Feasibility of data linkage in the PARAMEDIC trial: a cluster randomised trial of mechanical chest compression in out-of-hospital cardiac arrest. Issue 7 (28th July 2018)

• Record Type: Journal Article
• Authors: Ji, Chen; Quinn, Tom; Gavalova, Lucia; Lall, Ranjit; Scomparin, Charlotte; Horton, Jessica; Deakin, Charles D; Pocock, Helen; Smyth, Michael A; Rees, Nigel; Brace-McDonnell, Samantha J; Gates, Simon; Perkins, Gavin D
• Is Part Of: BMJ open. Volume 8:Issue 7(2018)
• Publication Date: 2018-07-28
• Subjects: adult cardiology -- clinical trials -- data linkage -- cardiac arrest; Medicine -- Research -- Periodicals; 610.72
• DOI: 10.1136/bmjopen-2018-021519
• Languages: English
• ISSNs: 2044-6055

Abstract

Abstract : Objectives: There is considerable interest in reducing the cost of clinical trials. Linkage of trial data to administrative datasets and disease-specific registries may improve trial efficiency, but it has not been reported in resuscitation trials conducted in the UK. To assess the feasibility of using national administrative and clinical datasets to follow up patients transported to hospital following attempted resuscitation in a cluster randomised trial of a mechanical chest compression device in out-of-hospital cardiac arrest. Methods: Hospital data on trial participants were requested from Hospital Episode Statistics (HES), the Intensive Care National Audit and Research Centre, and Myocardial Ischaemia National Audit Project and National Audit of Percutaneous Coronary Interventions, using unique patient identifiers. Linked data were received between June 2014 and June 2015. Results: Of 4471 patients randomised in the pre-hospital randomised assessment of a mechanical compression device in cardiac arrest (PARAMEDIC) trial, 2398 (53.6%) were not known to be deceased at emergency department arrival and were eligible for linkage. We achieved an overall match rate of 86.7% in the combined HES accident and emergency, inpatient and critical care dataset, with variable match rates (4.2%–80.4%) in individual datasets. Patient demographics, cardiac arrest-related characteristics and major outcomes were predominantly similar between HES matched and unmatched groups, in the linkage apart from location, response time and return of spontaneous circulation (ROSC) at handover. Conclusions: This study shows that it is feasible to track patients from the prehospital setting through to hospital admission using routinely available administrative datasets with a moderate to high degree of success. This approach has the potential to complement the trial data with the demographic and clinical management information about the studied cohort, as well as to improve the efficiency and reduce the costs of follow-up in cardiac arrest trials. Clinical trial registration: ISRCTN08233942; Post-results.