Regresar

Obesity cohorts based on comorbidities extracted from clinical notes

Abstract:

Clinical notes provide a comprehensive and overall impression of the patient’s health. However, the automatic extraction of information within these notes is challenging due to their narrative style. In this context, our goal was two-fold: first, extracting fourteen comorbidities related to obesity automatically from i2b2 Obesity Challenge data using the MetaMap tool; and second, identify patients’ cohorts applying sparse K-means algorithms on the extracted data. The results showed an average of 0.86 for recall, 0.94 for precision, and 0.89 for F-score. Also, three types of cohorts were found. The results showed that MetaMap can represent a good strategy for automatically extracting medical entities such as diseases or syndromes. Moreover, three types of cohorts could be identified based on the number of comorbidities and the percentage of patients suffering from them. These results show that hypertension, diabetes, CAD, CHF, HCL, OSA, asthma, and GERD were the most prevalent diseases.