Study design and population

A qualitative study was conducted in the Colombian cities of Cali (situated in the south-west) and Villavicencio (situated in central-east) in 2018. Both locations have a tropical climate, with average daily mean temperatures ranging from 23.4–24.2°C and 24.2–26.7°C throughout the year in Cali and Villavicencio, respectively (meteorological data from the Insituto de Hidrologia Meteorologia y Estudios Ambientales). The Colombian Departments of Valle de Cauca and Meta containing the cities of Cali and Villavicencio, respectively, experienced different Zika incidence between 2016–2017, with the Pan American Health Organization (PAHO) reporting upper estimates of 40 cases per 100,000 inhabitants in Meta and 605 cases per 100,000 in Valle de Cauca [23]. The study consisted of a series of 19 focus group discussions (FGDs) with female participants, and 14 one-to-one semi-structured interviews (SSI) with male participants. Participants fell into 3 categories: pregnant women, non-pregnant women of reproductive age, and their male partners. Pregnant and non-pregnant FGDs were held separately. In total, 10 FGDs were held with pregnant, and 9 with non-pregnant women. These FGDs took place in public institutions (such as libraries and health centers), contained between 2–11 participants (average of 7.5 per group) and lasted between 45–60 minutes. The focus group with only two participants was convened after the particular participants dropped out of their initially scheduled sessions. There were 14 one-to-one interviews with male participants, each lasting between 10–20 minutes. A total of 143 (71 pregnant and 72 non-pregnant) women, and 14 men consented to take part in the study. The female participants' ages ranged from 18 to 45 years, with an average age of 27.7 years. The male participants’ ages ranged from 19–58 with an average age of 36.7 years.

The study was designed to include participants from a range of lower and higher socio-economic strata [as defined by residential area, and designated according to the Colombian “estratos socioeconomicos” (socioeconomical strata) criteria] (Table 1). Recruitment was propositional, whereby organized groups containing potentially eligible participants were contacted by telephone, and all accepted agreed to attend. The groups contacted included: pregnant women’s groups, women's groups of sexual and reproductive health programs, and women's groups caring for children. For male SSI, recruitment priority was given to the partners of women participating in the FGDs. The inclusion criteria were as follows: pregnant women were required to be of legal age, were required to be pregnant (at any stage of gestational development), and needed to agree to participate in the study; non-pregnant women were required to be of legal and reproductive age, and needed to agree to participate in the study; males were required to be of legal age, and agreed to participate in the study. Written consent to take part in the study was obtained from all participants.

Interview structure

The interviewer in Cali was Carolina Mendoza, social worker and MSc student; and the interviewer in Villavicencio was Gloria-Isabel Jaramillo, PhD. Both interviewers were female and research scientists with a background in arboviruses at their respective universities. Both interviewers have had previous experience in performing FGDs for biomedical research. There was no relationship between the interviewers and the participants prior to commencement of the study. In Cali no people other than the interviewer, a field assistant and participants were present. In Villavicencio no people other than the interviewer, three observing medical students and the participants were present. Participants were introduced to the study and were told that the aim was to identify their concerns about Zika and their opinions around different personal protection methods. The FGDs were semi-structured and led by the moderator according to a predefined interview guide (see S1 Appendix). Interviewees were asked to speak freely, although prompts were provided if participants were reluctant to speak. They were conducted in Spanish, with the moderator asking the questions and allowing participants to answer directly and interact with the other participants. FGDs were either video recorded (Villavicencio) or tape recorded (Cali). Field notes were not taken. At the end of each session, recording was stopped and participants had the opportunity to ask the moderator questions if they had any concerns about Zika. They were also given a Zika fact sheet developed by the U.S. CDC.

Data analysis

Focus group transcripts were analysed in their original Spanish by researchers at Universidad del Valle and Universidad de Cooperativa del Colombia, sede Villavicencio. English translations were produced, and were provided to researchers at the London School of Hygiene and Tropical Medicine (LSHTM) for independent analysis. Transcripts were analysed by thematic coding according to a pre-defined coding scheme. Coding was performed using the Atlas.ti (version 8.0) (Scientific Software Development Gmbh, Berlin) and NVivo (version 11.0) (QSR International, Melbourne) software packages. There were 7 coding families in total, representing: common mosquito control methods, arbovirus concerns, Zika knowledge, sources of Zika knowledge, changes to control behaviours during the Zika outbreak, ideal mosquito control tools, and views around PPT. Each theme family was divided into a larger number of sub-families. Each interview was independently analysed by at least three individual researchers, and results were compared and discrepancies discussed until a consensus was reached. As the purpose of using the two study sites was to increase diversity and not to directly compare the two cities, comparisons have only been drawn when clear differences between study sites emerged.

Ethics

The study was approved by the ethical review boards of LSHTM (application number: 13592) and Universidad del Valle (Univalle: 004–018).

Study limitations

Whilst conducting a partial retrospective study was necessary in order to gain insights into the knowledge and perceptions of Zika virus before and after the outbreak occurred in Colombia, studies of this kind can be problematic due to the reliance on participant memory and response. Responder opinions of their actions and awareness may be tainted by experiences that have occurred since the period they have been asked to reflect upon or the pressures felt at the time of being asked specific questions. These factors are likely to introduce responder and/or recall bias. The validity of this study is also at risk of ‘social interaction threat’, which is expected to occur within a focus group setting due to the social relationships and interactions present within this environment [24]. To negate issues associated with confirmability, the same data were analysed by multiple researchers. Despite this, due to the level of interpretation required to analyse qualitative data, there is still likely to be an element of researcher bias introduced by personal idiosyncrasies. In addition, issues around transferability may be apparent, since findings concluded are specific to the participants and may not be generalisable to alternate group in Colombia. Field notes were not taken and hence non-verbal information could not be considered. We believe this should not affect the results as recordings were available and transcripts were analysed by teams including the individuals performing the field work.

The importance of Zika and other arboviruses

The magnitude of the perceived threat from the major arboviruses (dengue, Zika, chikungunya and yellow fever) varied considerably among study participants. Although the majority acknowledged the potentially serious consequences of arboviral infection, it was frequently described as less concerning than other infections, such as flu and diarrheal disease. However, some participants viewed arboviruses as a significant daily worry and this was often paired with a personal history of infection, or infection of a close family member or acquaintance.

“No, they’re [arboviruses] alarming. When my son had dengue, his platelet count decreased in less than two hours”. (FGD3, Non-pregnant, Cali)

Within the non-pregnant cohort, dengue, specifically severe dengue, was generally considered to be the arbovirus of greatest concern. Reasons given included a belief that dengue had the most serious acute symptoms, with a high probability of hospitalization, and the risks of mortality associated with severe dengue. Several respondents perceived a greater threat from dengue due to a belief that transmission was historically more consistent year-on-year, compared with the relatively emergent Zika and chikungunya epidemics, and, therefore, carried a greater risk of recurrence and infection. Similar to non-pregnant females, male participants were less likely to consider Zika as a primary arboviral concern, expressing greater fears over dengue and chikungunya.

“I had Zika but it wasn’t so serious. I had joint pain but it wasn’t so severe”. (FGD4, Non-pregnant, Cali).

“My feeling is that dengue is the most dangerous because it can be life-threatening.” (FGD13, Non-pregnant, Villavicencio).

“At home we’re more concerned about chikungunya, classic dengue, and hemorrhagic

Dengue [than Zika].” (Male interview 1, Villavicencio).

Pregnant women tended to be more likely than non-pregnant women to consider Zika the arbovirus of primary concern. Discussion among pregnant women often centred on concerns of the complications that can arise from congenital Zika transmission, particularly microcephaly and cognitive impairment.

I think we all worry about viral diseases but we worry even more about Zika because…it affects primarily the foetus. (FGD11, Pregnant, Cali).

Chikungunya and yellow fever were least reported as the arbovirus of greatest concern in both the pregnant and non-pregnant groups. However, fears were frequently raised over the joint pain and arthralgia associated with chikungunya infection, with a belief that it carried the most severe chronic symptoms.

Knowledge concerning Zika transmission and pathology

The majority of respondents in both the pregnant and non-pregnant groups reported an awareness of Zika as a disease threat and described mosquitoes as the source of transmission. However, the depth of knowledge beyond this varied considerably among participants, with some providing a description of Zika-specific symptoms and disease manifestations (e.g. birth defects), while others were able to give only a general arbovirus description (e.g. causes a rash, fever), or confused Zika with chikungunya, reporting a belief that they were alternate names for the same disease.

“I thought chikungunya and Zika were the same. I’ve just realised that they’re different” (FGD4, Non-pregnant, Cali).

“Very little [personal knowledge of Zika] …I can’t clearly make the difference between all of them [arboviral diseases] (Male interview 6, Villavicencio).

Women in the pregnant cohort were more likely than non-pregnant women to discuss congenital Zika, although awareness was not universal, with several participants, including some pregnant women claiming that the FGD was the first time they had be told about Zika-related congenital abnormalities.

“I didn’t [know about congenital Zika syndrome], I’m so scared now that I’m going to carry a mosquito net everywhere” (FGD6, Pregnant, Cali).

Pregnant woman (P1): “When I was hospitalized, a doctor told me that Zika is more dangerous in pregnant women because it mostly affects the unborn. The baby might be born with microcephaly: with deformed head and eyes.”

(P2): “Is it true? That’s scary.”

(Moderator): “What about the others?”

(P2): “I didn’t know.”

(P3): “We weren’t aware of all that.”

(Moderator): “What about you?” [to P4]

(P4): “Actually, I didn’t know that either”. (FGD18, Pregnant, Villavicencio).

Moderator: “What about Zika?”

Interviewee: “I don’t know what that is.”

Moderator: “Before I mentioned Zika, you have never heard of it?”

Interviewee: “No, I hadn’t” (Male interview 1, Cali).

Knowledge levels among the male partners were similarly variable, with some men showing familiarity with Zika-specific symptoms and transmission, while others claimed no knowledge beyond a basic awareness of it as an emerging disease.

Participants often acknowledged possessing a limited depth of knowledge of Zika. Several reasons were given for this, including a feeling of personal detachment from the disease:

“Sometimes you see flyers at Health Promotions Agencies. However, as you haven’t been through that disease, you don’t read them”, (FGD1, Non-pregnant, Cali)

and, a perception of insufficient community messaging on the part of the local health authorities.

“I think the information [distributed by local authorities] was good enough but too fleeting. During the disease peak, they did give information. It was basic, though”. (FGD13, Non-pregnant, Villavicencio).

Few details were offered on the behavioural and visual characteristics of the vector, although it was frequently described as being ‘white-legged’, likely a reference to the white banding on the legs of Ae. aegypti and Ae. albopictus. Only one respondent correctly described the Zika-transmitting mosquito as being a day-biter. Few participants mentioned sexual contact as a possible transmission route.

Participants frequently reported television news broadcasts, private conversations with friends and family, and local health center messaging as the principal sources of information about Zika. A smaller number of participants described using the internet and social media as an information source, although it was noted by several respondents that the internet was often misleading and was generally considered to be unreliable. Information originating in local health centers, largely in the form of discussions with health workers and distributed leaflets, was thought to be particularly valuable, carrying greater authority than other messaging.

Numerous respondents reported receiving information from health authorities concerning mosquito bite prevention, with several describing attending seminars at health centres or schools, where messaging was reported to focus on larval management and the use of bednets. A minority of women, largely in the pregnant cohort, reported receiving information concerning sexual transmission, and described condom use as an appropriate precaution.

Household mosquito abatement activities

Study participants were asked to discuss the mosquito reduction and bite prevention measures that they routinely implemented within their households. Measures were consistent between pregnant and non-pregnant cohorts, with most respondents reporting performing larval source management in and around their homes, including the regular emptying and cleaning of water containers to avoid stagnation, the covering of exposed water tanks, and the maintenance of a high level of general cleanliness, including the clearing from outside spaces of household refuse that could act as rainwater receptacles.

“We really avoid water stagnation where mosquitoes might breed. We also avoid keeping containers with water. For example, after watering plants, many people leave the pot dishes full of water. We try to avoid all that”. (FGD1, Non-pregnant, Cali).

The idea of cleanliness as being key to avoiding mosquito proliferation was widely held, with a majority of participants considering the prevention of household clutter as an effective method for reducing mosquito densities in houses. Women rarely reported routinely applying topical mosquito repellents to themselves, although regular use on young children was more common. A reluctance to use repellents was often accompanied by comments describing repellents as being excessively expensive, at least for routine use, and this was particularly common in the groups containing participants from lower socio-economic strata. Women from lower socio-economic groups, particularly in Cali, were also more likely to report the topical use of products not manufactured for the purpose of repelling mosquitoes. Most notably, Vicks VapoRub, a cream sold as a nasal decongestant, was mentioned frequently by respondents as being routinely used as both a method of repelling mosquitoes and for the soothing of bites. Johnson’s Baby Oil and various other moisturizing body creams were also reported being applied as a protective barrier to mosquito biting.

(P1) “I use Vicks VapoRub. I apply [it] on my skin. As it contains menthol, mosquitoes don’t approach”.

(P2) “I use it too, especially on my children. (FGD7, Non-pregnant, Cali).

Citronella oil was mentioned repeatedly throughout the FGDs, with participants reporting positive opinions of its efficacy and safety, with it being variously used as both a topical repellent and as a component of floor cleaner.

“There’s also citronella repellent. It has no poison and is very good.” (FGD15, Pregnant, Villavicencio)

Respondents reported burning various items of household refuse and dried plant material for the purpose of driving away mosquitoes from their homes. This practice appeared to be far more common in Cali, with participants describing the burning of egg cartons or eucalyptus seeds and leaves near windows on the street outside their homes. A smaller number of respondents reported burning specific insecticide-containing incense sticks and mosquito-coils.

“I burn eucalyptus leaves in the backyard. A week ago, I did it and it repelled all the mosquitoes”. (FGD2, Non-pregnant, Cali).

Participants also described the routine spraying of the inside of their houses with insecticidal aerosol products, particularly Raid spray, and the use of plug-in insecticide disseminators. Other mosquito control methods frequently mentioned included the use of bednets and the use of a fan during the night.

Participants were asked to describe any alterations to their regular mosquito reduction and bite prevention activities in response to the Zika outbreak. Several respondents in both the pregnant and non-pregnant groups reported being so alarmed by Zika that they incorporated more rigorous mosquito abatement activities into their routines. Most commonly this involved intensified larval source management, but also included increased use of bednets and the more frequent use of insecticidal sprays.

“Now [since the Zika outbreak], we sleep under a mosquito net, fumigate more often, and get rid of possible breeding sites like containers or bottles”. (FGD15, Pregnant, Villavicencio).

However, the majority of respondents reported making few changes to their regular mosquito abatement activities. This was often reported as being because the mosquito proliferation and bite prevention advice for Zika was the same as that disseminated in recent chikungunya and dengue outbreaks, and so people were familiar with the guidance and were, therefore, unmotivated to change their existing behaviours.

“I personally haven’t changed anything. I come from Bogotá so I have always feared mosquitoes. We’re used to fumigating the house and, when chikungunya was on the rise, we even took vitamin B to avoid mosquito biting. In any case, we didn’t change our habits due to Zika”. (FGD13, Non-pregnant, Villavicencio).

Protective clothing

Participants were asked to share their opinions on the personal use of clothing impregnated with mosquito repellents. Almost all were unaware that such products existed, although there was a strong sentiment from the majority of respondents that effective repellent clothing would be desirable and that they would wear such clothing, so long as it was safe, comfortable, and did not produce an unpleasant odor.

“I’d wear it [repellent clothing] unless it’s warm, has a strong smell of chemical, or is sticky. I’d like something I’d be comfortable with”. (FGD7, Non-pregnant, Cali).

“I would worry if I didn’t know the substance clothing was treated with. It could, for example, contaminate your blood or cause allergies. All skins are different.” (Male interview 5, Cali).

Few participants expressed a very strong dislike of the general concept of repellent clothing, with the majority of disapproving opinions resulting from a disinclination for additional contact with chemicals.

“Although we use chemical products in our daily life, I insist on the idea that the fewer chemicals we use, the better. We should try to find other alternatives”. (FGD7, Non-pregnant, Cali).

A distinction was often made between different classes of repellent chemical, namely natural chemicals (usually plant-derived), and synthetic chemicals (those arising from chemical engineering processes). Synthetic repellents were generally considered to be less safe than those perceived as derived from plants. Citronella was often discussed as an example of how naturally-derived repellents could be a benign alternative to synthetic compounds. Several respondents considered the source of the active compound (whether it was regarded as naturally-derived or not) as more important than the overall repellent efficacy. That is, a perceived natural product with lower efficacy was considered more desirable than a synthetic product with higher efficacy.

“I personally don’t like repellents. Rather than that, I prefer homemade natural products. They might not be as effective but, from my perspective, they’re less harmful.” (FGD13, Non-pregnant, Villavicencio).

Moreover, the smell of the repellent was a key concern. The presence of a perceived synthetic odour was considered by many women to be a major factor in determining whether a repellent product would be usable, with smells regarded as ‘natural’ considered far more acceptable.

“I would [wear repellent clothing] if it smelled like citronella, not like something weird”. (FGD7, Non-pregnant, Cali).

During a Zika outbreak, pregnant women would be the primary target group for a repellent clothing product. Several women in the pregnant FGDs expressed anxieties over potential allergenic effects of a repellent chemical, or impacts on fetal development, with an element of scepticism around supplier claims relating to product toxicity. Several pregnant women reported believing that the most trustworthy distributor would be the local public health authority, and there was the sentiment that health center distribution would foster trust and improve uptake and usage rates.

Repellent clothing is most effective when the arms and the legs of the wearer are fully covered. Participants were asked to discuss the likelihood of their wearing full-length clothing if it had the potential to provide high levels of protection from arboviral diseases. Respondents tended to be strongly disinclined to wear clothes that fully covered arms and legs. While the potential benefits of wearing long-sleeved repellent clothing were generally accepted, the majority of participants considered long-clothing to be prohibitively impractical in the day-time heat.

“In this heat, even if it’s true [that protection was provided], I’ll never wear long-sleeved clothes at home”. (FGD16, Pregnant, Villavicencio).

However, while some pregnant women explained that the heightened heat sensitivity during pregnancy would make long clothing too uncomfortable to bear, others suggested that the added protection would be the overriding factor and would make the discomfort of long-clothing worthwhile.

“When you’re pregnant or have babies, if something is good for your baby’s benefit, you keep it even though it’s uncomfortable for you. So, at home, I don’t take off clothes that protect me from mosquitoes.” (FGD5, Pregnant, Cali).

The financial implications associated with purchasing additional products for protection against vector-borne diseases was a concern amongst study participants, particularly as costs were expected to be recurrent. Pregnant participants of low socioeconomic status were less inclined to wear repellent clothing everyday due to the additional expense that this would incur.

“In a country like this, characterized by inequality, repellent should be included in the Mandatory Health Plan (POS) [Plan Obligatorio de Salud], at least during epidemic peaks so that it will reach the general public. I know repellent has a prohibitive price here”. (FGD14, Non-pregnant, Villavicencio).

Discussion around repellents by the female participants was occasionally framed in the role of a mother as having responsibility for the protection of her children. Some participants described being unable to afford repellents for themselves or their partners, but were willing to allocate money for repellents for their children.