Information Technology Reference
In-Depth Information
Concept Recognition Module.
This module extracts hypertension concept mentions
from inputted records using MetaMap and apache RUTA-based components. The
MetaMap component of this system needs to be installed separately as it requires
separate licensing. The configuration file has an option for MetaMap mmserver host
IP address and port which users need to configure based on their MetaMap installa-
tion configuration. Optionally, users need to configure word sense disambiguation
(WSD) server information if required. By default the HTNSystem is restricted to find
concepts relevant to SNOMEDCT_US only, however users can configure to use other
available vocabularies from UMLS. The configuration file also contains lists of com-
mon unique identifiers (CUIs)
7
for identifying Hypertension concepts, the CUIs
presented in Table 2 are based on our analysis of the corpus and it covers almost all
relevant hypertension concepts. However, if required, users can also add additional
CUIs to this list based on individual requirements.
The Ruta BP value extractor component extracts BP mentions and its associated
values based on Ruta scripts. HTNSystem infers a BP value as HTN, if the systolic
BP is greater than 140 or diastolic BP is greater than 90. HTNSystem also allows
users to configure these value ranges in configuration files. This component is a com-
pletely rule-based component and from our analysis of the corpus, Ruta BP value
extractor component is able to extract most of the BP values. The Ruta script shown
in figure 2 is capable of identifying mentions of BP values like "BP: 158/72","blood
pressure 149/96","Blood pressure is elevated at 188/92" and "BP unchanged at
145/70". The Ruta-based Hypertension abbreviation extractor extracts Hypertension-
related abbreviations like "ht" for hypertension and “hbp" for high blood pressure.
Table 2.
Default list of HTN relevant UMLS CUIs in HTNSystem
Term
UMLS CUI
Hypertensive disease
C0020538
Benign hypertension
C0264637
Essential hypertension
C0085580
Endocrine hypertension
C0264641
Malignant hypertension
C0020540
Systolic hypertension
C0221155
Diastolic hypertension
C0235222
Secondary hypertension
C0155616
Secondary benign hypertension
C0155620
Malignant secondary hypertension
C0155617
Secondary diastolic hypertension
C0264647
Post-processing Module.
Post-processing module in the HTNSystem is a combina-
tion of custom built Java based components. These components act like a filtering
system which filters out concept and mentions relevant to HTN. The CUI filter within
7
http://www.nlm.nih.gov/research/umls/new_users/glossary.html#c
Search WWH ::
Custom Search