En la Búsqueda de Asociaciones en Datos Clínicos para el Diagnóstico de Enfermedades Respiratorias

Abstract

In emergency rooms, it is common that several patients present symptoms associated with pulmonary diseases as pneumonia and pulmonary embolism, however, their initial diagnosis presents several challenges: a) both share symptoms, b) the lack of imaging studies at the first diagnosis stage and, c) the amount of data resulting from laboratory analysis are vast to be quickly analyzed. Thus, supporting medical staff through the use of computational tools for detecting clinical variables that are significant on the initial diagnosis, is crucial. Hence, a collaboration with the Centro Medico Nacional de Occidente (CMNO) of the Instituto Mexicano del Seguro Social (IMSS) was established to gather, process, and analyze patients' electronic medical records with a pulmonary embolism or pneumonia diagnosis who were admitted through the emergency room. Data extracting and processing of patients’ medical records (PDF) were performed. Thereafter, a statistical and associative analysis (Apriori) was performed looking for the determination of the clinical variables associated with a correct or incorrect prognosis. As a result, during the patient diagnosis several frequently present variables were identified, pointing out those that whose median values were significantly different for both diseases studied here. Finally, some laboratory variables are suggested to be carefully observed while a patient's initial diagnosis is performed.