Semantic Search & Document Classification

The Problem

During hospital stay, all information about a patient (e.g. diagnosis, therapy, check-in report, check-out report, etc.) is collected in a case history (electronic medical record). This is made of both structured and unstructured information (i.e., in textual format).
The set of all case histories represents a valuable repository of medical knowledge encompassing a large spectrum of past experiences about diseases, pathologies, diagnoses, treatment plans, lab tests, etc. One main objective pursued nowadays by many healthcare organizations is that of making all the collected knowledge, appropriately organized and analyzed, easily reachable and exploitable by both physicians and citizens.

Exeura solution

Exeura developed for an Italian health and social care structure a system capable of performing both semantic search of case histories and epidemiologic analyses.

Success story


Asolo: patient clinical records management

The Need

The ULSS of Asolo wants to create a patient case histories repository and make it easily accessible to both physicians and citizens. Further, it wants leverage the repository as the information base from which performing epidemiologic analyses.

The Solution

Semantic-based text mining technologies is what is needed to create a knowledge base where relevant documents and concepts can be detected with high accuracy. To this end, Exeura developed an expert system based on Rialto Text. The system is able to perform semantic classification and search by exploiting Mesh (Medical Subject Headings), a thesaurus containing over 26.000 medical terms created by the National Library of Medicine (NLM). The search engine provides a multi-dimensional search mechanism, whereby the user can select documents by combining key words, metadata, (Mesh) concepts and thematic categories. Further, it assists the user by suggesting additional search criteria.
Epidemiologic analyses are performed by extracting relevant data and concepts from clinical records through (semantic) pattern recognition techniques. Results identifying risk factors, geographical desease distributions, etc., are provided through an effective dashboard.

The Benefits

The solution developed provides physicians with improved insights into past experiences and best practices, along with a better understanding of the correlation between population diseases and risk factors. This should substantially contribute to the goal of improving healthcare policy decisions and enhancing citizen medical services.

Customer: ULSS Asolo
Market: Health
Product: Rialto Text
Task: Semantic access to electronic medical records and epidemiologic analyses

ULSS 8 Veneto