The Health Foundation is pleased to publish this guest blog, from Indra Joshi and Jess Morley of NHSX. Views do not necessarily represent those of the Health Foundation, but we welcome and encourage debate in this important and rapidly advancing field.
Last month, Adam Steventon (Director of Data Analytics at the Health Foundation) wrote an insightful blog, following a roundtable discussion about the pros and cons of the all-pervasive ‘data is the new oil’ analogy in the context of health care. It accurately notes that like oil, data is useless in its raw form but, unlike oil, it’s not a finite resource and asks the question ‘do we need a new analogy?’ We think not – and here’s why.
Let us reiterate that the ‘data is the new oil’ analogy doesn’t work in the context of health and care data for two reasons.
Firstly, it glosses over import and export issues. Oil that is produced in one place and used in another is still just oil (sometimes with terrible humanitarian and environmental consequences, but that is not the point here). But health data cannot be treated in the same way. Algorithms trained and tested on one dataset, or in one context, simply cannot be exported and used on an entirely different dataset, in an entirely different setting, without issues of accuracy, effectiveness or precision arising.
Secondly, it puts too much emphasis on individual benefit. In oil production, individuals (land owners), companies or governments often seek to maximise profit with little concern for the wider community or the environment. Health data is personal and should be used by individuals to help maximise their own health, but we must recognise that, while there are risks of data being exploited for profit, this data can and should be used as a source for good at a wider population level.
For example, there are considerable public health benefits to be gleaned from aggregated data sets. Data policy, including that which relates to AI, needs to take into account both individual impacts and group impacts (both positive and negative). This is something that it took the oil industry, and those governing it, far too long to wake up to, as the current climate crisis attests. We need to do better and be more proactive when it comes to health and care data use.
Instead we argue that it’s not possible to come up with a ‘data is the new….’ analogy in the context of health data because it is just new. It is in fact a paradigm shift.
Of course, data has always been important in health care. We’ve always recorded patient notes and looked for patterns, but we were primarily looking for strong signals (such as links between cause and effect). For example, if you drink contaminated water, you get sick.
But data is now used in fundamentally different ways, including recognising patterns that would have previously been completely unknown to us (known as ‘weak signals’), making inferences about us (such as what we respond to, sometimes subconsciously), and even influencing our behaviour (in both positive and negative ways). These new capabilities might be used for predictive analytics, personalised medicine, the discovery of new drugs or new understanding of specific diseases. Data therefore enables us to learn more about the health of individual people or the performance of the system and respond in real time.
Utilising these new data analytics opportunities in ways that are safe, ethical and responsible requires thinking in new and innovative ways. In order to do this, we must not be constrained by old ways of thinking about the art of the possible. This is exactly why we cannot rely on ‘data is the new….’ analogies, as these encourage us to think of new problems and prospects in old ways, and may therefore lead to missed opportunities or unintended harms.
We have to recognise that the scale of change we’re facing is just new, not a new version of something else, so that we can stay open-minded about it. This flexibility will enable us to be proactive in supporting innovation in data analytics, while also protecting patient safety.
You might also like...
Is it helpful to compare data and oil to make sense of how data can best be used to tackle health and care challenges?
A retrospective mixed-methods study
Real-time data analytics to improve mechanical ventilator wean – guiding clinical behaviour (ATTITUDE study)
Project that aims to understand the barriers to implementing evidence-based care in complex critically ill patients, and will...
Health Foundation @HealthFdn
Are you part of a #research team investigating the ways in which the physical & mental health of a population shape… https://t.co/VjUR3XwxZoFollow us on Twitter
Work with us
We look for talented and passionate individuals as everyone at the Health Foundation has an important role to play.View current vacancies
The Q Community
Q is an initiative connecting people with improvement expertise across the UK.Find out more