Down’s syndrome to be added to the School Census

A BACKGROUNDER

The DfE announced in January 2024 that it will add yet another sensitive new label to pupil records, adding “Down’s Syndrome” to the termly state School Census from 2025, Schools Week reported.  The school census is one of the 23 data collections that feeds every year, termly or annually, into the National Pupil Database, a melting pot of over 23 million records controlled by the Department for Education and from where identifying pupil data is distributed.

This has raised questions and concerns about why, how this label may be used by this or by future governments, and for what purposes. Down’s syndrome does not offer any explanation of what a child or school needs to support them, such as visual, auditory, speech, social or emotional needs. Those needs are already collected about each child in the termly census on a named basis. Down’s syndrome by contrast is an umbrella label with not much that is universally shared across each of the people who come under it, other than having an extra copy of chromosome 21. Will the new collection be on an aggregated total per school or on an individual pupil basis?  Most concerns would be alleviated if it is statistical only, but not entirely. We have received questions from individuals and discussed with various interested groups. There are number of reasons expanding the school census now, at this point in time, raises these questions and concerns:

1. Why would families of Down’s Syndrome expect this new label will be operationalised by the national Department for Education to any better support their children’s provision than existing data which does not? The government says that they recognise Special educational needs and disabilities (SEND), however is increasingly failing to meet them. There is a vast wealth of deeply detailed information collected about individual children already, including different categories of SEND that are misleading in their binary nature. An education, health and care (EHC) plan is for children and young people aged up to 25 who need more support than is available through special educational needs support and are a required part of the process that must be done before schools are given the funding to be able to provide the additional support time or additional provision to meet those needs. The DfE already has access to huge range of named data, including soaring SENDIST tribunal appeals, but it is clear that the SEND crisis fails to be addressed in any meaningful way, and unclear there is any appetite or intention in government to change this. As described at the SNJ, “There are twice as many statutory SEN plans as there were nine years ago – but the number of appeals has quadrupled in the same timeframe.” Having the label at local or LA level should not mean it would be required as more than statistics on a named basis at national level.

2. How will the label benefit the child from what it will tell the Department for Education or academic researchers that they do not know already? “For example, we know that children with Down’s syndrome should be supported with special educational needs provision as soon as they start school, but in some areas this support doesn’t happen until later on,” says the leading UK education and genomics researcher. It is claimed that research with administrative data is “life changing” but what will change at all for these children? Look at the example of research in children’s mental health care that over a decade has joined up education and health records. “The linked school and mental health data have many potential applications. They can provide detailed information on patient pathways and the extent of inequalities to services.” Yet CAMHS provision is at rock bottom while demand is at an all time high. The result is that desperate parents are trying all sorts of private products, the commercial and charity ‘wellbeing’ market has mushroomed without any health and safety standards, ethics review or oversight, and desperate parents are proposing national adoption of services that large-scale academic research found does not work, like a mindfulness in schools programme that didn’t improve young people’s mental health. (MiSP was tested over eight years, from 2015 to 2023, and involved more than 28,000 students, 650 teachers, 100 schools and 20 million data points).

3. Genomic data linked with pupil records means big risks and potential unintended consequences. The DfE and Government Office for Science have awarded a three-month £50,000 contract to look at the implications of future genomic technologies on the education sector, as reported by Schools Week.

This is not the first foray by the Department for Education into predictive genetics where some were wowed, and a former well-known Spad at the DfE was a big fan. In 2013 Professor Plomin described the role of genetics as “the elephant in the classroom” to the Education Select Committee and the Minister (Michael Gove) said […] “we need to do a bit more research to establish whether the professor is right or not. We do not, at the moment, have any solid international database, let alone a DfE database, that would allow us to establish whether he is correct […] In any case, I am not sure what policy implications it would have for us. We can see from places such as inner London the massive impact on young people you can make if you get the school system right.” Does the Department now intend building that database?

Some researchers who have been pushing for genetic testing to become routine for the population at large are once again being given public platforms to promote their ideas. Others routinely encourage linkage of education data and much more to genomic data all without consent. The same figures often overlap across all these policy spaces. Others argue that the “personalised medicine” pathways that genetic testing often promises is often portrayed in an overly optimistic light by the commercial marketing of the companies who sell it and also in mainstream media. As a result, when the public is surveyed a majority generally believe that the technology can tell you more than other researchers will say it does in reality. Genetic data are complicated, and can be easily misinterpreted or misunderstood outside of specialist knowledge and interpretation. Direct-to-consumer genetic tests are sold as providing answers and the public generally believes that “predictive” genetic scores are informative, but such tests rarely say that most people with apparent “positive” results do not go on to develop the related condition. (Read more https://www.bmj.com/content/367/bmj.l5688) Some neonatal researchers are keen on newborn genomic sequencing for all at birth and already testing it on a pilot basis about ‘developmental disorders’ and new frontiers.

Genetic research can also be used for a wide range of other purposes, and there are concerns that the boundaries between these are too weak. Some commercial antenatal testing companies are based outside the UK, and have more than one business-interest in genetic data. Journalists reported in 2020 that one test used by millions of pregnant women globally was developed by Chinese gene company BGI Group in collaboration with the Chinese military and has been used by the firm to collect DNA data from all over the world for a decade. That specific use-case comes with additional concerns about the future of state security.

This whole ‘genomics’ agenda is set against the backdrop of leading researchers who straddle the fields of genomics and education, who recently published that, “There is scant evidence on whether SEN provision improves health and education outcomes.” Are they implying SEN provision should stop? Ask parents and staff and they will tell what value and difference it makes that may not be measured in ‘outcomes’ data.

And what about the changes to Data Protection law around the corner? Genewatch warns that, “the Government’s Data Protection and Digital Information Bill rips up current safeguards for people’s genetic information (1,2). “This is a short-sighted and extremely dangerous attempt to tear up existing safeguards for people’s DNA and genetic information”, said Dr Helen Wallace, Director of GeneWatch UK, “If passed, these changes will damage people’s trust in health, research and police uses of their DNA, perhaps for generations”.

4. Groups ‘classified’ by shared characteristics have found those characteristics used to discriminate against them disproportionately in history, including disabilities but also by sexual orientation, religion, and ethnicity. 25.5% of pupils recorded as Traveller of Irish heritage ethnic group (1,760 pupils) had SEN support in 2023, and a further 6.1% had an EHC plan (418 pupils). The lowest rate for SEN is in the Chinese ethnic group, 4.9% of pupils in this ethnic group had SEN support (2,891 pupils) and 2.1% had an EHC plan (1,223 pupils). These groups are largely unchanged from 2022 but do these analysis give indicators of ‘why’, of wider experience in the education system or society and does knowing this for a decade or more, show any evidence that there has been systemic change to improve those children’s educational outcomes? Where does accountability for that rest in the system, and how long will we all simply accept the “lack of vital provision and support to meet the identified needs of children and young people, and the development of a genuinely inclusive educational system“? Or might it be normalised as acceptable if (as-is) SEN provision does not improve outcomes? What might that mean for children identified with Down’s syndrome or congenital conditions at birth? This concern would remain regardless of whether the data is at pupil-level or aggregated statistics.

5. These changes come when there are increasingly polarised, often narrowly economic focussed or religiously affiliated / moral views and values driving debate in UK politics. Two significant changes are proposed that attract opposed views; the first, whether abortion should be decriminalised for women, or made more restrictive by reducing the time limit from 24 to 22 weeks, for abortion to be legally be carried out in England, Scotland and Wales which must be approved by two doctors, with the health professionals agreeing that continuing with a pregnancy would be riskier for the physical or mental health of the woman than not. And the second not unrelated, the UK NSC is consulting right now until March 6th, 2024 on whether the quadruple screening test, which is offered to pregnant women who miss the first trimester combined screening test, should be used to screen for Edwards’ syndrome in addition to Down’s syndrome. How might this data be used in future if not to support children’s direct care or education but instead to combine with the LEO Dataset and “predict” a child’s cost to the state by joining up their education, HMRC and DWP ‘longitudinal education outcome records’*? Just because you can do it, does not mean you should. The LEO* dataset is an attempt at alchemy to turn data we already have into something valuable that you wished it could be. It is a fiction to suggest and despite this is used at face value as a predictive indicator of future earnings, for example as Liz Truss suggested in the House of Commons in 2018, saying “The LEO dataset shows you how much you can expect to earn by studying particular courses.” It is profoundly misplaced to look at what was several years ago, as predictor of anything for a different cohort, in different economic circumstances and the similarly-named courses don’t even often compare well and the dataset has a large number of caveats.

6. The personal identifying (not anonymous) individual level data about pupils is given away by the DfE including for commercial re-uses — families are not told this or given any choice to opt-in or out. Who else would find this label about children – which may sometimes be used rightly or wrongly to infer something of the parents’ genetic make-up as well — of commercial or other value? Parents will want an absolute right to have this label shared only for the direct education and care of their child during their education, and ensure there are no surprise secondary or indirect reuses. We’ve gone from ‘only for education purposes’ to give it away to everyone for free and now the DfE is thinking about “what it’s worth” to offer to AI commercial users. There are companies desperate to get hold of lots of longitudinal data that the UK admin datasets might appear to offer (ignoring data quality issues). But the government seems to have forgotten it does not own the data, it is a controller and custodian. Perhaps the reason it is seeking to change Data Protection law right now too.

7. The distribution of identifiable pupil data means families (and learners as children or throughout their adult lives) have no idea where their personal data may have been and has been kept, by whom or for what purposes, or how it could be used. This might be in ways they do not expect eg in deciding access to further education, jobs, or insurance. That’s already the case today, but the new data might provide ways to treat a whole sub-category of pupil data as a homogenous group and in new ways not yet seen. Access by gambling companies to information obtained from the Learner Records Service shows that this risk is not just a theoretical one but reality already. The 2020 ICO Audit found that, the DfE was not providing sufficient information to data subjects as required under the law and that has not changed. Parents will no more understand that there is a National Pupil Database or that their children’s sensitive data is given away from it than when we surveyed in 2018 via Survation. 69% said they did not know it existed.

8. In 2016 nationality and country-of-birth was added to the School Census and serves as a case study to show that what is said and what is true may differ. Ministers told Parliament and parents and teachers that it was for purposes that would benefit the children eg: to allocate more or better resources. In fact it had already been agreed and was written into a data sharing agreement with the Home Office that “Nationality (once collected)” would be shared for purposes specifically to “further the aims of the Hostile Environment” and immigration enforcement action, and to find and match pupils the Home Office was looking for. This was NOT what had been publicly said about the census expansion. (Read the detail here > https://defenddigitalme.org/my-school-records/boycottschoolcensus/) In fact in the BBC, the Department made very clear claims that ‘these data will not be passed to the Home Office.’ A week later we obtained a copy of the data sharing agreement that showed otherwise. The agreement was amended one week later, on October 7th 2016 to remove “Nationality (once collected)” from the agreed data. Other data continued to be handed over, including for punitive purposes listed in the agreement.

9. Police have come to the DfE for access to pupil records in bulk, including in 2019 for the pupil records of an entire school over a four year period, and there is no clear reason why they are not requesting the named pupil data from the school directly but using a more covert process. There is no public oversight of the process or outcomes. The DfE has begun allowing the National Pupil Database to be used as a de facto national ID database. But it is impossible to separate police access from other government departments ‘for criminal investigation’ where the data use register bundles them all together.

10. The DWP has started using the National Pupil Database again as a national ID check for welfare fraud detection — verifying if the information about families or children is genuine. However, the NPD only hold the data of state educated children at named level, and not necessarily every child who is privately educated (Electively Home Educated or in public schools) for example. This is therefore at best another indicator but not definitive dataset to confirm whether or not someone has the child they claim to or is making fraudulent claims. The DWP-DfE also already have a pupil-parent matched database.

How can families trust that data on disabilities in particular will not be used for such purposes, and with the risks of inacurate data or mistakes causing distress, what routes for redress would there be if harm is caused by data misuse? Educational settings send children’s personal data in good faith to the DfE daily, monthly and annually in over twenty-three collections for the purposes of education. And the DfE promised, “the data will only be shared for purposes that benefit children.” These new re-uses are not that.

---

 

SUMMARY

The ever growing scope creep of what was collected strictly for the purpose of promoting the education or well-being of children in England is now already stretched today far beyond that for what is framed as ‘criminal’ investigations; by police, by the Home Office, and for fraud detection, for commercial re-use and wide distribution to third-parties in research bodies and more. Already most of these are nothing to do with the aims of education and do nothing to uphold the rights of the child to reach their full potential and full flourishing into adulthood, in a manner consistent with the child’s human dignity. Throw in genetic research on educational ‘performance’ and academic advice that SEND provision does not improve outcomes, plus policy views that genetics outweighs teaching, backed by artificial measures of human value and costs tied to educational outcomes and tax and welfare costs to HMRC and DWP and we are far beyond what it means to discriminate using data — we could speculate and ask, without safeguards or independent oversight what prevents us heading back to a past where those in power thinking about economics, education and genetics together culminated in eugenics? Or what would the checks and balances be that prevent us fast forwarding towards the societal norms of Gattaca?

As The Special Needs Jungle has written, “on its own, the data produced by the school census:

• does not provide any information about the educational progress this group of pupils are making
• does not help anyone better understand their experiences of the education system or how included they’ve been.
• does not change the environment in which pupils who have Down’s syndrome are being educated
• and it certainly does not lead automatically to greater support.”

We will look forward to seeing more detail about what this label will be used for exactly and nothing more. We will want to see any plan to tell children or families answers to questions like these, not least how they can exercise their right to object to its collection which should be a choice with affirmative consent for such sensitive data. And that exercise must start well before it begins.