Data and AI: How healthcare should use it (or, Florence and the machine)

After the flood director Max Gadney wrote the below article for The Huffington Post on November 18, 2016.

Health services around the world are stretched.

Austerity economies have squeezed public health services while private suppliers compete on ever tighter margins. But there are ways to improve. One aspect of modern technology is particularly useful for improving health – beyond hardware and devices, there is data.

We throw off more data than ever before from our everyday activities – submitting it voluntarily when we book travel tickets, involuntarily when we’re filmed by CCTV, and somewhere between when we buy from Amazon and receive a recommendation based on our preferences. This data amasses over time and can come in very useful in areas where it’s perhaps not usually considered.

For instance, at After the flood we’ve been working with Guy’s & St Thomas’ Hospital on an app called 1-to-1, which aims to improve feedback between expectant women and midwives.

During this project, we reflected on the work of Florence Nightingale, who set up the first nursing college at St Thomas. Nightingale was a celebrated practitioner, but we particularly like her use of visual statistical evidence and how she drafted Isambard Kingdom Brunel to design and build a rapid-prototype hospital to treat the injured during the Crimean War.

We can match the breadth of Nightingale’s achievements thanks to the potential that data has to unlock value in our health systems. But just as Nightingale is trivialised with the ‘lady with the lamp’ epithet, so too are the arguments about data and artificial intelligence trivialised by over-simplification.

Data is too important for that – anyone working with data needs to look at how to capitalise on the opportunities and mitigate the risks.

Data is the Sixth Estate

Of course, there is some controversy surrounding data and how it is used—particularly in an industry as personal as health—and these concerns must be addressed.

Early Christian societies described the three estates of clergy, nobility and commoners. The press and then counter-culture were added as the fourth and fifth. I suggest that data is the sixth.

Nell Watson talks about the corporations as the third actors in the people/machine mix. For this, we can add any organisations, NGOs, companies, startups—all seemingly different entities—that share one new responsibility: to be transparent about how data is gathered and used. All these entities should ideally share a set of values – like journalists—and then more diverse stated aims within these values.

This is not to say that all transactions need to be open—much data is personal and needs privacy—but not all data from a person is personal.

Likewise, if the benefits of giving data to both parties are made clear, the person can choose if or how much they give.

Much of the fear-driven narrative derives from the assumption that data is an organisation’s to take. Actually, if they nuance the interfaces for collection in a people-centred manner, providing options and transparency of use, they can collect more data with greater diversity.

Improving the before, during and after for patients

Once people are happy for their data to be used, a whole variety can be gathered and utilised. The NHS excels at acute clinical care in hospitals, but it is more stretched in the before and after stages of an illness, injury or need for medical assistance. It’s here that data can really play its part.

Lifestyle data from mobile devices is already being used in preventive apps to help people behave in a healthier manner. The NHS key risk in the future is lifestyle inflicted diseases, and these can be prevented and pacified by ongoing monitoring and the data that results from it.

Furthermore, in hospital, the tightly choreographed dance of practitioners, carers, pharmacists, relatives and patients can be better coordinated by knowing who is doing what, where and when. The advent of cloud computing, Wi-Fi and mobile device proliferation means that clinicians can save logistical and operational overheads such as resource management and prioritisation and concentrate on making important clinical decisions.

The debate needs to be around how machines (a phone) will help humans (a nurse or doctor) do what they do best, not about replacing people. We already use thousands of machines to improve on our basic functions (bikes, cars instead of walking). The new machines will just be data-powered, rather than engine-powered.

If Nightingale could have built the hospital of her design in 2018, it would have a flexible architecture, enabling it to adapt to new systems and installations of hardware and connectivity.

She’d start by tearing out the old connectivity and installing ubiquitous, industrial Wi-Fi. She’d talk to her many influential peers and secure tech partnerships for hardware so all nurses and patients had robust mobile access to information when they needed it. She’d keep all her stakeholders up to date with beautifully designed visualisations of key KPIs and efficiencies being met.

The lady with the LAN.

This site is now archived. More here.