Background Emergency and urgent care in Ireland is changing under the Health Service Executive (HSE) transformation programme which aims to implement a ‘systems-based’ approach for discrete geographical regions. The SIREN (Study of the Impact of Reconfiguration on Emergency and Urgent Care Networks) project aims to describe and evaluate the development and performance of different emergency and urgent care systems (EUCS) in Ireland. As part of that study, the aim of this analysis is to identify variation in EUCS quality indicators over time and across geographical regions having adjusted for differences in population demographics.

Methods The quality indicator used for this analysis was in-hospital case fatality, with respect to the patient’s area of residence. Four serious emergency conditions (SECs) were selected – Stroke/CVA (ICD 10 codes: I61; I63; I64; I629), Myocardial Infarction (I21; I22; I23), Ruptured Aortic Aneurysm (I710; I711; I713; I715; I718) and Acute Heart Failure (I50). The number of patients resident in each county (including a division of North/South Dublin and North/South Tipperary) with an emergency admission to any hospital for these principal diagnoses between 2007 and 2012 was identified using Hospital In-Patient Enquiry (HIPE). Indirectly standardised in-hospital case fatality ratios (SCFRs) were constructed. Standardised results for each of the four SECs were combined to adjust for differences in case mix between regions.

Annual fatalities were modelled using an age-adjusted Poisson regression. Stability between case fatalities for the two time periods, 2007–2009 and 2010–2012 was measured using the Spearman rank correlation. A high correlation would suggest that between county differences reflected systematic rather than random variation. Findings were then compared with deprivation rates using the Pobal HP Relative Deprivation Index.

Results Deaths following emergency admissions (11,016) accounted for 40% of all deaths for these 4 conditions (27,788) between 2007–2012. A significant decline in annual fatality rates were observed; rates fell from 13% in 2007–2009 to 11.7% in 2010–2012, and the number of emergency admissions increased slightly from 44,506 to 44,848. Over this period the county level SCFRs varied from 81 to 121. However, ratios were not stable (Spearman rank correlation = 0.28, p = 0.14). From our preliminary findings, there was also no significant association between in-hospital case fatality and deprivation.

Conclusion Overall case fatality between the two periods has fallen slightly from 13% to 11.7%, however the ratios were not stable and we found no association with deprivation. Further multilevel analysis will model the variation in patterns, and change points, between regions from 2000 to 2012.