Home > About Fuji Xerox > Technology > New Technologies for Valuable Communication > Natural Language Processing Technology > Information Extraction Technology by Means of Language Processing > Adverse-effect statement aggregation

Adverse-effect statement aggregation

In the adverse-effect statement aggregation step, the system calculates the number of adverse effects that correspond to words listed in MedDRA/J and the number of drug names to which therapeutic category codes are assigned. The results are displayed in a contingency table.

Fig. 1: Configuration of the adverse-effect relation aggregation system

After assigning therapeutic categories and System Organ Class (SOC)-the highest level of MedDRA/J-respectively to drugs and symptoms, the system calculates the number of drug-symptom pairs extracted from the text. Then, as shown in Fig. 2, the system displays the results in a contingency table, with the horizontal axis representing SOC and the vertical axis representing therapeutic categories. Clicking the number of adverse reaction cases displays a pop-up window that shows a description of adverse effects in a discharge summary for each case. By clicking a therapeutic category, a user can see the situations where adverse effects appeared for each drug. Thus, this system enables users not only to refer to the total number of drug-symptom pairs but also to see more detailed information and directly access the text of a discharge summary by using a specific efficacy or symptom name as the key for searching.

Fig. 2: An example of a contingency table