In recent years, there has been a significant increase in interest in medical health applications using Wireless Body Sensor Networks (WBSN). Monitoring people health conditions is crucial activity of today's healthcare system. As the future state of health is a consequence of past state, we induce data mining approaches on the window of past signals to predict worst health condition coming in the future. We propose the combinatorial real-time clustering and classification module, in which the future health risk would be predicted in real time. Also to enable continuous monitoring of the patient in real time, we are introducing mobile healthcare system by which the patient analysis and future predictions are carried out and informed to the patient by physician. The proposed Online Distribution Resource Aware (ODRA) algorithm has demonstrated superior accuracy and a decreased rate of False Positives (FPR) in predicting the risk status of 64 patients with varying illnesses using their Heart Rate (HR), Oxygen saturation (SpO2), and Blood Pressure signals. This performance is in comparison to the Resource Aware High-Quality Clustering (RAH) algorithm. This indicates that ODRA can be a promising algorithm for healthcare professionals to better assess the risk status of their patients and provide more accurate diagnoses and treatment plans.