A Comparative Analysis of Determinants of Household Knowledge of Split and Grouped Mosquito Larval Source Management for Malaria Prevention and Control in Rural Malawi


 BackgroundLimited studies have compared determinants of household knowledge of split and grouped mosquito larval source management (LSM) methods. Thus, we compared determinants of household knowledge of split and grouped mosquito LSM methods for malaria control in Malawi. MethodsA cross-sectional study was conducted among 479 households in Nthache, Mwanza district, Malawi. Household knowledge of mosquito LSM methods; draining stagnant water, larviciding, clean environment and clearing grass/bushes were assessed by using a validated structured household questionnaire administered to an adult household member. Chi-square test was used to compare the distribution of household knowledge of each of the split and grouped mosquito LSM methods. Multivariable logistic regression was used to examine factors associated with household knowledge of each of the split and grouped mosquito LSM methods. ResultsHousehold knowledge of draining stagnant water was lower than high-level knowledge of grouped mosquito LSM methods (32.9% versus 83.5%, p=0.000). Household knowledge of clearing grass/bushes was lower than high-level knowledge of grouped mosquito LSM methods (8.2% versus 77.7%, p=0.000). Household knowledge of clean environment was lower than high-level knowledge of grouped mosquito LSM methods (21.8% versus 63.8%, p=0.000). No significant differences were observed between knowledge of larviciding and high-level knowledge of grouped mosquito LSM methods (4.1% versus 5.8%, p=0.421). Respondents without education had 57% less the odds of having knowledge of draining stagnant water than those with primary education (AOR=0.43, 95% CI 0.26-0.69). Those from iron-roofed households and the widowed had three and more than four times the odds of having knowledge of larviciding than those from grass-thatched households and the married (AOR=3.03, 95% CI 1.26-7.29 and AOR=4.73, 95% CI 1.34-16.73), respectively. ConclusionsExcept for larviciding, household knowledge of grouped mosquito LSM methods was significantly higher than split methods. Policy for integrated vector management should address determinants of split mosquito LSM methods to improve household knowledge.

comprehensive summary that compares and contrasts the current methods used to assess determinants of household knowledge of split and grouped mosquito LSM methods for malaria prevention and control in a rural Malawian community.

Study design
Data for this study were drawn from a cross-sectional study, Household Knowledge, Perceptions and Practices of Mosquito Larval Source Management for Malaria Prevention and Control in Mwanza district, Malawi, which was published elsewhere [10].

De nition of research variables
This study uses the methods of Kamndaya et al. and the methods description partly reproduces their wording [10].
1. De nitions a. Household knowledge of grouped mosquito LSM methods: Refers to the household knowledge of combined methods of draining stagnant water, larviciding, clearing grass and bushes as well as clean environment around the home.
1. Household knowledge of split mosquito LSM methods: Refers to the household knowledge of each of the following as a stand-alone method: draining stagnant water, larviciding, clearing grass and bushes as well as clean environment around the home. 2. High-level knowledge of mosquito LSM methods (yes): Refers to either having knowledge of two or more grouped mosquito LSM methods. 3. Low-level knowledge of mosquito LSM methods: Refers to either having knowledge of either none or one of the grouped mosquito LSM methods. 4. No: Not having knowledge of split mosquito LSM methods. 5. Yes: Having knowledge of one split mosquito method.

Statistical analyses
We conducted a comparative analysis of determinants of household knowledge between split and grouped mosquito LSM methods for malaria prevention and control. Chi-square test was used to compare the distribution of household knowledge of split and grouped mosquito LSM methods. Logistic regression model using enter method (contrast: indictor) was used to examine factors associated with household knowledge of each of the split mosquito LSM methods. The models adjusted for gender and selected socio-demographic factors (excluding pregnancy status). All statistical analyses used SPSS version 18. A detailed analysis of factors associated with household knowledge of grouped mosquito LSM methods are highlighted elsewhere [10].

Ethical considerations
Ethical approval for this study was obtained from the Malawi National Health Science Research Committee (NHSRC) (Approval number: 2158). Table 1 shows comparison of household knowledge of each of the split and grouped LSM methods among respondents in Nthache area, Mwanza district. We observed that knowledge of draining stagnant water was lower than high-level knowledge of grouped mosquito LSM methods (32.9% versus 83.5%, p=0.000).

Results
Our results suggested that knowledge of clearing grass/bushes was lower than high-level knowledge of grouped mosquito LSM methods (8.2% versus 77.7%, p=0.000). We observed that knowledge of clean environment was lower than high-level knowledge of grouped mosquito LSM methods (21.8% versus 63.8%, p=0.000). No signi cant differences were observed between knowledge of larviciding and high-level knowledge of grouped mosquito LSM methods (4.1% versus 5.8%, p=0.421).

Discussion
This study compared determinants of household knowledge of split and grouped mosquito LSM methods for malaria prevention and control in a rural community of Malawi. Knowledge is not the only factor that effect prevention and control practices [13], but it is an important prerequisite for the latter. Larval control is an important vector control methods that the majority of people in rural and urban communities need to be aware of. Literature suggests that mosquito larval habitats need treating, draining or lling as frequently as possible in order to disrupt the mosquito life cycle as it takes between 5 to 14 days for them to mature into adult mosquitoes [5]. It is, therefore, important to design and implement health promotion interventions that focus more on disrupting larval habitats in the rural areas.
We highlighted the lack of attention paid to the people without formal education in as far as awareness of draining stagnant water as well as clearing grass and bushes for malaria prevention and control is concerned in rural Malawi. In addition, our results showed signi cant differences between the knowledge of three (draining stagnant water, clearing grass/bushes and clean environment around the home) split mosquito LSM methods and high-level knowledge of grouped mosquito LSM methods. However, no signi cant differences between household knowledge of larviciding and high-level knowledge of grouped mosquito LSM methods were observed. It was further observed that the odds of having high-level knowledge of mosquito LSM methods were 62% less and 3.54 times higher among respondents without education and secondary education, respectively, than those with primary education. Overall, this shows that determinants of household knowledge for grouped mosquito LSM methods may have obscured those of split methods. Kamndaya et al. observed that twothirds (65.6%) and over half (53.0%) of the respondents showed knowledge of draining stagnant water as well as clearing grass and bushes for malaria prevention and control, respectively [10]. However, in our analysis of split mosquito LSM methods observed that respondents without formal education were less likely to have knowledge of draining stagnant water as well as clearing grass and bushes for malaria prevention and control than those who attained primary education. This was not an unexpected nding as schools in Malawi offer health education lessons on malaria prevention and control methods that include environmental management for larval habitats. It is worth noting, however, that an Iranian study observed a direct correlation between education levels of the respondents and their practice as well as participation in malaria programs as volunteers [14]. The result from this Malawian study, further, compares to ndings from a Cameroonian study that suggested that respondents with secondary/high school/ university level of education were more likely to have a good level of knowledge on malaria than their none/primary level colleagues [15]. Likewise, Kamndaya et al. observed that the odds of having high-level knowledge of grouped mosquito LSM methods were lesser among respondents without formal education as compared to those who attained primary education [10]. No wonder, Kimbi et al. recommended that new strategies for malaria prevention and control should be devised for sensitization messages targeting community members especially those without formal education [15]. Similar results were shared from a Tanzania study which observed that educational attainment was signi cantly positively correlated with both draining stagnant water and cleaning the environment for malaria prevention and control [9], and Malawian study which reported that uneducated women, women with limited levels of knowledge of malaria in terms of causes, symptoms and prevention measures, rural women and women whose households had no access to any form of media were more likely to have a lower knowledge score for malaria prevention [16]. It is against this background that Mukabana et al. emphasized personal interaction approach based on learning-by-doing when engaging in community-based educational programmes for larval control [7]. Thus, it is necessary to scale up awareness of malaria prevention including mosquito LSM methods among uneducated people in malaria endemic rural areas to eradicate the disease. Further, the odds of having knowledge of clearing grass and bushes around the home were higher among respondents that engaged in mixed farming (including pastoralists) and businesspersons (including the self-employed) as compared to crop farmers. According to [9], people of higher socio-economic status often perform environmental management techniques around their homes. Indeed, this nding should inform authorities to be innovative in nding ways of supporting rural communities with resources to engage in environmental management techniques for malaria prevention and control.
As low as 5.2% of the respondents knew larviciding for malaria prevention and control [10]. The current analysis observed that the odds of having knowledge of larviciding for malaria and prevention were higher among respondents living in households roofed with iron sheets as compared to those living in grassthatched households. Again, the odds of having knowledge of clearing grass and bush around the home were lesser among respondents living in mudwalled households than those living in brickwalled households. In general, people who live in grass-thatched and mudwalled homes are considered poor [6], as such these results re ect the differences in socio-economic status among respondents in the study area and elsewhere. Another Malawian study highlighted the importance of incentivizing community members to get them involved in larval control [17]. Interestingly, the odds of having knowledge of larviciding for malaria prevention and control were higher among respondents who were widowed as compared to those who were married. It is not clear why this was the case, as such we recommend that further qualitative studies should be conducted to better understand the reasons behind this situation. This recommendation, notwithstanding, it is vital for authorities to work with the widowed in raising awareness of this less known method among the community members including the married residents in the study area to in uence practice. Kamndaya et al. observed that less than half (48.9%) of the respondents knew clean environment for malaria prevention and control (environmental sanitation) [10]. This knowledge was low considering the effectiveness of this method in malaria prevention and control. We, however, observed that the odds of having knowledge relating to clean environment for malaria prevention and control were higher among housewives and respondents that engaged in mixed farming (including pastoralists) than those engaged in crop farming. This result is partly consistent with a previous study that found that women were more likely to clean environment for malaria prevention and control, as compared to men [9]. Culturally, women/housewives look after their homes in SSA. This should, therefore, provide an opportunity for health workers to work with them in raising awareness for this method in areas where malaria is endemic. However, high illiteracy levels among uneducated mothers of children of local Malawian farmers exacerbate malaria cases among their children. Indeed, a Malawian study observed that children of mothers without formal education were more likely to be diagnosed with malaria than children of mothers who had received secondary education [18]. Indeed, the comparison analysis of determinants of household knowledge for split methods may have been obscured or exaggerated by grouped methods as exempli ed by marital status and household roof type. This clearly came out in the analysis of determinants of household knowledge for larviciding for malaria prevention and control. It has to be pointed out that a larger sample size is highly recommended when conducting analyses of determinants of household knowledge of larviciding to increase the statistical power as very few respondents showed knowledge of this method in the current study.
In the analysis of determinants of household knowledge of grouped mosquito LSM methods that are reported in our previous publication [10], we identi ed educational status, occupational and socio-economic status as determinants of high-level knowledge of mosquito LSM methods. However, a comparative analysis shows that determinants of some split mosquito LSM methods were either obscured or exaggerated by those of grouped mosquito LSM methods. For example, an analysis of split mosquito LSM methods revealed that marital status and household roof type were obscured as determinants for knowledge of larviciding for malaria prevention and control when grouped mosquito LSM methods were analysed. Similar observations were reported when draining stagnant water was analysed, as education attainment was the sole determinant for household knowledge of draining stagnant water. We, further, observed that occupation was the sole determinant of clean environment around the home. Further, our study showed signi cantly higher household knowledge of grouped than split mosquito LSM methods of draining stagnant water, clearing grass/bushes and clean environment among the study population. It is important to note, however, that no signi cant differences between household knowledge of grouped mosquito LSM methods and knowledge of larviciding were observed. Except for larviciding, we observed signi cantly higher household knowledge of grouped than split mosquito LSM methods. It is, therefore, recommended that policy for integrated vector management should address determinants of split LSM methods to improve household knowledge.
This study had some limitations which have been clearly outlined elsewhere [10]. In addition to that, it is also important to note that this study used 2018 population projections for Nthache area because at the time this study was being conceptualized and implemented the results of the national census were not yet out. These limitations, notwithstanding, a comparative analysis of determinants of household knowledge between split and grouped mosquito LSM methods was done and will help in designing and implementing the environmental management strategies for malaria prevention and control in rural areas of Malawi.

Conclusions
We identi ed educational status, occupational and socio-economic status as determinants for household knowledge of grouped mosquito LSM methods. Determinants of household knowledge of split mosquito LSM methods such as larviciding, draining stagnant water and clean environment were either obscured or exaggerated by those of grouped mosquito LSM methods. For example, an analysis of split mosquito LSM methods revealed that marital status and household roof type were obscured as determinants for knowledge of larviciding for malaria prevention and control when grouped mosquito LSM methods were analysed. Similar observations were reported when draining stagnant water was analysed, as education attainment was the sole determinant for household knowledge of draining stagnant water. We, further, observed that occupation was the sole determinant of clean environment around the home.
Our study showed signi cantly higher household knowledge of grouped than split mosquito LSM methods of draining stagnant water, clearing grass/bushes and clean environment among the study population. It is important to note, however, that no signi cant differences between household knowledge of grouped mosquito LSM methods and knowledge of larviciding were observed. Except for larviciding, we observed signi cantly higher household knowledge of grouped than split mosquito LSM methods. Policy for integrated vector management should address determinants of split LSM methods to improve household knowledge. Availability of data and materials The dataset is available on reasonable request to the corresponding author.

Competing interests
The authors declare that they have no competing interests.

Funding Not Applicable
Authors' contributions DM was involved in the design, development of the study protocol, data management, analysis, and manuscript preparation. MK reviewed the design, study protocol, manuscript, supported with data management and analysis. KL reviewed the draft manuscript. All authors read and approved the nal manuscript.