Emergency Department Data

4.2 Data Collection: ESSENCE data are automatically extracted (via a Secure File
Transfer Protocol) from participating surveillance site databases. For example, in an
emergency room, demographics, chief complaints and diagnosis data are all
automatically sent from a hospital’s information system to the ESSENCE server.
Data are grouped into syndromes and then analyzed, with warnings and alerts
produced based on statistical algorithms.
4.3 Confidentiality: Data collected in ESSENCE do not have identifiers, and access to
the data is limited to ensure security and confidentiality.
4.4 24/7 Coverage: Data are collected 24 hours a day, all days of the week. DHMH
Office of Preparedness and Response epidemiologists are available to receive alerts
by phone whenever they occur. The epidemiologists’ contract phone numbers are:
410-767-6745 and 410-767-2074
4.5 Testing 24/7 System: The availability of 24/7 contacts is tested on a quarterly basis
using call-down drills through the Maryland Health Alert Network (MD HAN) (See
Appendix B). All members of the DHMH Office of Preparedness & Response staff
are included in the call down drills, including the epidemiologists responsible for
24/7 coverage of the ESSENCE system (See 4.4).
4.6 Health Department Staff Alerts: The epidemiologists at DHMH are responsible for the
ESSENCE program review alerts daily, and if an ESSENCE alert is determined to
have public health significance, the appropriate individuals within DHMH and/or
local health departments are contacted for follow up. Alerts are conducted through
MD HAN. Members of senior staff at the Health Department, including the Health
Officer, Deputy Health Officer, Emergency Preparedness Director and Director of
Communicable Disease, receive real-time alerts from HAN.
4.7 Retrieving ESSENCE Data: The Harford County Health Department has one
username and password for accessing the web-based system
(https://essence.dhmh.maryland.gov/). This system provides the Health Department
with access to the following information:
• Alert lists
• Data queries
• Statistical analysis of data sets
• Reports in the form of maps, time-series charts and tables
HCHD Policy # ADMIN 00-08
Page 2 of 2
Appendix A
Maryland Department of Health and Mental Hygiene,
Office of Preparedness & Response
ESSENCE User Guide
Version 2.0 (October 2012)
This should be considered an evolving guide due to the periodic changes made to
Maryland’s ESSENCE system. It will be updated as necessary, and as resources allow.
TABLE OF CONTENTS
INTRODUCTION ...................................................................................................................................................... 3
PURPOSE ......................................................................................................................................................... 3
OVERVIEW ....................................................................................................................................................... 3
LIMITATIONS AND OTHER CONTEXT...................................................................................................................... 5
A. ALERT LIST ............................................................................................................................... 9
Summary Alerts .......................................................................................................................................9
Region/Syndrome.................................................................................................................................. 10
Hospital/Syndrome ………………………………………………………………………………………… 15
ED Spatial Alerts........................................................................................................................................... 16
Region Caller Site (Alert List for Poison Control Data) ....................................................................... 17
Hospital/Sub-Syndrome Time of Arrival ................................................................................................. 18
School Absenteeism Alerts ………………………………………………………………………………19
Reviewing Alerts in ESSENCE................................................................................................................... 20
B. MY ALERTS ............................................................................................................................. 20
Creating myAlerts and Records of Interest ............................................................................................ 20
D. OVERVIEW PORTAL .............................................................................................................. 25
Using the overview portal, practical example:………………………………………………………….27
E. QUERY PORTAL ..................................................................................................................... 28
Free text chief complaint queries .............................................................................................................. 32
Using the query portal, practical example:............................................................................................. 33
F. MATRIX PORTAL .................................................................................................................... 35
G. WEEKLY PERCENT ................................................................................................................ 37
H. MAP PORTAL .......................................................................................................................... 39
I. BOOKMARKS .......................................................................................................................... 40
J. QUERY MANAGER ................................................................................................................. 42
Multi-Series time series graphs: ................................................................................................................ 43
K. myESSENCE…………………………………………………………………………………………..45
APPENDIX 1: Flowchart for Analysis and Response to Syndromic Surveillance Data………………....50
APPENDIX 2: FREE TEXT QUERY EXAMPLES ..............................................................................................52
EMERGENCY DEPARTMENT SECTION REFERENCES ..............................................................................54
2
Maryland Electronic Surveillance System for the Early Notification of
Community-based Epidemics (MD ESSENCE)
Introduction
This user manual is intended as a self-study guide to help the Electronic Surveillance
System for the Early Notification of Community-based Epidemics (ESSENCE) user
navigate through some basic functions in ESSENCE. As changes are made to
ESSENCE we will try to update this document as quickly as possible. Please contact
ESSENCE Help (mdessence@maryland.gov) at the Maryland Department of Health and
Mental Hygiene (DHMH), Office of Preparedness & Response (OP&R) with questions
about ESSENCE.
It should be noted that access to some features described in this training guide might
not be available to all individuals. For example, hospital personnel can only view data
for their hospital. Please contact ESSENCE Help (mdessence@maryland.gov) for
questions regarding access to data.
Purpose
Public health surveillance is often defined as “the ongoing systematic collection,
analysis, interpretation, and dissemination of data regarding health-related event for use
in public health action to reduce morbidity and mortality and to improve health” (1). A
second definition of syndromic surveillance more specifically is “an investigational
approach where health department staff, assisted by automated data acquisition and
generation of statistical signals, monitor disease indicators continually to detect
outbreaks of disease earlier and more completely than might otherwise be possible with
traditional methods for reporting disease” (2). The purpose of the ESSENCE system
within this context is as a system that provides an intuitive environment for
epidemiologists to conduct routine descriptive epidemiologic analysis, to monitor
morbidity and mortality trends over time, geography, and across multiple data sources,
and thereby providing information that can assist with making decisions on how to
improve population health.
Overview
In the mid-1990s the Johns Hopkins University Applied Physics Laboratory (JHU/APL)
began a collaborative biosurveillance project with the Maryland Emergency
Management Agency (MEMA) and the Maryland Department of Health and Mental
Hygiene. In 2000, the team realized that the Walter Reed Army Institute for Research
(WRAIR) was conducting a similar effort, called ESSENCE. These two projects soon
merged, and JHU/APL and Walter Reed entered into a collaborative research and
development agreement to continue working on the technology and expand system
access to both military and civilian public health authorities across the region. In 2001,
a worldwide United States military version of ESSENCE was implemented by JHU/APL.
3
In 2004 the military assumed the responsibility and implementation for the military
version of ESSENCE. JHU/APL has continued to maintain and develop the civilian
version of ESSENCE.
In 2005, the Maryland ESSENCE system was deployed. At this time, approximately 15
hospitals from the National Capital Region (NCR) and the Baltimore metro area were
contributing data to the ESSENCE system. By the summer of 2009 acute care hospitals in
the state of Maryland would be participating in the program. Currently the system in
Maryland includes four different data sources, each with its own module; 1) deidentified emergency department data from all acute care hospitals in Maryland and
two urgent care centers (n=47), 2) de-identified poison control center data, 3) deidentified over-the-counter (OTC) medication sales data from two large chain
pharmacies, and 4) de-identified school absenteeism data. All data sources are
updated daily. Including multiple streams of data in ESSENCE permits the visualization
and descriptive epidemiologic analysis of several data sources in one location. Each of
these data sources is briefly described below.
Emergency Department Data Module:
The data elements included from the EDs are patient age, sex, chief complaint, date of
visit, time of visit, zip code of residence, discharge disposition (e.g. admitted, discharged,
etc), discharge diagnosis, patient resident zip code, a unique identifier for the patient
and visit, and hospital name. Unfortunately, not all hospitals provide a discharge
diagnoses and discharge dispositions at this time. Automated parsing of the free text
chief complaints place these data into 11 syndrome categories, >100 sub-syndrome
categories, and allow for free text querying of the chief complaint text. These data are,
with very few exceptions, always one day behind. When users log in today the ED
visits for the previous 24 hour period are available for review. The data are loaded
automatically on weekends and holidays on the same schedule as noted above. ESSENCE
analyzes and displays the findings for use at the state, county, zip code, and hospital
level. Note: As a result the HITECH Act and the syndromic surveillance Meaningful Use
criteria, the number of requested variables may increase over time, and the data will
become more real-time.
Poison Control Center Data Module:
The Poison Control Center (PCC) data in ESSENCE is transmitted daily from the two
poison control centers that serve Maryland. This data feed provides coverage for all 24
jurisdictions in Maryland. A large number of variables are included in the ESSENCEPCC interface, including demographic data to orient the data by time, person, and
place, as well as information on substances, substance categories, clinical effects,
information on the exposure, and how it was managed clinically. These data have been
used to detect outbreaks associated with exposures to various chemical substances,
detecting individual cases of some reportable diseases (e.g. carbon monoxide
poisoning, pesticide poisoning), and providing situational awareness as it relates to
chemical substance exposures in Maryland.
Over-the-counter (OTC) Medication Sales Data Module:
The OTC data in ESSENCE is transmitted daily from two large pharmacy chains that
provide coverage for all 24 jurisdictions in Maryland. Data from these two large
4
pharmacy chains can be viewed separately or combined. Thermometer sales are
included along with other OTC medication sales. Variables included in the ESSENCEOTC interface include information on the product purchased such as a product
description, product type, product category, and a product UPC code. Other
information provided includes the date of purchase, a unique store identifier, the zip
code of the store, as well as a promotion indicator. These data have been used in
concert with other data sources to maintain situational awareness during large scale
events. They are also an indicator of disease activity including gastrointestinal illness
and influenza.
School Absenteeism Data Module:
The school absenteeism data in ESSENCE is transmitted on a daily basis from all 24
local public school systems in Maryland. Raw absenteeism numbers are provided for
all public schools. For each school, data is provided on the number of students
enrolled and the number of students absent, and these numbers are used to generate
a percent of students absent. Other variables included in the data are the date, zip
code of the school, jurisdiction/system the school is located in, the type of school (i.e.,
elementary, middle, or high school), and the type of day (e.g., normal, exam, snow,
etc.). The 2012-2013 school year will be the first full school year this data will be
received and utilized for surveillance purposes.
Limitations and Other Context
An understanding of the limitations of a particular data source enables users to make
correct interpretations of the data. This is especially the case with syndromic
surveillance data where the focus of the surveillance is on reporting timeliness and not
necessarily the specificity and completeness of the data. Limitations for specific data
sources are described below:
Emergency Department Data:
The syndromic classification of emergency department (ED) data in ESSENCE is based
on the patient’s chief complaint, which is generally some variation on what the patient
said was the reason for visiting the ED. Some things to consider include:
1) Chief complaints may be a comprehensive free text statement that mirrors the
patient’s actual statement very closely,
2) Chief complaints may be an abbreviated free text statement that includes only the
primary reason for the ED visit, or
3) Chief complaints may be a selection from a standardized “pick-list” of possible chief
complaints from within a hospitals data system that best fits the patient’s actual
statement.
4) Chief complaints may include misspelled words and/or medical short-hand.
5) Another limitation is that chief complaint data vary in their diagnostic precision, which
is also affected by the type of chief complaint text provided by the hospital. Systems
like ESSENCE tend to prefer options 1 and 2 above, because they often include
additional information that can be helpful for free text querying as well as with the
syndrome categorization. These kinds of chief complaints lend themselves well to
impromptu free text query development, and systems with free text chief complaints
may be flexible enough to add particular phrases to a patient’s chief complaint in certain
circumstances (e.g. adding the word “Haiti” in a patient’s chief complaint). Whenever
5
available we ask that hospitals send us the free text chief complaints, but this isn’t
always possible. Regarding misspellings and medical short-hand, the language parsing
algorithms in ESSENCE include a large number of common misspellings and medical
shorthand terms and will show what it thinks is the correct version of the chief
complaint.
Statistical Analyses:
Automated statistical analyses in alert lists, my alerts, and of time series data assists
the user by generating alerts when observed counts exceed a particular threshold.
While this information can be very helpful with focusing our attention on specific data,
users should not interpret statistical significance to necessarily mean public health
significance. Epidemiologists should conduct additional review of the data details (line
list) and distribution of the data to determine whether additional investigation is
warranted for a given circumstance. Appendix 1 provides an example of how to review
emergency department data. Additional information is available on the statistical
algorithms in the ESSENCE Help section and the Detector Algorithms section, both at
the top of the webpage once you have logged in.
Outbreak detection has always been one of the major goals of syndromic surveillance.
In Maryland we have had success in detecting otherwise unreported outbreaks using
both the ED data and poison control data. In this context there are a few things to
understand:
1) It is important to keep in mind the data source (ED and poison control data) and the
various kinds of disease outbreaks that you might expect to see given the context of
these data. For instance, in EDs you will see some percentage of patients who lack
another source of health care and as a result utilize the ED for that purpose, and those
who truly have a health emergency (or are at least very concerned about their health for
whatever reason).
2) The way patients present to the ED (for example) has to be in such a way that our
attention is drawn to these visits. One way is for the number of observed ED visits for a
given syndrome to exceed the expected value by some statistically significant margin.
For follow up to occur we also often like to see some other kind of clustering in the data
(e.g. by zip code, or age, or geographic location) in addition to the increased overall
count. A second way to draw attention to specific visits is related to what is said in the
chief complaint text. For instance, if a specific disease name is mentioned, or if a group
of seemingly related patients all mention “food poisoning” or something similar.
6
Getting Started
The secure ESSENCE website can be accessed by going to:
https://essence.dhmh.maryland.gov
Please add this website to your list of trusted websites.
Logging on:
1. Click on the link in the middle of the page to begin accessing ESSENCE
2. When prompted by the dialogue box enter your username and password. To
obtain ESSENCE access, please contact ESSENCE Help
(mdessence@maryland.gov).
3. The ESSENCE homepage will appear. All of the major functions of the software
are accessible using the main toolbar. We will describe most of the application
tabs (e.g., query portal, alert list) in the next pages. The remote data and user
admin tabs will not be discussed since these functions are not often utilized by
ESSENCE users in Maryland.
Prior to examining specific data outputs, you may want to look at the items on the very
top of the tool bar which describe: 1) history and background of ESSENCE and its
relationship to syndromic surveillance, 2) syndrome and sub-syndrome definitions, 3)
detector algorighms, a description of statistical techniques for outbreak detection, 4) the
data dictionary, a glossary of common ESSENCE terms and 5) the help section, which
contains FAQs, additional background information, and useful links. Please be aware
7
that many of these items have not been updated to reflect recent changes in the
ESSENCE system.
The homepage displays key information about the ESSENCE data including the total
number of hospitals per county that are reporting ED data for the previous day, as well
as the total number of Maryland hospitals reporting their ED chief complaint data for the
previous day. The most recent ED data in ESSENCE is usually the previous day’s data.
Please note that if only some of the hospitals in your geographic area have reported
their data at a particular time, this will have a direct influence on how you interpret the
ED data.
Monitoring of data completeness is one of the first activities each morning at OP&R. If a
hospital has not sent a file, or another problem is discovered, we will work with you and
the hospitals to resolve it. If you observe a problem with ESSENCE, please contact
ESSENCE Help (mdessence@maryland.gov).
The next sections of this manual provide a basic step by step introduction to navigating
through the ESSENCE system, and shares examples of the data analysis and
visualization tools. For the purposes of this manual, all tools will be explained using the
ED data. It is expected that by learning to use all tools with the ED data, the user will
be able to apply these tools to the other data sources contained within the ESSENCE
system. If further assistance is required to work with these other data sources please
contact ESSENCE Help (mdessence@maryland.gov).
8
Emergency Department Data
A. ALERT LIST
The alert list provides a tabular view of alerts for 10 of the 11 syndromes in ESSENCE
(it excludes the “other” category), and gives the user the ability to drill down into the data
from the alert list page. The alerts are the result of statistical analyses performed within
the ESSENCE system. ESSENCE uses the last 30-days as a baseline, minus the 2
most recent days, and compares the current data’s counts to that baseline to see if there is
a statistically significant increase. A yellow flag indicates a statistically significant pvalue between 0.01 and 0.05 while a red flag indicates a p-value of <0.01.
The statistical alerts presented on the Alert List page should be interpreted simply as
there being a larger than expected number of ED visits for a particular syndrome
category on a given day. This is an attempt to create an efficient way to scan through
all the data automatically and tell the user where to focus additional review. However,
statistical significance does not necessarily equal public health significance, and
inevitably there are many alerts generated that do not warrant additional follow up with
the hospital. It is up to the user to analyze the data and try to determine when
additional follow-up and investigation is warranted. Appendix 1 provides an example of
how to review alerts for the emergency department data in ESSENCE.
To begin navigating through the alert lists, click on the “alert list” icon.
Note: For a more detailed explanation please see the paper by Howard Burkom in the
Johns Hopkins APL Technical Digest (Volume 24; No. 4; 2003) “Development,
Adaptation, and Assessment of Alerting Algorithms of Biosurveillance”. A google search
of the title will return a few locations where the paper can be downloaded.
Summary Alerts
The “Temporal Alerts Summary” page is the first page displayed after clicking on the
Alerts List button. It shows a graphical summary of syndromic alerts for the state of
Maryland and individual regions. Please note: ESSENCE uses the term “region” in
different ways throughout the system. In the summary alerts, region means health and
medical region, but in the region/syndrome, hospital/syndrome and hospital/subsyndrome time of arrival alert lists, region means county/jurisdiction.
9
Each cell of asterisks (*) represents the past nine days for a given syndrome listed
chronologically with the current day being on the far right. Asterisks will appear as grey,
yellow or red. A yellow asterisk indicates a warning and a red asterisk indicates an alert
that may warrant further investigation. When looking at the data across regions, please
keep in mind that ED coverage can vary considerably.
The top tier of asterisks provides a sense of syndromic alert activity over time. By
clicking on individual asterisks, you can further investigate the data used to generate the
data point. The bottom tier of asterisks shows whether any event list entries have been
posted in a particular region. In Maryland, no events of public health importance are
recorded in the Event List feature, so clicking on these asterisks will provide no
information. For further details about the event list, see the “event list” section.
Region/Syndrome
Click on “region/syndrome” to review alerts by county (region = county).
10
In the Region/Syndrome view, and as in most tables like this in ESSENCE, you can sort
alerts by any of the column headings. A maximum of three columns can be sorted at a
time.
Note: The region column in the example above (and in other sections of this
manual) has been blocked for confidentiality reasons. When you log into
ESSENCE, you should be able to view this information.
Other useful features within the “Region/Syndrome” view
a. Configuration Options:
Clicking on the plus sign next to “configuration options” in the region/syndrome view
will allow you to view alerts for selected regions, syndromes, age ranges, and start
and end dates. Leave the data source field as “all data sources” and the sex field
as “all sexes”. After you have made the necessary changes, click on “change
configuration” to generate your new output. In the region/syndrome alert list,
syndromes are often presented in an aggregated view for age. Within the
configuration options, age can be delineated into subgroups.
b. Time Series:
11
Clicking on “time series” for an alert displays a time series graph of the number of
cases over time for the particular date and syndrome you selected. This is often
called “drilling down” into the data, and is a good way to see how an alert compares
to baseline data and previous alerts. Here is a time series graph that was
generated by clicking on “time series”. The title of the graph and other details can be
changed by clicking on graph options below the time series graph.
c. Data Details:
There are three places where you can access the data details (line list) view.
Directly below the time series graph generated in b (time series) above, you can
click on “data details” to see a line list of the data in your query for the entire range
of dates included in your time series. If you would like to see only one particular
day’s data details, then you could either click on the data point in the graph or click the
“data details” hyperlink in the data table below the time series graph.
12
The following are examples of figures and a line list that can be generated by
clicking on data details. Columns are blocked out for confidentiality reasons.
The line list information can also be sorted using the 3-sort function. Each case
record has fields for date, admit time, zip code, age, chief complaint, zip code,
etc. All of these data are essential for evaluating alerts. The field “MedRecNo”
is a unique identifier for the patient and the patient’s visit. This field can be
used to determine if duplicate records occur in the dataset. It can also be used
to facilitate communications among public health partners and hospital staff
during follow up or investigations
d. Map View:
There are three locations where you will find map view hyperlinks. 1) beneath
the time series graph, 2) in the data table below the time series graph, and 3) on
the data details (aka line list) page. Depending on your query, clicking on the
map view will create a map of your data by either county or zip code.
13
Once in the map the user can zoom in by double clicking on the map, and zoom out
using the “slider” on the left side of the map. At the top-center of the map above are 5
tools the user can select to drill down into the map for more information. These include
select by square, select by polygon, select by path, select by point, and map grabber.
The first 4 of these allow you to select parts of the map in different ways, and the
selections you make will appear in a table below the map. The table will include some
information about what you selected, and a time series link so that you can produce a
time series view of the selected data for that geography. On the right-hand side you
can see a column with “Layers” and “Global Parameters” selections via a “+” or “-“sign.
These function allow the user to manipulate the map in various ways, including adding
layers to visualize, changing the date range of the data displayed in the map, changing
the algorithm that creates the color ranges, change the number of ranges, and lets you
download the shapefile of the map you have created (click the picture of a disk under
results). Once the shapefile is downloaded the user can import it directly into ArcMap
for further manipulation.
14
Hospital/Syndrome
Clicking on “Hospital / Syndrome” will allow you to view the previous 8 days of alerts by
reporting hospital. This option offers additional focus (vs. Region) by further
individualizing the data. If an elevated number of cases in a particular syndrome are
reported in a hospital’s data, a yellow or red alert will appear and can be further
investigated by clicking on the time series hyperlink.
Clicking on the plus sign next to configuration options in the region/syndrome view will
allow you to view alerts for selected regions, syndromes, age ranges, and start and end
dates. Leave the data source field as “all data sources” and the sex field as “all sexes”.
After you have made the necessary changes, click on “change configuration” to
generate your new output. In the region/syndrome alert list, syndromes are often
presented in an aggregated view for age. Within the configuration options, age can be
delineated into subgroups. See the graphic below for a better idea of the selections a
user can have with this option.
15
ED Spatial Alerts
The spatial alert list is different from the temporal alert lists in that it only looks for
clusters over space, in this case by zip code centroids. If a particular combination of
zip codes has abnormally high counts of ED visits for a given syndrome as compared to
many combinations of the surrounding zip codes, it may be noted as a red or yellow
alert. The alert list will present information on the p-value, the observed count, the
number of zip codes in the cluster, and the center zip code in the cluster. The interface
has the option to click a “map view” and “time series” view so that you can drill down into
these data.
Note: For a more detailed explanation please see the paper by Howard Burkom in
the Johns Hopkins APL Technical Digest (Volume 24; No. 4; 2003) “Development,
Adaptation, and Assessment of Alerting Algorithms of Biosurveillance”. A google search
of the title will return a few locations where the paper can be downloaded.
16
Region Caller Site (Alert List for Poison Control Data)
These alerts are for the Poison Control (PCC) data. They provide a snapshot of exposure
calls made to the PCC by caller site. The alerts show if the number of calls on a
particular day exceeds the baseline number of calls for a health care facility, workplace,
school, residence, restaurant, or public area in a particular county. You can examine a
time series graph of the data by clicking on the time series link.
County/Generic Substance (Alert List for Poison Control Data)
This feature also provides alerts for the PCC data. It gives users a snapshot of exposure
calls made to the Poison Control Center by chemical substance.
17
Hospital/Sub-Syndrome Time of Arrival
The hospital/sub-syndrome time of arrival alert list looks for clustering of patients based
on the time they arrive at the hospital ED. Analysis of these data is by hospital, patient
sub-syndrome, and the time the patient arrived at the ED. A minimum of 4 visits is
required to produce an alert (i.e. p-value ≤ 0.0001). This tool will help find clusters of
patients who arrive at a hospital with similar symptoms and within a small time interval,
but whose total volume is not enough to create a statistical alert in the daily total alert
lists. The development of this tool was based on known outbreaks that were detected
using a manual version of this approach.
Make the selections you would like from the available fields in the data configuration
view, and then click “Change Configuration.” When the table loads you will see on the
left side Region (i.e. county), Hospital, and Alert Count. Along the top are the 30 minute
time intervals, which is the minimum time interval used in the time of arrival analysis.
You can change this table view by clicking on the “+” sign next to “Table Configuration
Options”, and then selecting sub-syndrome in the row field drop down. Open the “+”
next to “Configuration Options” and click change configuration to update the table view.
The red boxes in the table are the time intervals where a statistical alert was produced.
To drill down, click on the red box. A table will load at the bottom of the page with the
information on the alert you clicked on.
Follow the time series or data details link for additional information on these data.
18
School Absenteeism Alerts
School absenteeism alerts are alerts for the school absenteeism data. An algorithm
developed specifically for school absenteeism data is utilized to generate these alerts.
This algorithm uses the day of the week to determine whether or not the observed
absenteeism proportion for any given day is higher than expected. That is to say, if you
are looking at absenteeism data for a Monday, the algorithm uses all historical Monday’s to
determine if the absenteeism proportion for the Monday in question is higher than
expected. Using the data configuration button will allow you to choose individual school(s)
as well as the date range.
There is also an alerting mechanism for school absenteeism data based on thresholds.
Since the 2009 H1N1 pandemic, each Maryland public school system has had in place a
school absenteeism threshold. When this threshold is exceeded, the public school
system contacts their respective local health department. These thresholds have been
programmed into ESSENCE, and ESSENCE will generate an alert when the threshold is
exceeded. These alerts can be seen by clicking on the “Schools Threshold” link on the
Alert List page.
19
Reviewing Alerts in ESSENCE
Please refer to Appendix 1 to see a flowchart of how to analysis syndromic surveillance
alerts in order to identify possible events of public health importance.
B. MY ALERTS
The alert list feature (described in A - alert list above), provides alerts for 10 of the 11
syndrome categories in ESSENCE. “myAlerts” on the other hand, allows users to customize
which stratifications of the ESSENCE data they are interested in monitoring for routine
daily surveillance. It also enables users to set criteria for alerting that include statistical
thresholds, minimum counts, and consecutive days of alerting. Alerts can be created for
the standard syndrome and sub-syndrome categories, free text queries of the ED data,
as well as for any of the other data sources available in ESSENCE.
Creating myalerts and Records of Interest
To create a my alert for a sub-syndrome category (for example),
1) Click on query portal, and then select your data source. Data sources include
emergency room data, Poison Control Center data, OTC medication sales data,
and school absenteeism data.
2) In the next screen select your geography system and the medical grouping
system “chief complaint sub-syndrome”. You may also select one of the five time
20
resolution options to examine daily, weekly, monthly, quarterly, and yearly time
resolutions of the data. Statistical alerting algorithms only work on daily and
weekly time resolutions. For weekly time resolutions consider using CDC-C2.
When done, click submit.
3) In the next screen, choose the sub-syndrome for which you wish to create the alert
from the “select chief complaint sub-syndromes” field, make all other desired
selections and click on time series at the bottom of the page.
4) In the next screen, you will notice at the top of the time series graph a “query
name” field. Type in the name of your query and click “save query”.
21
After you click “save query,” a box will appear that will allow you share your query with
other registered ESSENCE users. Select the users you would like to share your query
with and then hit “save.” If your query was successfully saved, you will see a dialog box
similar to this:
Click OK. To create your sub-syndrome alert, click on the “create myalert” button on the
same page. The “create myalert” dialog box will pop up. Here, you can check the box
next to detection, choose a variable by which to stratify the detection algorithm, select
threshold p-values (or leave as default), and/or input the minimum count of ED visits for
which the alert should be created, then click save my alert to save the alert. You could
also check the box next to “records of interest” (discussed below) if you are interested in
monitoring rare events with small counts. Your myalerts may be accessed and
monitored routinely through the “my alerts” feature in the main toolbar. You also have
the option of sharing your myalert with other registered ESSENCE users.
22
Alternatively, once you have clicked on the save query button and your query has been
successfully saved, you can click on the “query manager” tab in the main toolbar where
you can create the alert. If you have some previously saved queries, your new query
will appear at the bottom of the list. Check the check box next to your query name and
click on “create myalert” at the top of the page.
23
Creating alerts for your free text queries:
To create a my alert for your free text query, select “chief complaints” as the medical
grouping system in step 3 above, click submit and in the next screen, enter your free
text query in the “select chief complaints” field.
Proceed to steps 4 and 5 as outlined above.
Records of Interest:
“Records of interest” allows ESSENCE to produce a line list of ED visits in ESSENCE
that match a specific query. It is useful in those instances where the user wants to
review each ED visit for a particular query. For instance, if you want to look at any ED
visits with specific reportable disease names, you would use the record of interest
feature. The “records of interest” feature is useful for monitoring rare events with small
counts.
To create a record of interest, follow the same steps as you would to create an alert
(steps 1 through 5 above). In the “create myalert” pop up (shown in step 5 above),
check the box next to “records of interest” and click “save myalert”.
24
C. EVENT LIST
The purpose of the event list feature in ESSENCE is to allow users who find alerts or
other information in the system that may warrant further attention to complete an entry
describing their findings or recommendations. The event list is not currently promoted
for use in Maryland because most ESSENCE activities are conducted by OP&R
epidemiologists who work in close concert with one another. However, if users would
like, they are welcome to use the event list if they feel it would benefit them locally. Contact
ESSENCE Help (mdessence@maryland.gov) if you would like more information.
In states where this feature is utilized, users can view events within the summary alerts
feature in alert list. These events would show up in the second tier of asterisks (please
see section A on alert list above). In Maryland ESSENCE, clicking on any of these
asterisks would not return any meaningful results.
D. OVERVIEW PORTAL
Overview portal allows a user to examine multiple time series graphs for the previous
three months (a default time frame). Some of the parameters for which these time
series can be created are region of the hospital (county), syndrome, and medical
subgrouping. The medical subgrouping option enables the user to view graphs by subsyndrome. Sub-syndromes are clinically defined subgroupings of the 11 main
ESSENCE syndromes.
There are two main ways to get to the overview portal. One way is through the query
portal, and the other is through the overview portal tab in the main toolbar.
To access the overview portal through the query portal,
1) Click on “query portal” in the main toolbar
2) Select your desired data source option
3) Next, choose a geography system, a medical grouping system, and a time
resolution
4) In the next screen, make all necessary selections and change the date range if
need be, then click on “overview portal” at the bottom of the page
5) Once the graphs appear on the screen, you may go back and change your
previous selections by clicking on the plus sign next to configuration options.
25
The other way to get to the overview portal, clicking on the overview portal icon in the
main toolbar, provides a quick way to visualize time series graphs stratified on a
specified parameter (hospital, region, syndrome, medical subgrouping, age group, sex,
or disposition category).
1) Click on “overview portal” in the main toolbar
2) Select a data source
3) Select an overview parameter in the next screen and click submit.
You can click on any portion of one of the small graphs (or click on “switch to interactive
view” at the bottom right corner of the graph) to activate the graph and examine the data
more closely.
26
ESSENCE also allows you to bookmark your queries so that you don’t have to repeat all
of these steps each time. At the top of each page is the bookmark page button:
Type in the name you would like to give the query and click on “bookmark page”.
Bookmarks can be accessed by clicking on the bookmarks link in the main toolbar.
Using the overview portal, practical example:
One use of ESSENCE is for situational awareness. You may want to monitor the
ESSENCE syndromes and sub-syndromes to identify potential events of public health
importance. The simplest way to do this is to use the overview portal to obtain small
graphs stratified on your variable of interest. In this example, we will look at graphs of
sub-syndromes to see if there are any alerts that may require further investigation.
1) Click on the overview portal tab in the main toolbar
2) For this example, choose your data source as either “emergency room data by
patient location” or “emergency room data by hospital location”. The first option will
provide the ED data for patients based on their zip code or county of residence. The
second option, “emergency room data by hospital location”, will give you information
on all ED visits irrespective of the patient’s county of residence.
3) Select “medical subgrouping” from the list of overview parameters and click submit.
This should return graphs of all the sub-syndromes in ESSENCE. You can click on
any portion of any of the graphs, and then click on a data point on the graph to
examine the data for a particular day more closely. You can also change your date
range and other parameters by going to the configuration options. In the screen
shot below, only 4 of the <100 sub-syndrome graphs generated by ESSENCE are
shown.
27
E. QUERY PORTAL
The query portal allows users to perform various searches of Maryland data in ESSENCE.
Based on your level of access, you may view and analyze emergency department data,
PCC data, OTC sales data, and/or school absenteeism data.
This section focuses on emergency department (ED) data. Subsequent sections will
explain how to analyze PCC data, OTC Sales data, and school absenteeism data.
To access the ED data,
1) Click on “query portal” in the main toolbar.
2) Select data source.
28
The first three options in the data source list provide access to the ED data.
“Emergency room data by patient location”, categorizes the ED visits by the patient’s
location (zip code and county). This option will provide ED data based on the patients
zip code or county of residence. The second option, “emergency room data by
hospital location”, will give you information on all ED visits irrespective of patient zip
code or county of residence. The next several options will allow you to access the
PCC data, OTC sales data, and school absenteeism data which will be discussed
later in this manual. In the example below, we will select the “emergency room data
by hospital location” option.
3) In the next screen, select a geography system, a medical grouping system, and a
time resolution option.
For the geography system, you can select “hospital region”, which will allow you
to select a county or grouping of counties. You can also select “hospital” if you
would prefer to see graphs of the ED data for a particular hospital or hospital
grouping.
Time resolution allows you to look at daily, weekly, monthly, quarterly or yearly
time resolutions of ESSENCE data.
For the medical grouping, you have the option of selecting one of three medical
groupings: ESSENCE syndromes, chief complaint sub-syndromes, or chief
complaints. The first two options bring you to pre-categorized groupings of chief
complaints. The ESSENCE syndromes include 11 syndrome categories, and the
sub-syndrome groupings include over 100 sub-syndrome categories. For more
information on the syndrome definitions, please click on the syndrome definitions
link at the top of the page.
We will focus on “ESSENCE syndromes” for now and discuss the other two
options later. Once you have made all the necessary selections, click submit.
29
4) In the next view, you can select your variables of interest. Please note that in the
query portal “region” means county/jurisdiction. If the age group you want to look
at is not one of the pre-set age categories, click on the “select age range” drop
down menu to customize the age or age range.
To make multiple selections of any of the variables, hold the control key down on your
keyboard and click on your choices. For example, to select both the Fever and GI
syndrome, hold the control key down and click on Fever and GI. Select a minimal
number of syndrome categories from the “select syndrome” field to avoid duplication
of data. Once you have made all your selections, you can click on “time series” at the
bottom of the page to see a time series graph of the data.
30
Alternatively, you could click on “table builder” to see a tabular view of the data. Once
you have clicked on table builder, you would need to select a column variable and a row
variable in the next screen and click on “create table”. The tables that ESSENCE will
generate can contain up to 30,000 cells at a time and will look similar to this:
Clicking on “data details” would produce pie charts, bar graphs and a line list of the
data.
Clicking on “overview portal”, will take you to the following screen:
Here, you could click on the drop down menu to select a variable by which to stratify the
data. For example, if you select “region of the hospital” ESSENCE would stratify the
data by county and produce 24 graphs, each representing a county/jurisdiction in
Maryland.
31
Creating a bookmark for your query:
Once you have run a query, you can bookmark that query so that you don’t have to go
through all of the steps each time. You can do this by typing in the name you would like
to give the query and clicking on “bookmark page” in the top left corner of the page
below the main toolbar.
You can access your bookmarks later by clicking on the bookmarks tab in the main
toolbar.
Free text chief complaint queries
ESSENCE data can also be queried using words or strings of words known as free text
queries. To query the free text chief complaints,
1) Click on “query portal” in the main toolbar.
32
2) Select ED data source.
3) Select a geography system, a medical grouping system, and a time resolution. To
query the free text, you will need to select the “chief complaint” medical grouping
and then click submit.
4) Next, in the “select chief complaints” field, you can type in your free text chief
complaint query using wildcards (^) and Boolean operators, examples of which are
“and”, “andnot”, “or”. The purpose of the “^”, or wildcard symbol, is to tell the system
that you want to see chief complaints with the word(s) enclosed in the “^”, plus any
other words. If you don’t use the “^”, then you will only get chief complaints with that
specific word(s). Wildcards and Boolean operators must be separated by commas.
Examples of a few free text queries include:
Carbon monoxide poisoning:
^carbon^,or,^carbon monoxide^
Fever and stiff neck:
^fever^,and,^stiff neck^,or,^mening^
Pneumonia without mention of aspiration pneumonia:
^pneumonia^,andnot,^aspiration^
To make multiple selections of any of the variables, hold the control key down on your
keyboard and select your choices, which will all be highlighted in blue.
You can also write more complex queries based on your specific needs. Please see
appendix 2 for additional examples of free text queries. After you have typed in the free
text query and made all the necessary changes to the other fields, you can click on
“time series” at the bottom of the page to see a time series graph of the data.
You can click on the “data details” hyperlink directly below the time series graph to see
a line list of ED visits for your query for the entire range of dates included in your time
series. If you would like to see only one particular day’s data details, then you could
either click on the data point in the graph or click the “data details” hyperlink in the data
table below the time series graph.
To change your previously made selections, click on the plus sign next to configuration
options above the time series graph. You can also click on the plus sign next to
description to see a summary of all variable selections you have made to generate your
results.
Using the query portal, practical example:
Suppose you have heard that unsafe attempts at home heating in Maryland lead to an
increase in carbon monoxide poisoning in January, and you want to explore this issue.
33
1) First, click on “query portal” in the main toolbar.
2) Next, select “emergency room data by hospital location” as your data source option
and click submit.
3) Then select “chief complaints” as your medical grouping system to allow you to
specifically query ESSENCE for ED visits pertaining to carbon monoxide. Since you
would like to look at trends for the entire state, you can leave the geography system
as “hospital region”. To examine the data for a particular hospital or hospital
grouping, you would select “hospital” as a geography system.
4) In the next screen, type your free text query in the “select chief complaints” field. For
this example, we will use ^carbon^,or,^carbon monoxide^ to find ED visits related to
carbon monoxide. Also leave the “select region of the hospital” as the default, “all
regions of the hospitals” and leave all the other fields as the default, unless you want
to look at specific stratifications of the data. The time period in this example is
September, 2009 through June, 2010.
As you can see from the time series graph, there were significantly higher numbers of
ED visits associated with carbon monoxide poisoning than expected in the month of
January. You can click on the “data details” hyperlink directly below the time series
graph to see a line list of ED visits for your query for the entire range of dates included
in your time series. If you would like to see only one particular day’s data details, then
you could either click on the data point in the graph or click the “data details” hyperlink
in the data table below the time series graph. This information can be used along with
supporting information from other sources to determine whether interventions should be
aimed at a particular geographic area and/or demographic group. A map view might
also be helpful in this case (not shown).
This example is only one of many ways the query portal can be used. As you explore
this feature, you will identify other utilities that will benefit your geographic area.
34
Again, please remember that the nature of chief complaint text may vary by hospital,
and the free text chief complaint queries may change based on language used in the
chief complaints from your particular jurisdiction. It is therefore important to be familiar
with the chief complaints from your area to find the best combination of terms that will
provide the best results for a query.
F. MATRIX PORTAL
The matrix portal allows you to create a custom report for ER data by patient location
along with a bar chart or line graph for your county of choice. You can also create
custom reports for OTC medication sales data. To look at aggregate data for all the
counties, please select all the counties in the region field. Also, select all the age groups
if you are not interested in looking at the data for any one age group. Below is one
example of a line graph, but many different line graphs and bar charts can be created in
the matrix portal.
1) To create a line graph that shows all ED visits for Maryland by syndrome for a
particular time period in matrix portal, first click on “matrix portal” in the main toolbar,
and then select “emergency room data by patient location”. Holding the control key
down on your keyboard, click on your syndromes and age groups of choice. To
create a graph for the entire state, please select all regions. In the columns category,
select “syndrome” and in the rows category, select “date range”. Select “line graph”
from the graph type field. In the start date and end date section enter the time period of
interest and click on “change configuration” below the date fields.
35
2) At the top of the matrix portal page, you will see a description of everything included
in the data table. Scroll down to see the data matrix and line graph. The matrix
shows all of the cells in which a red or yellow alert has occurred. The graph is a line
graph that shows the trend in each syndrome during the specified period for the
state of Maryland. Here, only a small section of the data table is shown.
3)
3) By clicking on a cell within the data matrix, you will see a table with hyperlinks for
ESSENCE to create a time series graph, data details (line list), map view, or an alert
list.
4) By selecting the time series option, you can see the time series for the last three
months of available data, including the time period selected within the matrix portal.
By selecting data details (not shown), ESSENCE will produce pie charts, bar charts,
and a line list of the chief complaints data for the specific day you clicked on in the
data matrix. This view would be similar to any data details page. By selecting the
map view (not shown) you will see a map for the specified syndrome and day you
36
clicked on in the data matrix. By selecting the Alert List option (not shown), you can
view more details about the syndrome alerts for the day you picked in the data
matrix.
G. WEEKLY PERCENT
Weekly percent allows a user to view graphs of percent ED visits for one of the
available medical grouping options by week. This function also works for OTC
medications sales. You can choose one or several options in each category (year,
category or medical grouping, data source, geography, and age group). For geography,
region stands for Maryland’s health and medical regions. Selecting “aggregate” by any
of the fields will combine the different options for that particular category into one,
producing one trend line instead of several for the same variable. For example, if you
click on aggregate next to age, this will collapse the different age categories (00-04,
05-19, etc.) into one group for your analysis.
As an example, let’s look at a graph of influenza-like illness in Maryland for the past three
years. To do this,
1) Click on the “weekly percent” tab in the main toolbar.
2) Holding the control key down on your keyboard, mouse over 2012, 2011, and 2010
in the “year” field.
37
3) Select “ILI” from the “category” field.
4) Click on the “emergency room data by patient location” data source option. This is
the only available data source in weekly percent.
5) Select “Maryland” from the geography field.
6) Holding the control key on your keyboard down, mouse over all the age groups, and
then check the “aggregate” check box and click on “show graph” below the page.
The graph produced will show the proportion of total weekly ED visits that are due to ILI
for the state of Maryalnd for the three year period. Weeks run from Sunday through
Saturday in weekly percent.
38
To change your previously selected options, click on “select a different graph” above the
time series graph and this will allow you to make your changes.
H. MAP PORTAL
The map portal is where the user can map statistical alerts (temporal and spatial)
39
across the different data sources. When you click on the map portal link a map
configuration page loads. From this page the user can select what data source they are
interested in seeing in the “alert list” selection list. The outcome selections then change
as appropriate when the user selects a given data source. Examples include:
1) Region/Syndrome: this option shows the user counties where there were temporal
alerts for a selected syndrome. If the user selects more than one day in the date range,
the map will map the alert with the lowest p-value.
2) Hospital/Syndrome: this option shows the user hospitals where there were temporal
alerts for a selected syndrome. If the user selects more than one day in the date range,
the map will map the alert with the lowest p-value.
3) Spatial: this option shows the user the zip codes that were included in a spatial alert
for a selected syndrome.
4) Region/Caller Site: this option shows the user the county where temporal alerts
occurred for poison control center calls associated with the origin of the call. If the user
selects more than one day in the date range, the map will map the alert with the lowest
p-value.
5) Region/Generic Substance: this option shows the user the county where temporal
alerts occurred for poison control center calls associated a particular substance. If the
user selects more than one day in the date range, the map will map the alert with the
lowest p-value.
6) Hospital/Sub-syndrome time of arrival: this option shows the user the hospitals that
were included in a sub-syndrome time of arrival alert.
7) Region/OTC Category: this option shows the user the county where temporal
alerts occurred for specific categories of OTC medication sales. If the user selects
more than one day in the date range, the map will map the alert with the lowest pvalue.
8) School Absenteeism: this option shows the user the county where temporal alerts
occurred for school absenteeism. If the user selects more than one day in the date
range, the map will map the alert with the lowest p-value.
I. BOOKMARKS
The bookmarks feature allows the ESSENCE user to save searches and use the same
criteria to recover data on a daily, weekly or as-needed basis without doing the stepwise
work. To save a search, you type in the name you would like to give the query and click
on “bookmark page” in the top left corner of the page below the main toolbar.
You can access your bookmarks at a later date by clicking on the bookmarks tab in the
main toolbar. ESSENCE will bring you to the following page, and you can click on the
40
name of the bookmark or “select today” to re-run your query.
Working with your bookmarks, “select” versus “select (today)”:
Once your bookmarks have been saved, you can click on the name of the bookmark
in the bookmarks page to run the query as you created it, without a change in dates. For
example, if your selected date range for the query was from January 1, 2010 to March
30, 2010, this option will produce results for that same date range. If you click on “select
(today)”, you will see the results of your query for the most recent three month period, up
to the most recent day.
Editing or deleting your saved bookmark:
1) Click on the bookmarks tab in the main toolbar
2) Click on the check box next to the bookmark you wish to edit or delete (or delete if
you would like to permanently delete the bookmark). Edit and delete are above the
list of bookmarks.
41
3) Follow the prompts to edit or delete your bookmark. Editing bookmarks also
includes grouping them into custom named groupings (sort of like email folders)
J. QUERY MANAGER
42
The query manager allows users to save and manage queries from sessions
in the query portal (Described in Query Portal). This saves the user from
having to re-write chief complaint free text queries should it become
necessary to re-run the same query or a similar one. Query manager is similar
to bookmarks, but it also has other useful features that are not available in the
bookmarks. For example, you can create your “my alerts” in the query
manager (described in B. Alert List). You can group your saved queries into
custom named groupings just like bookmarks. You can also share queries
with other ESSENCE users. Also, you can create multi-series time series
graphs, which are described below.
Multi-Series time series graphs:
Multi-series time series graphs allow a user to overlay trend lines from
different time series graphs, including from different data sources. For example,
to compare trends in influenza-like illness (ILI) activity for two counties, two
separate time series graphs can be generated in ESSENCE, one for each
county. From the two graphs, a single multi-series time series graph can be
produced in ESSENCE to compare trends in ILI activity for the two counties.
To create a multi-series time series graph of saved queries in the query
manager,
1) Click on “query manager” in the main toolbar.
2) Check all check boxes next to the names of the queries you would like to
create the multi-series time series graph for.
3) Click on “multi-series time series” at the top of the page.
4) This should bring you to the “configure multi-series time series display”.
For style, select “single”, and the system will plot all your selected queries
from step 2 on the same chart. “Multiple large” will plot each query on its
own large chart, and “multiple small” will plot each query on its own small
chart.
For the date alignment, “actual dates” uses the dates saved with each
43
individual query to run detection on that query. “Global dates” uses the dates
provided in your queries to run detection on all selected queries. “Start
together” uses each query's individual dates, but aligns them so that they all
start at the leftmost side of the graph. And “end together” uses each query's
individual dates, but aligns them so that they all end at the rightmost side of the
graph.
Further instructions are provided in the text below the configuration options.
The chart below is an example of a multi-series time series graph
comparing rash illness in one county to rash illness for the entire state of
Maryland. Separate queries were run to generate two time series graphs, and
the multi-series time series graph was created by going into query manager,
selecting the queries of interest, and proceeding as described above to create
the graph.
44
Deleting your saved query:
1) Go into the query manager
2) Click in the checkbox next to the query you would like to delete
3) Click delete at the top of the page to permanently delete the query.
You can share a saved query with other ESSENCE users by checking the query
you wish to share and then clicking on the share button at the top of the page.
K. myESSENCE
The myESSENCE feature allows the user to create multiple “dashboards” using
the various ESSENCE data sources available in the Maryland system. Each
dashboard is contained in a tab. These tabs work much like the tabs in an internet
browser. The tabs can be configured in various ways (different column numbers
and sizes) and may contain numerous widgets including time-series graphs,
myAlerts, maps, rich text labels, etc. The myESSENCE feature allows the user to
see data for multiple data sources using multiple visual tools in a quick and easy to
understand format. The graphs and other widgets contained on each tab can be
configured within the tab, and can be made to automatically refresh so that queries
etc. do not have to be re-run.
To create a myESSENCE tab:
1. Click on the myESSENCE button in the main toolbar.
45
2. Click on “Add New Tab” at the top left of the screen.
This will bring up a menu which allows you to create a name for the tab, as well as
select the column configuration for the tab.
3. Once you name the tab and select the column configuration, you can begin
adding widgets to the tab. To do this you click on “Add New Widget” at the
top left of the screen. This brings up a drop-down menu of widget options.
46
4. Choosing Time Series Graph will bring up a menu which allows the user to
choose from all of the saved queries in the Query Manager. At the bottom
left of this menu there is another drop-down list which allows the user to
adjust the time frame for the time-series graph. The user can choose the
dates that the query was saved with by choosing “Default Saved Query.”
The following is an example of a myESSENCE “dashboard” for influenza-like
illness surveillance.
47
Each widget has several buttons at the top right corner which allow the user to
minimize the widget, refresh the widget, adjust the configuration of the widget,
or to delete the widget.
myESSENCE tabs can be exported to a PDF file by clicking on the “Export to PDF”
button at the top left of the page. They can also be shared with other ESSENCE
users by clicking on the “Share Tab” button at the top left of the page.
48
49
Appendix 1: Flowchart for Analysis and Response to Syndromic Surveillance Data
50
One way to approach the review of syndromic surveillance data is to begin with those categories that produced statistical alerts.
However, in-depth analyses of syndrome categories and/or more specific free text chief complaint queries, regardless of alerting
status, have been used successfully to discover previously unknown outbreaks. For example, say you use the query
^vomit^,or,^diarrhea^ and the normal number of cases in your county is ~30 per day. Then on X day you find that there are 25
patients in this query (there is no alert produced), but when you review the line list by time of ER visit and zip code you find that
8 patients visited the ED very close in time and lived in the same zip code. This clustering of cases is worth following up on
even though no statistical alert was generated.
It is important to become somewhat familiar with what a patient’s chief complaint looks like in your county, and to understand the
limitations. Generally the chief complaint is what the patient says is wrong with them to the triage clerk at the ED. The triage
clerk could be a nurse, or someone who does not have much medical training. This can impact how much medical
interpretation goes into the chief complaint text. A hospital might also utilize a standardized drop down selection of symptoms
rather than typing in a free text description of the patient’s statement, which often makes the chief complaint less descriptive.
51
Appendix 2: Free text query examples
Note: These examples are shown to illustrate how various queries can be structured
within the ESSENCE system. These queries are not necessarily used by surveillance
epidemiologists who use the ESSENCE system.
Display the total ER census
^ (you can also simply leave the default “all” text, and click submit)
Carbon monoxide query
^carbon^,or,^carbon monoxide^
Rabies query (people visiting the ER for rabies shots)
^rabies^
Animal Bite query
^cat bite^,or,^catbite^,or,^puppy^,or,^kitten^,or,^dogbite^,or,^dog bite^,or,^animal
bite^,or,^raccoon^,or,^racoon^,or,^fox^,or,^bobcat^,or,^bat
bite^,or,^bitten^,andnot,^bugs^,andnot,^ants^,andnot,^child^,andnot,^fireants^,andnot,^i
nsect^,andnot,^person^,andnot,^snake^,andnot,^rattle snake^,andnot,^pygmy
rattler^,andnot,^red ants^,andnot,^red fire
ants^,andnot,^something^,andnot,^spider^,andnot,^wasp^,andnot,^leg spider^
GI query 1
^blood^,and,^stool^,or,^blood stool^,and,^blood urine^
GI query 2
^fever^,andnot,^congestion^,andnot,^sore^,andnot,^cought^,and,^diarrhea^,andnot,^sor
e^,andnot,^back^,andnot,^cough^,or,^fever^,andnot,^sore^,andnot,^cough^,and,^vomit^
,andnot,^congestion^
GI query 3
^diarrhea^,or,^vomit^,andnot,^preg^,andnot,^pregnancy^
GI query 4
^vomit^,andnot,^headches^,andnot,^headache^,or,^diarrhea^,or,^abdominal
pain^,andnot,^polyps^,andnot,^headache^,andnot,^high blood suger^
Food poisoning query
^food poison^,or,^food^,andnot,^NO FOOD^,andnot,^fatty food
dinner^,andnot,^UNABLE TO EAT SOLID FOOD^,andnot,^CANT DIGEST
FOOD^,andnot,^EATING
LESS^,andnot,^PNEUMONITIS^,andnot,^stuck^,andnot,^choking^,andnot,^choke^,and
not,^hot^,andnot,^esophagus^,andnot,^allergic^,andnot,^allergy^,andnot,^tray^,andnot,^
gagging^,andnot,^trapped^,andnot,^tolerate^,andnot,^throat^,andnot,^bolus^,andnot,^c
onsumption^,andnot,^intake^,andnot,^tube^,andnot,^impaction^,andnot,^swallow^,andn
52
ot,^in throat^,andnot,^down^,or,^food poisoning^,andnot,^EATING LESS^
Skin infection query
^MRSA^,or,^staph infection^,or,^spider bite^,or,^skin lesion^,or,^staff infection^
Fever and Rash query
^rash^,and,^fever^,or,^chickenpox^,or,^chicken pox^,or,^measles^
Viral query
^virus^,andnot,^BRONCHIOLITIS^,andnot,^upper^,andnot,^SYNCYTIAL^,andnot,^stom
ach^,or,^viral^,andnot,^SYNCYTIAL^,andnot,^stomach^,andnot,^gastritis^,andnot,^gast
roenteritis^,andnot,^upper respiratory infection^,andnot,^uri^,andnot,^HUMAN
IMMUNODEFICIENCY^
ILI query
^fever^,and,^sore
throat^,or,^fever^,and,^cough^,or,^flu^,andnot,^shot^,andnot,^leaking^,andnot,^reflux^,a
ndnot,^flush^,andnot,^fluid^,andnot,^flushed^,andnot,^flutter^,andnot,^diarrhea^,andnot,
^nausea^,andnot,^vomit^
Possible Reportable Disease query
^crypto^,or,^cryptosporidiosis^,or,^cyclos^,or,^cyclosp^,or,^mercury^,or,^encephalitis^,o
r,^botulism^,or,^smallpox^,or,^shigellosis^,or,^shigella^,or,^salmonellosis^,or,^salmonell
a^,or,^hepatitis^,andnot,^alcoholic^,or,^anthrax^,or,^ciguatera^,or,^dengue^,or,^malaria
^,or,^measles^,or,^mumps^,or,^varicella^,or,^chicken pox^,or,^lead
poisoning^,or,^pertussis^,or,^meningitis^,or,^whooping
cough^,or,^campylobacteriosis^,or,^cholera^,or,^creutzfeldt^,or,CJD,or,^diphtheria^,or,^
ehrlichiosis^,or,^ecoli^,or,^escherichia
coli^,or,^glanders^,or,^haemophilus^,or,^leprosy^,or,^hansen^,or,^hantavirus^,or,^legio
nell^,or,^leptosp^,or,^listeriosis^,or,^listeria^,or,^lyme^,or,^melioidosis^,or,^meningococ
cal^,or,^meningococcemia^,or,^mercury^,or,^pesticide^,or,^plague^,or,^polio^,or,^psitta
cosis^,or,^q fever^,or,^rabies^,or,^rocky
mountain^,or,^rubella^,or,^streptococcal^,or,^toxoplasm^,or,^trichinosis^,or,^tularemia^,
or,^typhoid^,or,^typhus^,or,^vibrio^,or,^yellow fever^,or,^hemorrhagic^,or,^expos^
53
Emergency Department Section References
1) Centers for Disease Control and Prevention (2001). Updated Guidelines for
Evaluating Public Health Surveillance Systems: recommendations from the guidelines
working group. Morbidity and Mortality Weekly Report, 50(RR-13), 1-35. Available at:
http://www.cdc.gov/mmwr/preview/mmwrhtml/rr5013a1.htm
2)Centers for Disease Control and Prevention (2004). Framework for Evaluating Public
Health Surveillance Systems for the Early Detections of Outbreaks: recommendations
from the CDC working group. Morbidity and Mortality Weekly Report, 53(RR-05), 1-11.
Available at: http://www.cdc.gov/Mmwr/preview/mmwrhtml/rr5305a1.htm
54
Appendix B
Excerpt of the Maryland Department of Health & Mental Hygiene
Health Alert Network User Guide
Revised: June 9, 2010
INTRODUCTION TO THE MARYLAND HEALTH ALERT NETWORK
When an event threatens the health of Marylanders, fast, efficient, and reliable communication to
those responding to the event can prevent illness and save lives. The Maryland Health Alert
Network (MD HAN) provides a uniform alerting and notification system capable of rapid
distribution of health alerts and secure collaboration among a defined set of public health
professionals working to protect the public during a disease outbreak, environmental threat,
natural disaster, or act of terrorism.
The MD HAN system from the Maryland Department of Health and Mental Hygiene (DHMH) is
provided with federal funds from the US Centers for Disease Control and Prevention (CDC) and
is in compliance with federal alerting and notification requirements as outlined in the CDC
Public Health Information Network (PHIN) Partner Communications and Alerting (PCA)
requirements. The program intends to ensure that each community has rapid and timely access to
emergent health information; a cadre of highly trained professional personnel; and evidencebased practices and procedures for effective public health preparedness, response and service on
a 24/7 basis.
HCHD Policy # ADMIN 00-08