A journal article published in The Lancet Digital Health has set out recommendations to support data infrastructure transformation, with the authors also producing a website intended to “promote transparency and showcase NHS data assets” following a mapping exercise involving “all electronic patient data flows across England, from providers to more than 460 subsequent academic, commercial and public data consumers”.
The authors begin by noting that although the importance of big health data is widely recognised, there are numerous known challenges including a complex data-sharing landscape, privacy and analysis capabilities. Through their mapping exercise, the authors aimed to follow recommendations in the NHS strategy review by Goldacre and Morley for mapping bulk data flows, in order to generate an understanding of the privacy risks, capabilities and positioning of secure data environments.
After completing their exercise, the authors noted that the current landscape indicates failures in transparency and privacy that risks comprising public trust. The authors acknowledge that data usage “most often occurs two or three interactions down a chain”, which places patients a “considerable distance” from data usage. They also found that reporting data use was often incomplete or lacked specificity, “which risks violation of the no surprises principle within data protection legislation”. The study also noted the risks around data breaches, highlighting that they found physical data transfers outside of secure data environments to be “the majority occurrence”.
Another finding was that “data volume hides insufficient diversity in information and population”, with the authors pointing out that an imbalanced landscape throws a question mark over the concept of effective and equitable data use.
In addition, the study highlights value implications across data flow chains. For the largest databases, the authors point out that costs are borne by researchers or companies wishing to access the data, with multi-study licensing fees between £75,000 and £330,000. “These costs cover infrastructure and administration, but could also produce net income, particularly for commercial brokers. For consumers, data access might additionally support revenue-generating services,” the authors state. “By contrast, value return to patients, care providers, and the NHS is a minority proportion of this landscape.”
As such, the authors put forward six recommendations “pertinent to the NHS and national data initiatives in England”.
Firstly, they say, “public transparency should not require investigative discovery”, requiring better reporting of distal data uses to increase transparency and safeguard against data breaches.
Secondly, “opt-out conditions should be set at the level of distribution to types of consumers, rather than at extraction”.
The third recommendation is that “improving and widening usage of existing infrastructure is a priority”, which could enable linked assets in NHS Digital and OpenSAFELY to be used for general purposes.
Fourth, the authors recommend that “new data infrastructure must focus on expanding capabilities for extracting untapped secondary care electronic health record data, and increasing multi-modal data availability”.
As the fifth recommendation, the authors suggest that infrastructure should focus on intervention rather than on analysis capabilities, in order to increase value return to patients and providers.
The final recommendation is that “monetary value transfer across the entire data flow landscape should be assessed to quantify value returned to the health-care system”.
As a result of their research, the authors provide an interactive public-facing dashboard designed to “visualise data use and to assist with discovery of NHS real-world data assets by the global research community.” It can be found here.
Citation: Zhang, J., Morley, J., Gallifant, J., Oddy, C., Teo, J. T., Ashrafian, H,. Mapping and evaluating national data flows: transparency, privacy, and guiding infrastructural transformation. The Lancet Digital Health. 5, 10 (2023) https://doi.org/10.1016/S2589-7500(23)00157-7.