Sistema de laudos de eletrocardiograma: a importância de ferramentas de suporte à decisão

Abstract

Introduction: The electrocardiogram (ECG) is a low-cost, easy to access and non-invasive test that is a screening exam for heart disease. For the electrocardiographic report, specialized knowledge and adequate training are required. Despite the development of systems aimed at ECG management, the use of assistive tools for medical decision-making in the report was rarely addressed. In a tele-electrocardiography service that reports an average of 3,000 exams per day, the inclusion of automated decision trees can help and reduce medical errors. Objectives: To present the ECG reporting system developed in the Telehealth Network of Minas Gerais (TNMG) and evaluate its impact on the rate of disagreement between medical reports by the audit after the incorporation of decision-making tools in the software. Methods: A new ECG analysis system was designed and built to assist the specialist and meet the need to expand the RTMG. For this, a web solution was developed, which uses resources from decision support systems (DSS), this restricts access to diagnostics that do not have their requirements met, in addition to evaluating ECG measurements and alerting abnormalities. The system generates a descriptive report based on the selected diagnoses, for this purpose a nomenclature standard was created that allows a combination of several diagnoses in the same exam. The DSS was built with the differential of following the specialist's clinical reasoning process, in order to complement his actions in the examination evaluation process. Results: An ECG system was developed that is part of a complex Teleassistance platform. In addition to the tracing measurement tools, the system provides automatic measurements that, in case of abnormality, notify the specialist. To define the diagnosis, the system uses a hierarchical list of classes, based on AHA / ACC / HRS / ISCE standards, which have intrinsic rules that define whether or not it can be selected. A full descriptive report is automatically generated based on the selected diagnoses. The system has been extensively tested and refined until ready in May 2018, and is constantly evaluated through periodic audits by a senior cardiologist. More than 1.000.000 ECGs were analyzed after the new system was implemented. The validation base consists of 4313 audits in the old system and 2213 in the new one. The disagreement rate is 14,7% (Q1-Q3: 11,0-17,3), in the old system (2015-2017) and 8,6% (7,2-9,5) in the new system (2018-2019), with p-value = 0,004. Conclusion: A new system of ECG reports on a Web platform with decision-making tools was developed and implemented at TNMG. It was associated with a reduction in the rate of disagreement between the medical reports.