CONECTA-PEME (IN852A 2016/53) PROJECT DraMEKNOW
Allenta Consulting, S.L., Antón Campos Consultores, S.L. (freelanceOracleDBA) and Health in Code, S.L. are developing “Proyecto DraMEKNOW: Análise, Deseño e Prototipado dun Sistema Intelixente de Apoio a Decisión Clínica no Ámbito da Cardioloxía”, whose goal is to develop an expert system for a comprehensive management of medical knowledge that will improve medical decision-making in daily clinical practice.
Project funded by the European Union’s ERDF and by Axencia Galega de Innovación, within OT1 “Promover el desarrollo tecnológico, la innovación y una investigación de calidad”, within the framework of the operational programme FEDER Galicia 2014-2020, with received funding for its execution of 537 636.27€. Supported by Consellería de Economía, Emprego e Industria de la Xunta de Galicia
PROGRESS 2016:
In the first year, we worked mainly on the following work packages:
Package A1: Requirements analysis
This task was the most advanced in the project. It was practically closed the question of the identified requirements and detailed the use cases that handle them.
Package A2: Functional design of the system
To complete the functional design, we started with the already validated requirements and defined the use cases that handle them. These use cases specified the functionalities to satisfy the requirement.
Package A3: Technical design of the system
The technical design advanced in defining the components involved in the management of catalogues of entities that allow to represent knowledge: knowledge sources (such as clinical guidelines), clinical variables, calculation rules of derived clinical variables, rules of knowledge, etc.
PROGRESS 2017:
Package A1: Requirements analysis
Initial requirements were closed and prioritized for development.
Package A2: Functional design of the system
Main components that are part of the system architecture were identified, as well as their dependencies and interfaces. The prioritized use cases were those that affected the back-office part, in which the clinicians define the ontologies and rules that the inference engine will use to execute the reasoning. These ontologies were defined using a repository of clinical variables, taxonomies that classify them, and rules that using the variables and taxonomies allow to reason with the facts.
Package A3: Technical design of the system
Designed the knowledge management module. Designed the inference engine and the interoperability layers between the back office, the engine, and the system portal. Designed the consultation portal for use by clinical specialists.
Package A4: System construction
The first versions of the entire repository for knowledge registration were implemented. The use cases for the management of the different entities that are part of the knowledge together with the user interfaces were implemented. The inference engine was implemented, allowing the uploading of knowledge, the execution of reasoning, and the execution of individual tests or of rules defined in the knowledge management module. A first version of the access portal to dynamic forms generated from the catalogue of clinical variables and taxonomies defined in the knowledge management module was implemented.
Package A5: System validation and verification
The work carried out was focused mainly on the selection of field (within cardiology) to do the tests, corresponding knowledge upload in the system, and revision by the users in terms of usability.
The first functional versions of the software were available in the last four months of the year.
PROGRESS 2018:
This year we worked mainly on the upload of clinical knowledge related to heart failure and hypertension as well as on the validation and testing of the pilot by the clinical experts who collaborate on the project.
Important performance improvements were made in the system’s query and result interpretation tool, as well as usability analysis and safety and performance tests in the different system modules.
Although the project initially began to accumulate delay in some of the work packages, it could be recovered as it progressed in its development, and it eventually was completed in due time.
The different objectives of the project are considered achieved, on which the consortium team will continue working to finish packaging the project’s result as a product and to be able to lay out its commercialization.
