Access to insecticide-treated bed nets and household-level implementation of IRS contributed to significant reductions in self-reported malaria prevalence among women of reproductive age in Ghana. This finding reinforces the need for a comprehensive malaria control response to contribute to the elimination of malaria in Ghana.
Data for this study are drawn from the Ghana Malaria Indicator Survey (GMIS). The GMIS is a nationally representative survey conducted by the Ghana Statistical Service from October to December 2016. In this study, only women of childbearing age aged 15-49 years participated in the survey. Women who had data on all variables were included in the analysis.
For the 2016 study, Ghana’s MIS used a multi-stage cluster sampling procedure across all 10 regions of the country. The country is divided into 20 classes (10 regions and type of residence – urban/rural). A cluster is defined as a census enumeration area (CE) consisting of approximately 300–500 households. In the first sampling stage, clusters are selected for each stratum with a probability proportional to size. A total of 200 clusters were selected. In the second sampling stage, a fixed number of 30 households were randomly selected from each selected cluster without replacement. Whenever possible, we interviewed women aged 15–49 years in each household [8]. The initial survey interviewed 5,150 women. However, due to non-response on some variables, a total of 4861 women were included in this study, representing 94.4% of women in the sample. Data include information on housing, households, women’s characteristics, malaria prevention, and malaria knowledge. Data were collected using a computer-assisted personal interview (CAPI) system on tablets and paper questionnaires. Data managers use the Census and Survey Processing (CSPro) system to edit and manage data .
The primary outcome of this study was self-reported malaria prevalence among women of childbearing age 15–49 years, defined as women who reported having at least one episode of malaria in the 12 months preceding the study. That is, self-reported malaria prevalence among women aged 15–49 years was used as a proxy for actual malaria RDT or microscopy positivity among women because these tests were not available among women at the time of the study.
Interventions included household access to insecticide-treated nets (ITN) and household use of IRS in the 12 months preceding the survey. Families that received both interventions were considered joined. Households with access to insecticide-treated bed nets were defined as women living in households that had at least one insecticide-treated bed net, while households with IRS were defined as women living in households that had been treated with insecticides within 12 months before the survey of women.
The study examined two broad categories of confounding variables, namely family characteristics and individual characteristics. Includes household characteristics; region, type of residence (rural-urban), gender of household head, household size, household electricity consumption, type of cooking fuel (solid or non-solid), main floor material, main wall material, roof material, source of drinking water (improved or not improved), toilet type (improved or non-improved) and household wealth category (poor, middle and rich). Categories of household characteristics were recoded according to DHS reporting standards in the 2016 GMIS and 2014 Ghana Demographic Health Survey (GDHS) reports [ 8 , 9 ]. Personal characteristics considered included the woman’s current age, highest level of education, pregnancy status at the time of the interview, health insurance status, religion, information about exposure to malaria in the 6 months before the interview, and the woman’s level of knowledge about malaria issues. . Five knowledge questions were used to assess women’s knowledge, including women’s knowledge of the causes of malaria, symptoms of malaria, methods of malaria prevention, treatment of malaria, and awareness that malaria is covered by the Ghana National Health Insurance Scheme (NHIS). Women who scored 0–2 were considered to have low knowledge, women who scored 3 or 4 were considered to have moderate knowledge, and women who scored 5 were considered to have complete knowledge about malaria. Individual variables have been associated with access to insecticide-treated nets, IRS, or malaria prevalence in the literature.
Women’s background characteristics were summarized using frequencies and percentages for categorical variables, whereas continuous variables were summarized using means and standard deviations. These characteristics were aggregated by intervention status to examine potential imbalances and demographic structure that indicate potential confounding bias. Contour maps were used to describe self-reported malaria prevalence among women and coverage of the two interventions by geographic location. The Scott Rao chi-square test statistic, which accounts for survey design characteristics (i.e., stratification, clustering, and sampling weights), was used to assess the association between self-reported malaria prevalence and access to both interventions and contextual characteristics. Self-reported malaria prevalence was calculated as the number of women who had experienced at least one episode of malaria in the 12 months before the survey divided by the total number of eligible women screened.
A modified weighted Poisson regression model was used to estimate the effect of access to malaria control interventions on women’s self-reported malaria prevalence16, after adjusting for inverse probability of treatment weights (IPTW) and survey weights using the “svy-linearization” model in Stata IC . (Stata Corporation, College Station, Texas, USA). The inverse probability of treatment weight (IPTW) for intervention “i” and woman “j” is estimated as:
The final weighting variables used in the Poisson regression model are then adjusted as follows:
Among them, \(fw_{ij}\) is the final weight variable of individual j and intervention i, \(sw_{ij}\) is the sample weight of individual j and intervention i in the 2016 GMIS.
The post-estimation command “margins, dydx (intervention_i)” in Stata was then used to estimate the marginal difference (effect) of intervention “i” on self-reported malaria prevalence among women after fitting a modified weighted Poisson regression model to control. all observed confounding variables.
Three different regression models were also used as sensitivity analyses: binary logistic regression, probabilistic regression, and linear regression models to estimate the impact of each malaria control intervention on self-reported malaria prevalence among Ghanaian women. 95% confidence intervals were estimated for all point prevalence estimates, prevalence ratios, and effect estimates. All statistical analyzes in this study were considered significant at an alpha level of 0.050. Stata IC version 16 (StataCorp, Texas, USA) was used for statistical analysis.
In four regression models, self-reported malaria prevalence was not significantly lower among women receiving both ITN and IRS compared with women receiving ITN alone. Moreover, in the final model, people using both ITN and IRS did not show a significant reduction in malaria prevalence compared with people using IRS alone.
Impact of access to anti-malaria interventions on women-reported malaria prevalence by household characteristics
Impact of access to malaria control interventions on self-reported malaria prevalence among women, by women’s characteristics。
A package of malaria vector control prevention strategies helped significantly reduce the self-reported prevalence of malaria among women of reproductive age in Ghana. Self-reported malaria prevalence decreased by 27% among women using insecticide-treated bed nets and IRS. This finding is consistent with the results of a randomized controlled trial that showed significantly lower rates of malaria DT positivity among IRS users compared to non-IRS users in an area with high malaria endemicity but high standards of ITN access in Mozambique [19 ]. In northern Tanzania, insecticide-treated bed nets and IRS were combined to significantly reduce Anopheles densities and insect vaccination rates [20]. Integrated vector control strategies are also supported by a population survey in Nyanza province in western Kenya, which found that indoor spraying and insecticide-treated bed nets were more effective than insecticides. The combination may provide additional protection against malaria. networks are considered separately [21].
This study estimated that 34% of women had had malaria in the 12 months preceding the survey, with a 95% confidence interval estimate of 32–36%. Women living in households with access to insecticide-treated bed nets (33%) had significantly lower self-reported malaria incidence rates than women living in households without access to insecticide-treated bed nets (39%) . Similarly, women living in sprayed households had a self-reported malaria prevalence rate of 32%, compared with 35% in non-sprayed households. The toilets have not been improved and the sanitary conditions are poor. Most of them are outdoors and dirty water accumulates in them. These stagnant, dirty bodies of water provide an ideal breeding ground for Anopheles mosquitoes, the main vector of malaria in Ghana. As a result, toilets and sanitation conditions did not improve, which directly led to increased transmission of malaria within the population. Efforts should be intensified to improve toilets and sanitation conditions in households and communities.
This study has several important limitations. First, the study used cross-sectional survey data, making it difficult to measure causality. To overcome this limitation, statistical methods of causality were used to estimate the average treatment effect of the intervention. The analysis adjusts for treatment assignment and uses significant variables to estimate potential outcomes for women whose households received the intervention (if there was no intervention) and for women whose households did not receive the intervention.
Second, access to insecticide-treated bed nets does not necessarily imply the use of insecticide-treated bed nets, so caution must be used when interpreting the results and conclusions of this study. Third, the results of this study on self-reported malaria among women are a proxy for the prevalence of malaria among women in the past 12 months and therefore may be biased by women’s level of knowledge about malaria, especially undetected positive cases.
Finally, the study did not account for multiple malaria cases per participant during the one-year reference period, nor the precise timing of malaria episodes and interventions. Given the limitations of observational studies, more robust randomized controlled trials will be an important consideration for future research.
Households that received both ITN and IRS had lower self-reported malaria prevalence compared to households that received neither intervention. This finding supports calls for integration of malaria control efforts to contribute to the elimination of malaria in Ghana.
Post time: Oct-15-2024