2014.ISDS.Abstracts.Final.pdf


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

Development of Mental Health Classification Related to 
Severe Weather Events

Alvin F. Chu*1, Stella Tsai1, Teresa Hamby1, Elizabeth Kostial2 and Jerald Fagliano1

1Communicable Disease Service, New Jersey Department of Health, Trenton, NJ, USA; 2Health Monitoring System, Pittsburgh, PA, 
USA

Objective
To describe the development and validation of a mental health 

classification to track emergency department visits for potential 
needed public health response during severe weather events.

Introduction
Real-time emergency department (ED) data are currently 

received from 78 of 80 New Jersey acute care and satellite EDs 
by Health Monitoring Systems Inc.’s (HMS) EpiCenter system. 
EpiCenter collects, manages and analyzes ED registration data for 
syndromic surveillance, and provides alerts to state and local health 
departments for surveillance anomalies. After the 2012 Superstorm 
Sandy devastated parts of New Jersey, NJDOH initiated a plan to 
develop severe weather surveillance using EpiCenter to provide the 
Department with the ability to track both health and mental health 
concerns during adverse weather conditions to alert the public about 
emerging health hazards.

Methods
The severe weather classifier consists of the following 

classifications: carbon monoxide, heat related illness, hypothermia/
cold related Illness, disrupted outpatient medical care, cardiovascular, 
tree-related injury, motor vehicle accident, gastrointestinal 
illness, respiratory illness, and mental health related illnesses. In 
collaboration with New York City Department of Health and Mental 
Hygiene (NYCDOHMH) and New York State Department of Health 
(NYSDOH), a group of specific mental health concerns including 
anxiety/adjustment, mood, psychotic disorders, suicide/self-inflicted 
injury, alcohol, and methadone/opiate/heroin use are evaluated. The 
evaluation process consists of two steps: 1) using cases where ICD 
codes are available that meet case definition to find possible inclusion 
keywords and text patterns not already included and 2) evaluating 
cases with keywords of interest but not containing ICDs meeting case 
definition to determine text patterns for exclusion. The sensitivity and 
positive predictive value (PPV) statistical measures are computed for 
each classification.

Results
During December 1, 2013 to March 31, 2014, eleven emergency 

departments frequently provided ICD diagnosis codes were used for 
this study. Based on this validation project, the anxiety/adjustment, 
mood, psychotic disorders and alcohol use classifications achieved 
a moderate-high sensitivity of 58.6% (PPV: 77.6%), 57.1% (PPV: 
88.6%), 63.2% (PPV: 52.6%) and 93.7% (PPV: 66.6%), respectively. 
Suicide/self-inflicted injury had a low sensitivity of 37.6% (PPV: 
51.2%). Methadone/opiate/heroin use achieved a moderate sensitivity 
of 64.7% but a low positive predictive value of 14.4%. Based on the 
validated keywords, seen in Figure 1, the anxiety/adjustment disorders 
showed slightly higher ED visits during the first week of Sandy; mood 
disorders also increased slightly several days after the storm; alcohol 
related ED visits followed the weekday/weekend pattern; suicide/
self-inflicted ED visits rates were elevated after approximately one 

week post-storm; ED visits for methadone/opiate/heroin use increased 
slightly during several days in the storm’s aftermath.

Conclusions
Validation results indicated a moderately-high sensitivity and PPV 

for anxiety/adjustment, mood, psychotic, and alcohol classifications 
and a low sensitivity for suicide/self-inflicted ED visits; methadone/
opiate/heroin ED visits had a low positive predictive value. Validated 
keywords from the chief complaint field of ED visits can provide 
meaningful information for NJDOH syndromic surveillance staff to 
track mental health related illnesses (e.g. anxiety/adjustment, mood, 
psychotic, and alcohol) before, during and after a severe weather 
event. For those two classifications with low sensitivity and PPV, 
additional validations are needed. Currently not all facilities provide 
ICD-9 codes in EpiCenter for keyword validation. However, NJDOH 
continues to recruit hospitals to provide ICD information for future 
validation and evaluation.

Keywords
Syndromic Surveillance; Classification; Severe Weather Event; 
EpiCenter; New Jersey

Acknowledgments

New York City Department of Health and Mental Hygiene; New York 
State Department of Health; Drs. Petros Levounis and Yvonne Farnacio 
(New Jersey Medical School)

*Alvin F. Chu

E-mail: alvin.chu@doh.state.nj.us    

Online Journal of Public Health Informatics * ISSN 1947-2579 * http://ojphi.org * (1):e120, 201