Data & Difficulties: Jane Austen & AI in Healthcare

tipu's_tiger_front_view_2006ah4173

Tipu’s Tiger – Captured during the fall of Mysore, and now in the V&A

In 2014, Lucy Worsley claimed that the early Georgian kings had laid the foundations for a modern Britain. Broadly, I think her argument is correct. Specifically though, I want to look at the life of Jane Austen (1775 – 1817), and talk about the changes in her world, what they meant for her, and what they mean for us.

It is a truth universally acknowledged, that a single man in possession of a good fortune must be in want of a wife

When Jane Austen was born, George III had been king for 15 years. Georgian England saw the creation of what we now recognise as the modern United Kingdom. It saw the first Prime Minister (driven by the need for leadership while George I spent most of his time in Hanover), the Industrial Revolution, the Napoleonic wars and loss of the American colonies; the abolition of the trans-Atlantic slave trade. This was a world with coffee-houses, satire and debates over womens’ rights. It saw the creation of the modern United Kingdom – the Act of Union with Ireland was signed in 1800 – and we can trace much of modern politics back to choices made in Georgian England.

It is worth looking back; Charles I had been executed ~120 years before Austen was born, and there had been a (partially successful) invasion of England by his descendants 30 years earlier. But by the time she was born, the Stuart world was truly gone.

Medicine was less modern. Guy’s hospital was built in 1721, and James Lind led the first randomised trial in 1747, but lemon juice was not in routine use in the Navy until 1800. Vaccination was introduced in 1796, but major changes in medicine – Semmelwiess’ work and antisepsis lay in the future (1847 & 1867 respectively).

In light of all of these changes, one of the notable questions around Austen concerns her silences. The Napoleonic wars barely feature in her books (although Mr. Whickham was presumably called up because of the war), and the other changes impact even less. Yet something persists. Her lines resonate even now, even with all of the differences between our worlds.

In an earlier post, I talked about some of the problems with viewing data as oil, and the tragedy of the anti-commons. In particular, I was concerned with the way in which existing models of law and ethics make it easier for large organisations – which can negotiate these thickets – to access data even when patients might not want them to, while work that patients might want to support can find it difficult to access data.

Austen’s characters are often described by their incomes, and these certainly govern their fate. These incomes came from either farm rents or government bonds yielding 4 – 5%. Inflation had been low, and so it was reasonable to describe a man largely in terms of his income:

Mr. Darcy soon drew the attention of the room by his fine, tall person, handsome features, noble mien, and the report which was in general circulation within five minutes after his entrance, of his having ten thousand a year

with the expectation that the value of 10 000 would have a stable value over time. One thing that leaps out from thinking about Austen is how true Amara’ law is:

We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.

And during her lifetime, these certainties began to fade: in 1797 the Bank of England suspended specie payments (i.e refused to exchange bullion for notes) and in 1801, inflation ran at > 30%. These were directly the result of the war against Napoleon, but this was part of a more general change in the economy: before 1700, the economy in England was driven mainly be agricultural cycles; by the 1800 there were major drivers from other areas of the economy.

Much of the change in the economy was driven by changes in land ownership. The act of inclosure in 1773 offered local landowners the opportunity to enclose and privatise what had been common-land. Linked with improvements in farming technology, this increased land yields – but it also led to the population being forced into cities. Less than 20% of the population lived in towns (>20k inhabitants) in 1801; by 1897 it was >50%. At the same time, the banking and financial sector exploded: the relationship between capital-intensive industries, large sources of capital and a regulatory environment best described as somewhere between helpful and corrupt, led to substantial feedback loops.

This didn’t just benefit the new industries; it also changed existing companies, and let them grow. One example is the East Indian Company. Founded in 1600, it started as late-comer, competing with established Dutch and Portugese counterparts. It wasn’t until the mid-1700’s that, in part because of competition with French counterparts, it began a program of Empire-building. By 1803, it had a private army twice the size of the British army, and was the de facto ruler of large areas of India.

These companies changed the world. They did so by finding solutions to the tragedy of the anti-commons – they enclosed common land, bought seats in parliament and acted as rulers of countries. The work they did changed the world, but at huge cost, and Austen is largely silent on these. In fact, her silence is telling: the changes that were unfolding during her lifetime led to the destruction of her world. The urbanisation and financialisation of England shifted the balance of power from the country to the town, and inflation destroyed the guarantee of stable income.

That story, of huge global companies, in some ways better thought of being countries in their own right, that benefited from a linked, integrated system of regulatory and financial support may seem somewhat familiar. The benefits of the new economy flowed to a small number of owners and financiers. The corn laws (1815) and Peterloo massacre (1819) are testament to how heavily biased thing were against urban inhabitants. This only came to an end when urbanisation and labour organisation began to improve the lives of workers. The Great Reform Act in 1832 resolved some of these; the East India Company collapsed in 1857, after the first Indian war of Independence, and was taken over by the government.

In an earlier post, I suggested that we should think of healthcare data as being more like music than oil. Irrespective of what analogy we use, there is clearly value in the data. But the current processes act as a block on what we would like to do (most patients expect us to share far more data than we do) without stopping work that many would object to. The growth of large companies, based in part on the ‘enclosure’ of user data follows a pattern we saw in the Georgian era. My question is how we make sure that the benefits of the data, and the work built on that data, benefit patients and the healthcare system, rather than just benefiting those who are able to make use of them. I have suggested that this might be a use case for medical blockchain – it would at least allow patients and carers to assign and control use of the their data in a robust, tracable way. Another alternative would be to democratise the access to the data – and this is something that the Simulacrum was designed to solve.

The current model of building and using large healthcare datasets for medical AI is likely to be unsustainable. The idea that we will take data from publicly funded health services, and allow companies to build valuable and expensive products on top of those to sell back to the NHS is unlikely to be viable in the longer term. Many individual trusts are aware of these problems, but there is a major imbalance between the power and expertise of NHS Trusts and large technology companies.

At the moment, patients have little say on what happens with the tools built on their data. That is unlikely to last, and is not something we should fear – the consistent message is that patients want us to use their data to benefit both them and other patients. How we we do that is a challenge, but there are options. One model is to build models based on patient data, and then make those models available in an open-source format. Clinically-assured models (i.e. those that have regulatory approval) would still require commercial support, in the same way that the NHS doesn’t make its own drugs or hip replacements. But the combination of open-sourced models with commercial validation might both keep costs down, and help spur further innovation & research. Whether the final model is this ‘licensed open source’, blockchain, public data, or something else remains open.

The changes that happened in Georgian England fundamentally changed the nature of England and the rest of the world. The scale of that disruption is difficult to imagine, but the picture and quote give us two different models. ‘Tipu’s tiger‘ was captured in 1799, during the British conquest of India, in a war that heavily involved Arthur Wellesley, later the Duke of Wellington. The capture and transport of that model, which now sits in the V&A in London is a story of Georgian globalisation. At the other end, we have Jane Austen’s world – already vanishing even as she wrote about it – which for all its deliberate silence on the changes going on around her, captured something timeless about people and their relationships. That should be a reminder that many of the elements of clinical medicine will remain unchanged, even as the world around it evolves.

 

A version of this talk was delivered as part of the ‘SmartHeart’ meeting on 16th January 2019

Categories: Uncategorized

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s