P103 Benefit to Patients and the NHS of Cardiac Magnetic Resonance Imaging after Primary Percutaneous Coronary Intervention: Data Challenges Within a Routine Data Registry

Background Cardiac magnetic resonance imaging (CMR) is a medical imaging technology for the non-invasive assessment of heart structure and function. Although the long term prognostic value of CMR is uncertain for some disease indications, uptake has been rapid among cardiologists with a 44% increase in the number of CMR scans in the UK between 2008 and 2010 (Antony et al. J Cardiovasc Magn Reson. 2011). A study was conducted to test the feasibility of setting up a UK registry using routinely collected NHS data to document CMR use in patients with suspected heart attack who activate the primary percutaneous coronary intervention (PPCI) pathway and compare outcomes between those who do/do not receive CMR. Methods Consecutive eligible patients who presented to a “24/7” PPCI centre in four UK hospitals (two with CMR facilities in England and two without in Wales) between May 2013 and August 2014 were invited to participate. Routinely collected clinical, biochemistry and radiology data were requested on all patients from all sites. Resource use data (Hospital Episode Statistics, HES, England; and Patient Episode Database Wales, PEDW) were also requested. Anonymised datasets of non-consenting patients were requested for comparative purposes. Results Of 2500 eligible patients, 1670 (66.8%) provided consent. Anonymised data were provided for 705/830 (84.9%) patients who did not consent, who were broadly similar to those who did. The proportion having CMR within ten weeks of the event was 18.1% (187/1035). Extracting routine data from different hospital databases was challenging; two sites could not easily provide biochemistry data and three could not provide details on medications on discharge. Bedside echocardiograms were not routinely recorded electronically. Other essential data fields were incomplete, clinical data were unavailable for 7.0% of patients and CMR information was recorded as free text, preventing simple electronic extraction of key data. HES data are not yet available for the two English sites but PEDW data have been obtained for the Welsh sites. Linkage of Welsh hospital data with PEDW admission data resulted in a date-match for the index event for 88.0% (559/635), with congruent documentation of PPCI and diagnosis of acute coronary syndrome for 73.2% (465/635). Conclusion Many required electronic data were not available routinely or promptly. Due to challenges with data extraction, collation and accuracy we concluded that it is not currently feasible to set up a registry using routinely collected data in this population.