Open access Original research BMJ Open: first published as 10.1136/bmjopen-2019-031637 on 5 December 2019. Downloaded from Improving data sharing between acute hospitals in England: an overview of health record system distribution and retrospective observational analysis of inter- hospital transitions of care Leigh R Warren ,1 Jonathan Clarke,1,2,3,4 Sonal Arora,1 Ara Darzi1 To cite: Warren LR, Clarke J, ABSTRACT Strengths and limitations of this study Arora S, et al. Improving data Objectives To determine the frequency of use and sharing between acute hospitals spatial distribution of health record systems in the English ► This study used a large national-level administrative in England: an overview of National Health Service (NHS). To quantify transitions of health record system distribution data set to identify transitions of care between acute care between acute hospital trusts and health record and retrospective observational National Health Service (NHS) hospital trusts. systems to guide improvements to data sharing and analysis of inter- hospital ► The use of recent administrative data ensures that interoperability. transitions of care. BMJ Open the findings from this study are relevant to current Design Retrospective observational study using Hospital 2019;9:e031637. doi:10.1136/ practice and policy. bmjopen-2019-031637 Episode Statistics. ► Health record system information relating to each Setting Acute hospital trusts in the NHS in England. Protected by copyright. Prepublication history for acute NHS hospital trust in England facilitated a ► Participants All adult patients resident in England that this paper is available online. comprehensive overview of health record system had one or more inpatient, outpatient or accident and To view these files, please visit distribution at this level of care. emergency encounters at acute NHS hospital trusts the journal online (http:// dx. doi. ► Analysis of care transitions was limited to acute between April 2017 and April 2018. org/ 10. 1136/ bmjopen- 2019- hospital- level care. 031637). Primary and secondary outcome measures Frequency ► The use and distribution of electronic health record of use and spatial distribution of health record systems. (EHR) systems is dynamic and, at a national-level, Received 14 May 2019 Frequency and spatial distribution of transitions of care changes in EHR systems are frequent. Revised 02 October 2019 between hospital trusts and health record systems. http://bmjopen.bmj.com/ Accepted 19 November 2019 Results 21 286 873 patients were involved in 121 351 837 encounters at 152 included trusts. 117 (77.0%) hospital trusts were using electronic health records (EHR). There care between multiple primary, secondary and was limited regional alignment of EHR systems. On tertiary care settings.1–4 To make informed 11 017 767 (9.1%) occasions, patients attended a hospital and safe decisions for patients negotiating using a different health record system to their previous this complex system, clinicians need the hospital attendance. 15 736 863 (73.9%) patients had two right information about the right patient in or more encounters with the included trusts and 3 931 255 5 (25.0%) of those attended two or more trusts. Over half the right place at the right time. However, on February 10, 2020 at Imperial College London Library. © Author(s) (or their (53.6%) of these patients had encounters shared between contemporaneous, accurate patient informa- employer(s)) 2019. Re- use just 20 pairs of hospitals. Only two of these pairs of trusts tion is often not available when it is required. permitted under CC BY. used the same EHR system. This results in ineffective care, duplication of Published by BMJ. 6 1 Conclusions Each year, millions of patients in England tests and medical errors. For over a decade, Department of Surgery and attend two or more different hospital trusts. Most of the Cancer, Imperial College London, the development and use of electronic health pairs of trusts that commonly share patients do not use the London, UK records (EHR) has been suggested as a key 2Centre for Health Policy, same record systems. This research highlights significant solution to the rising demands on healthcare Imperial College London, barriers to inter- hospital data sharing and interoperability. systems.7 Compared with paper records, which Findings from this study can be used to improve electronic London, UK have been the mainstay of medical record 3Centre for Mathematics of health record system coordination and develop targeted keeping for centuries, electronic health Precision Healthcare, Imperial approaches to improve interoperability. The methods used 8 College London, London, UK in this study could be used in other healthcare systems records have several potential advantages. 4 Department of Biostatistics, that face the same interoperability challenges. Central to this is the ability to more easily Harvard University, Boston, share digital records with other stakeholders United States involved in caring for an individual patient. Correspondence to INTRODUCTION Around the world, healthcare policymakers Dr Leigh R Warren; Many patients experience a fragmented have attempted to improve the adoption and leigh. warren@ imperial. ac. uk healthcare journey that involves transitions of use of EHRs through both incentivisation and Warren LR, et al. BMJ Open 2019;9:e031637. doi:10.1136/bmjopen-2019-031637 1 Open access BMJ Open: first published as 10.1136/bmjopen-2019-031637 on 5 December 2019. Downloaded from legislation.7 9–11 Although EHR usage in both community- We included organisations listed by NHS England as level and hospital- level care has dramatically increased,7 12 acute trusts20 in November 2018. Acute trusts are defined implementation, integration and interoperability have by NHS England, and in this study, as those providing been challenging.6 7 9 11 12 Open EHR standards such as acute and emergency care to patients.20 This includes Fast Health Interoperability Resources and Application inpatient, outpatient and accident and emergency (A&E) Programming Interfaces (API) have improved interop- care. Trusts often provide hospital services in one or more erability in recent years,13 but progress towards true hospital sites administered by that organisation.21 Acute semantic interoperability between EHR systems has been trusts include several types of organisations including slow. regional or national specialised care centres, general In the National Health Service (NHS) in England, a hospitals and teaching hospitals attached to universities. convoluted interplay of policy and technology changes These trusts are most relevant to the problem of acute over the last two decades has resulted in a complex care inter- organisational data sharing as they provide ecosystem of patient health records.7 11 13–15 Several active most of the care for patients outside a primary care policies and programmes are attempting to better link up setting. Organisational change during the study period records at both a regional and national level16 17 but data due to closure, merging or separation of providers were 7 sharing and interoperability challenges remain. Many managed by treating those organisations as a single patients still have their records fragmented between provider across the whole period. multiple systems that are unable to effectively share For the patient- level analysis, we included HES data patient information. Data sharing between healthcare for all adult patients resident in England that had one or organisations that use the same EHR system remains more more inpatient, outpatient or A&E encounters at acute 18 achievable than those that use different systems. There NHS hospitals in England between April 2017 and April are several examples of local and regional alignment of 2018. EHR systems that aim to capitalise on this.7 19 Despite these positive examples of inter-hospital data sharing, Variables the burden of information gathering and transfer still Outcomes Protected by copyright. often falls to general practitioners, care coordinators and Outcomes included the frequency of use and spatial patients themselves. distribution of health record systems and the frequency Policymakers, service managers and researchers need and spatial distribution of transitions of care between better methods to measure and understand the existing acute hospital trusts and health record systems. conditions that underlie health record system coordina- tion and interoperability at a hospital level. An accurate, Data sources contemporaneous overview of the current use and spatial Hospital trust health record system usage distribution of health record systems in the NHS in England http://bmjopen.bmj.com/ is required. Overlaying the use and distribution of these Manual collection of details of the health record system record systems with empirical data on patient movement type and vendor used at each hospital trust was under- between healthcare organisations can provide a valuable taken to establish a comprehensive and up-to- date trust- tool to guide better data sharing where it is most needed. level data set at November 2018. We followed several steps This study initially aimed to identify the frequency of to obtain the required data through open access sources. use and spatial distribution of health record systems in This initially included an online web search of published the English National Health Service. Combining this data information pertaining to health record usage from each