PERSIST Sensing Network: A Multimodal Sensing Network Architecture For Collection of Patient-Generated Health Data In The Clinical Workflow

Patient gathered health data (PGHD), including self-reports and patient-reported outcomes (PROs) and data based on biometry collected from the wearables, represent an important source of context into patient's life and experiences. However, in light of recent technological boom and privacy concerns, the integration of sensitive data sources, with the knowledge extraction (data processing) and the health-care information systems (data sinks) represents a challenge. In this paper we highlight PERSIST Multimodal Sensing network (P-MSN) as a light-weight enabler for privacy-aware and patient-centric collection and integration of data from the edge. It is designed as an aggregator of data sources. It connects them with sinks and can deliver data extraction in form of multimodal sensing. The framework is designed to support 160 patients in a prospective multicentre clinical study of the PERSIST Project. The main building blocks consist of Apache Camel and ActiveMQ Artemis to deliver interconnectivity, and Kafka to deliver multimodal micro-service network. For evaluation of the system, we applied load tests and the results showed that the lightweight infrastructure can support well over 1000 simultaneous users.