Factors influencing malaria infection in Rwanda 2010: a cross-sectional survey study using generalized structural equation modeling

Background Malaria is one of the primary public health concerns in the world and an important cause of morbidity and mortality in sub-Saharan Africa. Malaria morbidity is associated with poverty and vulnerability as it is not easy for the poor people to access preventive treatment and protective measures. In Rwanda, malaria prevention has become a major problem against the double-barreled burden of an overstretched health system and strained financial resources. This work was a cross-sectional survey study design based on data from Rwanda collected in 2010 through the Malaria Indicator Survey as part of the Demographic and Health Survey. The primary outcome variable was an ordinal variable with three categories; no malaria, probable malaria, and confirmed malaria cases. The outcome variable was formulated by combining rapid malaria test and confirmatory blood smear laboratory test. Statistical analysis was done using survey ordinal logistic regression modelling adjusting for random effects for direct effects and generalised structural equation modelling (G-SEM) to obtain total (direct and indirect) effects of malaria morbidity.

3 the number of rooms for sleeping.

Conclusion
Poverty is still the core issue to the morbidity patterns driving the malaria epidemic in Rwanda. Access to health insurance has a high positive impact on decreasing disease as such a special focus on some regions can be an effective intervention strategy. A better understanding of the drivers of morbidity directly and/or indirectly can better target interventions to be more efficient in those affected areas.

Background
Malaria is a major public health issue in the world [1]. It is a preventable and treatable infectious disease transmitted by mosquitoes, yet it kills more than one million people every year in sub-Saharan Africa. Globally, it was estimated that there were 660,000 deaths due to malaria of which a large proportion over 86% were in children (less than five years age) and older people (above sixty) [2]. In sub-Saharan Africa, malaria is the leading cause of death in children less than five years of age and accounts for about 91% of deaths occurring in children (less than five years of age) and older people (over sixty) [1]. (, #5;, #2) The African Leaders Malaria Alliance (ALMA), in collaboration with the nine countries (Angola, Cameroon, Chad, Congo, Gabon, Equatorial Guinea, Central African Republic, DR Congo and Sao Tomé-et-Principe) developed an African Roadmap to eliminate malaria by 2030 [1].
According to the World Health Organization (WHO) country-specific statistics, 2.2 million inhabitants out of 11.5 million people are at risk of malaria in Rwanda [3,4].
Malaria is mesoendemic (area in which a disease incidence is sufficiently high) in the low-lands and hypoendemic (area in which a disease incidence is sufficiently 4 low) in highlands of Rwanda [5]. Globally, in endemic areas where transmission occurred in long regular seasons, infection rates were highest among children less than five years of age who had not yet established immunity to the disease, contrary to epidemic areas where malaria transmission took place in short seasons; malaria infections in all age categories was high [6]. Furthermore, a study conducted in Ruhuha region of Rwanda concluded that there was also a high risk of malaria in older age groups during long regular malaria season [7]. The reason may be that older people did not sleep under treated bed nets as compared to young ones. Another reason might be because older people stayed out longer than younger ones which made them more likely to be bitten by mosquitoes [8,9].
Historically, it is believed that the most deadly malaria specie Plasmodium falciparum is prevalent in sub-Saharan Africa [10]. Consequently, malaria is hard to control in Africa due to the effectiveness of vector species and the predominance of most severe species Plasmodium falciparum [11]. Risk of malaria infection is high in rural areas of developing countries can be attributes to poverty and poor lifestyle [12]. Factors which play the major role in disease risk includes proximity to the vector breeding sites, age, socioeconomic status, altitude, moderate use of control measures, low income, illiteracy, land use near pools, and open houses [13][14][15][16].
A recent global report shows that due to good political commitment and better utilization of funding, a 54% of the above reduction has been experienced in African (WHO regions) [17]. As is the case in other sub-Saharan countries, in Rwanda the use of Long-Lasting Insecticidal Nets (LLIN), Indoor Residual Spraying (IRS) and treating malaria cases with Artemesinin-based combination therapy (ACT) had reduced malaria infection up to 50% [18,19]. Therefore, it has a significant influence on the health and socioeconomic well-being of people. Therefore, this 5 work focuses to determine the direct and indirect determinants of malaria morbidity in poorer households and at an individual level in Rwanda in 2010 using malaria indicator survey (MIS) data that was accessible from the Demographic and Health Survey (DHS) website.

Methods
This study was conducted as a cross-sectional survey design. The dataset used in this work is known as the Malaria Indicator Survey (MIS) dataset of Rwanda which was part of the Demographic and Health Surveys (DHS) 2010.

Inclusion Criteria
This study includes both males and females tested for malaria diagnosis (rapid malaria test and confirmatory blood smear test) either positive or negative result.

Study area
This work was conducted in Rwanda, a Central African country located South of the equator between latitude 1°4' and 2°51' South and longitude 28°63' and 30°54' East. It has a surface area of 26,338 square kilometers and is bordered by Uganda to the North, Tanzania to the East, the Democratic Republic of the Congo to the West and Burundi to the South. Landlocked, Rwanda lies 1,200 kilometers from the Indian Ocean and 2,000 kilometers from the Atlantic Ocean [20]. Rwanda is divided into five geographically-based provinces-North, South, East, West and the City of Kigali, with the provinces, further subdivided into 30 districts, 416 sectors, 2,148 cells, and 14,837 villages [20].

Two-stage sampling of Rwanda malaria indicator survey (MIS)
The data used for the analysis was obtained from the 2010 Malaria Indicator Survey 6 (MIS) conducted by Rwanda Demographic and Health Survey (RDHS) program. The previous study collected the data elements on basic demographic and health indicators, malaria prevention, treatment and morbidity.
Sampling in previous study was done in two stages. In the first stage 492 villages which formed the clusters were selected with probability proportional to the village size. The village population size also indicates the number of households in the village. The mapping and listing of all households in the selected villages was done.
The resulting list served as the sampling frame for the second stage of sample selection. All of the 492 clusters were selected for the modeling as surveyed for the 2010 RDHS. The selected data contained 11,865 households consented to participate in the study and completed the individual's questionnaires. Data for children less than five years of age was collected from their mothers [20,21].

Dependent variable
The malaria outcome variable for this study was defined according to the Center for or exogenous (explanatory variable) or both, which can be modelled by using generalised structural equation models. Details of these models provided in Figure 4 and Figure 5.

Power computation
There was no need for sampling for this study; instead power computation was done using a STATA ado-file. [28] A sample prevalence of 2.1% comprised of 1.4% children and 0.7% adults adopted from the Rwanda Malaria Indicator Survey [3] and a 3% assumed population prevalence at an alpha level of 0.05. A design effect of 10 was used and intercluster correlation (ICC) of 0.071 with a total of 492 clusters and average households per cluster of 115. Power was obtained an 89%.

Descriptive analysis and bivariate analysis
Descriptive analysis was conducted using survey chi square test (Rao-Scott adjustment) adjusted for cluster effect. For of categorical variables weighted percentages (proportion) with adjusted F-statistic are reported as shown in Table 1.
For continuous variables mean and confidence interval are reported as shown in  Table 2.

Multivariate analysis
Multivariate analysis was done using step-wise forward selection survey ordinal analyses that assume that all individuals are independent, the multilevel modelling approach accounts for the fact that people that live in the same area may have some characteristics in common. It should be noted individuals residing in the same household are not independent. However, this clustering and the household effect were included in this analysis given that it is common that multiple individuals from the same household were interviewed.
Another important feature of the random-effects model is that it gives information on the proportion of total variation that was explained by the cluster-level, individual level and household level predictors. The model sets with cluster number and household number as random effects to account for inter cluster correlations between individuals from the same cluster and household. 10 The models were also tested using the ordinal regression diagnostics i.e. goodness

Ethics approval
This study was granted ethical approval by the University of the Witwatersrand's

Descriptive statistics
The total number of individuals used in this paper was 11,865 with a mean age and   Table 1 shows the results of the descriptive statistics survey weighted percentages of independent variables with the significant test statistic of survey adjusted chisquare test. Table 2 presents the results of survey ordinal univariate analysis, 13 survey ordinal logistic regression modelling adjusting for random effects for random effects for direct effects and generalised structural equation modelling (G-SEM) to obtain (direct and indirect) effects of malaria infection.

Results of direct effects
The direct effects from G-SEM are shown in Table 3 Table 3 adjusting for other household level variables.
The number of rooms for sleeping was modeled as an exogenous variable, directly impacted negatively on malaria infection as shown in Table 3 with total effect as shown in Table 3.

Results of indirect effects
The indirect effects found from G-SEM are shown in Table 3. The connected arrows show indirect pathways that were statistically significant. Moreover, variable like the number of rooms for sleeping was indirectly influenced by least poor and least poor was indirectly affected by education gain in years. Education gain in years also affected the health insurance which showed the positive relationship as shown in Table 3. 14 Furthermore, when health insurance was treated an endogenous variable, SES variable category "least poor" indirectly impacted malaria infection and had a protective effect as shown in Table 3. Education in years as an exogenous variable had no direct influence on malaria infection, but indirectly affected malaria infection through least poor as shown in Table 3 and had a protective effect on health insurance as shown in Table 3.
When the number of rooms for sleeping was an endogenous variable, SES variable category least poor was indirectly negatively impacted the malaria infection.
However, it positively effected when the no of rooms for sleeping effect was direct as shown in Table 3.

Total effect
The total effect is an accumulative value of direct and indirect effects of each respective variable. The results of the total effect are shown in Table 3. Least poor category showed a significant total effect of as shown in Table 3 Health insurance also showed a total effect of as shown in Table 3. Education gain in years showed an indirect effect and resulting total effect was as shown in Table 3.

SEM)
Generalised structural equation modelling (G-SEM) was based on variables which were statistically significant in regression analyses were chosen for G-SEM pathways as shown in Figure 4 and   Figure 5.
Education gain in years indirectly impacted the malaria infection in two ways least poor (protective) and health insurance (a risk) as shown in Figure 5.

Education => health insurance => malaria infection = Risk
Eastern region showed greatest total effect on malaria infection as shown in Table   3, and cluster altitude in meters showed least total effect on malaria infection as shown in " Table 3.
The result of G-SEM showed both direct and indirect effects on endogenous variable malaria infection. Figure 4 shows the direct G-SEM model and indirect G-SEM model is shown in Figure 5. The endogenous variables; least poor, the number of rooms for sleeping, and health insurance are shown in Figure 5.

Discussion
The study revealed that malaria infection was influenced by the combination of variables such as behavior, household condition and ecological factors. In 2015, 214 million deaths were recorded due to malaria and 91 % were in sub-Saharan Africa [31]. Social-economic status was a fundamental factor of malaria morbidity in several studies [32,33]. The studies also provided a positive malaria diagnosis.
As mentioned earlier, malaria is associated with poverty [32]. The findings of this 16 work complement the results of the previous studies on malaria morbidity. Previous studies have shown that low socio-economic status increases the prevalence of malaria infection in developing countries [34]. According to the United Nations Accounts (UNA) central aggregated database 2015, Rwanda has low Gross Domestic Product (GDP) [35] which translates into higher malaria infections. This study also confirms the finding. It has also been indicated that the malaria infection is a disease of developing countries putting an additional burden on the healthcare facilities.
Logically the nutritional level is linked with GDP. Therefore, it is obvious that the poorer countries have inadequate resources to improve nutritional levels resulting in poor immunity against malaria disease [15]. Poor SES results into an insufficient use of health-care facilities, therefore increasing the vulnerability of the population to the risk of malaria. Governments of the poorer countries need rational, optimum and affordable policies to control malaria infection and treatment. This could be achieved by investing in research and data collection to investigate the trends of malaria to make effective and efficient policies and proper utilisation of resources.
Regions (Kigali, south, north, east and west) place-of-residence (urban or rural) and cluster altitude in meters may be interrelated. However, fewer studies indicated an effect on malaria prevalence dependence on altitude [36][37][38]. Researchers also revealed that individuals living in low altitude areas are more at risk as compared to higher altitude areas [39]. In this work, altitude has shown a marginal reduction in malaria infection.
A study conducted in Thailand concluded that school children could be a better source of anti-malaria education for the family members in contrast to disseminating messages by newsletters [40].
To achieve significant reduction in malaria morbidity in Rwanda, there is a need of improvement in the education status of the targeted population [41]. Knowledge about malaria prevention might be conveyed in the community by the students, who facilitate the household and families by applying and following influence in the targeted settings.
In this study, Age has shown a significance (p-value = 0.001). Previous studies have shown that children over the age of five were less at risk of malaria infection [8,42]. In this work, increase in age has shown a reduced tendency of malaria infection hence it is less likely to have positive malaria diagnosis as the age increases.
Existing literature also confirms that older people (above 50) have lesser chance of malaria infection due to developed immunity.

Conclusions
This study shows the importance of socio-economic status as well as influence of education in the fight against malaria. To eliminate malaria morbidity in the population, it is important for the governments to empower the community economically, intellectually and ensure the health education is a part of the efforts to fight the endemic. Access to health insurance has a positive impact on decreasing malaria infections. Therefore health insurance could be therefore focused as an effective tool in the intervention strategy especially in relatively high 19 income sectors to significantly reduce the infections. This will assist in the fight to eliminate malaria.
It is important to ensure that the resources are channeled to optimize prevention strategies that are put in place. Once the population is empowered, the preventative strategies can then be implemented successfully. If the population is Approval to use the MIS data was obtained from the Measure DHS website. The primary study, where the data was collected, verbal informed consent for testing of children was obtained from the child's parent or guardian at the end of the household interview and ethical clearances with the Rwanda authorities before the study started.

Consent for publication
Not applicable

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Tables   Due to technical limitations, Tables 1-3 are available in the manuscript file.