Data are the lifeblood of health research, and the UK government is claiming that data collected in the course of NHS clinical care are available to reputable researchers for the purposes of improving health and healthcare. However, the reality is rather different, and it is becoming increasingly difficult to obtain research data.
Anonymised patient level Hospital Episode Statistics (HES) data on inpatient admissions have been available for research and other secondary uses for many years, and their use has led to major improvements in quality, from identifying higher mortality at weekends to sounding the alarm about institutional failures such as those at Mid Staffordshire Foundation Trust. This was achieved with reasonable information security standards and a proportionate approach to the risks and benefits by the Information Centre for Health and Social Care, as it was called then.
However the advent of care.data in 2014 and its plan to extract clinical data from all primary care electronic health records (EHRs) was opposed by some groups and set loose the English media, heavily armed with Freedom of Information requests. The Daily Mail’s investigations (“NHS to carry on selling patients’ medical data to insurance firms despite history of blunders…”) led to an internal inquiry by Sir Nick Partridge of the Information Centre’s successor, the Health and Social Care Information Centre. The Partridge review recommended that the HSCIC develops “one clear, simple, efficient, and transparent process for the management of all data releases,” and that it “implements a robust audit function, which will enable ongoing scrutiny of how data are being used, stored, and deleted by those receiving it.” This seems very reasonable, but the implementation process has led to a virtual halt in the use of NHS clinical information for research.
We submitted a request for HES data back in January, to replace our previous datasets which we had been required to destroy after they reached their expiry date. We attached the required documentation—including information security policies and justifications that had been acceptable in the past—but our request was rejected for a list of reasons, including requests for copies of documents already sent, letters from research funders, and more detail on purposes and “fair processing.”
We discovered our experience was universal. Other research teams that had achieved NHS Information Governance Toolkit accreditation, the rigorous standard required for NHS trusts, who of course use patient-identifiable data, had had their requests turned down as well. In a teleconference last week organised by the Royal Statistical Society, another researcher described responding to 24 separate requests from HSCIC regarding a single data extraction. Others related how HSCIC was questioning the need for research using HES data which had already been funded by National Institute for Health Research, and how NIHR was threatening to withdraw funding because of delays in data extraction, putting researchers’ jobs at risk.
To be fair, HSCIC participated in the teleconference and is now aware of what the adverse impacts of its new policies are. We can only hope that a more balanced approach to research data access eventuates. Any comments from other researchers experiencing similar problems would be welcome.
Michael Soljak is a clinical research fellow at the department of primary care and public health at Imperial College London.
Competing interests: None declared.