During the last few years, the aging of the people in society and the increasing cost of healthcare arose the necessity of e-health systems as a suitable solution to maintain cost-effective high quality healthcare. The usage of e-health allows the real time collection and analysis of heterogeneous medical data on a large scale, since such data come from numerous patients. Usually, the analysis of such medical data needs to be fast to address time constraints of treatment, specially in severe sepsis or septic shock instances. Severe sepsis is the major cause of hospitalization, reaching high mortality rate, where early identification can reduce this mortality rate significantly. Information about infections, risk factors, patients demographic should be considered to develop comprehensive sepsis prevention tool for early recognition, and treatment strategies. Within this context, we propose an e-health system, called BE-SYS, to identify patients at high risk for septic shock based on the medical data collected as Big Data and rapidly analysed. Additionally, BE-SYS applies an iterative clustering approach to reduce both the scale of the problem and the amount of data analyzed, increasing the accuracy of diagnoses and complying with rapid responses to meet time constraints. The evaluation of the proposal using a real patient dataset shows that BE-SYS reaches high accuracy with very low response time analysis.