Pysmennyi I. Electronic health care system: continuous patient monitoring

This thesis is focused on solving actual problem of improving healthcare system’s efficacy by introducing Electronic Healthcare System (EHS), which will enable collection, processing, and persistence of patient’s data, integrate existing PHR solutions and provide pluggable connectivity for third parties. Existing approaches to digitalization of healthcare domain were analysed, and synergy of applying them together caused by increased amount of data available in the system and, as a result, extracted knowledge was shown. Advantages of applying Service-Oriented Architecture to given task were shown. Using ontology-based service discovery was proposed aimed to improve the efficacy of new service development by reusing already existing functionality with matching criteria being defined as sum ontology-based distance of respective API models and convolution of QoS requirements’ conformity. Classification of threats of various EHS components was performed and methods to mitigate found vulnerabilities were proposed. Utilization of blockchain with combined proof-of-authority, proof of stake and proof-of-conformance consensus is suggested to guarantee patient’s data integrity and ensure regulatory compliance requirements. To mitigate performance and privacy drawbacks of shared Hyperledger it is suggested to use it as an index pointing to data stored on the third-party services with an additional benefit of precise authorization boundaries. Suggested methods of system design can be used in the implementation of EHS specified in the Law of Ukraine "On state financial guarantees of health care" (from 19.10.2017 № 2168-VIII) and expand its potential by closer integration of patient’s interactions layer. Usage of fog computing for continuous patient monitoring is introduced with an example of computer appliance for user’s breath detection. Advantages of processing body sensor network’s signal on the edge were shown including increased system autonomy, reduced feedback latency, improved security of sensitive data due to restrictions on transferring measurements to the cloud resulting in lower chance of man-in-the-middle attacks. Novel graph-based fog computing network model, which allows distribution of computational task in the edge networks with the regard to the varying capabilities and resources of each node was developed and evaluated in the simulated environment. In this model smart sensors are regarded as services allowing other system components to consume their capabilities, improving networks robustness and redundancy. Algorithm of preservation of most meaningful data based on cumulative moving average for caching measurements on storage-restricted intermediate fog nodes with an algorithm to mitigate consequences of edge network fragmentation. The combination of edge computing inside body sensor networks and cloud-based EHS modules provides effective collection, processing, persistence, and access to user’s data bound with continuous patient monitoring; provides increased volume of meaningful information for diagnosis and treatment to the physicians; allows to reuse gathered information for scientific purposes; makes possible for governmental agencies to collect statistics and make data-driven decisions.