According to the 'Istituto Superiore di Sanita (ISS), hospital infections are the most frequent and serious complication of health care. This constitutes a real health emergency which requires incisive and joint action at all levels of the local and national health organization. Most of the valuable information related to the presence of a specific microorganism in the blood are written into the notes field of the laboratory exams results. The main objective of this work is to build a Natural Language Processing (NLP) pipeline for the automatic extraction of the names of microorganisms present in the clinical texts. A sample of 499 microbiological notes have been analysed with the developed system and all the microorganisms names have been extracted correctly, according to the labels given by the expert.

A NLP pipeline for the automatic extraction of microorganisms names from microbiological notes

Mora S.;Attene J.;Gazzarata R.;Giacomini M.
2021-01-01

Abstract

According to the 'Istituto Superiore di Sanita (ISS), hospital infections are the most frequent and serious complication of health care. This constitutes a real health emergency which requires incisive and joint action at all levels of the local and national health organization. Most of the valuable information related to the presence of a specific microorganism in the blood are written into the notes field of the laboratory exams results. The main objective of this work is to build a Natural Language Processing (NLP) pipeline for the automatic extraction of the names of microorganisms present in the clinical texts. A sample of 499 microbiological notes have been analysed with the developed system and all the microorganisms names have been extracted correctly, according to the labels given by the expert.
File in questo prodotto:
File Dimensione Formato  
Mora_et_al_21.pdf

accesso aperto

Descrizione: Articolo in volume
Tipologia: Documento in versione editoriale
Dimensione 372.59 kB
Formato Adobe PDF
372.59 kB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11567/1083028
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 2
  • ???jsp.display-item.citation.isi??? ND
social impact