our message is that it can at best be equivalent. ”
#whjnews #whjdigitalhealth #whjuk