Journal explores impact of primary to secondary care data sharing on quality of care in NHS hospitals

A journal published in npj Digital Medicine explores the landscape of data sharing across recent years in the NHS along with the impact of primary to secondary care data sharing on care quality, finding that data sharing capabilities are associated with reductions in breaches for the A&E four-hour decision time threshold and identifying positive links between these capabilities and patient experiences.

Through the study, the research team analysed 135 acute NHS trusts to characterise the landscape and progression of data sharing networks, testing association of data sharing capabilities with clinical care quality indicators.

The authors report that they identified three stages of infrastructure change in England, noting that prior to 2019, data sharing networks tended to be developed on an individual basis between secondary care trusts and local primary care providers through provision of remote access to primary care data via an EHR or through centralisation of local primary care data into a hospital-controlled HIE data warehouse. By the end of 2020, commissioning groups were procuring the majority of solutions and administering them to trusts and providers within a region. The authors note that this resulted in centralisation into unified local care records, supported by a primary technology vendor, “with increasing availability of hospital record sharing through common data standards”.

The latest stage, 2023 onwards, the authors write that there is “expected unification” of most local care records into consolidated local health and care records exemplars”, which is “expected to result in complete population coverage over England”.

On the impact of data sharing, the authors share that “ability to share data from primary to secondary care was associated with lower A&E breach percentage in three tested years”. Additionally, on the impact on patient experiences of emergency care, the authors looked into the NHS national patient survey and comment that data sharing capabilities “demonstrated significant adjusted association with better patient experiences”. Trusts also reported that “positive interoperability functionality” was linked with positive patient experience.

However, the study states, no association was found between primary to secondary data sharing and “either a standardised hospital mortality index, or incidence of patient safety events”. The study suggests that it is possible “that such mortality indicators, while worth exploring, are too multi-factorial to be considered useful measurements of outcome from broad digital interventions.”

The authors then turn to the implications of new regional data sharing agreements which are “opening the door to secondary uses”. They highlight a number of implications for healthcare systems reaching this level of interoperability, including the need to focus on system usability and workflow to support providers in achieving uniform impact; that impacts of data sharing interventions “may be marginal when competing against other, more prosaic determinants of pathway efficiency”; that investment in interoperability will “see the most gains” when building upon resourcing and staffing requirements; that direct interfacing with clinicians can incentivise better quality data entry; and that robust information governance procedures are essential in mitigating increased privacy risk.

To read the study in full, please click here.

Citation: Zhang, J., Ashrafian, H., Delaney, B. et al. Impact of primary to secondary care data sharing on care quality in NHS England hospitals. npj Digit. Med. 6, 144 (2023). DOI:

In research from elsewhere, a journal published in The Lancet has explored the challenges of representativeness in AI interpretation of medical imaging, and introduced the ‘Medical AI Data for All (MAIDA)’ initiative, described as a “framework for global medical data sharing to address the shortage of public health data and enable rigorous evaluation of AI models across all populations”.