Artificial intelligence, Big Data, and blockchain turn doctors into superheroes

Doctors + AI = Super Heroes

In recent years, artificial intelligence (AI) systems have been developing quickly. Today, they beat human experts in a number of fields, from playing the complex game of Go, optimising advertising strategies, to driving cars.

Of course, AI is taking over medicine, too. For example, computer systems are better at diagnosing heart problems or lung cancer than experienced doctors. Physicians supported by AI can significantly improve their diagnostic capabilities to detect illnesses earlier and more accurately.

Imagine doctors feeding patients’ symptoms into a system that will come up with the most likely diagnosis and suggest the best known treatment based on data from millions of other patients and all of the world’s medical research. The likelihood of misdiagnosis and sub-optimal treatments would hugely decrease. As such a system could be accessible through a smartphone while the actual computing happens in the cloud, it would be available even in rural or remote areas.

AI tools will grant doctors abilities in saving lives that will appear as super powers when compared to today’s medical standards.

AI gets better with more data

The progress of AI systems is mainly based on one factor: more data. Although the algorithms have been slightly improved and computing power has become cheaper and cheaper, the main driver behind improving AI is the availability of Big Data. Why is this the case?

The most successful branch of AI is Machine Learning (ML). ML works by training simulated digital neural networks on vast amounts of training data. Training data can be, for example, sensor data from driving cars to train autonomous driving systems. Or medical assistance machines could be trained with medical data including diagnoses, therapies, and their outcomes.

Just as (human) animals, ML algorithms ‘learn’ by detecting those patterns in the input data that lead to the desired or undesired outcomes. For example: Patient with fever → Aspirin treatment → patient no fever. However, with their pure mathematical approach, they are able to detect patterns that human minds can’t detect. And they can digest exponentially larger amounts of input than humans – the more, the better.  

There is another advantage of expert systems over human doctors: Once trained, AI can be easily copied. So, for example, people with weak hearts can carry around the world’s best cardiologist in their pocket to monitor them at all times at practically no cost.

Training data is locked away

However, AI will only enable superhero doctors who give us longer and better lives if it can be trained on vast amounts of patient files and clinical data.

Yet, most of that data needed to train AI systems is locked away. On the one hand, combining and sharing data is technically difficult because it is stored in thousands of incompatible systems owned by different organisations. On the other hand, in most countries, data protection regulations do not allow hospitals, doctors, and other healthcare providers to easily share patient related information.

Even worse, the data generated from new sources such as wearables like the Apple Watch gets aggregated in the data silos of a handful of big American companies. And their mission is not primarily to improve people’s lives, but to generate profit for their shareholders.

We will only be able to provide doctors with AI superpowers, if we get access to the data currently stored in incompatible data silos. And we need to put patients back in control of their own information, so they can decide to share it for their own good.

Blockchain to liberate medical data for AI

Blockchain, the revolutionary technology behind Bitcoin, leads the way to data sharing in healthcare across all sectors and sources of information.

If patient data were stored on a blockchain or similar distributed ledger technology (DLT), individuals could control who can read or add to the information of their patient files. Just as today, all healthcare stakeholders could use their own special applications to do their job. But they would use a shared, secure healthcare blockchain to store encrypted information about patients, treatments, and research.

Whenever stakeholders need to share data, the owner of the data (in many cases the patients themselves) can decide to grant access to interested parties. Further, with smart contracts, most administrative and invoicing processes could be automated across organisational boundaries on the healthcare blockchain.

With blockchain technology, we could build a medical database that spans the globe and at the same time puts individual patients back in control of their data.