Having extra medication and medical supplies on-hand, referred to here as personal medical preparedness (PMP), is a key part of personal emergency/disaster preparedness that can prevent excess morbidity and mortality during and after disasters.1 Increases in chronic disease morbidity and mortality following disasters have been well documented; for example, rates of myocardial infarction were elevated for 6 years following Hurricane Katrina.Reference Babaie, Asl, Naghipour and Faridaalaee2-Reference Waddell, Jayaweera and Mirsaeidi4 Some of the excess disease burden may be due to disruptions in medication access and medication adherence during disasters, leading to uncontrolled disease. After Hurricane Katrina, antihypertensive medication adherence decreased compared to pre-storm baseline.Reference Krousel-Wood, Islam and Muntner5 In 2018, during Hurricane Florence, one-third of regional emergency department visits were for medication refills.Reference Tanz6 Personal medical preparedness involves having necessary medication and supplies, important medical documents, medical equipment, oxygen, and support for electronic equipment (i.e., batteries or generator) on hand for self-care during and after a disaster when outside support may be limited.Reference Heagele, McNeill and Adams7 PMP may prevent medication adherence disruptions, reverse the trend of post-disaster chronic disease exacerbation, and increase individual and community health resilience.
Individuals with chronic conditions experience challenges related to preparedness for disasters and are less likely to be generally prepared relative to healthy individuals, but studies vary in collecting data specific to PMP.Reference Bethel, Foreman and Burke8-Reference Kohn, Eaton and Feroz10 Although hypertension is prevalent and a considerable burden during disaster medical responses both in the US and abroad, data are limited on the baseline PMP of people with hypertension in high-risk disaster regions.Reference Keasley, Oyebode and Shantikumar11 The goal of this study is to identify factors underlying baseline PMP in a sample of adults with hypertension in the Gulf South, an area with frequent exposure to natural disasters and high burden of hypertension and other chronic diseases.Reference Samanic12, Reference Colbert13 Understanding the baseline PMP of adults with chronic disease in this region can inform resilience and response strategies. Additionally, studying baseline preparedness of individuals with a high level of prior disaster experience may inform understanding of broader disaster preparedness behaviors.
Methods
Study Design and Sample
A cross-sectional survey was collected from a purposive sample of 120 adults with hypertension. Participants were identified as a sub-sample of prospective and enrolled participants (from both study arms) of the Medication Adherence Program (MAP) study (R01HL153750), an ongoing randomized controlled trial testing a behavioral intervention for improving antihypertensive medication adherence. All participants met the following criteria: fully insured by single payer for at least 1 year (either via commercial plan or Medicare Advantage), age 50+, 2 or more pharmacy claims for an antihypertensive medication in the same class within the most recent 12 months, and low adherer to antihypertensive medication refill, defined as proportion of days covered (PDC) <0.8.Reference Loucks, Zuckerman and Berni14
Data Collection
Data collection occurred during the 2023 Atlantic Hurricane Season from September 25-November 30. Prospective and current MAP participants were invited to complete the survey during MAP screening and follow-up visits (in person or over the phone). All surveys were interviewer-administered; responses were recorded using REDCap electronic data capture tool.Reference Harris, Taylor and Thielke15, Reference Harris, Taylor and Minor16 The survey took approximately 15 minutes to complete.
Study Measures
The survey used both published validated tools and investigator generated questions. The investigator generated questions were included to capture novel domains and were based on extensive literature review and interviews with content experts.
Primary outcome: personal medical preparedness via HEPI-AFN
Personal medical preparedness (PMP), the primary outcome, was captured using the Household Emergency Preparedness Instrument Access and Functional Needs section (HEPI-AFN).Reference Heagele, McNeill and Adams7, Reference Heagele, Adams, McNeill and Alfred17 The AFN section of the HEPI is applicable to individuals who have a disability, are over age 65, take at least 1 daily medication, or are pregnant. Because this study focused on PMP for individuals who take at least 1 daily antihypertensive medication, the 9-item AFN section was universally applicable. The scale score represents the proportion of applicable preparedness items to which the participant answers “yes” and ranges from 0-1, with higher values indicating increased PMP. For analyses of the primary outcome, the HEPI-AFN was dichotomized to compare those with a score of 1, indicating full preparedness in all applicable PMP categories, versus those who scored less than 1.
In addition to medical preparedness, participants also self-reported general disaster preparedness on a single item from the Centers for Disease Control Behavioral Risk Factor Surveillance System (CDC BRFSS).Reference Barbato, Bryie and Carlisle9 The item was adapted to ask, “How well prepared do you feel your household is to handle a hurricane or other disaster?”
Participants who noted that they did not have 2 extra weeks of medication or medical supplies on hand were asked: “In a few sentences, tell me why you don’t have extra medications or supplies set aside for emergencies”, with probes, “Have you ever thought about getting extras? Are there barriers stopping you from getting extras?” Common themes were identified, and responses were categorized by theme and reported as frequencies.
Sociodemographic and health status
Standard demographics were collected, including age, sex, race, ethnicity, education, relationship status, and smartphone ownership. Disability status was measured using the Washington Group Short Set on Functioning (WG-SS), a validated tool that asks participants to rate difficulty in 6 functional domains.Reference Bourke, Nichols-Dunsmuir, Begg, Dong and Schluter18 Participants were characterized as having a disability if they responded “some difficulty” in at least 2 domains or “a lot of difficulty” in 1 domain. Financial strain was measured using a 4-item tool commonly used in primary care settings and among individuals with chronic disease, which asks participants to indicate whether they can afford common expenditures; responding “no” to any expenditure indicated financial strain.Reference Friedman, Conwell and Delavan19 Health literacy was measured using the Single-Item Literacy Screener (SILS).Reference Morris, MacLean and Chew20 Participants were health literate if they reported “never” or “rarely” needing help reading health care information. Self-reported antihypertensive medication adherence was measured using the Krousel-Wood Medication Adherence Scale (K-Wood-MAS-4), a validated, open access 4-item screener associated with pharmacy fill, blood pressure control, and cardiovascular outcomes.Reference Krousel-Wood, Joyce and Holt21, Reference Krousel-Wood, Peakcock and Joyce22 The score ranges from 0-4, with higher values indicating worse adherence; poor adherence is defined as a score ≥1. Polypharmacy was defined as taking ≥5 prescription and non-prescription medications on a regular basis (if a medication was to be taken on a regular schedule, i.e., daily, every other day, or once per week, this counted as a “regular” medication).
Hurricane experience, planning, and risk
Hurricane risk perception was measured using a 3-item tool that was previously used to study at-risk communities in Florida.Reference Trumbo, Meyer and Marlatt23 Responses ranged from 0-3, with 3 indicating highest perceived risk from hurricanes; individuals with the top 25% of scores were considered to have high hurricane risk perception. Additional questions were added to measure prior hurricane experience and approaches toward preparedness and evacuation. These included, “Do you have a plan for how your household would handle a hurricane or other disaster?,” “Thinking about your current plan, how much is it influenced by your medical needs?,” and “Thinking about your health, does your health impact your likelihood to evacuate during a hurricane?”
Statistical Analyses
Of the 120 surveys collected, 100% had complete data and were included in the analyses. Participant characteristics were reported as percents or means as appropriate. A multivariable logistic regression analysis was completed to assess the independent associations between the primary outcome, total PMP, and the following covariates: age, sex, race, education, relationship status, financial strain, region, disability, polypharmacy, K-Wood-MAS-4 medication adherence, hurricane experience, disaster preparedness, and hurricane risk perception.
An exploratory analysis assessed the associations between key participant characteristics and other hurricane preparedness outcomes that were not part of the HEPI-AFN but may be important to both general preparedness and PMP for this population. The outcome variables included hurricane risk perception, having a plan, self-reported preparedness, having a 2-week supply of extra medications, and agreeing that “my health makes me more likely to evacuate.” Explanatory variables included financial strain, hurricane experience, comorbidities, disability, and medication adherence. Bivariate associations were tested using Pearson’s Chi-squared tests.
All statistical analyses were completed using R version 4.3.1 (June 16, 2023) and the significance level was set at α<0.05.Reference Team24
Data Confidentiality and Ethics
This study was approved by the Tulane University IRB (Study No. 2020-2221-CTU/TCCR) and all procedures were in accordance with human subjects guidelines. All participants provided verbal informed consent and were provided a $10.00 gift card for participation.
Results
Participant Characteristics
Two-hundred ninety-four (294) adults living in Southeast Louisiana were contacted about the study; 43 refused, 131 did not answer phone calls, and 120 were enrolled. Of those enrolled, 82 (68.3%) lived in the greater New Orleans area, and the rest lived in the greater Baton Rouge area. Participants were 50.0% women; 43.3% Black, 55.8% white; average age 62.6 years (SD = 8.1).
Over half of the sample (n = 71, 59.2%) had at least 1 comorbidity in addition to hypertension, with the most common being diabetes (n = 35, 29%), history of cancer (n = 20, 16.7%), and “emphysema, asthma, or COPD” (n = 19, 15.8%); 69 individuals (57.5%) self-reported low medication adherence on the K-Wood-MAS-4; and 9 participants (7.5%) had low health literacy (Table 1).
Table 1. Demographics, health, and hurricane factors among survey respondents, n = 120
SD – standard deviation.
1 Defined as reporting “some difficulty” in at least 2 domains or “a lot of difficulty” in 1 domain in Washington Group Short Set on Functioning (WG-SS).
2 Defined as score ≥1 in financial strain screener described by Friedman et al. 2017.
3 Defined as taking ≥5 medications on a regular basis.
4 Defined as score ≥1 on 4-item Krousel-Wood Medication Adherence Scale.
5 Measured by the Single-Item Literacy Screener (SILS).
6 Calculated via adaptation of 3-item tool described by Trumbo et al. 2014.
7 Meets all 3 of the following experience metrics: (1) has experienced home damage from a hurricane (2) was living in the area impacted by Hurricane Katrina during Hurricane Katrina (3) has lived in a hurricane-impacted area for ≥10 years.
8 Centers for Disease Control Behavioral Risk Factor Surveillance System, single-item emergency preparedness tool.
9 Percentage calculation utilizes denominator of total participants with a plan, excludes those with no plan.
Personal Medical Preparedness Measure, HEPI-AFN
The mean HEPI-AFN score was 0.79 (SD = 0.21, range = 0-1), indicating that, on average, participants were prepared on 79% of HEPI-AFN items that applied to them. Forty-two participants (35%) scored 1, the highest possible score, indicating that they were prepared on all applicable items (Figure 1). More participants had 1 week of medication saved (n = 95, 79.2%) compared to 2 weeks (n = 69, 57.5%). Medical supplies were used by 42 participants (21.1%), though only 24 (57.1%) of those individuals used electricity-dependent supplies. Refrigerated medication was used by 23 (19.2%) of participants.
Figure 1. Responses to the HEPI Access and Functional Needs section.
*Sample size <120 because question was only asked if applicable to participant.
Most participants had stored a medication list, list of doctors, and their medical history; the most common method of storage was online or in the cloud (Figure 2). Most did not have an advance directive.
Figure 2. Storage methods for important medical documents.
Total values add to >120 because participants who utilized multiple storage methods are re-listed for each respective method.
Frequencies of Hurricane Factors
Experience
Participants had lived in a hurricane-impacted area for an average of 51.3 years (SD = 18.1), including lifelong habitation for 72 (60%) participants. Most participants had high self-reported hurricane experience (n = 81, 67.5%), defined as meeting all 3 of the following criteria: (1) experienced home damage from a hurricane, (2) lived in the area impacted by Hurricane Katrina during Hurricane Katrina, (3) lived in a hurricane-impacted area for ≥10 years.
Nine participants (7.5%) reported that they have had trouble accessing their medication or medical supplies due to a disaster within the past 15 years (excluding Hurricane Katrina). Three participants (2.5%) had been to the emergency department or admitted to the hospital during or within 1 week following a hurricane or other disaster.
Evacuation
The mean number of evacuations in the time since Hurricane Katrina (i.e., in the preceding 18 years) was 1.4 (SD = 1.5, range = 0-6), including a mean of 1.8 (SD = 1.5, range = 0-6) for New Orleans residents and 0.7 (SD = 1.0, range = 0-4) for Baton Rouge residents (P<0.001). Health impacted the decision to evacuate for 31 (25.8%) participants. Among all participants, 28 (23.3%) stated that their health makes them more likely to evacuate, while only 3 (2.5%) stated that their health makes them less likely to evacuate. Two of the 3 individuals less likely to evacuate reported a disability as defined by the WG-SS.
Preparedness
Most participants (n = 92, 77.3%) had a plan for a hurricane or other disaster. Of those with a plan, 40 participants (43.5%) reported that the plan was at least slightly influenced by medical needs. Only 6 (5%) participants stated that they were “not prepared at all” (Table 1).
Eighty-three participants (69.2%) agreed that the following statement described their behavior: “I don’t keep extra medications set aside for emergencies, but if there is a hurricane threatening my area in the next few days, I’ll check how much I have left and try to get an early refill.”
When asked, “As far as you remember, has a health care provider or someone who works at your health care clinic or hospital ever talked to you about developing a hurricane or disaster plan specific to your medical needs?,” 107 (89.2%) answered no.
Factors Associated with Personal Medical Preparedness
Both unadjusted and adjusted associations between covariates and the primary outcome, PMP, are presented in Table 2. While there was a trend for higher PMP among women versus men; adults who were more educated versus less educated; and adults living in New Orleans versus Baton Rouge, no factors were statistically associated with PMP.
Table 2. Factors associated with personal medical preparedness
* Model is adjusted for all variables listed in table.
OR – odds ratio; CI – confidence interval.
1 Defined as score ≥1 in financial strain screener described by Friedman et al. 2017.
2 Defined as reporting “some difficulty” in at least 2 domains or “a lot of difficulty” in 1 domain in Washington Group Short Set on Fun Functioning (WG-SS).
3 Defined as taking ≥5 regular medications.
4 Defined as score ≥1 on a 4-item Krousel-Wood medication adherence scale.
5 Meets all 3 of the following experience metrics: (1) has experienced home damage from a hurricane (2) was living in the area impacted by Hurricane Katrina during Hurricane Katrina (3) has lived in a hurricane-impacted area for ≥10 years.
6 Centers for Disease Control Behavioral Risk Factor Surveillance System, single-item emergency preparedness tool.
7 Calculated via adaptation of 3-item tool described by Trumbo.
Exploratory Analyses
Results of exploratory analyses of associations between key participant characteristics and hurricane preparedness outcomes are presented in Table 3. Financial strain was associated with having high hurricane risk perception, not having a plan, health impacting likelihood to evacuate, and low self-reported general preparedness. Having no comorbidities (excluding hypertension) was associated with having a plan. Those screening positive for a disability had a significant association with high hurricane risk perception and health impacting likelihood to evacuate (i.e., making them either more likely or less likely to evacuate). High hurricane experience was associated with health impacting likelihood to evacuate. High medication adherence was associated with self-reported general preparedness.
Table 3. Associations between key participant characteristics and exploratory hurricane preparedness outcomes
P values with Yates’ continuity correction from Chi-squared tests of independence comparing across participant characteristics.
1 Calculated via adaptation of 3-item tool described by Trumbo.
2 Centers for Disease Control Behavioral Risk Factor Surveillance System, single-item emergency preparedness tool.
3 Defined as score ≥1 in financial strain screener described by Friedman et al. 2017.
4 Meets all 3 of the following experience metrics: (1) has experienced home damage from a hurricane (2) was living in the area impacted by Hurricane Katrina during Hurricane Katrina (3) has lived in a hurricane-impacted area for ≥10 years.
5 Defined as reporting “some difficulty” in at least 2 domains or “a lot of difficulty” in 1 domain in Washington Group Short Set on Fun Functioning (WG-SS).
6 Defined as score ≥1 on a 4-item Krousel-Wood medication adherence scale.
Barriers to Preparedness
Of the 53 participants (44.2%) who were missing either 2 weeks of extra medicine or 2 weeks of extra supplies, 49 (92.5%) gave responses to the open-ended follow-up question. Categories of responses are presented in Figure 3. Other responses included: participant recently forgot to pick up their refill (n = 2); refill is currently being complicated by pharmacy shortage (n = 1); participant is concerned about medicine expiring and being wasted (n = 1); participant states that medicine is only preventive and not actually necessary to survive (n = 1); and has thought about getting extra, but does not know how to request this (n = 1).
Figure 3. Barriers to having 2 extra weeks of medicine or medical supplies, n=49 responses.
Limitations
These findings should be considered in the context of study limitations. This study had a limited sample size of fully insured adults with hypertension and low antihypertensive medication pharmacy fill adherence and was conducted in a hurricane-prone area in Southeast Louisiana; results may not be generalizable to other populations. It is a cross-sectional study and causality cannot be determined. Larger longitudinal studies in more diverse populations should be conducted.
Discussion
In this population with a high prevalence of both hurricane experience (81.7% impacted by Hurricane Katrina) and self-reported general disaster preparedness (95% at least “somewhat” prepared), the average individual was prepared in the majority (79%) of personal medical preparedness (PMP) domains, though only 35% of individuals had total PMP. Despite nearly 90% of participants reporting that they had never had a conversation with their health care provider or a staff member about PMP, the participants had considerable PMP. PMP in this sample could be due to personal experience from prior hurricanes, support from social networks, active public safety messaging, or successful work by community organizations and others.Reference Pollock, Wennerstrom and True25–Reference DeSalvo, Lurie and Finne28 In this population, and even more so in less disaster-experienced populations, communication from health care providers and organizations may present an opportunity for preparedness-oriented messaging to optimize PMP and health outcomes.Reference Icenogle, Eastburn and Arrieta29 Health care providers have been and should continue to be a trusted source of risk communication messaging for the patients they serve. Free printable resources are available for providers to use during wellness visits to support prevention for disaster-related morbidity.1, 30
Having a two-week supply of extra medications was the PMP domain for which the largest percentage of participants were unprepared (79.2% had a 1-week supply, 57.5% had a 2-week supply). Only 10 participants had never thought about getting extra medication. Most (48%) were restricted from having extra medication on hand by policies and practices of pharmacies or insurers. The wording, “Do you have a 2-week supply of extra medications?” may have been ambiguous and could be interpreted as either currently having 2 weeks in their medication bottle or as having 2 weeks of medication set aside explicitly for emergencies, separate from the daily medication supply. As found following earthquakes in Japan, just having extra medication was not linked to bringing medication during evacuation; participants had to have extra medication packed in a separate go-bag to increase the likelihood of evacuating with it. Yet hurricane and earthquake evacuations have different timelines.Reference Ochi, Hodgson and Landeg31, Reference Tomio, Sato and Mizumura32 Over 2/3 of participants agreed with the statement: “I don’t keep extra medications set aside for emergencies, but if there is a hurricane threatening my area in the next few days, I’ll check how much I have left and try to get an early refill.” Public health and safety officials should consider clarification of the 1 versus 2-week recommendation and work with providers, pharmacies, and insurers to ensure that individuals are able to obtain 1 or 2 extra weeks of medication for ongoing PMP, not just during an immediate hurricane watch or warning.
Most participants had stored appropriate medical data, including medication list, doctor list, and medical history, online or in the cloud rather than on paper. This indicates a high reliance on patient portals and phone applications, which may be problematic in the event of a large-scale power outage or network failure, such as happened during Hurricane Ida in 2021.Reference Shultz, Trapido and Kossin33, Reference Fenton, Flitter and Kessler34 High reliance on the patient portal also demonstrates passive preparedness; a participant with a medication list in a patient portal did not actively put it there for the purpose of disaster preparedness. Other opportunities for passive preparedness safeguards should be explored. Most participants did not have an advance directive, which indicates an opportunity to increase preparedness for other types of large-scale disasters which might overwhelm the health care system such as the COVID-19 pandemic.Reference Berning, Palmer and Tsai35, Reference McAfee, Jordan and Cegelka36
Many participants reported that their plan is influenced by their medical needs (43.5% of those with a plan) or stated that perceived health challenges make them more likely to evacuate (23.3%). The 3 respondents who stated that their health makes them less likely to evacuate reported difficulty in at least 1 functional area (vision, hearing, mobility, cognition, self-care, or communication). These results reaffirm that health plays an important role in disaster decision-making for adults with chronic disease. This is especially true for individuals reporting disabilities, who had a statistically significant higher hurricane risk perception. Policies, programs, and guidelines for supporting individuals with access and functional needs before, during, and after disasters should be supported and expanded.Reference Mace, Doyle and Askew37-40 Similarly, the findings surrounding financial strain mirror previous studies showing lower levels of preparedness among low-income individuals and reinforce the need for protective policies and programs focused on this group.Reference Zamboni and Martin41
Strengths
The strengths of this study include its specific focus on personal medical preparedness (as opposed to general disaster preparedness) among community-dwelling adults with hypertension. The survey was collected during a time period in which there were no active disasters; thus, results may represent baseline preparedness, responding to the call many experts have made for developing disaster research capacity prior to the onset of disaster.Reference Montesanti, Walker and Chan42 Phone-based administration allowed for access to community members who would not otherwise be reached by virtual or clinic-based recruitment methods – this includes people with low technology literacy or transportation concerns and older adults.Reference Forsat, Palmowski and Palmowski43 Recruitment of an insured population reduces threats of confounding by health care insurance status.
Future Research and Applications
Further research with larger sample sizes and including more vulnerable populations (e.g., underinsured and uninsured) is needed to elucidate and confirm the exploratory associations found in this paper. In addition, policymakers in government, insurance, pharmacy, and health care systems should consider the real-world viability of preparedness recommendations. If experts recommend for patients to have 2 extra weeks of medication and medical supplies on hand, the health care ecosystem must enable and empower them to do so. Health care providers should take a more active role in helping patients to understand the relevance of PMP to their health outcomes, including practicing a “teach back” method to ensure patient understanding of personal medical preparedness. Though policies exist that allow for additional refills in the period immediately before and after disaster, these have limited helpfulness in baseline preparedness (La. Admin. Code tit. 46 § LIII-2521). Potential policies include insurance coverage for 2 extra weeks of medication once per year at the start of hurricane season or coverage for 90-day supplies of medications during hurricane season, a strategy which has been shown to increase medication adherence and decrease medication-taking disruptions compared to 30-day supply.Reference Rymer, Fonseca and Bhandary44 However, any consideration of policy change must be accompanied by further study of individual behavior. While 90-day filling options are a potential solution, co-pays and packing medications into go-bags are barriers. More work must be done to bridge the gap between preparedness policies and practices. Additional policies dedicated to increasing environmental resilience among populations with special medical needs have been suggested in the US and internationally.Reference Bean, Snow and Glencross45, Reference Wulff, Donato and Lurie46
Conclusion
Chronic disease exacerbations play a key role in post-disaster morbidity and mortality; this relationship is clear in cardiovascular disease complications following hurricanes in the Gulf South. By measuring personal medical preparedness (PMP) among individuals with hypertension in Southeast Louisiana, this study provides data about baseline PMP. This population exhibited reasonable PMP even though most had never spoken to a health care professional or ancillary staff about PMP. Having a 2-week supply of extra medications was the most missed preparedness category, with restrictive refill schedules enforced by pharmacies or insurers being the most common barrier to getting extra medications. While policies are currently in place to allow for emergency medication refills during a disaster, these data suggest that current status quo falls short in addressing PMP at baseline or before a disaster is imminent. If experts recommend that individuals with chronic disease have 2 extra weeks of medication on hand to be “prepared,” further research on improvements to policy and practice is warranted.
Author contribution
Claire Romaine was primarily responsible for study design, data collection and analysis, and manuscript preparation; TL1 Scholar for TL1TR00310.
Laura Perry assisted in study design, validated the analysis, and assisted in manuscript preparation.
Erin Peacock assisted in study design, oversaw data collection, supported analysis, and completed manuscript revisions.
Stephen Murphy, expert in disaster management, contributed to survey design and translation of the study results into key points for the discussion section. He also completed manuscript revisions.
Marie Krousel-Wood, expert in medication adherence, contributed to both study and survey design and provided senior oversight to data collection and manuscript preparation, Co-director of TL1TR00310 and PI of R01HL153750.
Funding statement
Research reported in this manuscript was supported, in part, by the National Institutes of Health under Award Numbers TL1TR00310 and R01HL153750. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Competing interests
None declared.
Disclaimer
Marie Krousel-Wood MD is a member of the United States Preventive Services Task Force (USPSTF). This article does not necessarily represent the views and policies of the USPSTF.
Open access
