The big data era: The usefulness of folksonomy for natural language processing.

ABSTRACT

BACKGROUND: A huge amount of clinical data is generated daily and it is usually filed in clinical reports as natural language. Data extraction and further analysis requires reading and manual review of each report, which is a time consuming process. With the aim to test folksonomy to quickly obtain and analyze the information contained in media reports we set up this study. METHODS AND OBJECTIVES: We have used folksonomy to quickly obtain and analyze data from 1631 discharge clinical reports from the Nephrology Department of Hospital del Mar, without the need to create a structured database. RESULTS: After posing some questions related to daily clinical practice (hypoglycaemic drugs used in diabetic patients, antihypertensive drugs and the use of renin angiotensin blockers during hospitalization in the nephrology department and data related to emotional environment of patients with chronic kidney disease) this tool has allowed the conversion of unstructured information in natural language into a structured pool of data for its further analysis. CONCLUSIONS: Folksonomy allows the conversion of the information contained in clinical reports as natural language into a pool of structured data which can be further easily analyzed without the need for the classical manual review of the reports.