Lord Darzi: Follow the 'digital breadcrumbs' to better health

Over the last 50 years we have seen immense advances in medicine and genetics. We now understand the molecular basis of many diseases and the sequencing of the human genome represented another dramatic leap forward.

But progress has been more uneven in translating those advances into new treatments and, most importantly, into better outcomes.

‘By following the trail of “digital breadcrumbs” we are already learning how to improve health, cut costs and even halt epidemics’

Slowly it has dawned that the huge variability of human behaviour has blocked improvement and we need a much richer understanding of how individuals live their lives.

That is why there is increasing excitement in medical circles about the potential of a new mechanism for monitoring health. “Reality mining” has already been cited as one of the 10 emerging technologies that could change the world by the MIT Technology Review.

It is a new way of observing human behaviour by monitoring the stream of data generated by the electronic devices on which we rely such as mobile phones, security cameras, social networks and loyalty cards.

These devices generate billions of digital traces − collectively known as “big data” − which offer a far more accurate and complete picture of our collective mental and cultural life than has ever been available before.

By following this trail of “digital breadcrumbs” we are already learning how to improve health, cut costs and even halt epidemics.

Insight or intrusion?

The idea of monitoring what people do generates alarm because it suggests intrusion into our private lives on a grand scale.

Big data has echoes of Big Brother − and it is not only governments that have an interest in what it may yield. Big data analysis has reached every sector of the global economy.

‘The key is the mobile phone, which is already being used to track movements, record interactions and measure vital bodily functions’

Protecting individuals’ privacy is vital but at the same time a means must be found to use the vast quantities of data now becoming available for the benefit of global health.

At this month’s World Innovation Summit for Health, an initiative of the Qatar Foundation and partners including Imperial College London, experts called for an international charter for open data sharing combined with updated privacy and data ownership policies.

They told the conference of 800 leading figures in government, business and the health sector from around the world that the charter should commit nations to sharing health data for their mutual benefit and set principles of best practice.

The key is the mobile phone, which is already being used to track movements, record interactions and measure vital bodily functions.

By monitoring their use, along with millions of others, it is possible to detect patterns that help understand and tackle diseases.

Mobile tracking

Researchers at the Massachusetts Institute of Technology (MIT) in the US have created a map of their own campus showing where the risk of contracting infections is highest.

They first interviewed by mobile phone a small sample of individuals about their health and then tracked their movements − by tracing their mobiles − to reveal the likeliest “hotspots” for transmission.

‘Top of the list of conditions that increased the chances of a patient being readmitted was their mental state’

A similar approach was used in Lahore, Pakistan, in 2011 to tackle the worst outbreak of dengue fever in the city’s history. Dengue fever is spread by mosquitoes, which breed in pools of stagnant water.

First, the location of cases of the disease was plotted and high risk areas for infection were identified. Then government employees were dispatched with smartphones to photograph and record the position of the pools of water.

These breeding grounds were then dealt with either by removing the water or by introducing tilapia fish, which eat mosquito larvae. The result was a sharp decline in dengue cases in 2012 and 2013. 

Reducing readmissions

A hospital in Washington, US, made a surprising discovery when it analysed several years of anonymised medical records seeking ways to reduce readmission and infection rates.

‘As individuals fall ill they turn to the internet to look for remedies and by tracking searches it is possible to follow an epidemic’

Top of the list of conditions that increased the chances of a patient being readmitted was their mental state.

One of the best known applications of big data for health purposes is Google Flu Trends − a means of predicting flu outbreaks, not by tracking the virus but by tracking people’s responses to it.

As individuals fall ill with flu they turn to the internet to look for remedies and by tracking those searches it is possible to follow the path of an epidemic as it spreads across a continent.

Reality mining also promises to broaden research. Much of our information about the link between chronic disease and diet, physical exercise and smoking has come from questionnaire based studies in North America and Europe.

But the spread of mobile phones across Africa has opened up the continent to studies that are urgently needed. Africa is experiencing an epidemic of chronic disease that will soon overtake the burden of infectious disease.

A project led by Harvard Public School of Health in Nigeria, South Africa, Tanzania and Uganda, which is currently in the pilot phase, plans to enrol 500,000 participants from whom information will be collected by mobile phone.

Cutting costs

Big data analysis may also be used to reduce costs. In the UK, a study of drug costs revealed areas of the country where expensive branded drugs were being prescribed in place of cheaper generic alternatives because of variations in doctors’ prescribing patterns.

‘As aggregated data has a high commercial value, the cooperative could invest proceeds on behalf of members’

The analysis suggested the NHS could save more than £1bn a year if doctors changed the way they prescribed.

To win the confidence of the public, it is crucial that personal data is secure. Although individuals own their health data, doctors, hospitals and insurance agencies control it.

Experts proposed at the summit in Qatar that cooperatives might store and manage health data on behalf of members, who would decide what part of it they wanted to share.

As aggregated data has a high commercial value to pharmaceutical companies and research institutions, the cooperative could reap the benefits and spend or invest the proceeds on behalf of members.

In this manner members of the cooperative would enjoy true citizen empowerment − and the full potential of the health data would be released to improve the health services on which we all depend.

Lord Ara Darzi is director of the Institute of Global Health Innovation at Imperial College London and executive chair of the World Innovation Summit for Health