Data can save lives – so how can we make sure it actually does?
In the 21st century, the opportunities to use the best of human knowledge are growing every day. The internet contains 4.84Bn (and growing...) indexed pages and it is estimated Google has only indexed 0.004% of the total internet. Everyday, we connect data seamlessly to generate new insights to solve old problems. If someone has learned something in Japan, once on the internet, another person could apply that knowledge for something else in Spain.
So WHY can’t we use this approach to improve our health and treat illnesses more reliably, cheaply and more quickly? Everyday, patients in the NHS give up opportunities to find the right lifesaving treatment for their cancer, based on data which might exist from a trial. Or perhaps find the right combination of medicines and treatments to slow down or reverse diabetes in Manchester, which might have been tried in South London.
When it comes to our health, which is perhaps our most important asset, whilst a lot of data is collected it sits in individual silos. For instance, an audit showed that a single NHS Hospital Trust in London had 200 separate databases, none of which talked to each other. And I am sure that only a handful of different people are able to connect to each one! If data can save lives, then this example must surely be costing lives as must many others.
So what do we do? How can we use our data more effectively? I believe the answers are here today but that no single stakeholder has taken ‘ownership’ of the collective problems to create a viable solution. To explore this, lets look at the problems.
The first problem is one of trust. Today people trust others with their employment backgrounds on Linkedin and personal information on Facebook and Instagram for example. Yet when it comes to our health we’re nervous of handing over information. One of the solutions to this is to create a social benefit data partnership, which, like Facebook or Big White Wall, allows patients to share information for their own and others benefit to save lives and improve health. Such a partnership MUST be in sole existence for the good of the patient, and should be able to work with multiple database holders to pool data for common benefits.
The second problem is of permissions management. It is essential that patients’ control and ownership of data is respected. For example in France patients are expected to have their own patient records, rather than the GP.
In a perfect world, each of the member databases that become part of a data partnership, should log what level of permission the patient has given, against each patient. Today, quite often, this permissions information is incomplete and exists in filing cabinets rather than as part of the patient record. Addressing this simple issue would enable some patients to give permissions for research use, some to give permission for only clinical use and for the right record to only available for ONLY the use specified.
The third problem is one of quality. Without addressing this, it is possible the data we are using will create inaccurate insights. Whilst data quality has many dimensions, it can, and should be rated, for each database in the partnership. This doesn’t mean that lower quality or incomplete data can’t be used, it certainly can. If there is enough of it, it might not be used for validating a diagnostic test, but with enough volume it might be usable for forming clinical hypothesis. The key is that quality must be rated, and there are already tools and organisations that do this.
The next problem is getting the applications to access and use the data. In the consumer world we have Google, Yahoo, Bing to search the internet, and we have ‘Apps’ like Google Maps which might interface with raw data which we’ll never see directly. Equally, a data partnership, needs to allow multiple organisations who follow the right rules, respecting patients’ permissions, to develop apps and search engines that allow proper utilisation of the data which sits within it. This would allow apps to be developed to help clinicians to reliably check whether a patient is likely to respond to a steroid inhaler, or whether it is likely to be toxic based on hundreds or millions of records. Search engines might be used by doctors or researchers to understand possible susceptibility of a patient to a virus, or to better understand which combination of cancer medicines might treat a patient’s pancreatic cancer right first time.
However, all of this is useless if, in the end, it is impossible for doctors and researchers to get access to the apps and search engines. An interesting approach in Scotland has been used to collate data and then make it accessible in regional centres. Without using the term, effectively, the Scottish are ‘accrediting’ a small number of government centres to be access points. Extending this idea of accredited analytics centres or organisations could be very powerful. Few people would want data to fall into the hands of the wrong people, who could abuse it. However, through accreditation of organisations and centres, this data, and the route to using it could be controlled rather than constrained as we do today. As pointed out at the beginning of this article, keeping useful knowledge assets in the hands of very few, unwittingly harms patients, costs lives and quality of health, which could otherwise be enjoyed. As a reformer, I am passionate about this, but as a patient, this failure is akin to hubris.
Finally, and perhaps most controversially, such a partnership must have a role which allows database holders, to be rewarded through royalties for data that is used. Why? The answer is simple. I often say that “if people had to build our homes for free, then we would all be living in mud huts.” Just like our homes, creating high quality datasets, which can be used to improve lives, does not come for free. Maintaining the quality of data, implementing better tools to capture more accurate data, none of these activities can be delivered by people who can’t be paid for the activity, else, who will pay their mortgage? You might think this can be legislated. Wrong! It will still come at a cost, and even if the government paid NHS IT departments to do it, still some will do it to ‘tick a box’. Rather it is better to create a way to enable the people who see the opportunity for patients and themselves. And if royalties management sounds far fetched, then the music industry has been doing it for years, and with online payment methods and algorithms to calculate value, this is possible.
Any approach such as this will need to be gradual and incremental rather than a big bang. Putting together all datasets irrespective of quality, rules, with no apps, without standards for accreditation is unlikely to be successful. But like the growth of the internet was unpredictable, which started as a small collaboration for research, today, has become an incredible interconnected network. And as late as 1995 cybersecurity expert, astronomer and author Clifford Stoll, predicted that nobody would buy anything from the internet, which has clearly turned out to be wrong. So the key is to start small, but start in a way that works and can be scaled-up, iteratively getting better. The aim would be that in five or more years, this network will have grown beyond all expectation, so that results found in one part of the country, are immediately factored into a clinical app improving treatments on the other side of the country. Equally this data will lead to yield insights we never expected, lead to the discovery of new treatments and new technologies – and ultimately the data will save lives.
Anoop Maini, Empower: Data4Health's spokesman on health data
Anoop is a social reformer and healthcare entrepreneur, with a background in social impact. He recently wrote the first plan, accepted by Innovate UK to set up the Precision Medicine Catapult and its UK Centres of Excellence. Before this he worked with cities across England to establish an approach and methodology to drive bottom-up change in local health systems. Prior to this, Anoop was Senior Policy Advisor to Monitor, the sector regulator for the NHS in England, where he led NHS Futures; the programme across Monitor, NHS England and the NHS Trust Development Authority on the routes to innovation.