The study data comprise survey results from a convenience sample of English- or Spanish-speaking adult residents and workers who lived, worked, or volunteered in Louisiana parishes with flooding and building damage from Hurricane Ida, collected at the request of the Louisiana Department of Health. Residents and workers ≥18 years of age were recruited in-person at recruitment sites (including 13 home improvement stores, three not-for-profit organization and mold remediation training organization work sites, two farmers markets, one FEMA Disaster Recovery Center, and one rural health clinic) between December 4–14, 2021, during recovery and renovation efforts (Hurricane Ida made landfall on August 29, 2021). Additional detail on data collection is described elsewhere [14]. This activity was reviewed by CDC, deemed not research, and was conducted consistent with applicable federal law and CDC policy.¹

Survey administrators asked three open-ended questions: 1) among those responding “yes” or “no” to the stem question, “Do you think you are personally at risk of getting sick from mold?”, respondents were asked the open-ended question: “Why do you think you are/aren’t at risk of getting sick from mold?”, 2) If respondents reported having personally cleaned up mold, and after specifying their PPE use, they were asked, “Why do you use/not use this protective equipment?”, and 3) “How do you know that a mask² fits?” To ensure that participants would consider the full spectrum of respirators, despite the simplified terminology of “mask,” respondents were shown a display of face-worn products that included (1) NIOSH Approved^® respirators and respirators meeting international standards, (2) cloth masks and disposable face masks, and (3) face coverings such as bandanas. The face-worn products shown included those considered to be protective from a mold hazard when used properly (i.e., respirators) and those not considered to be protective from a mold hazard, even when used properly (i.e., cloth masks, disposable face masks, and face coverings). Participants self-reported other types of PPE without visual aids. Survey administrators took detailed notes to record the verbal responses to each open-ended question.

Given the straightforward nature of the open-ended questions and field notes, which included brief responses, we selected a qualitative descriptive content analysis approach guided by the HBM. Qualitative descriptive analyses focus on “low-inference description” [15, 16]. Thus, the analysis aimed to develop codes that stayed close to the ideas expressed by respondents, rather than more interpretive connections.

Two researchers separately completed initial coding of the responses. Initial codes [17] identify recurrent ideas in the data. The researchers then reviewed their initial codes together to inform one comprehensive list of final codes and to group codes into categories. This process utilized consensus-building to identify, agree upon, and finalize categories with deductive assessment regarding how codes related to HBM constructs [12, 17]. For example, separately, the two researchers identified codes related to using PPE to protect oneself or others. After meeting to discuss, the researchers agreed that these codes related to the “perceived benefits” of PPE use during mold cleanup.

After all categories were decided, the researchers referred to the frequency of the codes occurring within each. Because these data were collected as part of an emergency response, saturation of the open-ended responses was not continuously assessed. Therefore, codes assigned to only a single response and responses with unclear meanings were dropped. Although the frequency of some codes might not have reached code saturation (i.e., some codes had few instances reported) the researchers were most concerned that the pattern of codes within each category met the criteria for meaning saturation [18].

Despite relatively similar patterns of responses by workers and residents, frequencies of categories and codes are reported separately for workers and residents. “Workers” self-reported that they were involved in mold clean-up activities as part of their job, on a volunteer basis or for wages. “Residents” resided in the survey catchment area but were not involved in mold clean-up activities as part of their job. With variation depending on their employer or volunteer organization, workers might be more likely than residents to fall under respiratory protection programs (RPPs [19, 20]), to have access to fit testing, to be provided NIOSH Approved respirators, and to receive information about mold exposure and PPE from employers [20].

Because sample respondents who are current or former smokers were more likely to report perceived risk from mold compared with never smokers, supplemental materials also display responses to “Why do you think you are/aren’t at risk of getting sick from mold?” stratified by ever smoking (Table S1, http://links.lww.com/JOM/C83), as well as responses to “How do you know a mask fits?” stratified by whether the respondent endorsed only using recommended respirators while cleaning mold (Table S2, http://links.lww.com/JOM/C83).

The survey asked, “Have you ever heard or seen any public messages from the State, like the health department, or government about cleaning up mold? (for example, public service announcements on radio, TV, billboards, flyers, or social media).” Those who answered affirmatively were asked, “Was this information helpful to your cleanup efforts?” Respondents selected their main source of information about mold, and their perception of the best channels to disseminate mold-related messages to the public. They were also asked if they had previously seen any of three existing federal communications materials relevant to mold cleanup (Figure 1). We describe frequencies and proportions of responses, stratified by resident versus worker respondents.

For residents, the highest percentage (27%) reported that their main source of information about mold was from the internet. A similar proportion of workers (26%) reported their main source of information was from their employers (Table 4). Few residents or workers reported government websites or the health department as a source of mold information.

Most residents (63%) and workers (59%) reported that they had never heard or seen any public messages from the state about cleaning up mold. Among those who had, about two thirds of residents (67%) and workers (64%) attested that the information was useful to their cleanup efforts.

Most residents (62%) reported that they had never seen any of three existing federal communications materials relevant to mold, mold exposure, and safety practices for cleaning up mold that were shown to them (Figure 1), while 41% of workers reported never seeing any of these materials.

Most residents (55%) and workers (69%) responded that social media would be the best way to get messages about mold to the public, followed by television (44% and 41%, respectively) and radio (23% and 31%, respectively).

Exposure or lack of exposure to mold was the dominant explanation respondents offered for why they did or did not feel at risk of sickness from mold. Perceived susceptibility, and to a lesser extent perceived severity, have shown associations with individuals’ likelihood of wearing respiratory protection in prior studies of infectious disease [13]. This aligns with research on HBM constructs and health-protective behaviors such as vaccination initiation; perceived susceptibility was significantly associated with behavior more often than perceived severity [12]. Messaging to improve recommended PPE use during mold cleanup could focus on connecting mold exposure to susceptibility to sickness from mold, including how to recognize mold or indoor moisture sources [21]. Additionally, messaging could highlight that mold spores are a health hazard whether viable or nonviable [2].

Some who did not feel at risk described their perceived low susceptibility as an inherent trait (e.g., being “healthy” or “lucky”). Evidence suggests that individuals with certain health conditions (such as those who are immunocompromised or have asthma) are indeed more susceptible to severe health effects of mold exposure [8, 22]. However, PPE is universally recommended when cleaning mold, regardless of individual risk factors [6]. Some prior research suggests that narrative methods of health education might be more effective than traditional didactic methods, even in the context of recipients’ possible over-optimism about their own risk level [23, 24].

Similar to the current study, prior qualitative work found that Houston, Texas residents after Hurricane Harvey also identified mold as a threat to their health [10]. Unlike the Gandhi and colleagues’ study, which was conducted three years after a hurricane, our study was conducted only 14 weeks after a hurricane. They used focus groups of a mostly older population with semi-structured discussion guides, rather than simple open-ended questions, to investigate perceptions about mold cleanup. They also identified themes of perceived susceptibility and severity of mold-related health issues that overlap with the current study. Disseminating information that alerts hurricane-affected residents and workers of their susceptibility to health effects from mold might increase residents’ and workers’ perceived threat of health effects from mold exposure. For those conducting mold clean-up in the future, perception changes might improve mold cleanup practices.

Findings indicate that HBM constructs might play a role in decisions about PPE use during mold cleanup, aligning with prior research [13]. Respondents who used PPE during mold cleanup commonly described wanting to avoid mold exposure, perceived benefits of PPE such as personal protection, and concepts relevant to perceived threats such as avoiding sickness. Perceived barriers to PPE usage included discomfort and inconvenience. A few workers perceived low access to PPE, including through their employer. An HBM-guided approach does not robustly model structural factors [12] which might have implications for access. Consequently, a more comprehensive evaluation of residents’ and workers’ access to PPE for mold cleanup, including access to employer-provided PPE and fit tests for workers, could be informative when selecting interventions intended to increase proper PPE use during mold cleanup. Finally, some residents noted that PPE disuse sometimes occurred because they simply did not think of it. These responses suggest that messaging campaigns might be cues to action. A prior study of coal miners suggested that cues to action including reminders from family or a doctor was associated with increased likelihood of workers using respiratory protection [25].

Responses to the question “How do you know a mask fits?” could be relevant to respondents’ self-efficacy in using of respirators for mold clean-up [26]. Prior research has shown that self-efficacy around use of respirators was positively associated with intention to use respirators on high air pollution days in a longitudinal survey of young adults in Beijing, China [27]. Respondents to the current study showed understanding of several important concepts related to user seal checks and/or fit tests. Some responses indicated ignorance or misunderstanding about concepts related to respirator use, and messages could reinforce recommendations and counter misunderstandings. For example, facial hair must be limited enough not to interfere with a respirator’s seal in order to allow for proper fit [20, 28]; workers using a respirator under an RPP are required to be clean-shaven [19, 20]. Importantly, several respondents reported that they “don’t know” how to assess whether a respirator fits, which might indicate low self-efficacy surrounding respirator use. While there is no formal system analogous to a RPP to guide non-occupational respirator use [20], disseminating messages for the lay public about how to conduct a user seal check [29, 30] could help those without access to fit testing increase their sense of self-efficacy, which in turn might increase their likelihood of using recommended respirators during mold cleanup and in other scenarios, such as wildfire smoke events.

Results suggest that many residents and workers have not heard or seen messages from the state about cleaning up mold; fewer respondents reported having heard public messages about mold in this survey compared to one conducted after Hurricanes Katrina and Rita [9]. Few respondents endorsed that the health department or a government website was their main source of information about mold, and most residents (62%) had never seen any of the offered federal communications materials about mold. These findings suggest that there is opportunity for public health practitioners to reach a wider audience of residents and workers in hurricane-affected areas.

Diverging from some prior work [10, 31], most participants endorsed social media as a good way to disseminate mold-related messages to the public. Emerging research on public entities’ social media use for emergency response reveals both opportunities and challenges. For example, new artificial intelligence-based tools could be used (with attentiveness to the challenges of emerging tools [32]) to analyze or generate social media content quickly, but misinformation is also easily spread and difficult to counter on this medium [33–35]. As well, in Hurricane Ida’s aftermath, residents and workers faced varying levels of Internet access [36, 37]. Communicators could consider multiple strategies, such as message dissemination ahead of the hurricane’s landfall.

Resident and worker responses to the open-ended questions showed some similar patterns. Mold exposure was a common theme in both groups’ responses to questions about why they do or do not use PPE, and about why they do or do not feel at risk of health effects from mold.

However, some patterns of worker responses diverged from residents. Unlike residents, workers never mentioned mold remediation as a reason that they did not feel at risk, possibly because workers are engaged in mold remediation as part of their occupation.

A smaller proportion of workers than residents gave responses to the question “How do you know a mask fits?” indicating lack of awareness or misunderstanding of important concepts related to respirator use. This might be because some workers are covered under an RPP, which requires workers who need to use respirators undergo medical clearance, respirator use training, and fit testing, among other components [20]. However, some workers’ responses still indicated ignorance or misunderstanding of concepts related to PPE use. Of workers, 7% said they “did not know” how to determine if a mask fits. Some workplaces with a small number of employees, self-employed workers such as freelance handy-people, and some other workers are not covered by RPPs [20]. Reaching out to larger and smaller employers, as well as volunteer organizations, could help ensure that more workers benefit from respirator use, thus protecting them from possible occupational hazards.

This study’s data relied on a convenience sample, which might not be representative of southeast Louisiana or of all hurricane-affected populations. The predefined sample of respondents, recruited during emergency response, limits our ability to assess for saturation in qualitative responses [18]. Additionally, the survey was designed primarily to identify knowledge, attitudes, and perceptions, and the HBM framework was not incorporated in its design. The mapping of responses onto HBM constructs is post-hoc and exploratory, and not a formal test of the HBM. This study can inform messaging relevant to individual perceptions; however, these findings might be less informative regarding how to improve access to PPE. While a strength of the data collection was its effort to incorporate worker respondents, there may be unmeasured differences in working conditions relevant for perceptions about PPE use across respondents depending on the type and size of their employer; and if unpaid volunteers, what kind of volunteer organization they worked for, if any. Due to the sample size of workers, distinctions in patterns of response between paid workers and volunteers could not be explored; future research could investigate whether these types of workers differ in terms of their knowledge, attitudes, and practices about mold cleanup and PPE use. Future research might consider collecting more detailed information about working conditions and how it might relate to perceptions about PPE use.

The findings from this study indicate that government messaging about PPE usage is not reaching target audiences and that use of social media to disseminate such messages could be a possible solution. The results affirm that HBM constructs might be relevant to perceptions regarding PPE during mold cleanup, and that perceived susceptibility to sickness from mold could be a fruitful focus for messaging. Knowledge gaps remain regarding proper usage of respirators for occupational and non-occupational settings. Findings could inform message development for promoting recommended PPE use behaviors among those in hurricane-affected areas.