Content uploaded by Patricia Araki
Author content
All content in this area was uploaded by Patricia Araki on Oct 20, 2015
Content may be subject to copyright.
ISDS Annual Conference Proceedings 2014. This is an Open Access article distributed under the terms of the Creative Commons Attribution-
Noncommercial 3.0 Unported License (http://creativecommons.org/licenses/by-nc/3.0/), permitting all non-commercial use, distribution,
and reproduction in any medium, provided the original work is properly cited.
ISDS 2014 Conference Abstracts
A Comparison of Fever Classified Chief Complaints and
Diagnoses with Recorded Body Temperatures
Patricia Araki*, Emily Kajita, Kelsey OYong, Monica Z. Luarca, Bessie Hwang and Laurene
Mascola
County of Los Angeles, Department of Public Health, Los Angeles, CA, USA
Objective
The Los Angeles County (LAC) Emergency Department (ED)
Syndromic Surveillance System (SSS) classifies patients into
syndrome categories based on stated chief complaints. In an effort
to evaluate the accuracy of patient- stated chief complaints and
final diagnoses, both “fever” chief complaints and diagnoses were
compared with patient body temperature readings.
Introduction
The LAC SSS has been in existence since 2004. Currently, the
system collects data from over 50 hospitals daily and performs a chief
complaint-based syndrome classification analysis of all ED visits.
The keyword “fever” is of special interest due to its inclusion within
several syndrome category definitions such as influenza, meningitis,
etc. However, inclusion of such terms in syndrome definitions may
be a disadvantage as such keyword searches would depend upon the
consistency in which the term “fever” is reported. In 2014, several
LAC syndromic surveillance hospital data connections were upgraded
to include notes recording patient body temperature. To evaluate the
newly added temperature information, analyses were conducted on
those observations that included body temperature, chief complaint,
and diagnosis information.
Methods
For this study, emergency department admitting data from 9
hospitals were reviewed. A total of 24,402 observations from a
five month period were categorized into groups by patient body
temperature, those with fever classified chief complaints (n=1441)
that included the terms “fever” or “febrile”, or patients with a similarly
defined fever classified diagnosis (n=970) and/or corresponding ICD9
code. Binary classification tests were conducted on this population
to observe fever classified chief complaint and diagnosis outcomes
against selected body temperature ranges.
Results
The sensitivities for fever classified chief complaints were 28.5%,
59.8%, and 73.9% for body temperatures *99°F, *100°F, and *101°F,
respectively. Corresponding positive predictive values (PPV) were
60.7%, 43.2%, and 27.8% for the same groups. For fever classified
diagnoses, sensitivities were found to be 22.2%, 47.5%, and 56.9%
while PPVs were at 70.3%, 51.0%, and 31.8%, respectively, for
the same temperature groups. The majority of fever classified chief
complaints (81.4%) and diagnoses (75.7%) were categorized into
the respiratory syndrome category with 54.3% of all fever classified
chief complaints resulting in a consistent diagnosis. Furthermore,
55% of observations with both a fever classified chief complaint and
diagnosis also recorded body temperatures at or above 100°F.
Conclusions
These preliminary findings have provided a basic understanding of
the utilization of the term “fever” within the LAC SSS. In addition,
the evaluation of body temperature versus fever classified chief
complaints and diagnoses will be utilized to determine whether,
and how, the body temperature variable may be included in future
iterations of syndrome classification algorithms.
Keywords
fever; syndromic surveillance; temperature; chief complaint
*Patricia Araki
E-mail: paraki@ph.lacounty.gov
Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * ϳ(1):e105, 201ϱ