Community Knowledge and Acceptance Of

Magaço et al. Malar J (2019) 18:27 https://doi.org/10.1186/s12936-019-2653-x Malaria Journal

RESEARCH Open Access Community knowledge and acceptance of indoor residual spraying for malaria prevention in Mozambique: a qualitative study Amílcar Magaço1*, Carlos Botão1, Pedroso Nhassengo1, Mohomede Saide1, Arminda Ubisse1, Sérgio Chicumbe1 and Rose Zulliger2

Abstract Background: Malaria control remains a leading health challenge in Mozambique. Indoor residual spraying (IRS) is an efective strategy to control malaria transmission, but there are often barriers to reaching the coverage necessary for attaining maximum community protective efect of IRS. Mozambique recorded a high number of household refusals during the 2016 IRS campaign. This study sought to evaluate household and community factors related to the acceptability of IRS to inform strategies for future campaigns in Mozambique and the region. Methods: A cross-sectional, qualitative study was conducted in eight urban and rural communities in two high malaria burden provinces in Mozambique. Data were collected through in-depth interviews with community members, leaders, sprayers, and representatives of district health directorates; focus group discussions with community members who accepted and who refused IRS during the 2016 campaign; systematic feld observations; and informal conversations. Data were systematically coded and analysed using NVIVO-11®. Results: A total of 61 interviews and 12 discussions were conducted. Community participants predominantly described IRS as safe, but many felt that it had limited efcacy. The main factors that participants mentioned as having infuenced their IRS acceptance or refusal were: understanding of IRS; community leader level of support; characteristics of IRS programmatic implementation; environmental, political and historical factors. Specifcally, IRS acceptance was higher when there was perceived community solidarity through IRS acceptance, desire to reduce the insect population in homes, trust in government and community satisfaction with past IRS campaign efectiveness. Participants who refused were mainly from urban districts and were more educated. The main barriers to acceptance were associated with selection and performance of spray operators, negative experiences from previous campaigns, political-partisan conficts, difculty in removing heavy or numerous household assets, and preference for insecticide- treated nets over IRS. Conclusions: Acceptance of IRS was infuenced by diverse operational and contextual factors. As such, future IRS communications in targeted communities should emphasize the importance of high IRS coverage for promoting both familial and community health. Additionally, clear communications and engagement with community leaders during spray operator selection and spray implementation may help reduce barriers to IRS acceptance. Keywords: Malaria, Indoor residual spraying, Acceptability, Mozambique

*Correspondence: [email protected] 1 Sistemas de Saúde, Instituto Nacional de Saúde, Ministério da Saúde, Moçambique, Maputo, Mozambique Full list of author information is available at the end of the article

© The Author(s) 2019. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Magaço et al. Malar J (2019) 18:27 Page 2 of 12

Background protection. Individual and household level protection Malaria is one of the leading public health problems in is largely derived from community level coverage [10, Mozambique where the national prevalence in children 11]. Tus, the WHO has emphasized the importance of between 6 and 59 months of age was 40% in 2015 [1]. community acceptance and cooperation for successful Prevalence was even higher in the two most populous IRS implementation [9]. Tis is critical as a study from provinces of the country, Zambézia and Nampula, where diverse epidemiological settings found that malaria pro- approximately two-thirds of children under fve were tection from IRS was much more strongly associated infected [1]. Malaria accounts for approximately 45% of with high community level coverage than with household outpatient clinic visits, 56% of pediatric ward admissions acceptance [12]. and 29% of all deaths in hospitals [2]. Malaria is endemic Mozambique has a long history of IRS implementa- throughout the country and considered to have stable tion dating back to the 1940s [13], but the NMCP has transmission, despite epidemic peaks generally related struggled in recent years to achieve the desired cover- to the rainy season [3]. Important progress has been age in areas targeted for IRS in parts of the country. For made in reducing malaria mortality in Mozambique, but example, in 2016 the reported IRS coverage of structures the overall burden of disease reported in routine health found by spray operators in Zambézia, the province with service data has steadily increased in recent years. In the highest burden of malaria in the country, ranged 2017 there were over 9.5 million malaria cases reported between 72 and 89% [14]. Given the inherent challenges throughout the country [4]. in identifying the houses that should be included in the Te cornerstone of malaria control in Mozambique is spray coverage denominator [10], the true community early diagnosis and treatment of malaria cases with efec- coverage was likely even lower. Tis challenge is not, tive drugs and vector control [4]. Te primary malaria however, unique to Mozambique as community level vector control interventions are universal distribution coverage was found to be poor throughout sub-Saharan of insecticide-treated mosquito nets (ITNs) through- Africa [11], underscoring the importance of identifying out the country and indoor residual spraying (IRS) with factors to facilitate better coverage. non-pyrethroid insecticide in districts with evidence of Evidence from other African settings demonstrate that pyrethroid resistance, continued high transmission of numerous factors motivate and inhibit IRS acceptance. malaria, and/or that are targeted for malaria elimina- For example, a 2010–2011 qualitative study in Tanza- tion by regional initiatives [4]. In Mozambique, the main nia found that level of education, lack of understanding vectors of malaria are endophagic (feeding indoors) and of IRS efectiveness, fears of increases in other insects, endophilic (resting indoors), which makes these vectors potential side efects of IRS, embarrassment at removing vulnerable to ITNs and IRS [2, 4, 5]. Tere is now wide- limited household possessions, and association of the IRS scale resistance to the pyrethroid insecticides on the ITNs campaign with politics infuenced IRS acceptance [15]. A used in Mozambique [2, 4–7] which has underscored the 2015 quantitative study from Uganda noted that the most importance of using more efective insecticides, whether important barrier to IRS acceptance was fear of adverse delivered with IRS or ITNs. Presently, IRS ofers a greater efects from insecticide and that age, socio-economic sta- variety of efective insecticides for malaria control than tus, prior IRS experience, belief that IRS was benefcial ITNs and is a critical tool for management of insecti- and roof composition were associated with IRS accept- cide resistance. As such, understanding of social factors ance [16]. afecting coverage and acceptance of IRS is needed. Prior research in Mozambique from 2006 to 2008 in a IRS for malaria prevention has the potential to be a district with a low burden of malaria found limited per- highly efective strategy for malaria control, particularly ceived efcacy of IRS for malaria prevention; however, in the presence of pyrethroid resistance, but its efective- participants accepted IRS in their homes based on trust ness is contingent upon achieving high coverage through and compliance with community leaders and health vectorial mass efect [8]. Te World Health Organiza- authorities, the perceived efect of IRS on other insects, tion (WHO) recommends coverage of greater than 85% and trusted relationships with spray operators [17, 18]. of households in a community in order to have commu- Another study in southern Mozambique of malaria pre- nity-level protective efect [9]. IRS works to interrupt vention preferences during pregnancy noted that IRS was malaria transmission through killing and repelling mos- the least preferred prevention strategy [19]. quitoes, thereby reducing vector density, longevity and Te relevance of these data to high burden areas and transmission. Importantly, it is most efective in prevent- in the context of next generation insecticide products ing transmission of malaria after feeding, underscoring is unknown. As such, this study sought to determine the importance of high IRS coverage at the community factors associated with IRS acceptance in the highest level in order for there to be household and individual burden provinces of Mozambique in order to guide Magaço et al. Malar J (2019) 18:27 Page 3 of 12

development of strategies to improve acceptance of Data collection IRS in future campaigns. It also sought to build upon Tis study used an ethnographic approach to data col- the limited global evidence on factors that facilitate or lection. Tis approach allowed for capturing statements impede IRS acceptance. and perceptions about malaria prevention and IRS, while also documenting the context in which such perspectives were expressed. In each of the study districts, the study Methods team conducted a combination of methods, namely IDI, Study population FGD, participant observations, and informal conversa- Tis qualitative study included in-depth interviews tions with community leaders and members during walks (IDI), focus group discussions (FGD), systematic feld around the communities to identify potential study par- observations, and informal conversations. Data were ticipants. During these participant observations, team collected in districts that were purposively selected in members collected feld notes on behavior and com- Nampula and Zambézia provinces based on malaria ments related to the latest IRS campaign which were sub- prevalence and IRS experience during the 2016 cam- sequently triangulated with IDI and FGD data. paign. Tese two provinces are the most populous in At the start of each IDI and FGD, study researchers the country and have the highest malaria prevalence and research assistants administered a short, structured (66% and 68%, respectively) and case burden [1]. Each questionnaire to all participants to collect data on soci- district was targeted for spraying of the whole district odemographic characteristics. Tis questionnaire also IRS in 2016 and 2017 and had relatively high levels of documented participants’ IRS acceptance or refusal. Data refusal during the 2016 campaign. All districts used were collected by trained feld workers using established a micro-encapsulated pirimiphos-methyl insecticide interview guides. Interviews were conducted in Portu- in 2016 after use of the pyrethroid, deltamethrin, for guese or the local language and lasted approximately the 2014 and/or 2015 campaigns. Deltamethrin was 1 h. Interviews allowed participants to discuss perspec- phased out in 2016 due to resistance. Te IRS campaign tives, values and experiences in response to a series of in Zambézia province was conducted by the Ministry open-ended questions. Tese included questions about: of Health (MoH) in partnership with the President’s malaria knowledge, understanding of IRS and associated Malaria Initiative while the IRS campaign in Nampula procedures, perceived importance of IRS, experience was independently implemented by the MoH. One with acceptance or refusal during the 2016 campaign, the rural and one urban study site was selected in each of role of household heads in an IRS program, and motiva- the four included districts for a total of eight study sites. tors of IRS acceptance and refusal. Separate guides were Te two districts included in Zambézia were Mopeia used for the IDI which focused on individual perspec- (2017 population: 136,520 inhabitants) and Mocuba tives and for the FGD which explored community per- (422,681). Te two districts included in Nampula were ceptions and experiences. All interviews were recorded, Nampula City (743,125) and Monapo (413,694). Data transcribed and translated prior to analysis. IDIs and were collected in May 2017, 6 months after the end of FGDs were conducted until the study team concluded the 2016 IRS campaign. that saturation had been reached and there was no need Participants in the study sites were purposively selected to interview more respondents. to include diverse perspectives on IRS and the IRS campaign. Participants were community members, community leaders, spray operators, and representatives of Data analysis district health directorates. IDI were conducted with Data analysis was performed in two stages. Pre-analysis community members, community leaders, district repre- of the feld notes led to the development of analytic cat- sentatives, and health service staf directly linked to the egories that were analyzed sequentially during feld work. malaria prevention program, as they were considered Tis procedure helped inform the fnal categories that to have critical and varying perspectives on IRS imple- were used during the coding and analysis of full interview mentation and acceptability. FGD were conducted with transcripts. NVIVO-11® software was used to extract, individuals who refused IRS during the 2016 campaign code and sort interview text [20]. Some categories were and with individuals who accepted IRS during the 2016 determined prior to data collection by the interview campaign. A minimum of 8 individuals with the same guides, but data analysis was iterative and categories spray status were required for a FGD in each of the sites were added, removed, and amended during data analy- so some study sites only had either a spray or a no spray sis based on emergent study fndings. Te coded data FGD. FGD participants were diferent from those that were then organized into overarching themes related participated in the IDIs. to the study objectives. Te present analysis focuses on Magaço et al. Malar J (2019) 18:27 Page 4 of 12

knowledge of IRS and motivators for IRS acceptance and Table 1 Characteristics of in-depth interview participants, refusal described by participants. Participant statements Mozambique 2017, showing number of participants are included that are illustrative of the broader category and associated percentage, except where otherwise noted content to contextualize study fndings. Accepted IRS Refused IRS Total

Ethical approval N 28 (46) 33 (54) 61 Median age (range) 37 (19–66) 35 (19–69) 35 (19–69) Participation in the study was voluntary and the study Province was ethically approved by the Institutional Committee on Nampula 17 (50) 17 (50) 34 Bioethics in Health of the National Health Institute (012/ Zambézia 11 (41) 16 (59) 27 CIBS-INS/2016). Tis study was reviewed by the Centers Community role for Disease Control and Prevention (CDC) and deter- Community member 16 (33) 32 (67) 48 mined to be human subjects research with non-engage- Community leader 12 (92) 1 (8) 13 ment by CDC staf. Sex Results Male 16 (47) 18 (53) 34 Female 12 (44) 15 (56) 27 A total of 61 IDI, including community members (n 48) = Education level and community leaders (n 13) were conducted in the = No education 1 (33) 2 (67) 3 eight study sites in June–July 2017. Additionally, 12 FGD Primary education 15 (52) 14 (48) 29 were contemporaneously conducted with 65 community Secondary/higher 12 (43) 16 (57) 28 members who accepted and 64 who refused IRS during College 0 (0) 1 (100) 1 the 2016 campaign. Marital status Not married 9 (50) 9 (50) 18 Participant sociodemographic characteristics Married 17 (50) 17 (50) 34 Sociodemographic characteristics of the IDI participants Cohabitating 1 (14) 6 (86) 7 are summarized in Table 1. A total of 33 IDI participants Divorced/widowed 1 (50) 1 (50) 2 refused IRS in their homes during the 2016 campaign. Occupation Te median participant age was 35 years old and most Formally employed 4 (36) 7 (64) 11 respondents (79%) were self-employed in the informal Informally employed 24 (50) 24 (50) 48 market. Most of the interviewees had a primary level Unemployed/student 0 (0) 2 (100) 2 (n = 29) or secondary/higher level (n = 28) of education, but three respondents had no schooling. Participant characteristics were similar between those that accepted and those that refused IRS, but 92% of community leader associated benefts was not necessarily associated with participants accepted IRS. IRS acceptance as the majority of IDI participants (25 A total of 129 community members participated in of 33) who refused IRS were able to accurately describe twelve FGD (46 male/ 83 female). Te median participant the procedures for spraying houses. For example, one age was 34 years. Te most common duration of resi- urban man who refused IRS clearly explained the steps dence in the study area was fve or more years (Table 2). of IRS:

Knowledge about IRS procedures “Tere you have to take your things from home. After the smell goes down you need to sweep and In general, knowledge about IRS procedures and ben- tidy things up again. What they put that day [the efts among study participants was mixed. A consid- insecticide] is something that even kills, but it erable number of IDI participants reported that they doesn’t kill for long.” (IDI/Man/Refused/Urban). understood IRS to be a method to prevent malaria and mosquito bites. Despite describing a basic understand- Tus, it was not always a lack of knowledge of IRS ing of IRS rationale and implementation procedures, procedures that led to the refusal of IRS. participants frequently reported limited knowledge One barrier to IRS knowledge mentioned by par- on how the insecticide worked, on the importance of ticipants was that information that the IRS campaign spraying the house, and on the probable side efects or was going to occur was often shared with community expected residual efcacy of the spray. Some respond- members very late, limiting their ability to prepare for ents stated they did not trust and did not accept spray- the campaign. Many participants noted the need for at ing because of the limited knowledge of these factors. least a few days’ notice before the campaign in order Yet, higher knowledge of IRS procedures and the for community members to plan and prepare for IRS Magaço et al. Malar J (2019) 18:27 Page 5 of 12

Table 2 Characteristics of focus group discussion including participants such as teachers, drivers and other community member participants, Mozambique 2017 public ofcials. Tese participants were skeptical of IRS Accepted IRS Refused IRS Total and shared a desire to see a tangible impact in terms of reduced mosquitoes and malaria transmission, some- N (%) 65 (50) 64 (50) 129 thing they did not think occurred with IRS. Te main Median age—(range) 37 (18–66) 30 (18–60) 34 barriers to acceptance reported by participants were Province associated with spray operator selection and perfor- Nampula 31 (42) 42 (58) 73 mance, negative experiences from previous campaigns, Zambézia 34 (61) 22 (39) 56 political-partisan conficts, difculty in removing house- Sex hold assets, and preference for ITNs over IRS. Male 27 (59) 19 (41) 46 Female 38 (46) 45 (54) 83 Duration of residence in the study area Spray operator selection and performance Native 8 (53) 7 (47) 15 Many participants expressed dissatisfaction with the < 2 years 0 (0) 2 (100) 2 selection of spray operators as having infuenced their 3–5 years 1 (25) 3 (75) 4 acceptance of the IRS campaign. According to partici- > 5 years 56 (52) 52 (43) 108 pants, not all members of the communities that applied to work during the campaign were approved and selected for the position by the MoH and/or its implementing compliance. Tis may have been more pronounced if partners. Tere was little transparency as to how and why a participant worked away from their home such as in certain individuals were or were not selected. Because distant agricultural felds or in formal employment. of this unclear process, the spraying campaign was described as having predominantly been implemented by IRS diferences by geographic area individuals whose provenance was unknown to members Most IDI interviewees (n = 19) who accepted the spray- of the community. Community members and leaders ing reported being happy with the results of the inter- expressed reticence in allowing these unknown individu- vention and were glad that something was being done to als into their home because they were not trusted to not combat malaria in their communities, but participants damage or rob the homes. in some geographic areas felt it had limited efectiveness Dissatisfaction with the selection process also led to or safety. Participants from rural areas and participants the spread of rumours against the campaign. Commu- whose homes had not previously been sprayed prior to nity leaders noted that when young members of the com- their acceptance of the 2016 campaign were generally munity were not selected, these same individuals created more satisfed and perceived IRS as more efective than rumours about IRS, encouraging residents to refuse the those participants from urban areas. Additionally, par- campaign. ticipants from the study sites in Nampula and from the “Here the problem for some community leaders is district of Mopeia in Zambézia described the 2016 cam- that their children were excluded in the selection paign as efcient and improved relative to the spraying of of sprayers. Tese children plus the leaders became previous years’ campaigns. For example, one participant upset and started to spread rumours for people not from Mopeia reported being satisfed with the results of to accept” (IDI/Community Leader/Rural). the last spray campaign: Additionally, some study participants from the district “We here in Lualua [community name] liked this of Mocuba in Zambézia described perceived poor imple- year’s [IRS campaign] and are thankful for the spray- mentation by IRS spray operators as one of the reasons ing because since they have done it, so far mosqui- for the limited spray efectiveness and low acceptance of toes are few, they are not as many as it was before. spraying. Study participants alleged that spray operators It killed not only mosquitoes, but also feas, cock- diluted the insecticide used to spray the houses and then roaches and ants.” (FGD/Woman/Accepted/Rural). later sold the remaining insecticide. Tis community perception of poor spray operator performance afected Barriers to IRS acceptance trust in and acceptance of IRS implementation. In order Despite some participants’ satisfaction with IRS, over to minimize selection of untrustworthy or unknown half of participants refused IRS during the 2016 cam- spray operators, participants recommended that commu- paign. Participants who refused were mainly from urban nity-based leadership be involved at all levels in the selec- districts with secondary and higher education level, tion of sprayers to ensure better acceptability of IRS. Magaço et al. Malar J (2019) 18:27 Page 6 of 12

Negative experiences from previous campaigns the current administration. In this context, many areas An important factor that infuenced IRS acceptance have both ofcial government party community leaders was experience in prior campaigns. Refusal was greater and opposition party community leaders. IRS campaigns among participants whose houses were sprayed in pre- typically included government party community leaders, vious campaigns because they did not feel that the past but not other local community leadership. Te leaders campaigns were as efective as had been promised or they who started the rumours were described as individuals described collateral efects such as skin irritation and with a strong infuence in the community who were not allergy in the nostrils. Additionally, many participants involved in IRS mobilization because they were afliated who refused the 2016 IRS expressed suspicions about the with opposition political parties. use of a new IRS product. Participants complained about One infuential politically-oriented rumour was related the strong smell of insecticide and the fact that it created to the fact that IRS was ofered as a free health interven- stains on their walls. Participants also reported experi- tion by the ruling government party. Tere was a belief encing frequent side efects after accepting IRS in their that the government would not ofer a benefcial cam- homes in prior campaigns such as skin allergy and irrita- paign for free in areas that did not vote for the ruling tion in the nostrils. One of the respondents who refused party. One participant expressed this doubt and asked, to spray during the 2016 campaign explained how his “Te government can bring this product here for free while home was a month after spraying in a previous campaign: they know that we do not vote [for their party]?” (IDI/ Man/Refused/Urban). As such, some community mem- “When they do the work they dirty the walls with bers believed that IRS was only provided free of charge stains, and after 4 weeks the mosquitoes and other by the government in order to hide the fact that the gov- insects have increased in my house, it seems that this ernment was actually attempting to bring other diseases spraying only makes things worse” (Man/Refused/ to reduce the population in groups opposed to the ruling Urban). party. Tese rumours were particularly common in areas Other participants who had homes sprayed during where there was strong opposition party presence such previous campaigns did not accept spraying in the 2016 as in Nampula City and in Mocuba district. Community campaign because of perceived lack of IRS efectiveness. leaders stressed their important role in either promoting Tis included the perception that IRS only chases away or dispelling rumours related to IRS and in IRS accept- insects, but does not kill them. For example, a man who ability within their communities: refused the 2016 campaign said he did not accept IRS “Our problem here is more political because the because prior years’ IRS attracted many insects to his population only follows what we recommend. If I say home. Additionally, participants noted concerns related that the product is not a good thing, then many will to the residual efcacy of IRS in previous campaigns as not join” (IDI/Community Leader/Urban). an infuence on acceptance. One participant explained, “Tat product does not kill anything, mosquitoes only fee. Tus, by not including all appropriate local leadership After 3 days or a week at the most many mosquitoes come in promoting the IRS campaign, acceptance was blocked back” (IDI/Man/Refused/Urban). by some community leaders, thereby afecting overall IRS While participants were pleased that mosquitoes and acceptance. As such, community leaders emphasized the other insects declined shortly after spraying, there was need to include community-based leadership at all levels some frustration with the duration of the efect of the and from various political and religious groups to dispel product against mosquitoes. rumours and to facilitate higher acceptance.

“Tat product helped a lot, even the cockroaches and Difculty in removing household goods feas died too, but after 3 months other mosquitoes Another barrier to IRS acceptance frequently mentioned are already coming back, it only feas have not yet by participants was related to the need to remove house- seen” (DGF/Women/Rural). hold goods for spraying. Tis barrier was mentioned across settings, yet important diferences were noted Political‑partisan conficts between the rural and urban areas. In rural areas, par- Another frequently reported barrier to IRS acceptance ticipants frequently mentioned that they found it dif- was politically-motivated rumours against IRS. Tese fcult to accept that the sprayers would see the level of rumours were perceived to have been created by reli- poverty inside their houses when they removed their gious and community leaders not afliated with the IRS household assets for the spray. One woman explained, campaign. Nampula and Zambézia are both provinces “I was ashamed that they saw my poverty, if I have noth- with high levels of support for the party in opposition to ing inside it [my house]. I only have a mat and a bucket of Magaço et al. Malar J (2019) 18:27 Page 7 of 12

water.” (Female/Refused/Rural). Tis was in stark contrast protection. Community members also infuenced one- to urban areas where participants described difculty in another’s acceptance by describing their positive expe- removing their extensive or heavy household goods from riences. For example, individuals who refused spraying the houses. For example, one man stated that, “I at this when spray operators frst came to their homes found age I can’t carry things from inside to the outside, so I that many of the rumours, particularly those regard- denied” (Man/Refused/Urban). ing efcacy and side efects, were false after talking to Te level of efort associated with removing all house- neighbours. Tis led to acceptance when spray opera- hold assets in wealthier areas was considered to be par- tors returned. ticularly cumbersome by participants who did not feel “At frst I was afraid, but after hearing the neigh- that there was a long-lasting beneft of IRS as explained bours talking about killing all cockroaches, feas by one participant: “We cannot take our fridge and furni- and mosquitoes, then I asked for it too. When it ture outside only for something that will not last for a long was frst sprayed, no one died, no one had side time.” (Female/Refused/Urban). efects, so when they came to spray, no one was against it.” (FGD/woman/accepted/urban). Preference for ITNs over IRS Te IRS campaign was accompanied by the involvement Lastly, participants mentioned preference for other, non- of community leaders who disseminated messages that IRS malaria prevention methods such as ITNs as one of emphasized the elimination of the mosquito that causes the barriers to the acceptability of spraying. Participants malaria in communities. Tese messages were seen as from both rural and urban areas expressed reticence to efective, but their impact, as noted above, was infuenced accept IRS because they believed that by accepting such a by whether it was a trusted community leaders who deliv- method of prevention, they would not receive ITNs. One ered the messages. Participants and communities who community leader explained, “Here the people want more had positive experience with spraying during previous mosquito nets instead of spraying, so they don’t accept the campaigns were more likely to accept. Other interview- spray, stating that ‘we are waiting for the nets’.” (Commu- ees whose homes were sprayed in previous rounds, but nity Leader/Rural). Te preference for ITNs was gener- missed the most recent campaign, expressed their regret ally because they were felt to be more efective than IRS. and the desire to have their home sprayed, particularly Another leader stated. when their neighbours reported positive experiences. “Te people do not accept spraying, out of fear of not receiving nets. Tey say that ‘Tis product doesn’t Desire to reduce insect population of homes work. We want mosquito nets’.” (Community Leader/ Participants in the rural districts covered by the study Urban). reported having accepted IRS because of a desire to Additionally, participants frequently reported that reduce the number of mosquitoes and others insects they perceived ITNs to be safer than IRS and, as such, inside houses. Even though IRS was promoted by the preferable. government as a malaria prevention tool, participants were often motivated to accept because of perceived collateral benefts of IRS such as the killing of feas. Facilitators of IRS acceptance “I was very happy because there were lots of feas Te main factors infuencing the acceptance of spray- and mosquitoes in my house. All the mosquitoes ing were the perceived community solidarity through and feas and cockroaches have already died from IRS acceptance, desire to reduce the insect population this spraying.” (Woman/Accepted/Rural). in homes, trust in government and health workers and community satisfaction with IRS efectiveness. Focus group discussions, especially in rural areas, revealed that the main motivation for many participants to accept spraying was to eliminate other insects. Community solidarity Nevertheless, participants were also motivated by the Te most frequently mentioned motivation for accept- desire to provide protection from malaria for their chil- ance of spraying was solidarity among neighbours. Te dren by killing mosquitoes. study participants reported that when a house was not “My kids used to be sick often, but now I’m happy. sprayed, mosquitoes from that house could go to the Since they came to do spraying they aren’t getting neighbors’ homes and infect residents with malaria, sick like before.” (Man/Accepted/Rural). underscoring the importance of IRS for community Magaço et al. Malar J (2019) 18:27 Page 8 of 12

Trust in government and health workers acceptance of IRS [11, 12]. High acceptance is critical Another motivation for IRS acceptance in the more when IRS programs use more expensive insecticides, remote rural study areas was the role of the government such as the use of organophosphate in Mozambique, in delivering the intervention and the involvement in the because as programs become more expensive they must campaign of infuential health workers in the community. provide high community-level coverage and benefts in “We accept yes, but also because it is a government order. order to remain cost-efective [22]. Tus, the relatively And the government is not crazy to spend money for some- low IRS acceptance experienced by Mozambique dur- thing that does not work.” (Woman/Accepted/Urban). Te ing the 2016 campaign demonstrated the importance of infuence of trust in the government was reinforced by understanding motivators of IRS acceptance and refusal another participant who stated, “We can’t deny a thing in order to achieve public health. that comes from the government, the government does not Tis study explored community and stakeholder knowl- think evil, if they are for this product then it will kill mos- edge, experiences, and perceptions of IRS in high malaria quitoes” (Man/Accepted/Rural). burden districts of Mozambique. It specifcally identi- Spray acceptance was also attributed to the fear of future fed a number of factors that facilitated IRS acceptance problems with district health authorities in case a house- or were associated with IRS refusal during the 2016 IRS hold member got malaria. In many cases, some of the campaign. Te main factors that infuenced IRS accept- respondents justifed their decision to accept the spraying ance or refusal in this study site and others were: level as an important perquisite for future government services. of education and understanding of IRS; socioeconomic status; geographic location; community solidarity, leader “We have to accept, if you don’t accept, and tomor- infuence, and preferences; IRS programmatic implemen- row go to the hospital with malaria they will send tation; environmental; political and historical factors. Te you away, they will say you are the one who denied factors identifed in this study in Mozambique and their the spraying, so you will treat yourself.” (Woman/ relevance to fndings from other settings are presented in Accepted/Rural). Table 3. Some of these perceptions may have been propagated In Mozambique, IRS refusal was often due to barriers by spray operators and community mobilizers as some such as distrust in IRS spray operators, negative percep- participants reported that these workers stated, “If you tions of IRS due to prior campaign experience or politi- do not accept it, later you cannot go to the hospital with cally-motivated community rumours, lack of community malaria” (Man/Accepted/Urban). leader buy-in, distrust in the government-run campaign, fear of potential side efects, and shame or difculty associated with removing household goods. Tese results are Community satisfaction with IRS efectiveness over time similar to barriers noted in studies from other countries Most participants who accepted IRS reported being satis- such as Tanzania, Uganda [15, 16, 23]. Interestingly, these fed with the outcome of the campaign because they had fndings are similar to those that were noted in lower not had to worry about so many insects or seek hospital transmission areas in Mozambique, providing some evi- services. While some participants complained of short dence that local malaria burden may not be the most residual efcacy, one participant said that the efect of important infuence on IRS acceptance [9–11]. the IRS in his house lasted more than 6 months. Another While some community members were motivated by man stated, “For me, the mosquitoes all died, so far I can’t perceived malaria-related benefts of IRS, others reported see anything, I can even sleep without a net, they don’t that they accepted IRS because of their desire to reduce even have mosquitoes” (Man/Accepted/Rural). other household insects like feas and to comply with With the exception of a few heads of households in governmental and community expectations. Motiva- more urban areas, most of the community members tors of acceptance of spraying in these high transmission covered by the study in the two provinces were willing areas are similar to those noted for acceptance of IRS to accept future IRS campaigns. Tese rural commu- with dichlorodiphenyltrichloroethane in a low transmis- nity participants stated that the last spray campaign was sion area of Mozambique [9–11] which found that adher- much better than prior years’ campaigns and was efec- ence was driven by trust in local health authorities and tive in combating mosquitoes and other insects. the infuence of community leaders. Tis fnding on the importance of community solidarity across transmis- Discussion sion settings has implications for IRS communication IRS has been integral to many advancements in global strategies and overall engagement with communities. and national malaria control, but this progress is fragile For example, key messages for future campaigns might [21] and is contingent upon continued high community better capitalize on this solidarity by emphasizing that Magaço et al. Malar J (2019) 18:27 Page 9 of 12 - - in previous campaigns were less likely to were campaigns in previous IRS in Uganda [ 16 ] accept future politically motivated [ 15 ] politically motivated Tanzania, Rwanda and Thailand felt that felt Thailand Rwanda and Tanzania, concernedIRS attracted insects and were [ 15 , 18 19 24 25 ] about side efects ment of spray operators and their correct operators ment of spray application of insecticide has been noted settings [ 15 , 18 24 – 26 ] across Tanzania with some participantspersisted in [ 15 ] the insecticidenoted smell [ 15 , 24 ] and in sufcient aware about not having Tanzania noted [ 15 ] and in Zimbabwe ness of spray concerts with stains left on walls [ 27 ] southern Mozambique [ 17 – 19 ] embarrassment of removal of limited and of limited embarrassment of removal quality [ 15 ] and in low household items household propertyUganda noted security concerns [ 23 ] likely accept IRS than individuals from to socio-economiclower classes [ 16 ] ], Tanzania [ 15 ] and Mozambique [ 17 ], Tanzania Uganda [ 23 ]. In was Uganda willingness with higher education [ 16 ] associated Individuals whose houses were not sprayed not sprayed Individuals whose houses were Some participants in Tanzania felt spray was spray felt Tanzania Some participants in Participants in southern Mozambique, in southernParticipants Mozambique, The importance recruit The transparent of local, of IRS efectiveness Skepticism regarding and Rwanda also Tanzania in Participants Preference for ITNs over IRS was also noted in IRS was also noted ITNs over for Preference Participants in Tanzania noted barrier noted of Tanzania in Participants Middle-class less individuals in Uganda were Lack of information was a barrierLack of information in southern [ 15 ]. educated more were Tanzania Refusers in Primary settings barriers from other community leaders but did not kill spray operators spray diluted and stain on walls calendar cation of spray items to remove to items efcacy Negative past experiences with IRS Negative Engagement of only government partyEngagement of only government Belief that IRS chased away mosquitoes, mosquitoes, Belief that IRS chased away efcacy residual low Perceived Selection of unknown or not trusted Concerns that insecticide was over- Use of insecticide smell that left strong Insufcient or lack of timely communi - Lack of community leader support ITNs for Community preference More urbanized areas More Families with too many/too heavy many/too with too Families items few with too Families Lack of information on spray schedule on spray Lack of information about residual Lack of information Higher education Primary study barriers from this - - with prior IRS experience in Tanzania Tanzania with prior IRS experience in [ 15 ] and Uganda 16 Mozambique [ 18 ] [ 15 ], southern Mozambique [ 18 ] and Uganda [ 16 , 23 ] IRS against malaria was noted across IRS against malaria across was noted settings [ 15 , 16 24 ] infuential in southern Mozambique WHO the by [ 18 ] and recommended [ 9 ] scored role of community in facilitat role scored noted Tanzania ing acceptance and in concerns about community level [ 15 , 17 18 ] efect important infuence on IRS uptake in and Uganda [ 15 , 16 23 ] Tanzania WHO [ 9 ] the by recommended IRS demand was higher among those Trust was also noted in southern was also noted Trust This desire was also noted in Tanzania Tanzania in was also noted desire This Infuence of perceived efectiveness of efectiveness Infuence of perceived Inclusion of known was operators spray Participants in Mozambique under Participants Community education on IRS was Primary facilitators from other from other Primary facilitators settings workers insects from home insects from - in reduc be efective to believed ing malaria neighbours IRS promotion Prior acceptance of IRS Prior Trust in government and health in government Trust Desire to remove other non-malaria remove to Desire IRS was accepted because it was IRS was accepted Rural areas Desire to protect neighbours protect to Desire experience of from Learning in Community leader involvement Primary facilitators from this from this Primary facilitators study Primary facilitators and barriers to indoor residual spraying acceptance in this Mozambique study and from other settings study and from other in this Mozambique acceptance spraying residual and barriers to indoor Primary facilitators Historical Political Environmental Programmatic Geographic location Geographic Community Socio-economic Education and IRS knowledgeEducation 3 Table Magaço et al. Malar J (2019) 18:27 Page 10 of 12

by accepting IRS, community members are being good Such messages may result in the short-term beneft of neighbours and helping to protect their communities, acceptance of the current campaign, but they may ulti- rather than only emphasizing individual or familial ben- mately be detrimental by raising false expectations that efts. Tis is particularly relevant given the importance of result in negative campaign experiences that inhibit high community acceptance for IRS to be efective. Addi- future acceptance. Te important infuence of historical tionally, it underscores the value of having strong buy-in IRS experiences on future IRS acceptance was found in from community leaders as these individuals can lever- this study and across settings [15, 16]. age community cohesion for the beneft of the campaign Spray operator selection and community leader buy-in when sufciently engaged, or can infuence widespread were two important factors associated with IRS accept- refusal when the leaders do not have sufcient buy-in ance or refusal, a fnding that may be relevant for both in the campaign. For example, studies from southern Mozambique and other settings. In this study, communi- Mozambique found that acceptance was motivated by ties in which the leaders were not sufciently engaged or IRS delivery by known individuals [9–11], underscor- where the spray operators were not known and therefore ing the important role of spray operator recruitment not trusted reported rumours related to IRS, perceptions and selection in IRS acceptance and community leader that the insecticide was diluted by spray operators and engagement. fears of letting a stranger into one’s home unattended. A notable fnding of this study is that many partici- Tese fndings are not unique to Mozambique (e.g. [15, pants did not perceive IRS to be very efective and many 18, 24–26]), and leader engagement and operators elec- reported preference for ITNs, yet they continued to con- tion are important areas that can be improved upon sent to spraying. Tis acceptance of IRS, despite the per- in future campaigns. Te results of other studies also ception of its lack of efectiveness has also been noted in showed that the conduct of unknown spray operators other African settings [23] and Mozambique [17, 18]. In in can be problematic and impede successful IRS cam- particular, participants in this study reported concerns paigns or can be important motivators of acceptance. For about the residual efcacy of IRS. Tis is an interesting example, the need for local, transparent recruitment of fnding given that residual efcacy of pirimiphos-methyl spray operators and their correct application of insecti- in the 2016 campaign in Zambézia was at least 5 months cide has been noted across settings [15, 18, 24–26]. More [28] and pirimiphos-methyl was considered by some par- extensive and meaningful engagement with leaders dur- ticipants to be more efective than the deltamethrin pre- ing campaign mobilization and a more transparent spray viously used due to the strong smell and residue left on operator selection process may enhance IRS acceptance. walls temporarily after spraying. It is, however, important to note that in many parts of Findings across countries underscore the importance Mozambique where there are multiple community lead- of education and IRS knowledge on IRS acceptance. For ers from diferent political parties, this approach may be example, participants in this study and others [15] were challenging to implement. able to explain the IRS procedures, but they often did Tis study found that another barrier to IRS acceptance not understand the residual efcacy of IRS or the fact in poor, rural areas was a desire to not have neighbours that it only killed mosquitoes that rest indoors. Te low see families lack or belongings. Tis fnding shows that perceived efectiveness and residual efect mentioned by the level of poverty of some households may infuence the study participants could be related to the fact that com- acceptance of IRS. It is important to note, that since the munity members often do not diferentiate between the majority of Mozambique’s population lives in poor, rural Anopheles mosquitoes targeted by IRS and other spe- areas, the National Malaria Control Programme should cies of mosquito, particularly Culex, which tend to have design specifc strategies for this type of population. greater levels of resistance to insecticides. Tus, families Tis study has a few important limitations that may may not perceive the beneft of IRS if they continue to have infuenced the transferability and trustworthiness have many mosquitoes in or near their homes following of the data. Study participants were volunteers from the campaign, as was also noted in other studies [10]. four high-burden malaria districts. Te study fndings Tis low perceived efectiveness underscores the com- from these areas are consistent with fndings from the plexity in communication promoting IRS acceptance. In low-burden south of the country, but their transferabil- order to promote IRS without raising false expectations, ity outside of Mozambique may be limited. Addition- communication should specify that IRS prevents the ally, the study was conducted by a team of investigators mosquitoes that cause malaria. Tis nuanced message is, from the Mozambique Ministry of Health. Many par- however, often lost since it is more complicated and less ticipants still mentioned concerns with the government salient than messages such as the common IRS mantra and the government-implemented IRS campaign, but it in Mozambique, “Mata o mosquito!” (Kill mosquitoes!). Magaço et al. Malar J (2019) 18:27 Page 11 of 12

is possible that the investigators’ background may have Consent for publication Not applicable. infuenced participant responses. Ethics approval and consent to participate The study received human subjects approval from both the Mozambique National Institute of Health and the United States Centers for Disease Conclusion Control and Prevention. Participants gave written informed consent before Te fndings from this study highlight important factors enrollment. associated with IRS acceptance and refusal in Mozam - Funding bique. Although the majority of respondents perceived This study was funded by the United States President’s Malaria Initiative, via the efectiveness of IRS as limited, a large proportion the African Network for Field Epidemiology (AFENET), under Cooperative accepted the intervention to combat malaria due to Agreement U2RGH000046 with the Mozambican government. diverse motivations such as desire to kill other insects and to comply with community and government expec Publisher’s Note - Springer Nature remains neutral with regard to jurisdictional claims in pub- tations. Tese fndings suggest that having trusted lished maps and institutional afliations. community leaders from all political parties and spray mobilizers communicate with households that IRS kills Received: 4 December 2018 Accepted: 16 January 2019 the mosquitoes that cause malaria, thereby providing protection for their families and for their communities, may increase intervention acceptance. Tis interper- References sonal communication can be challenging to implement 1. MISAU, INE.Inquérito de Indicadores de Imunização, Malária e HIV/SIDA at scale, but is necessary for ensuring household aware- em Moçambique (IMASIDA) 2015: Relatório de Indicadores Básicos. ness and gaining community trust. Additionally, clear Maputo, 2016. 2. MISAU. Programa Nacional de Controlo da Malária: Relatório Anual, 2015. communication with community and political leaders (Pública DNdS ed. Maputo, 2016. during spray operator selection processes and spray 3. Mabunda S, Mathe G, Streat E, Nery S, Kilian A. National Malaria Indicator implementation should help reduce barriers to IRS Survey, Mozambique (MIS-2007). Saúde Md ed. Maputo, 2007. 4. MISAU. Programa Nacional de Controlo da Malária: Plano Estratégico da acceptance across diverse settings. Malária 2017–2022. (Pública DNdS ed. Maputo, 2017. 5. President’s Malaria Initiative. Malaria operational plan FY 2018. Washing- ton: President’s Malaria Initiative; 2017. Abbreviations 6. Abilio AP, Marrune P, de Deus N, Mbofana F, Muianga P, Kampango A. FGD: focus group discussion; IDI: in-depth interview; IRS: indoor residual spray- Bio-efcacy of new long-lasting insecticide-treated bed nets against ing; NMCP: Mozambique National Malaria Control Programme; WHO: World Anopheles funestus and Anopheles gambiae from central and northern Health Organization. Mozambique. Malar J. 2015;14:352. 7. Glunt KD, Abilio AP, Bassat Q, Bulo H, Gilbert AE, Huijben S, et al. Long- Authors’ contributions lasting insecticidal nets no longer efectively kill the highly resistant AM, CB, RZ jointly wrote the frst draft of the manuscript. SC, PN, MS reviewed Anopheles funestus of southern Mozambique. Malar J. 2015;14:298. the paper and provided additional information. All authors read and approved 8. Pluess B, Tanser FC, Lengeler C, Sharp BL. Indoor residual spraying for the fnal manuscript. preventing malaria. Cochrane Database Syst Rev. 2010;4:006657. 9. WHO. Indoor residual spraying: an operational manual for indoor residual Author details spraying (IRS) for malaria transmission control and elimination. Geneva: 1 Sistemas de Saúde, Instituto Nacional de Saúde, Ministério da Saúde, World Health Organization; 2015. 2 Moçambique, Maputo, Mozambique. US President’s Malaria Initiative 10. Bridges DJ, Pollard D, Winters AM, Winters B, Sikaala C, Renn S, et al. and Malaria Branch, Division of Parasitic Diseases and Malaria, US Centers Accuracy and impact of spatial aids based upon satellite enumeration to for Disease Control and Prevention, Maputo, Mozambique. improve indoor residual spraying spatial coverage. Malar J. 2018;17:93. 11. Larsen DA, Borrill L, Patel R, Fregosi L. Reported community-level indoor Acknowledgements residual spray coverage from two-stage cluster surveys in sub-Saharan The authors are grateful to the Mozambican Ministry of Health, to the research Africa. Malar J. 2017;16:249. assistants who contributed to data collection, as well as to study participants 12. Rehman AM, Coleman M, Schwabe C, Baltazar G, Matias A, Gomes IR, for their dedicated contribution to the study. The team would also thank Cyn- et al. How much does malaria vector control quality matter: the epide- thia Baltazar and Sergio Gomane for their contribution to the planning and miological impact of holed nets and inadequate indoor residual spraying. implementation of this study. The realization of this study would not be pos- PLoS ONE. 2011;6:e19205. sible without the support of the American people through the US President’s 13. Mabaso ML, Sharp B, Lengeler C. Historical review of malarial control in Malaria Initiative. southern African with emphasis on the use of indoor residual house- spraying. Trop Med Int Health. 2004;9:846–56. Competing interests 14. PMI, Africa IRS (AIRS) Project Indoor Residual Spraying (IRS 2). Task Order The authors declare that they have no competing interests. The fndings and Six. Mozambique: 2016 End of Spray Report. Maputo, Mozambique, Abt conclusions in this report are those of the authors and do not necessarily rep- Associates; 2017. resent the ofcial position of the Centers for Disease Control and Prevention. 15. Kaufman MR, Rweyemamu D, Koenker H, Macha J. “My children and I will no longer sufer from malaria”: a qualitative study of the acceptance and Availability of data and materials rejection of indoor residual spraying to prevent malaria in Tanzania. Malar The datasets used and/or analyzed during the current study are available from J. 2012;11:220. the corresponding author on reasonable request. 16. Wadunde I, Mpimbaza A, Musoke D, Ssempebwa JC, Ediau M, Tuhebwe D, et al. Factors associated with willingness to take up indoor residual Magaço et al. Malar J (2019) 18:27 Page 12 of 12

spraying to prevent malaria in Tororo district, Uganda: a cross-sectional 23. Ediau M, Babirye JN, Tumwesigye NM, Matovu JK, Machingaidze S, Okui study. Malar J. 2018;17:5. O, et al. Community knowledge and perceptions about indoor residual 17. Montgomery CM, Munguambe K, Pool R. Group-based citizenship in spraying for malaria prevention in Soroti district, Uganda: a cross-sec- the acceptance of indoor residual spraying (IRS) for malaria control in tional study. Malar J. 2013;12:170. Mozambique. Soc Sci Med. 2010;70:1648–55. 24. Ingabire CM, Rulisa A, Van Kempen L, Muvunyi C, Koenraadt CJ, Van Vugt 18. Munguambe K, Pool R, Montgomery C, Bavo C, Nhacolo A, Fiosse L, et al. M, et al. Factors impeding the acceptability and use of malaria preventive What drives community adherence to indoor residual spraying (IRS) measures: implications for malaria elimination in eastern Rwanda. Malar J. against malaria in Manhica district, rural Mozambique: a qualitative study. 2015;14:136. Malar J. 2011;10:344. 25. Hongvivatana T, Leerapan P, Smithisampan M. An observational study 19. Boene H, Gonzalez R, Vala A, Ruperez M, Velasco C, Machevo S, et al. of DDT house spraying in a rural area of Thailand. J Trop Med Hyg. Perceptions of malaria in pregnancy and acceptability of preventive 1982;85:245–50. interventions among Mozambican pregnant women: implications for 26. Govere J, Durrheim D, la Grange K, Mabuza A, Booman M. Community efectiveness of malaria control in pregnancy. PLoS ONE. 2014;9:e86038. knowledge and perceptions about malaria and practices infuencing 20. QSR International Pty Ltd. NVivo qualitative data analysis Software. In malaria control in Mpumalanga Province, South Africa. S Afr Med J. Version 11, 2016. 2000;90:611–6. 21. Oxborough RM. Trends in US President’s Malaria Initiative-funded indoor 27. Chimberengwa PT, Masuka N, Gombe NT, Tshimanga M, Takundwa L, residual spray coverage and insecticide choice in sub-Saharan Africa Bangure D. Indoor household residual spraying program performance (2008–2015): urgent need for afordable, long-lasting insecticides. Malar J. in Matabeleland South province, Zimbabwe: 2011 to 2012; a descriptive 2016;15:146. cross-sectional study. Pan Afr Med J. 2015;20:27. 22. Chaccour C, Alonso S, Zulliger R, Wagman J, Saifodine A, Candrinho 28. PMI, Africa IRS (AIRS) Project Indoor Residual Spraying (IRS 2) Task Order B, et al. Combination of indoor residual spraying with long-lasting Six: 2017 End of Spray Report, Maputo, Mozambique. Maputo: Abt Asso- insecticide-treated nets for malaria control in Zambezia, Mozambique: a ciates Inc.; 2018. cluster-randomized trial and cost-efectiveness study protocol. BMJ Glob Health. 2018;3:e000610.

Ready to submit your research ? Choose BMC and benefit from:

• fast, convenient online submission • thorough peer review by experienced researchers in your ﬁeld • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year

At BMC, research is always in progress.

Learn more biomedcentral.com/submissions