The next step: linking medical records, DNA and genetic information

1. The next step: linking medical records, DNA and genetic information Dr Helen Wallace Director, GeneWatch UK www.genewatch.org

2. Source: Presentation by Tim Hubbard Wellcome Trust Sanger Institute at the Academy of Medical Sciences, 22nd February 2012 Source: Tim Hubbard, Wellcome Trust Sanger Centre at Academy of Medical Sciences, 22nd February 2012

4. This plan is government policy • Human Genomics Strategy Group (HGSG) report: National DNA database needed for personalised medicine drive. The Telegraph. 25th January 2012. http://www.telegraph.co.uk/health/healthnews/9038712/National-DNA-database-needed-for-personalised-medicine-drive.html • “The[HGSG]report has been welcomed by the Secretary of State for Health and the Minister for Universities and Science, who have asked for the recommendations to be implemented through a shared strategic framework”. Letter to Genetics and Ethics from Department of Health, 2012.

5. What is special about DNA? • DNA (contained in blood or saliva) can be analysed to produce a biometric i.e. a permanent link between your body and any stored data • Police DNA databases use a DNA profile based on parts of the (non-coding) sequence (STRs); medical researchers commonly use 100s to 1000s of SNPs (single chemical letters that differ between individuals); or single mutations (rare diseases); whole genomes include everything. • Unlike other biometrics, genetic data can also be used to identify relatives (including non-paternity) • Genetic data includes (some) information about health and/or potential to pass recessive genetic disorders to children

6. Anonymisation of whole genomes linked to medical records is impossible • De-identification can happen by deduction (e.g. combination of information in other databases) • Or, someone can collect your DNA from your beer glass or coffee cup and compare with your sequence (linked to your medical record) in a research database • And/or people working in the NHS may gain non-anonymised access anyway (at the point of care) or individuals may grant subject access online (including enforced subject access to insurers)

7. Collection and processing of DNA • Processing and storage of DNA by unauthorised persons without consent is unlawful under the Human Tissue Act 2004, but the Secretary of State may issue regulations to allow this for research. • The Human Genomics Strategy Group and the Academy of Medical Sciences have argued for amendments to the HTA that would exempt e.g. saliva from its provisions. • Babies’ blood spots are collected with consent for medical tests at birth: millions are being stored under guidelines which may allow research use. • UK Biobank has established a new system of “broad consent”: 500,000 people have opted in without knowing who will get access to their data. (However, only 7% of people invited opted in).

8. Storage and sharing of genomic data • Genomes or genotyping of SNP panels (which is cheaper) will be treated as health data • Attempts to store and share genome data without consent will be the same as for e-health records • Proposed weakening of Data Protection Regulation in EU would allow use by any “researcher” for any research or statistical purpose without consent, and indefinite storage • Role of DNA as biometric will allow (potentially) unique (and unbreakable) identification of any individual and their relatives

9. The WT/HGSG plan involves • Complete removal of people’s right to know who is using their health and/or genomic data for what purpose (including any conflicts-of-interest) as required by the Helsinki Declaration • Construction of a biometric database without consent: allowing tracking and categorisation of every individual and their relatives • A massive reallocation of resources towards collecting and storing data that is mostly not relevant to a person’s care • Abandoning of screening criteria in favour of individual feedback of personalised risk predictions • Significant scope for misuse of data for “personalised marketing” of healthcare products to individuals or their relatives

10. The vision • “…we can now see a future where the doctor will swab a few cells from inside your cheek, put them into a DNA-sequencing machine and a computer will spit out a complete reading of your unique genetic makeup – all 30,000 or so genes that make you who you are. From that, doctors could pinpoint flawed genes and gene products and predict what diseases you are likely to develop years in advance of any symptoms – and how to help you avoid them”. (Tony Blair, ‘Science Matters’ speech, 23 May 2002).

11. Genetic variants are poor predictors of common diseases • “To explore the clinical validity and utility of polygenic information, studies have been conducted for type 2 diabetes, coronary heart disease, breast cancer, prostate cancer, multiple sclerosis, and others.23–30 Collectively, analyses show that the contribution of a combination of multiple alleles at multiple loci will be limited in predicting disease for any given individual.…Furthermore, many of these studies show that adding polygenic information to risk-prediction models, when available, provide no or little additional discrimination (as reflected in analyses of the area under the curve) to current risk-prediction models based on traditional risk factors such as age, body”. Khoury et al. Genetics in Medicine. advance online publication 14 February 2013

12. “The overall conclusion based on these arguments is that the predictive value provided by genetic screening tests for either disease susceptibility or normal variation will be too low to have widespread medical or social application”. Wilkie A (2006) Polygenic inheritance and genetic susceptibility screening. Encyclopedia of Life Sciences. DOI: 10.1002/9780470015902.a0005638. • “Many authors have recently commented on the modest predictive power of the common disease susceptability loci currently emerging. However, here it is suggested that, for most diseases, this would remain the case even if all relevant loci (including rare variants) were ultimately discovered.” Clayton, DG (2009) Prediction and Interaction in Complex Disease Genetics:Experience in Type 1 Diabetes. PLoS Genetics, 5(7): e1000540.

13. Improving health or sinking costs? • Whole population becomes market for whole genome sequencing without any evidence of benefit to health • Investors in whole genome sequencing cash in • Taxpayers pay sunk costs (sequencing, infrastructure) and public provides samples and data (as “duty” to NHS) • Intermediaries (incl. Google) construct risk algorithms, which may be used for personalised marketing • Possible role for technology assessment (NICE) at feedback stage but this is likely to be over-ruled by “right to know” • Screening criteria abandoned (process of weighing up population benefits and harms) • Babies and children sequenced without consent

14. Necessary for public health? • Supporters of “Public Health Genomics” argue that storing genomes and allowing data-mining is necessary for public health • This is an attempt to undermine Art. 8 of the European Convention on Human Rights (right to privacy) • Public health genetics/genomics was originally invented by the eugenicists who went to work for the tobacco industry in the 1950s! Genetic screening of smokers was supposed to identify the 1 in 10 smokers who got lung cancer in advance so the rest could “smoke with impunity”. This false idea was promoted to get funding for the HGP.

15. Risk predictions will not be reliable or useful in most cases • New EU regulations for in vitro diagnostics (IVD) including gene tests and computer algorithms are now with the European Parliament: there is no pre-market assessment of clinical validity or utility (only quality assurance for the sequencing). • This will be used under the US-EU free trade agreement negotiations to undermine moves by the US FDA to regulate gene tests as diagnostics.

16. Whose vision of the NHS? • The WT/HGSG proposal has been promoted as a vision for the NHS since at least 1999 • There has never been a public consultation on the plan • There has never been an assessment of the costs and benefits • There is substantial commercial interest in the plan as a means to expand the market for healthcare products (using “personalised marketing” based on individual risks) • The idea of genetic screening was originally proposed by the tobacco industry (later backed by the food, chemical and nuclear industries) as a means to undermine public health measures

17. Who are the lobbyists? • Sir Mark Walport, new Chief Scientific Advisor, former head of the Wellcome Trust • Professor Sir John Bell (Wellcome/Oxford, Human Genomics Strategy Group, Academy of Medical Sciences, Office for the Strategic Coordination of Health Research) • Sir Richard Sykes (former chair of GSK, Chair of Imperial College Healthcare NHS Trust) • Sir Paul Nurse (head of the Francis Crick Institute and the Royal Society) • And many others…

18. Who are the “researchers”? • Google and its gene test company 23andMe • Private healthcare companies e.g. GE Healthcare (investing in “early health”) • GSK and other big pharma (the drug market is expected to at least double: Burrill & Co. 2009) • Food industry (to sell nutraceuticals, functional foods etc.) • Your insurer, your employer?

19. UK Biobank approved researchers (also includes academic in e.g.China)

20. Other vested interests • Computer hardware/software/infrastructure companies (e.g. Microsoft, IBM, cloud providers) • Genotyping/gene sequencing companies (Life Technologies, Illumina/Oxford Nanopore) • VC investors (may include Wellcome Trust?) • Universities esp. Oxford, Cambridge and Imperial (via spinouts, R&D tax credits, matched PPP infrastructure funding)

21. Implications • Access to medical records plus genomes (minus names) will be “sold” to private companies • Privacy will no longer exist • Every individual and their family can be tracked using their DNA (by the state or others) • Stigma and discrimination (insurers, employers) • Personalised risk assessments will be used to market medication and other products • Doctors replaced with computers • Massive expansion in the drug and “wellness”market; de-skilling in the NHS; commercial control over diagnosis and prognosis • Shift from public health controls on products/pollution/health inequalities to prevention as “personalised marketing”

22. Publicly acceptable? • Only 7% of people approached opted in to UK Biobank: should 93% be presumed to have given their consent unless they actively refuse? • People do not trust the system to keep their data secure or anonymised and, whilst supportive of research, want to be asked for their consent (WT/MRC 2007, CfH 2008) • There was significant public and professional opposition to previous attempts to share data without consent (Clause 152, based on the Thomas-Walport data-sharing report) • Retention of innocent people’s data led to significant loss of public trust in police use of DNA • Loss of public trust could damage legitimate medical research

23. Conclusions • Data-sharing plans for electronic medical records are a step on the way to sharing genomes (and/or genotypes of multiple SNPs) • This amounts to building a DNA database of the whole population by stealth in the NHS