The access and use of personal health data for research purposes is increasingly receiving press attention. Undeniably, there are renewed possibilities of conducting health research by processing and cross-referencing large volumes of data through sophisticated algorithms. Thus, there is a growing necessity to access health databases, including genetic data collections.
This necessity is intertwined with the need to develop artificial intelligence in the healthcare system as highlighted by a recent Nuffield Foundation report. It is also intertwined with the expected possibilities of reducing health disparities through the development of digital health technologies and the advancement of precision medicine which seeks to care for patients according to the similarities they share with some subgroups of the population.
As a result, health data has been labelled as a fuel, an asset with value. In the UK, recent publications focused on the commercial value of the NHS's database while in France, articles tackled the need to ensure that large institutional health databases are available for research.
In this context, health database producers receive a lot of attention. Some producers are private operators who collected personal health data for their commercial activities (eg pharmaceutical companies' clinical trials data). However, in France and in the UK, public institutions are at the centre stage since their databases are the largest. Indeed, the UK NHS database and the French SNIIRAM database on care and treatment reimbursement, each hold the personal health data of more than 65 million individuals. If data is labelled as an asset then, without a doubt, this massive volume of data represents a coveted treasure. Therefore, in France and in the UK the same question has been raised: how to make those databases available for research?
To answer this question, the French Parliament is, at the moment, discussing the 2019 Health Bill which includes provisions applicable to health data. Section 11 of the bill seeks to create a large Health Data Hub accessible to both public and private research. This hub will centralise access to different French health databases such as the care and treatment reimbursement database (SNIIRAM), the causes of death database, but also any other databases composed of data collected from the clinical care of patients, when this care has been publicly reimbursed.
Those latter sets of data refer to most GPs' data, the results of prescribed genetic tests, but also the databases of willing private companies collecting data within the healthcare setting in the course of their commercial activities, such as the producers of connected medical devices. Thus, the Health Data Hub has the possibility to gather a wide range of health data.
Moreover, section 12 of the 2019 bill provides for a comprehensive individual online health space for patients to manage their health data. This personal hub will gather, among others, the electronic health record, advanced care directives and any other data that the patient will desire to upload, such as data extracted from 'Internet of Things' devices such as connected activity trackers. Regrettably, at this stage of the parliamentary discussion of the bill, the online space is set to be systematically created for all, without the need for the individual to consent to it.
Pending the vote, this autumn, of the 2019 Health Bill, the French Government has already launched a test version of the Health Data Hub which selected, earlier this year, 10 projects that the hub will assist in their data access applications. At the same time, the Government had initiated, in December 2016, the Medicine French Genomics 2025 Plan which aims to advance high-throughput genomic sequencing. Such an enterprise will result in the collection of very large amounts of genomic data.
The overall ambition is straightforward: boost research based on the cross-referencing of very large and diverse databases. However, this ambition appears to be quite risky. Indeed, even though efforts are made to anonymise datasets, the processing of different databases creates the risk of identifying an individual (for example, by combining an address, an age and a pathology). To limit those risks, and the possibility of discrimination attached to them, any access to the hub will need to comply with the French Data Privacy Act. Furthermore, it is prohibited to use the processing's results to raise insurance premiums. Nevertheless, those safeguards cannot ensure that all patients will be informed of the processing made to some of their data.
In the UK, the Medical Research Council seems to lead the effort on health data research by providing support resources for researchers. In addition, the NHS has set up an opt-out rule for patients to express their refusal of their personal data being used outside clinical care. Such a rule, which amounts to presumed consent, is set to facilitate the use of personal health data for research purposes.
In both France and the UK, some trade-offs are still under negotiation and concern the conditions under which private actors, like large tech companies, will be able to access the publicly funded databases. Under French law, private operators may access the hub for projects of public interest but might be subjected to applications fees.
Ahead of growing international competition to be the leading country for health data research, no doubt that the legal mechanisms adopted in France and in the UK will be the object of scrutiny in order to appraise their effectiveness.