Affiliation:
1. Université de Versailles St-Quentin, UMRS 1018, Occupational Health Unit, Poincaré Teaching Hospital, Garches, France
2. Inserm, U1018, (Population based epidemiological cohorts) Research Platform, F-94807, Villejuif, France
Abstract
Introduction We sought to identify symptoms predictive of hospitalisation of patients having the acute allergic reactions managed by an Emergency Medical Dispatch Centre in France. Methods A prospective study was conducted from 20th August 2006 to 5th November 2006 on incoming calls to the Emergency Medical Dispatch Centre for the Hauts de Seine province, France. Statistical analyses were used to identify hospitalisation predictors. The validity of predictors identified was tested. We calculated equivalents of sensitivity (probability of presence of the sign when the reaction is severe), specificity (probability of absence of sign when the reaction is not severe), positive predictive value (probability of severe reaction when the sign is present) and negative predictive value (probability of non-severe reaction when the sign is not present). Results A total of 210 calls were included. The following clinical hospitalisation predictors identified had very good specificity and good positive predictive value: facial oedema, sense of choking, respiratory distress, cough, difficulty speaking, dysphagia, abdominal pain, dizziness, collapse, and chest pain. However, none of the selected predictors had a safe negative predictive value (and sensitivity). Conclusion This study has identified positive clinical predictors of hospitalisation for acute allergic reactions patients managed by a Medical Emergency Dispatch Centre. The next step will be a multicentre study in order to develop a severity score or a decision-making algorithm. (Hong Kong j.emerg.med. 2014;21:80-87)