Whenever an advert pops up on my Facebook page selling me exactly what I had been searching for the day before; or the supermarket gives discounts fitting my shopping profile, I quite rightly assume that ‘my data’ has been analysed, re-purposed and turned around to give me (and the company) new options – what could be more simple?
Indeed, the re-use of data to enhance value is a given in our informatics-centric 21st century. So why is it so difficult for researchers and hospitals to use routine medical data in the same way? Why are there fewer examples where data enhances the patient experience, makes it easier for health professionals and delivers the promised academic returns such that we clearly use it to improve health care?
In fact, as you dive into the problems, there are a myriad of reasons. Data in the NHS has largely been collected through software designed to deliver the specific needs of the service, the research opportunities were not considered. Stored data are often not in a form that can be used in research or indeed any data collection; take a patient episode in X-Ray, certain data e.g. date and time, might be recoverable but information such as the report itself is stored as the ‘words’ of the radiologist. Such semi-structured data needs further interpretation for it to be used in informatics-based research.
Furthermore, creating the correct data fields to collect the right data needs considerable insight into the medical processes and traditionally those creating IT systems have had inadequate focus from health professionals (there are exceptions of course) who will be using them. Likewise, I think there may be some ambivalence among health professionals regarding consent for use of data. Traditionally, in health care, consent to store data has been implied and based on the trust between the two parties and there is no culture to routinely support 3rd party researchers, let alone industry.
Indeed, there are no publically agreed rules that facilitate data use for research and most health care institutions quite understandably interpret the law conservatively. Another barrier is that the NHS informatics environment is not usually a place where high performance computing, of the power required for big data research, is available and legitimate data export is prevented because of the consent issue.
Anyway, the data are often hidden away in databases where only a handful of people know where they are and they are not the same people who necessarily understand the meaning! I could go on……. Informatics in hospitals is rather like our electricity supply, it only has prominence when it goes wrong and as such there is a reluctance to make any big change that incurs risk.
So is there a solution? Is the only option massive investment and time that are so hard to come by?
For a start, no one type of professional can sort the whole problem. We need new working practices such that those who know what the data means, where and how it is held, how it can be manipulated and what the important questions are, work together and influence those who have the power to invest. No such informatics departments exist in most institutions.
The new Southampton BRC, which starts in April, has a data science stream but we are by no means complacent as to how we will deliver on this vital area of research. We are bringing together key professionals from health, research, medical informatics and above all NHS facing informatics and are starting with exemplar respiratory and nutrition based projects.
Our ambitions includes the creation of a new ‘medical informatics innovation space’ at the heart of the academic health campus at Southampton General. It would be a place where we co – locate (UHS/FoM), train informaticians and meet collaborators – like the concept of the Pompidou centre, but making informatics visible to the outside!
We realise that data has the potential to transform medical care but it will never do it without new thinking, painstaking groundwork and bold action. So if you are up for the challenge, find out more by getting in touch.
Karen Temple
Professor of Medical Genetics
Head of the Academic Unit of Human Development and Health