Layout 1


ISDS Annual Conference Proceedings 2012. 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 2012 Conference Abstracts

Evaluating Syndromic Data for Surveillance of Non-
infectious Disease
Ramona Lall* and Marc Paladini

New York City Department of Health and Mental Hygiene, Queens, NY, USA

Objective
To evaluate several non-infectious disease related syndromes that

are based on chief complaint (cc) emergency department (ED) syn-
dromic surveillance (SS) data by comparing these with the New York
Statewide Planning and Research Cooperative System (SPARCS)
clinical diagnosis data. In particular, this work compares SS and
SPARCS data for total ED visits and visits associated with three non-
infectious disease syndromes, namely asthma, oral health and hy-
pothermia.

Introduction
Syndromic surveillance data has predominantly been used for sur-

veillance of infectious disease and for broad symptom types that
could be associated with bioterrorism. There has been a growing in-
terest to expand the uses of syndromic data beyond infectious dis-
ease. Because many of these conditions are specific and can be
swiftly diagnosed (as opposed to infectious agents that require a lab
test for confirmation) there could be added value in using the ICD9
ED discharge diagnosis field collected by SS. However, SS discharge
diagnosis data is not complete or as timely as chief complaint data.
Therefore, for the time being SS chief complaint data is relied on for
non-infectious disease surveillance. 

SPARCS data are based on clinical diagnoses and include infor-
mation on final diagnosis, providing a means for comparing the chief
complaint (from SS) to a diagnosis code (from SPARCS), for evalu-
ating how well the syndrome is captured by SS and for assessing if it
would be advantageous to get SS ED diagnosis codes in a more
timely and complete manner.

Methods
Syndromes previously developed by the DOHMH were used for

this work. Syndrome definitions are based on querying the cc field in
SS data for terms associated with asthma, oral health and hypother-
mia. The asthma syndrome consists of search terms for ‘ASTHMA’,
‘WHEEZING’ and ‘COPD’. The oral health syndrome uses
(‘TOOTH’ or ‘GUM’) and (‘ACHE’, ‘HURT’) and excludes visits
resulting from trauma (e.g., ‘INJURY’, ‘ACCIDENT’). The hy-
pothermia syndrome is limited to search for the word ‘HY-
POTHERMIA’. For the purpose of comparison of the SS data with
SPARCS data for the three syndromes, the following ICD9 diagno-
sis codes were considered in SPARCS: 493 for asthma, 521-523, 525,
528-529 for oral health and 991 for hypothermia.  

SS and SPARCS data for 2007 were used for this work as this was
the most recent and complete SPARCS ED dataset that was avail-
able. Overall city-wide daily counts and hospital-level annual counts
for total ED, asthma-, oral health- and hypothermia-related visits
were computed for SS ED data and SPARCS ED data. A comparison
of daily and hospital trends for SS and SPARCS for total and syn-
drome-related counts were conducted using correlation coefficients.

Results
There is a high correlation between total ED SS and SPARCS daily

counts (r=0.98, p-value<0.001). On average, SPARCS daily counts
are higher by approximately 75 visits (range: -674, 591) per day. Cor-
relations between SS and SPARCS daily counts for asthma, oral
health and hypothermia were 0.96 (p-value<0.001), 0.66 (p-
value<0.001) and 0.45 (p-value<0.001), respectively. Correlations
between SS and SPARCS hospital-level annual counts for asthma,
oral health and hypothermia were 0.89 (p<0.001), 0.87 (p<0.001) and
0.07 (p=0.61). In 2007, less than 8% of individual SS records had a
discharge diagnosis, and this was found to vary between hospitals (0-
69%); therefore, a comparison between SS discharge diagnosis and
SPARCS diagnosis data was not possible.

Conclusions
Overall, syndromic surveillance data was found to be a useful data

source for public health surveillance of non-infectious disease. Total
ED visits were found to be comparable between SS and SPARCS.
While direct comparison of counts for syndromes is not possible, the
daily syndrome counts between SS and SPARCS correlated well.
However, the strength of correlation varied depending on the syn-
drome, with a better correlation for syndromes with larger volume of
visits to the ED (e.g., asthma) and with more commonly used terms
in the cc search (e.g., ‘tooth ache’) compared to syndromes with very
specific search terms (e.g., ‘hypothermia’). 

In certain instances, it is hypothesized that SS discharge diagnosis
would provide more reliable and representative estimates than cc for
tracking non-infectious disease. Future work will consider a period
with more complete SS ED discharge diagnosis data for further com-
parisons and to test the hypothesis that more complete and timely SS
ED discharge diagnosis data could improve surveillance efforts.

Keywords
chief complaint; syndromic surveillance; New York City; non-infec-
tious disease; discharge diagnosis

Acknowledgments

Data Analysis and Syndromic Surveillance Unit, Bureau of Communica-
ble Diseases

*Ramona Lall
E-mail: rlall@health.nyc.gov

Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * 5(1):e163, 2013