Community Perspectives on Zika Virus Disease Prevention in Guatemala: A Qualitative Study

Guiding concepts.

The extended parallel processing model²⁰ posits that preventive behaviors are most likely to be adopted when perceptions of risk and efficacy are high. Perceptions of efficacy include self-efficacy, defined as the community members’ confidence in their ability to perform the preventive behaviors in a manner satisfactory to them, and response efficacy, defined as the behaviors’ effectiveness to avert the threat. Self-efficacy is influenced by a host of factors, including the level of skill to perform the behavior,²¹ ease and convenience, access to enabling commodities (e.g., larvicide, window screens, or skin repellent in this case), household and community collaboration and other societal factors,²² descriptive and injunctive social norms,²³ and periodic encouragement and feedback about the behaviors practiced,²¹ all of which increase community self-efficacy for continuing to practice these behaviors.²¹ Community self-efficacy is better described as collective efficacy, a concept that stresses the interdependent efforts and influences of community members to achieve desired results, whether working collectively or individually.²⁴ SBCC may be more effective in increasing people’s efficacy if it strategically addresses these influencing factors. We explored perceived seriousness of Zika illness in the community as a proxy for risk. As a proxy for efficacy, we explored the cultural salience, perceived effectiveness, collective efficacy, and feasibility of Zika prevention recommendations in the community context.

Selection of study sites.

In collaboration with the USAID/Guatemala and Asociación Pro-Bienestar de la Familia de Guatemala, a USAID-funded Zika-implementing partner in Guatemala, we selected two study sites: Mazatenango, Suchitepequez, a town in the Pacific lowlands 371 m above sea level, and Barberena, Santa Rosa, a town in the southeast 1,221 m above sea level. These towns are referred to in this study as the lowland town and the highland town, respectively. Zika cases had occurred in both towns according to the Ministry of Health reporting, and the partner organization was implementing Zika-related home visits and community-level intervention activities in both towns through community outreach workers or clinic staff who integrated Zika into their regular counseling on family planning, reproductive health, or antenatal care.

Selection of study participants.

In each town, a community outreach worker who engaged urban or peri-urban communities in the organization’s intervention activities extended invitations to interested clients through convenience sampling. Invitations were made according to the following selection criteria: women aged 18 years or older who were currently pregnant, male partners aged 18 years or older of currently pregnant women, and women aged 18–30 years with a steady partner and either no children or a single child. We considered the latter group as being likely to become pregnant in the near future.

Data collection.

Between July 27 and August 1, 2017, we conducted two focus groups with one population group per site, for a total of 12 focus groups, six per site, during mostly rainy weather. All data collection team members were experienced in qualitative methods, fluent in Spanish, and familiar with Zika. The focus group moderator was a Guatemalan national. We used a structured focus group discussion guide, touching first on the perceived seriousness of Zika, followed by three collective rank-ordering tasks of 19 cards, each depicting household or community actions. Many, although not all, of these actions represented a Zika prevention method recommended at the time. Several methods were informed by the Strategic Communication for Zika Prevention: A Framework for Local Adaptation.¹⁰ The moderator introduced the cards to the participants not as methods to prevent Zika but as a series of actions by community members and not directly by the participants themselves, to discuss during the rank-ordering tasks. Participants rank-ordered each card three times, according to its relative effectiveness for preventing Zika, according to its relative collective efficacy to be performed by community members, and according to its relative feasibility in the community context. The moderator promoted discussion on why a depicted action was ranked higher or lower than another and to shed light on disagreements among participants when applicable. All discussions were audio recorded.

Arriving participants (n = 68) were first asked to individually free list²⁵ all the actions that people in the community currently take to avoid Zika, and their responses were recorded in writing, in the sequence in which they were given. We probed further to gain more detail on the nature of the action mentioned and to ensure that the participants mentioned all potential actions they knew of.

Informed consent.

We obtained oral informed consent from all participants. The research protocol and the recruitment and consent scripts were approved by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board in the United States (protocol 00008014) and by Zugueme, an independent ethics committee in Guatemala (protocol OFZU 1387-17).

Data analysis.

A team of six researchers fluent in Spanish and familiar with Zika and the Guatemalan context analyzed the data, four of whom were also present for all data collection activities. We entered free-listing data into Excel and consolidated similar wordings of the same action into a single best wording. We calculated individual salience for each action mentioned by each participant and (mean) Smith’s salience²⁵ for each action across all participants in each target group and at each study site. Smith’s salience is a statistical accounting of rank and frequency which describes the relative importance of items included in a mental category, in this case, actions to prevent Zika.

We transcribed focus group recordings and conducted thematic analysis on the textual data providing the context on perceived seriousness of Zika illness and on why a certain action was ranked higher or lower than another action in the card set. We developed topical codes by action, effectiveness, collective efficacy, feasibility, and relevance to Zika seriousness; we then coded the data in MaxQDA 2007 (VERBI GmbH, Berlin, Germany).

The rank order of the 19 cards was photographed and entered into Excel. The measures ranked higher have the lowest scores, and those ranked lower have the highest scores. We calculated mean ranks for each card-depicted action, by effectiveness, collective efficacy, and feasibility, in each target group and at each study site. We calculated a shared mean rank in the case of tied-ranked cards. We considered the rankings as high (ranked from 1 to 6.33), variable or moderate (ranked from 6.34 to 12.66), and low (ranked from 12.67 to 19) by dividing 19 into three approximately even intervals. To summarize the emerging theory on efficacy as shaped by the data, we rank-ordered Smith’s salience and the three sets of overall mean ranks (effectiveness, collective efficacy, and feasibility) for each of the card-depicted actions, as well as condensed the qualitative data.

Sociodemographics of study participants.

Each focus group included five to nine individuals, for a total of 70 participants, who were of Mestizo and Indigenous ethnicities. The majority of the women were homemakers (15 of 20 were pregnant women and 17 of 28 women were likely to become pregnant). The other women included three domestic workers, two food vendors, three teachers, two students, a social worker, a secretary, and a sales associate. One woman did not provide any occupation. Male partners of pregnant women included three farmers, two security guards, two mechanics, two students, a lawyer, a cook, a salesman, a supermarket employee, a credit inspector, a commercial driver, a laboratory technician, and an ornamental flower grower. One man was recently unemployed. Table 1 summarizes participants’ demographics.

Perceived seriousness of Zika illness overtime.

A majority of women participants and many of the men described their community as being seriously concerned when they first heard about Zika because it appeared soon after an outbreak of chikungunya. In that outbreak, almost everyone had experienced severe symptoms of the disease, and they subsequently feared that Zika would be similar. But the milder symptoms or lack of symptoms associated with Zika reduced the public’s concern, and a parallel decline in messaging led participants to believe that the worst was over.

Men cited concerns about the impact that Zika could have on pregnant women and babies, but they placed the primary responsibility for preventing Zika on women Overall, uncertainty existed about the differences between Zika, dengue, and chikungunya and whether Zika was the same as the other two diseases.

Cultural salience of actions taken by people in the community to avoid Zika.

During the free-listing task, 68 participants mentioned 214 different actions taken by people in the community to avoid Zika. These actions were consolidated into 88 actions according to similarity. A mean of 7.06 actions were elicited per participant. Table 2 shows the first 32 elicited actions in the order of Smith’s salience.

Because of the large number of actions elicited, the salience data are more meaningful in comparative terms than as individual values (Table 2). Bed net use had the highest salience possibly because of its consistent promotion for malaria control, and it was more salient for the pregnant women than the other groups, likely reflecting antenatal counseling at the time. “Deschatarrización,” or the elimination of containers that accumulate “agua del cielo” or rainwater, had the second highest salience. Participants often explained that the mosquito larvae come with the rainwater or that the rainwater attracts the mosquito, which then places its eggs and larvae in the containers. The third most salient action was skin repellent use, which was more salient for the residents of the lowland town than those of the highland town. The lowland town residents explained that the coast generally has many more mosquitoes than the rest of the country.

Among other high-salience actions was general cleaning of the house and its surroundings (fourth), which included sweeping dust away, sweeping under beds, and keeping the yard clean. Condom use to prevent the sexual transmission of Zika (fifth) was more salient for the male partners of pregnant women, followed by the pregnant women, and less salient for the women likely to become pregnant, indicating that this new Zika-specific recommendation was coming across. Although lower in salience (31st), a vaccine for Zika was mentioned in two ways: some women thought that they had already received a Zika vaccine in the arm along with other vaccines received at antenatal care, whereas others termed an injection of strong pain medicine to calm the symptoms of chikungunya a vaccine. Male participants noted that clandestine vendors sold such “vaccines” during the chikungunya outbreak.

Perspectives on the relative effectiveness of Zika prevention recommendations.

Table 3 presents the ranking data of the card-depicted actions according to their relative effectiveness to prevent Zika, by participant group.

This finding, along with the high salience of cleaning the house and the area around it (Table 2), reveals people’s lack of an exclusive connection between the immature mosquito stages in the water containers and Zika prevention, or a perception that mosquitoes are also harbored in dry waste.

Family planning to delay pregnancy during the epidemic ranked generally low in effectiveness (13.13–16.5, Table 3) because most participants did not make the association with Zika prevention. However, a very small number of participants in two focus groups said that the methods were effective in helping to delay pregnancy during the Zika epidemic.

Elimination of exposed tires ranked moderately (7.13–8.31, Table 3), although participants were aware that tires led to mosquito infestation, describing the difficulty in eliminating them and coloring effectiveness by collective efficacy (Table 4). Emptying of buckets ranked variably (7.44–9.13, Table 3) because it did not solve the problem, according to participants—it only transferred standing water to puddles on the ground. This finding reveals people’s lack of awareness that Ae. aegypti lays its eggs exclusively on the inner walls of water containers, in contrast to other mosquito genera, which lay their eggs directly on the water (such as puddles on the ground). Water chlorination in the drum followed by covering the drum and a related card showing the hermetic covering of water storage drums ranked variably among the participants (6.50–14.00, Table 3) as being not consistently effective because mosquitoes often found their way in under the cover.

Despite the high salience of bed nets (Table 2), their use during pregnancy ranked low in some groups and moderately overall (7.38–12.13, Table 3), while individual participants disagreed with their group depending on whether they themselves already owned and used a bed net or not. When bed nets were assigned a low effectiveness ranking, it was because of their apparent inability to stop the mosquitoes. However, no one mentioned that Ae. aegypti is a day-biter, which suggests that people do not differentiate between mosquito types. Skin repellent use during pregnancy received a wide range of group rankings (2.50–17.00), averaging between 10.25 and 13.63 (Table 3). Despite being recognized as effective in most discussions, participants who assigned it low effectiveness were influenced by feasibility (Table 5), citing access problems, such as cost, and the need to constantly reapply it being impractical.

Perspectives on collective efficacy and feasibility to practice the Zika prevention recommendations.

Table 4 presents the ranking data of the card-depicted prevention measures for relative collective efficacy, whereas Table 5 presents the ranking data according to their relative feasibility.

Prevention methods that ranked quite high for both collective efficacy and feasibility were the detailed pila cleaning and outdoor sweeping of non-containers (Tables 4 and 5). Detailed pila cleaning ranked highest for collective efficacy (4.13), as it had for effectiveness (Table 3). Importantly, participants reiterated that there was a right way and a wrong way to clean the pila and that not everyone did it correctly; however, they expressed high collective efficacy for the correct way, according to the steps that the pila cleaning card illustrated, namely, with detergent, bleach, and a brush, and reiterating the final step of water disinfection with bleach on refilling if the water was not chlorinated. Cleaning the pila also ranked highest for feasibility (4.21), and many participants indicated that all their community members did it, from constantly and daily to three times a week to once a week—“Who doesn’t do it?” For some, cleaning the pila was considered cost effective, whereas for others, a certain cost was implied from having to buy soap and bleach, and due to an inadequate water supply. One group of pregnant women gave this measure a low feasibility ranking because pregnancy made it hard to bend and do the cleaning well, but the overall high rankings revealed strong underlying community norms of cleanliness and a sense of control. In reference to who actually cleans the pila, men generally thought the women should do it, although several men shared that they too will help with pila cleaning especially during a partner’s pregnancy.

The importance of maintaining a clean home was a recurring theme, and outdoor sweeping was its embodiment. Two participants indicated that “Cleanliness is the main thing.” However, for the general waste removal to be feasible, reliance on local municipalities was necessary. This reliance lowered the collective efficacy and feasibility of outdoor sweeping (6.83, Table 4; 6.46, Table 5).

Eliminating any container-like objects from public spaces as well as tires ranked higher in feasibility (7.88, 7.75, Table 5) and lower in collective efficacy (10.08, 10.00, Table 4) in contrast to the high effectiveness of both actions (Table 3). This was because public spaces were not under the participants’ exclusive control; broader community collaboration was needed to perform these measures in public spaces. Furthermore, participants mentioned that each town has locales known as “Pinchazos,” or tire repair shops that possess a large quantity of tires and take little or no action to store them properly. Households have little control over such situations and depend on the functionality of the local government’s waste services.

Emptying buckets ranked as a highly feasible action by some groups, but other participants indicated that they must store water and collect rainwater because of the lack of infrastructure supporting consistent delivery of water to their households. Emptying buckets was more palatable if it leveraged alternative uses such as outdoor washing/cleaning. It, therefore, received moderate rankings (7.00, Table 4; 6.88, Table 5).

Covering water storage drums had lower feasibility (7.00, Table 5) because of a lack of access to proper covers, lack of a proper cover, and the financial cost to cover all water storage drums. Households may opt to only cover the water that must remain clean for drinking or food preparation, as opposed to water used for less critical activities. But covering ranked higher for collective efficacy (6.44, Table 4), indicating a positive attitude that could be leveraged if the feasibility barriers could be overcome.

Bed net use during pregnancy received moderate rankings (7.63, Table 4; 8.00, Table 5), as with effectiveness (Table 3), despite its high salience (Table 2). In contrast to the women’s groups, the men in the highland town conveyed high collective efficacy (mean rank 3.5), whereas the men in the lowland town conveyed lower collective efficacy (mean rank 9.75) because of the excessive heat typical for their geographic location. The latter men pointed out how important the bed net was with so many mosquito-borne diseases to the extent that they were trying to get their children used to sleeping under a net from the start because it was much more difficult to get used to it as an adult.

Using skin repellent during pregnancy had moderate rankings (9.83, Table 4; 11.58, Table 5) because of the need for frequent reapplication and because of the cost implied. Furthermore, as a male participant explained, there was a culture clash: during pregnancy and after giving birth, new mothers are looked after by their own mothers who will not let them use repellent. Their mothers view it as a dangerously strong product that should not be used when a pregnant woman or a young mother and her new baby are “en dieta”^† because of their vulnerable state.