1. Streams – a secure clinical app

The Streams app, pictured below, was built by DeepMind Health and digital product studio ustwo to provide real-time information to clinicians, enabling them to more easily prioritise patient care in hospitals. Patient alerts will provide doctors and nurses with the data they need and the app will allow clinicians to communicate about next actions.

DeepMind says that ‘by freeing clinicians from juggling multiple pager, desktop-based, and paper systems, Streams will help redirect significant amounts of time from admin and towards direct patient care.’

Streams has been in use at the Royal Free London hospital since November 2016 helping to treat patients with acute kidney injury. Early feedback has been good, with nurses reporting time savings of up to two hours per day.

streams app

Image via DeepMind Health

2. Verifiable Data Audit – a ‘blockchain’ system to track health records

It’s perhaps worth noting some tension around the Streams app at the back end of 2016, as critics said that Google’s five year partnership with the NHS gives it too much access to NHS data without appropriate oversight. That’s where DeepMind’s next big project may come in.

DeepMind Health published a blog post in March 2017 about its work to bring scrutiny to the use of medical data. Verifiable Data Audit is a blockchain-style system that will give partner hospitals an additional real-time mechanism to check how data is being processed.

Where authorisation and patient consent is needed, the system will show who is accessing data. DeepMind explains further in its blog post:

For example, an organisation holding health data can’t simply decide to start carrying out research on patient records being used to provide care, or repurpose a research dataset for some other unapproved use. In other words: it’s not just where the data is stored, it’s what’s being done with it that counts. We want to make that verifiable and auditable, in real-time, for the first time.

As with blockchain, the system will not allow for any record of data use to be erased, with third parties able to verify data is uncompromised. However, the system will not be as complex or expensive as blockchain, nor will it be decentralized, as it will rely on hospitals and national bodies to act as verifiers.

3. The Study Watch – an everyday wearable

Verily Life Sciences, one of Alphabet’s research organisations, has developed a stylish health tracking watch which differs from wearables on the market.

As it is designed for every day use in medical studies, the watch is as wearable as possible, ensuring that study participants don’t neglect to wear it. That means it looks much like a stylish analogue watch, with only time and date displayed (participants do not have access to logged data), and the battery lasts for a week with syncing only required on a weekly basis, too.

Furthermore, the watch’s software can be updated wirelessly and a processor encrypts the data collected, which includes:

  • heart rate
  • electrocardiograms (low res versions taken by gripping the watch bezel)
  • movement data
  • electrical conductance of the skin
  • ambient light and sound

verily study watch

The Study Watch is used in several of Verily’s studies which aim to collect a wide variety of data. One such study will begin in June with the Parkinson Center at Radboud University Medical Center, Netherlands. 650 volunteers will each use one of the watches, as well as have brain scans and blood tests.

Bastiaan Bloem, a neurologist at the center tells MIT Tech Review the watch is “an exciting way to track people 24 hours in their homes…We know one of the early symptoms in Parkinson’s is heart rate variability. But we only measure it in the hospital.”

4. Project Baseline – a huge longitudinal study

Project Baseline was announced in April 2017 and its scope makes it the most eyecatching thing that Verily has worked on to date. 10,000 volunteers will be monitored over four years, building a whole range of datasets using in-clinic testing and continuous data collection via a Baseline app and the aforementioned Study Watch.

The study is predicted to take a decade to complete (due to the time needed to find enough participants) and cost more than $100m, with the ultimate aim of finding predictors of conditions such as heart disease and cancer.

Participants will undergo x-rays, heart scans, genome sequencing and blood tests, alongside the continuous tracking of activity data and sleep patterns (using an electric loop placed under a mattress).

With such a variety of data, Verily is ‘creating an infrastructure that can process multi-dimensional health data – much of which have never been combined for an individual.’ Other long term studies of note (e.g. the Framingham Heart Study) do not look at such a wide range of data, so Baseline’s results will be eagerly anticipated.


5. An eye-scanning algorithm to diagnose eye disease

One burgeoning area of medical research is the use of machine learning to assist with diagnosis, as well as some practical parts of medicine (such as the careful segmentation process during radiology).

DeepMind Health is innovating in these areas and received much publicity for its work in ophthalmology. Since July 2016, DeepMind has worked with Moorfields Eye Hospital in London in training computers to assist with screening for diabetic retinopathy (DR). The algorithm was trained using 128,000 retinal images that had already been classified by ophthalmologists. In testing, the algorithm then matched or exceeded the performance of experts, not only in identifing DR but also in grading its severity.

DR affects a third of diabetes sufferers and Moorfields assesses 3,000 scans every week. The development of such an algorithm allows expertise to be transferred to areas of the world that are lacking in retinopathy experts.

retinopathy research

6. Miniaturized continuous glucose monitors (CGMs)

Verily is working with Dexcom to develop miniaturized continuous glucose monitors (CGM) for those with Type 2 diabetes.

The monitors are predicted to be ready by 2020, and will help to make continuous monitoring less expensive and less disruptive, using wireless connectivity and precluding the need for measurement with the traditional finger stick.

Fewer that one in five people with Type 2 diabetes use continuous glucose monitors. This new technology may make the practice affordable and convenient for many more.


More on healthcare: