Sample Final Report

Attachment F: Sample Final Report
Community Assessment for Public Health Emergency Response (CASPER) after the major ice storms,
Kentucky 2009

BACKGROUND
On January 26, 2009, a massive ice storm hit Kentucky causing 36 deaths and leaving 770,000
people without power across the state. The storm continued for three days with ice over an
inch thick reported in many locations and snow accumulating across the state. Property
damage was widespread and basic communication was drawn to a halt due to fallen trees and
power lines weighed down by the ice. Public health officials were concerned of the health
status and needs of the western portion of Kentucky where communication with the state
health department was nonexistent due to the storm. To address these concerns, on February
2, 2009, the Kentucky Department for Public Health (KY DPH) requested assistance from the
Centers for Disease Control and Prevention (CDC) in conducting a needs assessment in the
severely affected areas in the Pennyrile District of Western Kentucky. The purpose of the
CASPER was to determine health and safety needs of residents living in Western Kentucky who
were severely impacted by the ice storm.

METHODS AND MATERIALS
CDC and KY DPH conducted a Community Assessment for Public Health Emergency Response
(CASPER) along portions of western Kentucky that were severely affected by the ice storm. On
the basis of regional and local information, the KY DPH leadership decided to conduct the
assessment in four severely impacted counties in the Pennyrile District of Western Kentucky
(i.e., Livingston, Caldwell, Lyon, and Crittenden counties) where communication had been
severely hindered due to the storm. CASPER is an epidemiologic technique designed to provide
household-based information about an affected community’s needs following a disaster.
CASPER rapidly obtains accurate and timely data in a relatively inexpensive manner through
precise analysis and interpretation. The information gained is then shared in a simple format


This is a sample report. While the information in this sample report is based on real events, it is not a complete
account of the work conducted by the Kentucky Department of Health and the Centers for Disease Control and
Prevention in response to the ice storms.

with decision-makers to inform response efforts. CDC developed a one-page questionnaire in
coordination with the KY DPH. The questionnaire was designed to capture 1) demographic
information; 2) concerns about injuries and illnesses, including mental health concerns; 3)
medication availability and access to care; 4) information about basic utilities, transportation,
generator usage, and risky behaviors for carbon monoxide exposure; 5) supply needs, such as
food and water; and 5) communication usage, including information on warnings and gathering
health advice (see questionnaire: Appendix D).

A two-stage sampling method was used to select a representative sample of 210 households to
be interviewed across the four selected counties. In the first stage, 30 clusters (census blocks)
were selected with probability proportional to the number of housing units within the census
block according to the 2000 Census. In the second stage, interview teams randomly selected
seven households from each of the 30 clusters. The interview teams were provided with
detailed maps of each selected cluster and instructed to select the housing units for the seven
interviews by use of a standardized method for randomization.

A three-hour training session on interview techniques, safety issues, household selection,
tracking methods, and referrals was given on February 5, 2009 to the 15 two-person interview
teams. Teams consisted primarily of Kentucky public health staff from the local western
Kentucky region and Epidemic Intelligence Service Officers from the Centers for Disease Control
and Prevention. Each team attempted to conduct seven interviews in each of the 30 census
blocks selected for the sample, with a goal of 210 total interviews. Residents of the selected
households who were at least 18 years of age or older were considered eligible respondents.
Additionally, the field team members distributed flyers provided by KY DPH, with information
regarding kerosene and chainsaw safety, food safety, and carbon monoxide poisoning. Teams
were instructed to complete confidential referral forms whenever they encountered urgent
needs and to forward the forms to the KY DPH for immediate follow up.
Epi InfoTM 3.5.1, a free statistical software package produced by the CDC, was used for data
entry and analysis. We conducted weighted cluster analysis to report the estimated number of
households affected in the assessment area. To account for the probability that the responding
household was selected, we created sampling weights based on the total number of occupied

houses according to the 2000 Census, the number of clusters selected, and the number of
interviews completed in each cluster. This weight was used to calculate all weighted
frequencies and percentages presented in this report. The contact rate was calculated by
dividing the completed interviews by the total number of households where contact was
attempted; the cooperation rate was calculated by dividing completed interviews by the total
number of households where contact was made; and the completion rate was calculated by
dividing the number of completed interviews by 210 (i.e., the goal for completed interviews in
this CASPER).

RESULTS
On February 7, 2009, the interview teams were able to conduct 187 interviews, yielding a
completion rate of 89% (Table 1). The 187 interviewed households were a sample of the
19,497 total households in Livingston, Caldwell, Lyon, and Crittenden counties. Unweighted
frequencies, percentages, and projected population estimates based on weighted analyses can
be found in Table 2 through Table 5.

Household demographics, evacuation behaviors, and utilities are shown in Table 2 and Table 3.
Twenty-two (22%) percent of households with pets claimed that owning a pet prevented them
from seeking alternative shelter or tending to their own health needs. At the time of the
CASPER, 12 days after the storm began, the majority of residents felt safe and secure in their
homes (97%) and neighborhoods (92%). However, 9% of households did not have any source of
electricity and 17% of households were using generators as their source of electricity.
Approximately half (52%) of households were using bottled water as their source of drinking
water. The remaining households were using well or municipal water, and 77% of those
households were not treating their water. Additionally, 5% of households reported they did not
have enough water or food for the next three days.

An estimated 58% of households reported using a generator at some point since the ice storm
(Table 4). Of those households, less than half (39%) had reported owning a carbon monoxide
detector and only 73% of those with detectors reported it was currently working. In addition,
approximately 1% of households were using their generator indoors, 3% in the garage, and 5%
outdoors but near an open window. However, no severe headaches or dizziness was reported

among these respondents (data not shown). Approximately 80% of households using a
generator reported running them within 25 feet of their house. Additionally, of the 38% of
households who had used a charcoal/gas grill since the storm, almost a third self-reported
improper usage, with 21% using grills inside, and 9% of those using grills outside reported using
them near an open window or a door.

Self-reported health status, current needs, and access to care are shown in Table 5. Since the
ice storm, 5% of households reported at least one person who was injured and 17% reported at
least one person in the household who experienced one or more illnesses. Nausea, stomach
ache, or diarrhea were the most commonly reported illnesses, with 4% of households reporting
at least one person experiencing symptoms (25% of households reported one or more
illnesses), followed by 3% of households reporting at least one person with cough with fever
(19% of households reported one or more illnesses), and 2% of households reporting at least
one person with severe headache with dizziness (11% of households reported one or more
illnesses). Three percent of households indicated that someone in their home needed
supplemental oxygen, and 6% reported they were not getting needed medication. In addition
to these injuries and illnesses, 15% of households reported at least one person with mental
health concerns.

DISCUSSION
The data presented here represent reports from the CASPER surveys conducted in Livingston,
Caldwell, Lyon, and Crittenden counties on February 7, 2009. To create sampling weights,
information from the 2000 Census was used to determine the household probability of being
selected. Some areas may have experienced significant population changes since 2000, and
thus, Census data may not be representative of the current population in those areas. Changes
in population since the previous census may result in less reliable generalizations of weight
analyses to the sampling frame. The discrepancy between the 2000 Census and the current
status would not, however, affect the unweighted frequencies presented in this report.
Further, local knowledge of the Pennyrile District report relatively minor changes in population
over the past decade. The contact rate, 54.4%, indicates that the field interview teams had to
sample more households within the clusters to complete the necessary number of interviews,

and this additional sampling might affect the representativeness of the results. In other words,
interview teams completed, on average, one interview for every two houses selected.
Additionally, there is no available information from a baseline or comparison group that can be
used to interpret the percentages of illnesses reported.

This CASPER met the stated purpose of determining health and safety-related needs of
residents living in the severely affected regions of Western Kentucky to inform KY DPH stormresponse efforts based on the public health recommendations (see below). Although these
communities were still recovering from the ice storm, overall, most residents felt safe and
secure in their homes and neighborhoods and reported having enough food and water for the
next three days. Additionally, most residents reported the capability to procure personal
medication and transportation, if needed. As of the day they were interviewed (i.e., February 7,
2009), many residents in the Pennyrile District were still waiting for power to be restored by the
electric company, but several of them were using generators to power their homes. More than
half the residents had used a generator at some time since the ice storm. There were many
reports of improper or unsafe generator use that should be addressed in public health
messages to prevent carbon monoxide poisoning. Further, the majority of those using well or
municipal water were not treating their water. Additional information is needed to determine
which counties were under boil-water advisories at the time of the CASPER to assess the
number of households who may have been unaware they were consuming water that should
have been boiled.

RECOMMENDATIONS On the basis of the results of the CASPER, the following actions are
recommended:

1. Emphasize carbon monoxide exposure risks and employ early communication of
prevention messages by using appropriate media.
Since a high percentage of households reported improper generator and/or charcoal grill usage,
messaging about carbon monoxide exposure should be communicated to the affected area.
These messages should be distributed through a variety of media because a lack of electricity,
poor road conditions, and questionable telephone service might prevent people from accessing
their usual sources of information. Therefore, carbon monoxide exposure prevention messages

communicated after the storm should include radio announcements for those who have
battery-powered radios and, if cellular service is available, providers should be contacted as
soon as possible about the possibility of sending mass public service text messages. Ideally,
these messages should be prepared or obtained in advance of the disaster and, for future
events, should be communicated before the storm hits and as soon as possible after the storm.

2. Emphasize safe generator use
Safe generator location should be the main focus of these prevention messages. However,
during the course of the CASPER, several people indicated that they were aware of
recommendations to use generators away from the house but did not do so because they were
afraid of having the generator stolen or ruined by rainfall. Also, short extension cords
sometimes limit the user’s ability to place the generator at a safe distance from the house.
Therefore, in addition to specific information about unsafe generator placement behaviors
(e.g., indoors, in a basement or a garage, near an open window), prevention messages should
include specific advice, such as procuring a lock and chain, to secure the generator to a tree or
other solid structure; placing the generator on a flat surface that is not subject to puddling;
sheltering the generator under an outside table to prevent water damage and electrocution;
obtaining a long extension cord rated for outdoor use; and obtaining a carbon monoxide
detector that is either battery-powered or AC-powered with a battery back-up. Additionally,
vendors should be encouraged to provide safety information at the point of sale and to advise
customers to purchase locks, chains, long extension cords for outdoor use, and carbon
monoxide detectors, along with the generator.

3. Discuss other unsafe heating sources
A number of unconventional heating sources (e.g., charcoal grills and gas stoves) were used
during the power outages. Prevention messages should include specific information to the
effect that these heating sources are unsafe.

4. Consider establishing pet-friendly shelters
Over 20% of respondents reported that they did not seek alternative shelter due to their pets.
Future efforts to provide pet-friendly shelters are encouraged.

5. Respond to the needs of oxygen-dependent people
Kentucky has a high rate of lung disease; a noteworthy number of households indicated the
immediate need for supplemental oxygen. State health officials should immediately respond to
these needs and, in the future, should be prepared to respond to the needs of oxygendependent residents (e.g., ensuring enough oxygen canisters are available for those in need,
providing a place with a source for charging oxygen devices, and developing a plan to identify
and contact those in need).

6. Communicate available mental health resources
Sixteen percent of respondents in Livingston, Caldwell, Lyon, and Crittenden counties reported
mental health concerns. It is unclear from these data what (if any) portion of these mental
health concerns can be attributed to the ice storm specifically. Nonetheless, county officials
should promote community awareness of available mental health resources.

TABLES

Table 1. Questionnaire response rates for the February 2009 Western KY CASPER
Questionnaire response
Percent
Rate
*
Completion
89.0
187/210
†
Contact
54.4
187/344
‡
Cooperation
97.9
187/191
* Percent of questionnaires completed in relation to the goal of 210
† Percent of questionnaires completed in relation to all households where contact was attempted
‡ Percent of questionnaires completed in relation to all households where contact was made

Table 2. Self-reported housing type, perceived safety of respondents and pet ownership, Western KY CASPER, February 2009
Characteristic
Frequency (n=187)
% of households Projected households
Projected %
Housing Structure
Single family
148
82.2
15451
82.6
Multiple unit
14
7.8
1336
7.1
Mobile home
18
10.0
1915
10.2
Safety
Feel safe in house
181
96.8
18797
97.0
Feel secure in area
171
92.9
17629
92.4
Pets
Have pets
114
61.0
11993
61.9
Pet prevented from seeking shelter
25
21.9%*
2660
22.2%*
* Of households with pets (n=114)

95% CI
74.4–90.8
0–15.0
5.0–15.5
94.4–99.7
87.4–97.4
52.0–71.8
--

Table 3. Self-reported basic utilities, Western KY CASPER, February 2009
Characteristic
Frequency (n=187)
% of households
Food and Water
Inadequate drinking water
9
4.8
Inadequate food
8
4.3
Source of Drinking Water
Bottled
Well
Public/municipal
Not treating water
Current Source of Electricity
Power company
Gasoline generator
None
Current Source of Heat
Electricity
Propane/gas
Wood
Kerosene heater
Other
Had working toilet
Had working phone

Projected households

Projected %

95% CI

923
887

4.8
4.6

1.0–8.5
0–9.2

91
11
83
70

49.2
5.9
44.9
74.5*

10000
1050
8110
7081

52.1
5.5
42.4
77.3*

41.3–62.9
1.5–9.5
32.1–52.6
--

137
29
19

74.1
15.7
10.3

14190
3200
1789

74.0
16.7
9.3

61.9–86.0
7.6–25.7
3.8–14.8

69
93
11
3
7
181
174

37.7
50.8
6.0
1.6
3.7
96.8
93.5

7066
9939
1026
286
668
18809
17879

37.2
52.4
5.4
1.5
3.4
97.1
92.8

27.7–46.7
43.2–61.5
1.5–9.3
--94.9–99.3
87.2–98.3

* Of households whose primary source of drinking water was well or public/municipal (n=94)

Table 4. Self-reported carbon monoxide exposure risk behaviors, Western KY CASPER, February 2009
Characteristic
Frequency (n=187) % of households Projected households
Generator Use
105
56.1
11248
Inside
1
1.0*
95
Garage
4
3.8*
382
Outside
97
92.4*
10481
Near open window
6
6.2†
549
Charcoal Grill/Gas Grill/Camp Stove Use
68
38.0
7300
Inside w/window open
7
10.4‡
922
Inside w/window closed
7
10.4‡
922
Outside
53
79.1‡
5487
Near open window
5
9.0**
465
Have CO detector
69
36.9
7609
Working CO detector
49
71.0††
5557
* Of the households who reported using a generator (n=105)
†Of the households who reported using a generator outside (n=97)
‡Of the households who reported using a grill/stove (n=68)
**Of the households who reported using a grill/stove outside (n=53)
††Of the households with a CO detector (n=69)

Projected %
58.1
0.9*
3.4*
93.2*
5.2†
38.3
12.6‡
12.6‡
75.2‡
8.5**
39.3
73.0††

95% CI
47.8–68.3
----30.0–46.5
----29.2–49.4
--

Table 5. Self-reported health status, current needs and access to care, Western KY CASPER, February 2009
Characteristic
Frequency (n=187)
% of households Projected households
Injury
Injured since storm
9
4.8
885
Illness
Become ill since storm
32
17.1
3372
Nausea/stomach ache/diarrhea
8
4.3 (25.0 of ill)
817
Cough with fever
7
3.7 (22.0 of ill)
668
Severe headache w/dizziness
4
2.1 (12.5 of ill)
382
Chronic illness worsened
2
1.1 (6.3 of ill)
191
Mental Health
Emotional concerns, anxiety, sleep
30
16.0
2926
problems, or memory problems
Medication
Not getting needed medication
11
5.9
1102
Special Needs
Home health care
1
0.5
95
Oxygen
6
3.2
627
Dialysis
1
0.5
95
CPAP
1
0.5
95
Breathing treatment
1
0.5
95
Transportation
Available if needed
185
99.5
19179

Projected %

95% CI

4.6

1.5–7.6

17.4
4.2 (24.2 of ill)
3.4 (19.8 of ill)
2.0 (11.3 of ill)
1.0 (5.7 of ill)

10.4–24.3
-----

15.3

8.3–22.1

5.7

1.7–9.7

0.5
3.2
0.5
0.5
0.5

------

99.5

98.4–100.5