SMART: Monitoring and Evaluation System of the National Telehealth Program Brazil Networks
Telehealth; Program Evaluation; Interoperability; SOA; Business Intelligence
The Telehealth Program was established by the Ministry of Health (MH) in 2007 with nine technical and scientific centers administered by public higher education institutions and, since its inception, no standards, processes, applications or quality indicators have been defined. The lack of these definitions, allied to the decentralization of the centers, made each one of them to develop their own Information Systems (IS), with different types of language and architectures, without any regulation and integration of information with the MH, hindering thus the evaluation of the program against the offered services . In this context, this paper describes a solution proposed to integrate the data produced by the various heterogeneous systems developed and maintained by the telehealth centers in Brazil in a central data store so that the MH can monitor and evaluate the results of the actions of Telehealth. This solution consists of specifying a National Interoperability Model and the specification, implementation and validation of an architecture, entitled SMART, based on Business Intelligence (BI) and Service Oriented Architecture (SOA) paradigms. SMART architecture consists of four main components: web tool for information manipulation with a interactive interface for data exploration; a component for receiving the data produced by the IS of the centers; a component responsible for saving the received data in decision support data and; a component that collects data from external sources to compose the data warehouse (DW). For the architecture validation, two experiments were performed. The first one executes performance tests under high and extreme workload and the second experiment evaluates if the proposed optimization strategies contribute to the performance and efficiency of the architecture. The results of the experiments were summarized to attest to the effectiveness of SMART. The analysis of the results obtained on real data showed that the performance of SMART remained stable over the considered workloads and the high quality was proven since no errors were recorded during the experiments.