Aims 1) To identify risk factors for death or unplanned readmission within one year following hospital discharge after cardiac intervention for congenital heart disease. 2) To characterise patient groups at highest risk who would benefit from targeted intervention.
Methods Records in the national congenital cardiac surgical audit (NICOR) pertaining to UK infants who had a cardiac surgery or intervention aged under 12 months between 01/01/2005 and 31/12/2010 were matched with intensive care admission records in Paediatric Intensive Care Audit Network (PICANET); linked records with known life-status were obtained for 7634 infants. Outcome measures were: Outcome 1 – death within 1-year following discharge; Outcome 2 – Outcome 1 or emergency readmission to PICU within 1-year following discharge. Potential risk factors available from either dataset were pre-specified and univariate and multivariate logistic regression used to investigate the effects of these on each outcome. Classification and regression tree (CART) analysis was used to identify distinct patient groups differentiated by risk of Outcome 2, each defined by a set of patient characteristics.
Results 3.2% (246/7643) and 6.7% (514/7643) of infants experienced Outcome 1 and 2 respectively. Fitted multivariate models for both outcomes were robust in risk factor selection (Outcome 1 - ROC AUC = 0.78, 95% CI [0.75, 0.82]; Outcome 2 - ROC AUC = 0.78 [0.75, 0.80]). Risk factors significant in the multivariate Outcome 2 model were: age at procedure, weight z-score, cardiac procedure, cardiac diagnosis, non-cardiac congenital anomaly, neurodevelopmental condition, prematurity (<37 weeks gestation), ethnicity, and length of stay in specialist centre (LOS). Clinical deterioration was additionally significant to Outcome 1 whilst neurodevelopmental condition and acquired diagnoses were not. Key defining characteristics of infants in the patient groups identified as higher risk were [% Outcome 2]: (1) neurodevelopmental conditions [24%]; (2) Hypoplastic left heart, single ventricle or pulmonary atresia [15%]; (3) Congenital anomalies and LOS > 1 month [24% risk]; (4) No congenital anomalies and LOS > 1 month [9% risk].
Conclusions Understanding patient risk groups should inform recommendations for improving services, support development of interventions to mitigate each profile of risk and facilitate evaluation of the priority and feasibility of targeting each group.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.