Understanding real world evidence and the GPDPR debate
On 12 May 2021, NHS Digital announced the General Practice Data for Planning and Research (GPDPR) system following legal instruction from the then Secretary of State for Health and Social Care, Matt Hancock.
The GPDPR system is designed to collate de-identified data from all GP practices in England to support research and service planning. Access to GP data on this scale holds great potential for improving our understanding of the distribution and determinants of disease, as well as informing treatment and public health policy. However, the rollout was paused in July 2021 until concerns raised by advocacy groups, experts in data/cyber security, the British Medical Association and the wider public had been addressed.
Real world evidence
Real world evidence (RWE) is an umbrella term for outcomes generated through the analysis of routinely collected health data, such as electronic health records, prescription data, disease registries and wearables. Randomised controlled trials are still the gold standard for evaluating interventions in medical/health research, although they have important limitations, including high costs, considerable resource requirements and long durations, and they often exclude patients with (multiple) chronic conditions. Furthermore, it is not always feasible or ethical to randomise trial participants to certain interventions (eg smoking vs not smoking). RWE is particularly useful in bridging the knowledge gap between clinical trials and practice, supporting the refinement of guidance as more data becomes available from routine clinical care.
Monitoring the rollout of Covid-19 vaccines among the public is a prominent example of RWE in action. National and international data has been combined to estimate the effectiveness for each type of vaccine and has enabled safety information to be updated.
Real world data, the resource from which RWE is generated, is also used to describe patterns of disease, identify and estimate the effect of disease risk factors (eg socioeconomic status, age and sex), and determine areas of greatest unmet need. This information is utilised by various professionals including:
- healthcare decision makers for allocating resources;
- researchers for refining the design of trials; and
- pharmaceutical companies for defining therapeutic targets.
Using data to improve patient care
The importance of research-accessible linked health data (ie data that comes from multiple sources to provide greater scope and depth) is highlighted by the Goldacre Review, commissioned by the Department for Health and Social Care, and numerous efforts are underway in the UK to explore how NHS data can be utilised to further improve patient care. For example, NHS DigiTrials was established to offer data services to support clinical trials in order to improve study efficiency and decrease the burden of research on participants by reducing the number of face-to-face visits.
Routinely collected health data has been widely used during the Covid-19 pandemic including in the RECOVERY trial, which reported that dexamethasone improved survival in hospitalised patients; this finding is estimated to have saved more than one million lives globally. However, the journey from real world data to evidence presents a number of non-trivial challenges and the debate that followed the announcement of GPDPR highlights many sensitive issues around consent, access and use of data. Although several barriers to GPDPR have been addressed, outstanding matters persist. A key component is a trusted research environment, where approved researchers can work securely within agreed parameters that do not compromise patient privacy. The OpenSAFELY platform, which has been used during the pandemic to explore factors associated with Covid-related outcomes, potentially provides an operational blueprint. There is also a need for medical researchers to conduct their work to appropriate standards to enable the strength of RWE to be evaluated.
The benefits that may arise from the GPDPR resource are considerable but establishing confidence and trust with the public and clinicians is vital for ensuring this potential can realise improvements in patient care and to do so equitably. Many members of the public are not aware of GPDPR and those that are often do not feel fully informed. Any plans to make health data available for research require appropriate planning and consultation if not to suffer the same fate as the failed care.data scheme. Furthermore, all data sources used to generate RWE need to be critically appraised for their strengths and weaknesses, and outcomes need to be reported to robust standards that are comparable with clinical trials.
Routinely collected data has the potential to make improvements to human health but its use must be ethical and within well-regulated frameworks - trust takes years to build, seconds to break, and forever to repair.