Data Mining: the Future of Healthcare?
Florence Nightingale’s Deep Mind
Life expectancy has doubled in Europe and North America over the last 150 years and is still increasing.
I will ask you: What is the single most important factor behind that remarkable fact? And you will probably cite the rise of pharmaceutical medicine, as in the invention of penicillin and the roll-out of mass vaccination; or the improvement in our diets thanks to advanced industrial agribusiness; or the decline of homicide to resolve disputes (though we’re not quite there, yet); or the containment of infectious disease by sanitation.
Now many historians of medicine think that the latter – sanitation – is the most important. In fact, the greatest advance in healthcare in the last two hundred years can be expressed in a single four-letter word: soap.
Personal hygiene was lauded by the Romans who lolled about in bathhouses, smothered themselves in unguents and shared clean communal loos at the agora. But after the fall of Rome to the Visigoths in August 410, no one had a bath again in Europe for about the next 1,400 years. Not even royals. If one could go back in a time machine to the Middle Ages one would meet people who pong. Until about the late 18th century it was still common to dispose of the contents of one’s chamber pot within the walls of the City of London by chucking it out of the top-floor window.
Soap had been around in limited quantities since antiquity, but industrially manufactured bar soaps only first became available in the late 18th century. Discoveries by two French chemists made this possible. In 1791, Nicholas Leblanc patented a method of making sodium carbonate (soda ash) from salt. And in 1811, Michel Chevreul formulated the chemical characteristics of fatty acids and glycerine. But it was an Englishman, Andrew Pears, who began manufacturing transparent soap in London in 1807. Pears Soap is a brand still with us – now made in India by Hindustan Unilever, a company in which Unilever PLC (LON:ULVR) has a controlling 67 percent stake.
As well as soap, consider toothpaste. After Sir Walter Raleigh brought sugar back from the New World in Elizabeth’s reign, the challenged dentition of the population became chronic. London hostelries displayed pliers on the walls for the removal of troublesome teeth until well into the 19th century. Colgate started to produce toothpaste (in jars) in the USA in 1873. Colgate is still around as a major, indeed iconic, healthcare and consumer products company in the form of Colgate-Palmolive (NYSE:CL).
Now a rising consciousness, indeed a culture, of hygiene is today universally recognised as a social and economic advance. Though there are still places in the modern world where sanitation is basic. I have written about India over the last year-and-a-half and it still shocks me that about half of India’s population of over 1.2 billion people does not have access to even basic latrines. Mr Modi has now made this a priority of his government. Future historians might conclude that Ghandian economics was finally subverted, not by the triumph of capitalism, but by the fact that China had manifestly better toilet facilities than India.
Hygiene is one thing: but how did the effects of hygiene stroke sanitation become evidence-based? Through statistics of course. Enter the Lady of the Lamp, Florence Nightingale. I was introduced to Florence at primary school as an über-nurse, the caring heroine of the Crimean War. But she was much more than that. She was a pioneering medical scientist who used the most advanced statistical mathematics available to demonstrate, beyond dispute, that most British fatalities in the Crimean War were dying, not of their war wounds, but of totally avoidable secondary infections[i].
She even correlated mortality with the number of times per month that soldiers’ shirts were washed. She inspected and analysed hospitals all over Europe, gathering and comparing statistics of all kinds – medical, demographic, meteorological – assembling a data base (though she did not call it that) which had never existed before. Thankfully, Nightingale got the attention of not just the War Office but of Queen Victoria too. And she got the resources that she needed to do the job for wounded British soldiers in Scutari Hospital[ii]. She even invented the concept of sample mortality compared with expected mortality (as in if they had remained at home). This concept is a key tool in medical research, particularly epidemiology, today. And her achievements are all the more remarkable when one considers that she was a woman living at a time when women were not even admitted to university – let alone to medical school.
Fast-forward to 2016. I recently learnt from New Scientist[iii] that Google’s (NASDAQ:GOOGL) Artificial Intelligence (AI) outfit, Deep Mind Technologies – about which, and the genius who founded it (Demis Hassibis), I have written before – have recently signed a contract with the UK’s National Health Service (NHS). The agreement gives Google/Deep Mind access to the anonymised records of 1.6 million NHS patients. These are people who use three famous London hospitals run by the Royal Free NHS Trust – Barnet, Chase Farm and Royal Free.
What I understand is that Google/Deep Mind is trying to generate algorithms to relate medical histories to medical outcomes. New Scientist states that they are also developing an app called Streams to help staff to monitor patients with kidney disease. But in my view that is not the key point. Let’s stick to the big picture. Google/Deep Mind will log all day-to-day hospital activity (even concerning hospital visits) and marry it to all medical data generated from tests (pathology, radiology – the lot). And they will link this with historic data stored in centralised records in order to develop patient profiles which have predictive outcomes. What are the chances that Patient X will develop Cancer Y over Period Z? And what can be done to minimise that risk? This would be a medical approach where prevention is at least as important as cure.
Google/Deep Mind is also developing a platform called Patient Rescue which appears to be a system which supports (influences?) diagnoses. But Google is not the only player in this game. The Computational Health Informatics Lab at Oxford University is doing something similar for the four hospitals across the Oxford University Hospitals NHS Foundation Trust. These players do not talk about automating clinical decisions – yet.
The medical profession is still notoriously resistant to performance measurement of almost any kind in the (overwhelmingly public funded) health sector. There are some who view the medical profession, as in Florence Nightingale’s time, as a kind of clerical elite which, instead putting all medical statistics into the public domain, seeks to use them so as to preserve its sacred status. Look at the very poor level of debate about the seven-day-a-week NHS. That may be harsh; but the tide of data transparency is such that much of the lower levels of the profession – General Practitioners in particular – might soon be replaced by expert systems.
If you went to the MI Conference on 23 April you may have seen a keynote address by Lorne Abony, CEO of Fastforward PLC. Amongst the many interesting things Lorne said, he mentioned that we can all now download an app to our smartphones which, with a low-cost sensor, enables us to perform an electrocardiogram (ECG or EKG) on ourselves. The resulting data could be beamed to a central monitoring computer-doctor at virtually zero marginal cost to the NHS. Why isn’t our cash-strapped Government rushing this into reality? The Millennials who all have smartphones anyway would love it.
The future of medicine is not necessarily about more advanced pharmaceuticals. We are probably already peak-pill; though the medics will always tell you to pop just one more. In ten years’ time we might regard Google as the number one global player in healthcare. But I’m researching some other players in the incredible space of healthcare data mining. Though good things come to those who wait…
Sometimes, a very basic technology – soap, for one, toothpaste for another – can have incredible results. Not long from now we shall understand that medical outcomes are not driven solely by doctors or nurses – or even pills. They are ultimately driven by data, as Florence Nightingale knew.
[i] Naomi Alderman has written and narrated an excellent BBC R4 programme on this subject, first broadcast on 18 May 2016, available at: http://www.bbc.co.uk/programmes/b07bdg6b
[ii] The British military hospital in Constantinople (Istanbul) where wounded British soldiers were taken.
[iii] Google knows your ills, New Scientist, 07 May 2016, page 22 by Hal Hodson.
Comments (0)