UK Biobank - why won't GPs share data?

Risk averse doctors are harming a world class health database

Imagine a huge database containing the health records of half a million people who had agreed to undergo all sorts of tests and be tracked over time. That would be an invaluable resource for researchers trying to discover new ways of treating life-threatening diseases or planning how to confront a pandemic. Sounds great doesn’t it? Well it’s not a fantasy - it exists and it’s called UK Biobank.

When I wrote at the beginning of this month of the challenges in getting this kind of medical data sharing project off the ground I knew vaguely about the Biobank but had no idea of its scale. What I also didn’t know was that this was another example of the endless struggles to get hold of primary care data for use in research - and one all the more frustrating because those half a million participants have all given explicit consent for that to happen.

“UK Biobank is absolutely unique,” its director Professor Sir Rory Collins told me. "There isn't anything on this scale, with this depth, with this duration, that is so readily accessible to researchers around the world.” The participants, aged between 40 and 69, were recruited between 2006 and 2010 with the aid of the NHS. Since then the data store has filled up with their hospital records, regular blood, saliva and urine tests, and their answers to questionnaires about their lifestyle.

What has been missing for much of the period since the launch is their GP records. They have been caught in a bureaucratic tangle with fears about data protection making everyone involved perhaps over cautious. The records are actually held by two technology providers EMIS and TPP - but when the Biobank came calling they pointed out that, under the law, the data controllers are the individual GPs: “In order for EMIS and TPP to make the primary care data available to UK Biobank each GP has to give their explicit consent,” explains Professor Collins.

Years were spent in meetings with the Royal College of General Practitioners and the BMA, letters were sent to GPs explaining that their patients had given consent, and that the Information Commissioner had confirmed that the data sharing scheme complied with the law. Despite all of this, most GPs wouldn’t play ball, even though they were not being asked to do anything complicated: “We never got above 20% of the GPs even responding and pushing the button - I mean literally pushing the button on their supplier’s system to say yes, the data can be provided to UK Biobank.” Sir Rory accepts that GPs are very busy but suspects that it is fear of getting the wrong side of data laws which deterred them from pushing that button.

Then along came the pandemic and suddenly with one bound the data was freed up. Using what’s known as a COPI (Control of Patient Information) notice the Health Secretary Matt Hancock simply instructed EMIS and TPP, along with other organisations, to share the GP records with UK Biobank for the purpose of Covid research.

What happened next is seen by Professor Collins and his colleagues as proof of how useful primary care data can be.  Over 800 groups of researchers accessed the Biobank data, leading to 200 published studies, many focused on identifying the lifestyle and genetic factors behind severe cases of Covid-19 and highlighting its prevalence amongst different ethnic groups. These kind of studies are far more difficult to do without GP records - relying just on hospital data means you miss out on all sorts of interactions between a patient and the NHS, notably in the area of mental health.

The Biobank could also be a great way of monitoring the arrival and progression in its participants of conditions such as Parkinson’s, dementia and depression but again only access to the primary care data will give an accurate picture.

Now that is under threat with the expiry of the COPI notice which took the responsibility for allowing the sharing of the data out of the hands of GPs. Sir Rory Collins thinks that if the Secretary of State tells the family doctors that it’s their decision again, they will revert to type:

“The GPs are very anxious about data protection. They can only get into trouble by making the data available and so apart from being very busy, their default is better not to do anything. There's no reward in it.”

The Royal College of General Practitioners told me it did recognise the crucial role that GP data has to play in research and gave me this statement:

“There may be a number of reasons why practices haven’t signed up. We know that GPs take their data controller responsibilities very seriously and this is important given the sensitivity of much of the data we collect in general practice. It may also be that extremely busy practices don’t know about Biobank or that they are unclear about what it involves. The College is continuing to have conversations with UK Biobank on the next steps to help identify the most appropriate mechanism to ensure important GP patient data, when consented by a patient, can be easily shared for research purposes.”

I wrote earlier this week about the slow pace of innovation in the NHS and how part of the problem was the difficulty of building a coalition behind good ideas across the many disparate parts of the health service. Well, here’s a prime example, a world class health database which could play a valuable role improving treatment and making the NHS more efficient but is hamstrung by a failure in collaboration. Surely this is one problem that shouldn’t be hard to sort out?