03 Data linkage across ambulance services and acute trusts: challenges and opportunities. Issue 10 (28th September 2017)

• Record Type: Journal Article
• Authors: Clark, Sophie; Porter, Alison; Halter, Mary; Dorning, Holly
• Is Part Of: Emergency medicine journal. Volume 34:Issue 10(2017)
• Page Start: 696
• Page End: 696
• Publication Date: 2017-09-28
• Subjects: Emergency medicine -- Periodicals; 616.02505
• DOI: 10.1136/emermed-2017-207114.3
• Languages: English
• ISSNs: 1472-0205

Abstract

Abstract : Background: Most callers to 999 ambulance services are transported to hospital emergency departments (EDs), but ambulance services receive no information on patient outcomes. PHED Data is a two-year mixed-methods observational study of the process and potential benefits of linking data from EDs with ambulance service data to allow analysis of patient outcomes. We report on our first aim, to examine the potential opportunities and challenges to routinely linking these data. Methods: We approached six acute trusts, selected to give a range of performance, location and size, from an English metropolitan area. We used a structured learning log to collect data on the process, time input and reflections. We analysed these data with descriptive statistics, and qualitatively for themes. Results: All six trusts we approached agreed to participate. Data were linked using an algorithm based on date, time and patient demographics. We achieved a dataset of 7 75 018 records covering 2012–2016, and a linkage rate of 81%. We identified five stages of the process: senior approval; exploring data availability; information governance agreement; data transfer and linking. The most intensive phases were; negotiating senior approval (mean research team input per trust of 8 hours 5 min [SD 8 hours 3 min] plus additional time from acute trusts), and data linkage (mean research team input per trust 12 hours 40 min [SD 7 hours 4 min]). The stage which took the longest was information governance (mean 19 weeks). Key themes included the positive attitudes of trusts to participating, the range of decision-makers involved, and the need for sustained input from the research team. Conclusions: We have found the process of data linkage to be feasible, but requires dedicated time from research and trust staff, over a prolonged period, in order to achieve initial set up. Linked data are now being analysed.