Big data: a friend or foe in healthcare?

William Pett discusses the growing suspicion that big data is being used to subvert rather than enhance democracy and explores the implications of  this for the healthcare industry. 

The digital age brought with it the promise of transparency, accountability and freedom of information on an unprecedented scale. However, there is a growing suspicion that data is being used to subvert rather than enhance democracy and this has increased the need for a new data framework in healthcare, where the potential benefits of the data revolution could arguably be the most significant.

Data scepticism

Last week, former Wall Street analyst turned data scientist Dr Cathy O'Neil gave a talk to the London School of Economics arguing that the ability to develop and apply algorithms to personal data on a mass scale has had damaging effects for society. Whether it be personality tests employed by companies that discriminate against those with mental conditions, or financial models that restrict black communities' access to credit, there is a growing perception that the world of data is opaque and malevolent.

Controversial data practices are beginning to come to light – the role of Cambridge Analytica, for example, in both the US election and the EU referendum has been put under the spotlight – and when it comes to an area as sensitive as healthcare it is unsurprising that the public is sceptical of how data is being used. There has consequently been a high level of scrutiny of the deal struck recently between the Royal Free Hospital in London and the data company DeepMind, owned by Google. The aim of the project – to use data to improve early detection of AKI (Acute Kidney Injury), a condition that kills up to 40,000 patients a year – is admirable and could well save lives, but the partnership has proved to be contentious. Last week, the Information Commissioner's Office (ICO) ruled that there were shortcomings in how the data was handled by the Royal Free Trust, noting that patients were not adequately informed that their data would be used as part of the project.

The ruling was not therefore a condemnation of a private company acting illegally or immorally, and indeed the ICO commented on the benevolent intentions of the project. It was a judgement that patients were left in the dark, and for those with an interest in health data this is an all-too-familiar situation.

Communication breakdown

In 2013 the government announced its intention to revolutionise the way that data is collated and used across the NHS. The project – care.data – would bring together patient records from across the health service into a vast central database, with significant potential benefits for patients. Not only would it make the NHS more efficient, with an end to paper records, but it would also allow weaknesses to be identified across England and across treatment areas, help doctors to better understand the cause of diseases, and enable better identification of the side effects of new drugs.

Despite having the backing of healthcare professionals, the academic community and the pharmaceutical industry alike, the project was scrapped in 2014 amid concerns about data confidentiality and how the project had been communicated to the public.  Patients were unsure about how their personal data would be used and, crucially, who would have access to it.

In response to the failed implementation of the care.data programme, and to establish a more coherent regulatory framework around data, a review on data security and consent was undertaken by the National Data Guardian Dame Fiona Caldicott. The recommendations of the review, including better protection of anonymised data and statutory opt-out rights, were set out in September 2016 and this week the Department of Health finally accepted these recommendations in its response.

Building public confidence

The long-awaited approval of new data standards has already been welcomed by a number of prominent health charities. There is hope that, in light of the upcoming implementation of the EU's General Data Protection Regulation (GDPR) in 2018, this will at least be the first step towards providing patients with some reassurance about how, and indeed why, the NHS, government and private sector are using personal records.

If we take two assumptions to be true – that there is huge potential to save lives by responsibly encouraging innovation in data and that patient privacy/authority must be central to data reforms – then a clear framework to build public trust and provide private sector stability is long overdue. The sooner the government acts on Dame Caldicott's recommendations, therefore, the better.