Loading...

Combining Machine Learning and AI with Deep Science

//Combining Machine Learning and AI with Deep Science

Combining Machine Learning and AI with Deep Science

The combining of technical progress with healthcare is an appealing pairing and Google’s DeepMind Health has been at the forefront of bringing together technical coding and research with the hope of supporting resource and rapidity challenges faced by the NHS. 

Streams, for example, is a secure mobile app built to help the right clinicians get to the right patients in time and subsequently prevent the ‘failure to rescue’ scenario where possible. The app holds the patient’s important medical information all in one place and allows this to be accessed and updated as needed. What is as interesting as the functionality of the app is how such functionality was decided upon. The importance of accessibility and end user feedback cannot be understated and the success of Streams has been very largely due to the clinician feedback that has gone directly to the makers of the technology. A seemingly obvious approach, yet a rare one to actually see implemented. Thanks to input direct from the doctors and nurses using the technology, the app knows not to simply show the numbers associated with patient results but clearly illustrate the order of such numbers so clinicians can better analyse them. 

It is certainly an exciting prospect when I.T. expertise comes together with healthcare advancement. A recent AI breakthrough for DeepMind Health in partnership with Moorfields Eye Hospital was a study that showed AI to be capable of diagnosing eye disease as accurately as some leading experts in the field. A machine was able to read eye scans and detect 50+ eye conditions.

The machine was able to diagnose cases as urgent, semi-urgent, routine and observation only. Each urgent diagnosis was correctly captured and there is a great deal of excitement that this type of AI and machine learning will help reduce delays as the number of patients awaiting assessment is a huge problem, particularly when conditions such as wet age-related macular degeneration can result in blindness if not captured and treated rapidly. 

Whilst this algorithm is a positive enhancement, as it should be seen that AI is not being pitted against humans but supporting humans to overcome some serious time and resource challenges, there are concerns over safeguarding patient data. Sharing health details with Google is a rather alarming thought and yet there are certainly persuasive arguments for AI to continue pushing forward in the realm of healthcare so it is able to go beyond interpreting scans but also mammograms and CTs – clinicians can then spend more time planning treatments and caring for their patients. However it is essential that the ethics around such changing processes be pushed forward too, at the same pace, if not faster. Undoubtedly the ethics behind the technology needs to be given as much importance and attention as the technology itself. 

Devneet Toor

2018-11-12T12:16:32+00:00