Artificial Intelligence in Healthcare: Rewriting Medical Textbooks and Closing the Value Loop

Earlier this year I had a fascinating conversation with Ed Godber, a Healthcare strategist and AI consultant, about data revolution and transforming role of Artificial Intelligence in the Healthcare sector.

A lot has been written about big data in various industries, and healthcare is not an exception. It has historically generated large amounts of data, driven by record keeping, compliance, regulatory requirements, and patient care.  With the digitalisation of medical records and the introduction of new methods of gathering and measuring data, these huge quantities of data hold the promise of supporting a wide range of medical and healthcare functions, including clinical diagnostics and decision support, disease surveillance, disease prevention and treatment, drug development, anti-ageing and many more.

According to Ed, data in healthcare faces numerous challenges. 

Due to the privacy and history of handling the medical knowledge, a lot of data is still inaccessible and fragmented. Relevant data is often slow and expensive to obtain.  There is a big push from internet giants such as Apple, Amazon, Microsoft and Google all trying to get involved in sorting out electronic health record infrastructure and get hold of valuable data. However, there is a big debate on who should own and control access to the data.

A lot of health data is simply not captured because we, as individuals, often cannot discover that something is wrong with us.  Many health issues do not manifest as pain and are hard to articulate.   If this data is not captured, then it is difficult to make use of recent discoveries in pharmaceuticals, devices, or nanotechnology.  This is being addressed by introducing new ways of measuring and collecting relevant data, often direct from patients.  Company 23andMe gathers DNA data and Viome collects bacterial genome data from consumers.  Facebook-like health communities have evolved to capture patient-reported data. Finally, there is a huge trend in collecting data through wearables and iPhones. 

Nevertheless, this is not only about Big Data.  In fact, the emphasis is on Smart Data.  Data in Healthcare is of the entirely different nature and has enormous value.  To put things into perspective, imagine you are browsing things on your phone, and in a few minutes you give away about 5000 pieces of data.  What you are getting in return is probably quite superficial: perhaps a bit of information and also some ads that want something from you. 

“In healthcare, with 5 pieces of information I can not only diagnose you as having cancer, but I can also work out what is going on in terms of your genes to be able to tell you which drug will save your life”.

There are lots of ways that artificial intelligence, data technologies and data measurement are rebuilding the way we can do diagnostics, treatment and disease prevention.  

The way Ed describes it, new technologies help to close the “Value Loop” more quickly than it was possible in the past.  For instance, if you diagnose an elderly person with Alzheimer’s disease, it will only have meaning if you not only understand the symptoms, but you also understand what is going on inside the body, understand the cause and know how to fix it.  Otherwise we are not closing the loop.  

This process can take 10-11 years for an average disease.  A good example of this is a life-saving drug Gleevec, for leukaemia. The underlying cause, a genetic mutation called Philadelphia Syndrome, was discovered in 1960s. However, what was going wrong with proteins was not discovered until 20 years later because we simply didn’t have means to analyse the relationship between the genes and proteins. Then the process of turning that into a cure took only 10 years, much shorter period of time than it took to analyse what was going on in the body. 

Now, with AI and particularly using deep learning and neural networks for image, voice and text analysis, we have more advanced ways to measure, analyse and understand causality.  For instance, in the last 5-6 years we have been able to measure how bacterial diversity in the gut affects health, and come up with the easy ways to prevent and cure Crohn’s disease, Irritable Bowel Syndrome and Liver Disease.  Another fast-developing area is cancer diagnostics and treatment.  Combining deep learning with imaging and genomics has proven to be successful in the early detection of cancers and identifying the best treatment for individual patients. Technological advances drive the shift from traditional statistics-based population science to personalisation in diagnostics and medicine. 

As we are gaining new understanding of the diseases and their causes with the help of AI technologies, medical textbooks are being rewritten and medical science is being re-examined.

A full talk with Ed Godber has been recorded for my training course “Introduction to big data, data science and artificial intelligence, which you can watch here.

Please contact me if you need a discount code for this course.