Prehospital triage of acute aortic syndrome using a machine learning algorithm

Background Acute aortic syndrome (AAS) comprises a complex and potentially fatal group of conditions requiring emergency specialist management. The aim of this study was to build a prediction algorithm to assist prehospital triage of AAS. Methods Details of consecutive patients enrolled in a regional specialist aortic network were collected prospectively. Two prediction algorithms for AAS based on logistic regression and an ensemble machine learning method called SuperLearner (SL) were developed. Undertriage was defined as the proportion of patients with AAS not transported to the specialist aortic centre, and overtriage as the proportion of patients with alternative diagnoses but transported to the specialist aortic centre. Results Data for 976 hospital admissions between February 2010 and June 2017 were included; 609 (62 center dot 4 per cent) had AAS. Overtriage and undertriage rates were 52 center dot 3 and 16 center dot 1 per cent respectively. The population was divided into a training cohort (743 patients) and a validation cohort (233). The area under the receiver operating characteristic (ROC) curve values for the logistic regression score and the SL were 0 center dot 68 (95 per cent c.i. 0 center dot 64 to 0 center dot 72) and 0 center dot 87 (0 center dot 84 to 0 center dot 89) respectively (P < 0 center dot 001) in the training cohort, and 0 center dot 67 (0 center dot 60 to 0 center dot 74) and 0 center dot 73 (0 center dot 66 to 0 center dot 79) in the validation cohort (P = 0 center dot 038). The logistic regression score was associated with undertriage and overtriage rates of 33 center dot 7 (bootstrapped 95 per cent c.i. 29 center dot 3 to 38 center dot 3) and 7 center dot 2 (4 center dot 8 to 9 center dot 8) per cent respectively, whereas the SL yielded undertriage and overtriage rates of 1 center dot 0 (0 center dot 3 to 2 center dot 0) and 30 center dot 2 (25 center dot 8 to 34 center dot 8) per cent respectively. Conclusion A machine learning prediction model performed well in discriminating AAS and could be clinically useful in prehospital triage of patients with suspected AAS.