Community Assessments for Public Health Emergency Response (CASPERs)—US Virgin Islands, 2017–2018

Abstract

Objectives. To assess preparedness levels of communities to help public health and others plan for disasters or emergencies and tailor messaging to increase community preparedness.

Methods. US Virgin Islands Department of Health conducted a Community Assessment for Public Health Emergency Response (CASPER) in June 2017, 2 hurricane response CASPERs in November 2017, and a recovery CASPER in February 2018. CASPER is a 2-stage cluster sampling method designed to provide household-based information about a community’s needs in a timely, inexpensive, and representative manner.

Results. Roughly the same amount of households reported having a 3-day supply of food and water before and 3 months after the hurricanes. During the response, approximately a third of households resupplied between 3 and 6 days and an additional approximately 40% between days 7 and 14.

Conclusions. On the basis of the CASPERs, we were able to track whether households had an emergency preparedness kit, whether they used it during the storms (and what was missing), and if they resupplied their kit in recovery.

Public Health Implications. CASPER is a promising tool to measure community preparedness to help state, local, tribal, and territorial jurisdictions plan for disasters or emergencies.

Community preparedness is essential to a successful response and helps mitigate loss of life, injuries, and illnesses immediately after a disaster’s impact.¹ Following a disaster, community members may be on their own for a period of time because of the ongoing response efforts, the size of the affected area, lost communication, and impassible roads.² The Federal Emergency Management Agency (FEMA) encourages community members to get a kit, make a plan, and be informed.^3,4 The kit should include a recommended 3-day supply of food and water and a 7-day supply of prescription medicines. This recommendation is largely for the general population of the continental United States; however, there are unique circumstances that may extend the length of time before response efforts are in place, such as being on an island. Although some local jurisdictions encourage families to make emergency supply kits for longer periods (e.g., 7–14 days),^5–7 this recommendation has yet to be consistently adopted within national guidelines.

Assessing the preparedness levels of communities can help public health and other agencies plan for disasters or emergencies and tailor messaging to increase community preparedness. This can include determining if households have emergency plans; supply kits; adequate food, water, and medicine; intended evacuation (or nonevacuation) plans; and preferred and trusted communication sources. The Community Assessment for Public Health Emergency Response (CASPER) can help assess these factors. CASPER is an epidemiological technique designed to provide household-based information about a community’s needs in a timely, inexpensive, and representative manner. The information generated can be used to initiate public health action; facilitate disaster planning, response, and recovery activities; and assess new or changing needs during the disaster recovery period.⁸ Although CASPER was originally designed to provide information during a disaster response, it has been increasingly used throughout the disaster cycle to obtain population-representative data.⁹ In the past 5 years, more than 50 preparedness CASPERs have been conducted in jurisdictions across the United States to assist with planning for emergencies by assessing household emergency preparedness, evacuation plans, and other activities.^9–11

The US Virgin Islands (USVI) Department of Health (DoH) conducted a CASPER in June 2017 as part of the ongoing Zika virus outbreak in the territory. Because the CASPER was conducted at the beginning of hurricane season, in addition to Zika-related questions, the CASPER also included a section on general disaster preparedness. Three months after the CASPER was conducted, 2 Category 5 storms hit USVI within 13 days of each other: Irma on September 6, 2017, followed by Maria on September 19, 2017. Both storms led to widespread destruction to homes, businesses, and the health care infrastructure, leaving the island without power and potable water for months.^12–14 The USVI DoH conducted 2 CASPERs in November 2017 to assess the experience and health of residents before, during, and after the hurricanes. Conducting a pre- and poststorm CASPER allowed us to compare preparedness levels and to quantify the use of recommended resources after an actual disaster. This was the first time CASPERs were conducted right before and after a major disaster. The USVI DoH also conducted a follow-up recovery CASPER in February 2018 to monitor recovery efforts and changing needs in the community.

METHODS

USVI DoH conducted a CASPER during June 26 to 29, 2017, as part of the response to the ongoing Zika outbreak on the territory (St Croix, St Thomas, and St John); 2 hurricane response CASPERs (St Croix during November 7–8, 2017, and St Thomas and St John during November 13–14, 2017); and a follow-up recovery CASPER during February 26 to March 1, 2018 (St Croix, St Thomas, and St John). USVI DoH decided to conduct 1 recovery CASPER for the territory as a whole as there were no major differences among the islands in the response CASPERs. Therefore, for comparative purposes, we combined the 2 response CASPERs into 1 data set to run the analyses and defined the sampling frame as all occupied households (n = 43 214) within USVI according to the 2010 US Census.

We applied standard CASPER methodology, which includes 2-stage cluster sampling to select a representative sample of households to be interviewed.¹¹ In the first stage, we selected 30 clusters (census blocks) with a probability proportional to the number of households within the clusters, and in the second stage, interview teams used systematic random sampling to select 7 households from each of the selected clusters. Eligible respondents were aged 18 years or older and resided in the selected household; for the response and recovery CASPERs, employees and volunteers deployed on island for the response and recovery efforts were not interviewed. Teams made 3 attempts at each selected household before substitution. Teams also provided public health informational materials to all potential respondents and interested persons (e.g., community members who approached interview teams to ask questions) and completed confidential referral forms when requested for urgent physical or behavioral health needs.

The Zika CASPER (June 2017; before the 2017 hurricanes) was primarily focused on the ongoing Zika virus outbreak in the territory with 1 section of the questionnaire dedicated to general disaster preparedness. For this analysis, we are only including demographics and general disaster preparedness questions. The response (November 2017) and recovery (February 2018) CASPERs used a similar 2-page questionnaire that included questions on demographics; household experiences since the hurricanes, including communications; vector concerns; physical and behavioral health status; and household needs. For all CASPERs, we conducted weighted analyses to report the projected number and percentage of households within the sampling frame that gave a particular response. We compared results from the CASPERs, where applicable. We determined significance based on 95% confidence intervals (CIs).

RESULTS

Interview teams conducted 192 to 201 interviews during the CASPERs (completion rates of 91.4%–95.2%). Of the households with an eligible participant answering the door, 62.6% completed an interview in June 2017 (Zika/Preparedness), 84.1% completed an interview in November 2017 (response), and 81.3% completed an interview in February 2018 (recovery).

Table 1 describes household demographics and hurricane-related damage. Overall, approximately two thirds of households lived in single-family homes, with the majority of households having 1 or more members aged 18 to 64 years. Fewer than 10% of households had 1 or more children aged younger than 2 years, and roughly 40% of households had 1 or more members aged 65 years or older. The mean number of household members was 2.7 to 2.9—comparable to the census average for USVI of 3.1¹⁵—with a minimum of 1 and maximum of 11 people living in a household. The primary language spoken within the household was English followed by Spanish. We found no significant differences among the 3 CASPERs for any demographic or other control variables. Almost 70% reported that their homes were either damaged or destroyed by the hurricanes, and, 5 months after storms’ impact, 10.4% still felt their home was unsafe in which to live.

TABLE 1—

CASPER Results for Household Demographics: US Virgin Islands

	Zika/Preparedness, June 2017		Response, November 2017		Recovery, February 2018
	Estimate	% of HH (95% CI)	Estimate	% of HH (95% CI)	Estimate	% of HH (95% CI)
Number of HHs with members in each age category
< 2 y	3 910	9.1 (4.4, 13.8)	3 168	7.3 (4.5, 10.1)	2 552	5.9 (2.8, 9.0)
2–17 y	13 438	31.2 (22.7, 39.8)	14 742	34.1 (27.9, 40.3)	13 026	30.1 (24.1, 35.9)
18–64 y	34 743	80.8 (73.6, 88.0)	32 408	75.0 (69.5, 80.5)	33 107	76.6 (70.2, 83.0)
≥ 65 y	17 780	41.5 (33.1, 50.0)	18 218	42.5 (36.9, 47.4)	18 236	42.2 (32.5, 51.9)
Primary language spoken at home
English	38 659	89.5 (85.3, 93.7)	39 014	90.3 (86.9, 93.6)	39 287	90.9 (85.3, 96.6)
Spanish	2 332	5.4 (2.4, 8.4)	3 077	7.1 (4.4, 9.8)	3 121	7.2 (1.8, 12.6)
Creole	1 811	4.2 (1.2, 7.2)	697	1.6 (0.2, 3.1)	. . .	. . .
Type of structure
Single-family home	29 495	68.3 (57.3, 79.2)	26 282	62.2 (54.6, 69.8)	27 832	64.4 (52.0, 76.8)
Multiple unit	10 804	25.0 (14.4, 35.6)	15 756	37.3 (29.7, 44.9)	14 765	34.2 (21.7, 46.8)
Damage to home
None/minimal	. . .	. . .	16 373	37.9 (32.5, 43.3)	14 302	33.1 (24.6, 41.6)
Damaged	. . .	. . .	24 821	57.4 (52.9, 62.0)	26 906	62.3 (54.2, 70.2)
Destroyed	. . .	. . .	2 021	4.7 (1.5, 7.8)	2 006	4.6 (1.9, 7.4)
Household feels home safe to live in
Yes	. . .	. . .	36 885	85.4 (80.8, 89.9)	38 087	88.1 (83.1, 93.2)
No	. . .	. . .	6 329	14.6 (10.1, 19.2)	4 476	10.4 (5.8, 14.9)

Note. CASPER = Community Assessment for Public Health Emergency Response; CI = confidence interval; HH = household. Percent may not add up to 100% if households refused to answer or responded “don’t know.”

Before the hurricanes, 67.0% of household reported they had an emergency supply kit (Table 2). However, during the response CASPER, significantly fewer households reported they had an emergency supply kit during the storms (47.9%), and, of those, 64.3% reported that they used their kit. Of those households who reported that they needed supplies that were not in their kit, medical supplies and batteries were the top responses (Table 3). Roughly the same amount of households reported having a 3-day supply of food (78.2%–84.4%), 3-day supply of water (84.5%–89.9%), and a 7-day supply of medication (53.5%–57.6%) before and after the hurricane. During the response, 28.5% of households had to resupply water between 3 and 6 days, 42.3% between 7 and 14 days, and 8.9% after 2 weeks. A similar trend was found with food with 37.3% of households needing more food between 3 and 6 days, 42.6% between 7 and 14 days, and 4.5% after 2 weeks. Roughly 30% of households ran out of medication they needed in the aftermath of the storms with 34.2% needing more medication within the first week. During the recovery phase, 59.3% of households (95% CI = 50.2, 68.3) reported having an emergency supply kit.

TABLE 2—

CASPER Results for Household Emergency Supplies: US Virgin Islands

	Zika/Preparedness, June 2017		Response, November 2017
	Estimate	% of HH (95% CI)	Estimate	% of HH (95% CI)
Household had emergency supply kit
Yes	28 953	67.0 (59.4, 74.6)	20 685	47.9 (39.8, 55.9)
No	13 437	31.1 (23.4, 38.9)	22 529	52.1 (44.1, 60.2)
Used emergency supply kit during storm
Yes	. . .	. . .	13 234	64.3 (55.9, 72.8)
No	. . .	. . .	7 340	35.7 (27.2, 44.1)
Needed supplies not in kit (n = 118)
Yes	. . .	. . .	5 897	44.6 (33.1, 56.0)
No	. . .	. . .	7 337	55.4 (44.0, 66.9)
3-d supply of water
Yes	38 865	89.9 (85.9, 93.9)	36 499	84.5 (80.1, 88.8)
No	3 256	8.2 (4.0, 12.3)	6 715	15.5 (11.2, 19.9)
3-d supply of food
Yes	36 464	84.4 (79.3, 89.5)	33 787	78.2 (72.6, 83.8)
No	6 544	15.1 (10.1, 20.1)	9 427	21.8 (16.2, 27.4)
7-d supply of medication
Yes	24 879	57.6 (50.3, 64.9)	23 105	53.5 (48.6, 58.4)
No	5 926	13.7 (7.7, 19.7)	6 532	15.1 (10.9, 19.4)
No prescriptions	11 791	27.3 (20.7, 33.9)	13 200	30.5 (25.5, 35.6)
Household ran out of medication
Yes	. . .	. . .	12 574	29.1 (23.8, 34.8)
No	. . .	. . .	16 701	38.6 (33.3, 4.0)
Don’t know	. . .	. . .	471	1.1 (0.2, 2.0)
None needed	. . .	. . .	13 312	30.8 (25.7, 35.9)

Note. CI = confidence interval; HH = household. November 2017 asked “Immediately after the storms, if household had. . . .” Percent may not add up to 100% if households refused to answer or respond “don’t know.” Ellipses indicate question was not asked or cell less than 5.

TABLE 3—

CASPER Results for Household Emergency Supply Needs: US Virgin Islands

	Response (November 2017)
	Estimate	% of HH (95% CI)
Needed supplies not in kit (n = 118)
No	7 337	55.4 (44.0, 66.9)
Yesa	5 897	44.6 (33.1, 56.0)
Yes: medical supplies	1 848	13.5 (8.5, 18.5)
Yes: batteries	3 057	22.3 (12.2, 32.5)
Yes: otherb	2 880	21.0 (12.9, 29.2)
Needed more water after . . .
3–6 d	10 297	28.5 (22.8, 34.2)
7–14 d	15 288	42.3 (35.6, 48.9)
> 14 days	3 208	8.9 (5.3, 12.4)
Never needed to resupply	6 588	18.2 (12.2, 24.3)
Needed more food after . . .
3–6 d	12 625	37.3 (30.6, 44.1)
7–14 d	14 759	43.6 (37.1, 50.2)
> 14 days	1 527	4.5 (1.2, 7.8)
Never needed to resupply	4 594	13.6 (8.7, 18.4)
Of those with medication, needed more after . . . (n = 267)
<  7 d	4 430	14.8 (10.3, 19.2)
After 7–14 d	6 024	20.1 (16.0, 24.2)
After > 14 d	2 147	7.2 (3.3, 11.0)
Never	16 701	55.3 (48.4, 62.2)
Don’t know	806	2.7 (0.6, 4.8)

Note. CI = confidence interval; HH = household. Question not asked in preparedness or recovery. Percent may not add up to 100% if households refused to answer or responded “don’t know.”

^aHouseholds could choose more than 1 response.

^bOther includes food, water, bug spray, cleaning supplies, etc.

Table 4 describes the general preparedness of households; these questions were only asked in the Zika CASPER. Before the storm, 82.1% had copies of important documents in a safe location, 63.2% of households had multiple routes away from their home in case evacuation was necessary, 37.5% had a designated meeting place outside of their neighborhood, and 32.1% had a designated meeting place immediately outside of their home or close by in the neighborhood. When asked during the recovery period how households would prepare differently for future storms, the top responses included board up their windows (19.7%), fix or build a stronger roof or house (16.6%), buy food or water (10.3%), leave the island (9.1%), and buy emergency kit supplies (8.9%).

TABLE 4—

CASPER Results for Household General Preparedness: US Virgin Islands

	Zika/Preparedness (June 2017)
	Estimate	% of HH (95% CI)
Copies of important documents
Yes	35 476	82.1 (75.7, 88.5)
No	7 532	17.4 (11.1, 23.8)
Multiple routes away from home
Yes	27 314	63.2 (54.2, 72.3)
No	14 048	32.5 (24.2, 40.8)
Emergency communication plan
Yes	26 614	61.6 (53.2, 69.9)
No	15 777	36.5 (28.7, 44.3)
Designated meeting place outside of neighborhood
Yes	16 216	37.5 (30.4, 44.6)
No	18 609	43.1 (35.9, 50.2)
Not applicable	7 772	18.0 (12.4, 23.5)
Designated meeting place outside of home
Yes	13 870	32.1 (25.6, 38.6)
No	20 338	47.1 (39.3, 54.9)
Not applicable	7 772	18.0 (12.4, 23.5)
Don’t know or refused	1 235	2.9 (0.6, 5.1)

Note. CI = confidence interval; HH = household. Percent may not add up to 100% if households refused to answer or responded “don’t know.”

DISCUSSION

Comparing household information before, during, and after the storms enabled us to gauge perceived household preparedness with preparedness during an actual response and recovery. Approximately 20% of households did not have the recommended levels of food and water before the storm. In addition, although 78% to 85% reported having the FEMA-recommended 3-day supply of food and water, roughly a third ran out between days 3 and 6 after impact. And, while about half of households reported having enough medication for 7 days, the majority ran out and needed to resupply less than 2 weeks after the storms, when medication on the island was still scarce. In addition, a quarter of households had difficulty obtaining prescription medication after the storms with the closure of pharmacies and clinics. Therefore, the FEMA recommendation to maintain a 3-day supply of food and water and 7-day supply of medications may not be adequate for island territories where getting supplies can be particularly challenging after a disaster. This information is useful to increase community education about having emergency supply kits and motivating households to prepare. It is also important for planning efforts. Although through news releases FEMA has since encouraged households to be prepared for the 2018 hurricane season for 14 days,^16,17 their official guidance on the Web site remains 3 days.

While food, water, and medication estimates were similar before and after the hurricanes, significantly fewer households reported having an emergency supply kit during the storm than did during the prestorm CASPER. Possible explanations of this discrepancy include the following: those that reported having a kit realized that their kit was inadequate when they went to use the kit, they did not have what they thought they had earlier, or they had exhausted kit supplies quickly. Despite not all households having an emergency kit, approximately two thirds of households did use their emergency supply kit during the storm. There were many items that households reported not having, such as medical supplies (e.g., adhesive bandages) and batteries, even though FEMA recommends these items for an emergency supply kit.¹⁸ Other items reported missing from household emergency kits included mosquito repellant and sunscreen, which may be more location-specific to the USVI and not often promoted by FEMA, especially because of the recent Zika outbreak and warm island weather. To our knowledge, this is the only such data on the use of preparedness kits following a public health disaster or emergency and, although people did report using their kits, there remains room for improvement. This information will be useful for future preparedness efforts and plans as the health department can tailor messaging about emergency supply kits.

Interestingly, though not significant, the number of households reporting having an emergency supply kit was trending upward during the recovery CASPER. Households may have realized the value in having a kit after experiencing the hurricanes. Continued follow up with these affected communities is important to find out if the number of households with a kit has increased and if households are maintaining their kits. Previous Behavioral Risk Factor Surveillance System data from the emergency preparedness survey module (2006–2008) in 6 states (Delaware, Georgia, Louisiana, Montana, Nevada, and Tennessee) found that 42% of households had preparedness kits that included water, food, battery-operated radio, and flashlight with batteries.¹⁹ Results from other more local surveys found that the number of households with kits varied between 22% and 67%.^20–24 The large range in the percentage of households with emergency supply kits nationally shows the variable nature in preparedness across different communities.

In addition to emergency supply kits, household preparedness plans are essential to help minimize the feeling of panic and enable the household to remain calm, collected, and, most importantly, safe.² While more than 80% of households had their important documents in a safe location before the storm, significantly less had the more intangible plans in place that help minimize chaos and stress such as multiple routes for evacuation and designated meeting places for their family members in case they are separated. When asked during the recovery period how they would prepare differently for future storms, top responses included purchasing preparedness items, boarding up windows and doors, and having a plan in place to leave the island. These data have been useful as the USVI DoH revises messaging about disaster preparedness.

Limitations

The assessments were subject to the following limitations. The data generated by the CASPERs represent a snapshot in time, which should be considered when interpreting the results of ongoing preparedness. In addition, although the sampling frame remained the same, the same households were not selected across the 3 time periods; however, though not the same households, the data represent the same population and we found no variability in demographics between the CASPERs. The age distribution of the sample population may be skewed, with a greater proportion of individuals aged 65 years and older represented in the CASPER than reported by the US Census (33.3%). Therefore, survey responses may not be representative of USVI as a whole. The CASPER demographic results, however, are comparable to each other, suggesting that households participating in these CASPERs may have been older, retired residents likely to be home during daylight hours when the CASPER was conducted.

Public Health Impact

As 2017 hurricane recovery is expected to be a long-term process, and multiple hurricanes came close to landfall on the territory in 2018, the continued preparedness of the community is essential. Using CASPER methodology, we were able to track whether USVI households had an emergency preparedness kit, whether they used it during the 2017 storms (and what was missing), and if they resupplied their kit (or bought a new one) during hurricane recovery. Therefore, CASPER is a promising tool to track community preparedness to help jurisdictions, including other places likely to be affected by hurricanes, plan for disasters or emergencies. USVI DoH shared CASPER findings within the various sections of the health department and other agencies to promote data-driven efforts to strengthen the emergency preparedness capacity on USVI. This included acquiring funding and support for recovery activities, tailoring health and clean-up messaging, and informing Virgin Islands Territory Emergency Management Agency and other response partners of household recovery status to continue to update and modify recovery planning on the islands. USVI DoH will continue conducting CASPERs over the next 2 years (approximately every 6 months) to gauge progress and estimate current levels of preparedness as the recovery efforts evolve.