Big Data, the Quantified Self, the Predicted Self and why the NHS is like a Basilisk

In a series of occasional posts I am going to be looking  at ‘data’ in the NHS and share some thoughts. These are some of the topics I will be covering:

  • ‘Big Data’ is the new industry buzzword. Hopefully it will allow us richer insights into what happens when a patient makes contact with the NHS but we need to avoid becoming overly preoccupied with gourging at the trough of and avoid Data Obseity. How much data do we really need to make a difference? Consider the 80/20 rule. And understand that Big Data does not provide the answers, it may however allow us to ask much better questions.
  • Investing in Big Data and making it available is of little use unless people are properly equipped to understand it and then have the skills to do something about the questions it highlights. A bit like giving a Kalashnikov to a child – a lot of spray and pray.
  • Currently NHS Big Data is about what happens when a person is in contact with the NHS. It helps us understand how we manage patients in the rear-view mirror. Unless combined with other socio-economic and personal data it tells us very little about their pathway before they came to NHS’s attention. It should help us manage patients better but the risk is that we simply focus on shifting the deck chairs when the iceberg is looming. It tells us little about keeping people out of the clutches of the NHS in the first place. A preoccupation with NHS Big Data, however worthy, could simply lock the service into its current mould.
  • The most important and promising opportunities for radically re-shaping the relationship of individuals with their own health (as opposed to their relationship with the Health Service) lies in the Quantified Self.  A simple ‘lean’ data stream that continuously monitors a few elements of a person’s lifestyle, using consumer technology design principles and provides feedback that uses emerging research in behavioural modification is likely to have a much greater long term impact than Big Data for a lot less cost.
  • The evolution of the Quantified Self is the ‘Predicted Self’. Evidence from some well established services is beginning to tell us that the absolute figures generated by monitoring are less important than the trends and patterns in that data.  We are able to use just a few continuously sampled data items generated by the Quantified-Self data to predict and anticipate potential problems and offer opportunities to avoid deterioration or crisis. There are already some highly successful industry analogues that are arguably dealing with far more complex challenges monitoring and prediction challenges very successfully.
  • The NHS central approach to Telemedicine and Telecare has been deeply flawed – akin to the Basilisk’s gaze (read Harry Potter and the Chamber of Secrets). The Whole System Demonstrator projects may have been well intentioned but were misconceived, suffered from industry capture and limited central imagination and have set back the ’cause’ several years. Although that in itself provides powerful learning. There is a very clear distinction to be made between Remote Care which is in the mould of current service modalities and the evolving field of Predictive Monitoring. It is unlikely that the NHS centrally has the culture, capability, capacity, appetite or imagination to make rapid advances in the use off the Quantified Self and Predicted Self fields. Bringing it into the research field risks ‘academic capture’ – an existence in a totally different time/space continuum that does not match the rapidly evolving market and the speed of approach of the health challenges we face. The market, private investment, private providers and private payers and just possibly some very imaginative CCGs are going to drive the first and second generations of services and absorb the risk. The NHS has to be ready to learn from the experience and ride the third wave.
  • Pieces of the Quantified Self and Predicted Self jigsaw are already available – the real challenge is the development of a viable service model that does not simply replicate current healthcare models but radically re imagines the nature of the relationship between self, family and services. This is where it gets contentious – because we are talking about taking the Doctor out of the Loop (DooL) and event the Clinician out of the Loop (CooL; and yes- you saw it here first!) at key stages in the process. This is about using pattern matching algorithms and advances in machine learning to spot trends, variations and to learn from the Big Data generated by Lean Data of the quantified self.
  • If NSA and GCHQ can do it already then it is going to be happening in the ‘real’ world sooner than you think. Come to think of it, if GCHQ is that short of money that it has to take funding from the NSA then perhaps there is an opportunity here for them.