Diagnose a Disease: A Fog Assisted Disease Diagnosis Framework with Bidirectional LSTM

Abstract

The Coronavirus (COVID-19) pandemic has created a huge havoc on a global scale including Pakistan and its surrounding regions in South Asia. The underdeveloped medical infrastructure and inadequate healthcare staff have become a dilemma during this pandemic fostering the need for digital health system. In this paper we propose Diagnose A Disease (DAD), a novel telehealth solution in Pakistan for remote patient monitoring and disease diagnosis. The three layered hybrid architecture of DAD comprises of data collection layer, analytics engine layer and cloud storage layer. In the first module, vital physiological signs of patients are measured and recorded through a set of wearable sensors. The next module makes use of fog enabled cloud framework for resource management of worker nodes. The analytics engine module also includes a trained Bidirectional Long Short Term Memory neural network model for heart disease, blood pressure and diabetes classification. Finally, the last module makes use of the cloud service for data storage, analysis and distributed secured health data sharing among medical authorities. The telehealth solution comes with emergency notifications, standard clinical guidelines and many advanced features with fog service to reduce latency and delays that becomes crucial in healthcare applications. PureEdgeSim, a simulation toolkit for fog environments is used to evaluate the proposed DAD model in terms of latency, bandwidth usage, power consumption, execution period and accuracy. Results depict that the proposed architecture performed well in handling real time requests, resource utilization and response time for healthcare decision making which further enhances its utility in real life situations.