Genomic data: to share is to care
7 October 2014
The genomic medicine revolution can only truly be realised within the NHS if all those involved in delivering genetic testing services share their data and knowledge freely and expansively with each other. Currently, vital information that could lead to the diagnosis of a patient with a rare genetic condition may exist within one molecular genetics laboratory, but not necessarily be available to all others offering the same diagnostic testing. This incomplete distribution of information around the genetic testing network in the UK means that patients’ chances of getting a diagnosis may well come down to the serendipitous choice of where their test is performed.
Our latest briefing note, Sharing clinical genomic data for better diagnosis, describes the barriers that have been preventing the widespread sharing of genomic data between NHS clinical genetic testing services. These include:
- a lack of incentives for sharing
- the existence of disincentives
- insufficient infrastructure
- ethical, legal and privacy concerns
Historically, different NHS genetics laboratories have focused on delivering genetics tests in distinct disease domains. They have therefore often not needed to share their data with one another, as they have tended to be the sole consumers of that information in the UK. This situation is rapidly changing as the nature of genetic testing expands from sequencing small sets of genes to entire genomes and as new providers of genomic medicine emerge.
One of the major disincentives for data sharing is a consequence of the competitive nature of NHS genetic testing services. Laboratories are often in direct competition with each other to attract NHS test referrals. For a laboratory to surrender hard-earned knowledge used in test interpretation without guarantee that their competitors will reciprocate would put them at a distinct disadvantage. Despite this situation, data sharing is encouraged by the clinical genetic scientist peer network (ACGS). Sadly for patients, this encouragement alone has not been sufficient to achieve compliance with existing data sharing guidelines set. While this may have been tolerable in the past, as more complex genome wide testing is undertaken, and interpretation becomes correspondingly more complex, lack of access to relevant data from other patients and populations will become a significant limiting factor in the utility of genomic medicine.
Hence, for adequate data exchange to occur between all laboratories, it must be mandated. This might for example involve making compulsory data submission an essential criteria for laboratories to obtain accreditation, but whatever mechanism is determined to be most effective, policy makers will need to take the lead in ensuring the principle is rigorously applied and adhered to.
Different data – same issues
The concept and value of data sharing is not something unique to the human genomics. Effective surveillance of infectious disease outbreaks in the genomic era will also require sequence data on infectious disease microbes to be aggregated promptly and shared with relevant public health authorities. In fact, many of the benefits of sharing (and risks of failure to share) invariably apply across a number of sectors, health-related and otherwise, as do the wider considerations when devising a data sharing strategy:
How to create a solution that is:
- Useful to end users - Any mandate to share data must be underpinned by adequate infrastructure and sustainable support to enable data collation and access. For human genomic data within the NHS, we believe a purpose-built NHS data repository is needed to achieve this. In the future, for this resource to adequately support, enable, and expedite genomic analysis, it should ideally contain not only those genomic variants suspected of underlying disease, but all other variants identified, and ultimately the entire sequenced genomic data itself, as well as clinical and phenotypic data on the patient.
- Acceptable to those whose data will be deposited - Given the scale, detail and personal nature of this information, privacy and patient re-identification concerns must be addressed appropriately. For example, clear guidance will be needed on which elements of the data can be safely shared in the public domains, and what should remain within a managed access network. Ultimately, collection of patient data and its downstream use must be consistent and compliant with the consent taken at time of testing.
- Not burdensome for contributors - Given the sheer scale and detail of data being discussed, ease of data deposition and ease of data mining will be critical to the success of a centralised repository. Automated data deposition would be the ideal resolution.
How to pay and who should pay?
The need to create new and/or revise existing infrastructure will require investment. Given current restraints on the public purse, the subject of funding new resources may seem contentious. Yet if the ambition is to bring genomic-medicine to the NHS, then the NHS will inevitably require funding to realise this ambition. The predicament of who should pay is a complex one. The benefits and savings arising from improved genomic data sharing will not be restricted to genetics alone. Patients with underlying genetic disorders are cared for within other clinical specialties but faster, accurate genetic diagnosis will help inform their management and eliminate the need for other, often expensive, diagnostic tests. The challenge, as always, is how to ensure a balanced allocation of both the costs and benefits (in terms of efficiencies) between the different parts of the health system involved in delivering genomic medicine.
The PHG Foundation is keen to see effective application of ‘big-data’ in genomics and me dicine to deliver widespread patient benefit in a responsible way, cognisant of the sensitivities surrounding data privacy. Alongside the clinical genome analysis project, the Realising Genomics project is undertaking an in-depth analysis of the ethical, legal and social challenges raised by the use of whole genome and whole exome technologies, and their final report will be available on our website soon.