An immune memory and negative selection to visualizing clinical pathways from electronic health record data

Clinical pathways indicate the applicable treatment order of interventions. In this paper we propose a data-driven methodology to extract common clinical pathways from patient-centric Electronic Health Record data (EHR). The analysis of patient's, can lead to better regarding pathologies. The proposed algorithmic methodology consist to designing a system of control and analysis of patient records based on an analogy between the elements of the new EHRs and the biological immune systems. The detection of patient profiles ensured by biclustering Matrix. We rely on biological immunity to develop a set of models for structuring knowledge extracted from EHR and to make pathway analysis decisions. A specific analysis of the functional data leds to the detection of several types of patients who share the same EHR information. This methodology demonstrates its ability to simultaneously processing data, and is able to providing information for understanding and identifying the path of patients as well as predicting the path of future patients.