Emergency Department Overcrowding Detection by a Multifractal Analysis

Sid'Ahmed Emine (1) , Guillaume Bouleux (1) , H Haouba (2) , Eric Marcon (1)

1 DISP - Décision et Information pour les Systèmes de Production
2 UNA - Université de Nouakchott Al-Aasriya

The observation of data which counts the arrivals at the ED in scales of hours or days reveals that these arrivals are characterized by the phenomenon of burstiness. The burstiness is a phenomenon that appears in the majority of Emergencies due to a batches of arrivals coming to the ED in a small time interval over a wide scales. The prediction of congestion in ED caused by the batched arrivals seems important to the medical staff. The proper modeling data like arrivals process that has this scaling property is relevant with the self similar time series. The goal of our study is to establish that the time series of patient flow displays fractal behavior, whose quantitative characteristics vary with time. To discover whether a Patients flow has begun to congest, we analyze the time series of patients arrivals data with a Multi-Fractal Detrended Fluctuation (MF-DFA) algorithm.

Mots clés: Signal Analysis; Detection; Data Flow Analysis; Fractals Systems; Decision Support Systems

Domaines: Statistiques [math.ST]; Traitement du signal et de l'image [eess.SP]; Applications [stat.AP]; Recherche d'information [cs.IR]; Modélisation et simulation

Sid'Ahmed Emine, Guillaume Bouleux, H Haouba, Eric Marcon. Emergency Department Overcrowding Detection by a Multifractal Analysis. 10th IFAC Symposium on Biological and Medical Systems (IFACBMS 2018), Sep 2018, Sao Paulo, Brazil. ⟨hal-01876444⟩