Persistência de dados clínicos baseados no openEHR: uma abordagem orientada por recursos limitados

• Autor(a) principal: Martins, Beatriz Proto
• Data de Publicação: 2016
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFG
• dARK ID: ark:/38995/0013000005fg3
• Texto Completo: http://repositorio.bc.ufg.br/tede/handle/tede/6645
• topic: Armazenamento e recuperação de informação; Registro eletrônico de saúde; Sistemas de informação em saúde; Modelagem multinível; Information storage and retrieval; Electronic health records; Health information systems; Multilevel modeling; CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO

Motivation: Electronic Health Records contain clinical data and are found in Health Information Systems. In this scenario, openEHR specification defines the record structure to allow systems to be interoperable, that is, to have a common understanding over exchanged data. A record comprises data modeled according to health domain concepts, called archetypes (knowledge level). An archetype, in turn, is composed by a subset of fixed entities from the Reference Model (information level). Due to the required detailing, the defined structure can be highly granular. Thus, the persistence of records, with the same format used during data exchange, can be hampered in terms of performance, especially in devices with a considerable resource limitation. Method: This work presents a strategy that serves as reference for the storage and retrieval of clinical data based on openEHR. Considering resources limitation, health services can persist their records in an optimized format, different from the format used for exchange. In this way, each service must implement a strategy for packing and unpacking that makes the conversions between both formats. Results: The persistence strategy presented in this work employs mapping rules between the objects graph of the Reference Model and a serialized data array. The rules range from primitive data types, such as an integer, to complex types, such as a hashmap consisting of objects with variable types and sizes. Conclusions: The strategy was designed considering the reduction of memory space occupied, but without turning the processing time unfeasible. Studies should be carried out for the strategy implementation and its experimentation.