ECHILD - a positive story about patient data?

The NHS has a mountain of hugely valuable patient data but every major attempt to use it for research and planning is beset by rows over privacy and ends in abject failure. Correct? That is certainly what I thought until I heard about ECHILD. It’s a database linking the school and hospital records of just about every child in England and so far - touch wood, cross fingers, hold your breath - it has attracted very little controversy.

I heard about this miracle when a civil servant got in touch after hearing me speak about the importance of data sharing during a panel discussion about technology and ethics at Westminster Abbey. He first told me I needed to know about the National Pupil Database , a major project which itself has been under the radar, then introduced me to Professor Ruth Gilbert, who’s one of the leaders of the project to bring together the school records and hospital data to create ECHILD.

Professor Gilbert, a paediatrician and academic at UCL, is co-director of the NIHR Children and Families Policy Research Unit. She explained that she’d long been interested in linking health data to education and social care records for her research.

The National Pupil Database started in 1996 and has been fully operational since 2002. It contains records for all children in state schools in England, and all those that take public exams, such as GCSEs, so capturing most privately educated children too. It also captures their interactions with the care system

It is combined with what are called hospital episode statistics, that is all admissions of children to hospital, all A&E attendance, and all outpatients visits - and crucially it includes births, so again it captures data about nearly all children in England.

“This is high quality, rich data,”says Ruth Gilbert. “It really is a complete picture, from conception - because we can follow their mothers in the health data so we can see the health of their mothers before they were conceived - through to their 20s.”

She says the real value is in the longitudinal nature of the combined databases - the fact that you can see what is happening to children at school and at hospital over time. “So you may have a child who has behavioural problems at two or three, recognised by the school in their special educational needs, and we might only see them in the hospital episode statistics at 12.”

She gives an example of where the data might help researchers - understanding the impact of the fall in the percentage of children getting some kind of special educational needs support at school from 41% to 29% between 2004 and 2014. “So who is in that 12% who don’t get it now? Is it the poorer kids? Is it the English as a Second Language kids? And then what is the impact on their health?”

But in getting the ECHILD programme up and running Professor Gilbert and her colleagues have had to tread very carefully around issues of privacy. They consulted young people and saw pushback against certain possible uses of the database: “They were certainly not happy for schools to be able to see the health data. The other way - specialists, paediatricians, being able to know a bit about their school records - was not so contentious.”

By the time it reaches analysts, who are allowed to see it only in what is called a trusted research environment, the data is all anonymised. But if someone really wanted to track down a child and had a few key pieces of information - say that they had a rare disease - then it would not be impossible to piece together their identity.

So it is vital, says Ruth Gilbert, that nothing is done to undermine public trust in the ECHILD programme. There are obvious opportunities for the data to be used for commercial purposes which could be a big earner for the NHS - allowing a tutoring firm to advertise to the parents of children with autism for instance. But that, she says, could cause huge reputational damage to a vital project: “You're losing trust, for these huge benefits for the population, for sensible evidence based policies, for better planning, and in quality improvement of services.”

I took part this week in a panel discussion organised by the Health Communications Association on building public trust in patient data. I made the point that the first time most people hear about projects like the GP data sharing scheme GPDPR is when there is a row over privacy and, instead of hearing of the potential benefits, they read headlines talking of an “NHS data grab”.

GPDPR has been on hold for more than a year and even if it is revived it seems likely that millions will opt out of having their data shared with researchers. That in turn means that certain communities may be underrepresented in the data, making it harder to frame policies which work in their interest.

ECHILD, probably the biggest single health records database made available for use by UK researchers, has attracted remarkably little attention so far. Ruth Gilbert hopes the strict, sometimes downright annoying restrictions on how researchers can access the data will help the project continue to avoid controversy. Let’s hope she is right.