Mass media exposure and its impact on malaria prevention behaviour among adult women in sub-Saharan Africa: results from malaria indicator surveys

Settings and data source

Data were obtained from the MIS conducted in Burkina Faso (2014), Ghana (2016), Mali (2015), Malawi (2014), Kenya (2015), Nigeria (2015), Sierra Leone (2016), Uganda (2014–15). The surveys were conducted with technical assistance from the National Bureau of Statistics (NBS) of each country. It was funded by ICF International through the Demographic and Health Survey (DHS) Program which is also a USAID supported project in the implementation of population and health surveys [24]. The countries were selected based on geographical diversity. The SSA countries included in this study were thus; Eastern (Kenya, Malawi and Uganda) and Western (Burkina-Faso, Ghana, Mali, Nigeria and Sierra Leone) countries respectively (Fig. 1) [25].

The MIS s employed a stratified two-stage cluster design. In the first stage, sampling strata were created from which clusters were selected by a probability-proportional-to-size technique. A complete listing in the selected clusters served as the sampling frame for the second stage. In the second stage, households were selected from each cluster by equal probability systematic sampling. Study participants were women aged 15–49 years residing in non-institutional places in urban and rural areas. Details of the survey procedure have been reported in the method section of a related MIS report [26].

The surveys were to provide quality data for measuring the progress of effective monitoring and assessing national malaria programs implementation. Specifically, the surveys focused on measuring the indicators of ownership and use of mosquito bed nets, coverage of the intermittent preventive treatment programme for pregnant women, treatment-seeking behaviour, measuring indicators of knowledge and attitudes related to malaria control, and determining the factors associated with malaria and anaemia. Besides these, the survey also captured women’s basic demographic and educational information, information on use of ITNs, malaria prevention practices during pregnancy, and knowledge of malaria. Further details on survey are available from the data sources [24].

Variables

The outcome variables were two self-reported indicators of malaria prevention behaviour among the participants: 1) Sleeping under ITN the previous night, and 2) Use of antimalarial drugs in last pregnancy. Both of the questions were answered with the options of ‘Yes’ and ‘No’.

The independent variables of interest were reports of receiving malaria information through mass media and interpersonal communication. The following sources reported by the participants were included in the analysis and were categorised as ‘Yes’ or ‘No’: TV, Radio, Religious institutions, Poster, Community event, Health worker.

To measure the independent associations of malaria prevention behaviour, the analysis was adjusted for some potentially confounding variables selected through evidence on their correlation with the outcome and explanatory variables. The following were found relevant to the analysis: Age: 15–19/ 20–24/ 25–29/ 30–34/ 35–39/ 40–44/ 45–49; Setting: Urban/ Rural; Education: No education/ Primary/ Secondary/ Higher; Religion: Islam/ Christian/ Other; Wealth index: Poorest/ Poorer/ Middle/ Richer/ Richest.

All the variables were self-reported except for the wealth index. For the calculation of household wealth status, the volume of durable goods (e.g. TV, radio, bicycle) possessed by the household as well as housing quality (e.g. type of floor, wall, and roof) were taken into consideration. Each item was assigned a factor score generated through principal component analysis (PCA) which were then summed and standardized for the households. These standardised scores placed the households in a continuum based on relative wealth scores. The scores obtained were categorized into quintiles to rank the household as poorest/poorer/middle/richer/richest [27].

Data analysis

Statistical analyses were performed with SPSS version 24. Firstly, the datasets were cleaned for outliers, checked for potential multicollinearity, and then merged to perform pooled analysis. Due to nonproportional allocation of the samples across the survey regions, they were not self-weighted. Furthermore, MIS surveys use cluster sampling design. To adjust for sample weight and cluster sampling techniques, the analysis was preceded by the preparation of complex survey file by accounting for primary sampling units, sample strata and sample weight. Prevalence of the outcomes variables for each explanatory variables were presented in percentages. In addition, the percentage use of ITN and antimalarial drug used during pregnancy was presented for each study country. Multivariable binary logistic regression model was used to calculate the odds ratios of the associations between ITN use and uptake of antimalarial drug for the last pregnancy with mass media exposure and interpersonal communication channels. Each outcome variable was regressed as a function of exposure to the six different types of mass media and to the binary logistic regression, while adjusting for various demographic and socioeconomic variables which were selected based on standard conceptual framework [28]. Results of regression analysis were presented as odds ratios along with their 95%CIs as an indicator of significance as well as precision of the OR values. For all associations P-value of < 0.05 was considered statistically significant.

Ethical approval

The protocol of DHS surveys was approved by the Ethics Committee of ORC Macro Inc. [16, 17]. The study was based on analysis of anonymized secondary data available in the public domain of DHS, hence, no further ethical approval was necessary for the study.

Description of sample population

The overall prevalence of sleeping under ITN in the night preceding the interview was 67.9% (95%CI = 66.6–69.2), and that of taking antimalarial drug during the last pregnancy was 72.8% (95%CI = 71.3–74.2). Table 1 further shows that the prevalence rates of both of these indicators tended to be higher among women aged 25–29 years, residing in urban areas, having higher level of education, followers of Islam, living in non-poor households, and women who reported getting malaria related information through television, radio, religious institutions, posters, community events and health workers.

The percentage utilization of ITN and antimalarial drug during pregnancy were presented in Fig. 2. The percentage use of ITN was less than 40% in Nigeria (39.3%). While Ghana and Kenya also reported low use of ITN accounting for 46.6 and 54.8% respectively. However, other countries which had improved use of ITN include Sierra Leone (96.3%), Mali (89.0%) and Burkina-Faso (83.4%). Furthermore, antimalarial drug use during previous pregnancy was slightly above 90% in Sierra Leone, Malawi and Ghana respectively. Nonetheless, Nigeria, Kenya and Uganda accounted for 54.3, 59.0% and about two-third respectively. These results showed large disparities in ITN and antimalarial drug use during pregnancy across various SSA countries.

Multivariable analysis on the association between mass media exposure and interpersonal communication with malaria prevention behaviour.

Main findings

The findings of this study showed that there is need for improvement in using ITN regularly and taking antimalarial drugs during pregnancy. Striking variances were observed in the prevalence of ITN use among the eight countries. While Sierra Leone, Mali, Burkina Faso, attained coverage of over four-fifth among the adult women population, the situation seems to be far more challenging for other countries including Nigeria, Ghana, Kenya, and Malawi where about two-fifth to three-fifth of the women are not using ITNs. Similarly, some countries are clearly far behind than others in achieving optimum coverage for utilisation of antimalarial drug during pregnancy.

Except for women aged 15–19 years and antimalarial drug use, younger women were more likely to use both the services compared to those aged 40 years and above. The increase in antimalarial drug use during pregnancy could be as a result of the fertility level among the young women, which is generally known to outscore the older counterpart. Again, it is probable that the younger generations are enjoying a higher exposure to mass media and health related information, which could help them to become more aware of their health needs and environmental conditions. Women in the rural areas, those with low educational qualification and poor households were found to be less adherent to malaria prevention behaviour. Majority of the African population are living in rural areas where grid connectivity is still a serious problem, and more so the communities remain deprived of the benefits of health communication offered by fast expanding digital media technologies [22]. As the results further describe, television and radio were the most commonly reported sources of malaria information after direct communication with healthcare workers.

Percentages of women sleeping under ITN and taking antimalarial drugs in last pregnancy were higher among participants who reported receiving malaria information from TV, radio, poster/billboard, religious institutions, community events and healthcare workers. In the model that adjusted for potential confounders, receiving malaria information through radio, poster, community event and health worker turned out to be significant predictors of using ITN. However, not receiving malaria information through the mass media including radio amongst others, significantly resulted to reduction in the odds of ITN and antimalarial drug use during pregnancy.

Previous studies

Country representative studies on the use of ITN among women are still scarce, however there are evidences of appreciable success in increasing ITN coverage in the general population for several countries. In Burkina Faso, for instance, household ownership of ITN has increased dramatically from 5.6 to 89.9% between 2003 and 2014, a success that is largely attributable to the free mass distribution of ITNs in the years 2010 and 2013 [29]. At sub-national levels the prevalence of ITN use varies from 56.3% in Western Kenya [30], 90% in Malawi [31], and 27% in northern Nigeria [17]. These variations are hard to account for in light of the present analysis. Some possible explanations might include socioeconomic inequalities and the coverage and efficacy of the respective programs. For example, although Mali introduced its first IPTp and ITN policy (in 2003 and 2006 respectively) only a few years after Kenya (in 1999 and 2001 respectively) [32], a considerable disparity can be seen in the prevalence of ITN and IPTp use between these two countries.

In the current literature the evidence on the association between mass media exposure malaria prevention practices are scarce. However, studies in Cameroon [33], Gambia [34], Ghana, Malawi Nigeria [22], highlighted the importance of targeting the dissemination of health knowledge through traditional media. In Gambia, songs performed by community members were found to be effective in encouraging people to repair bed nets [34].

General discussion and policy implications

In the last 10–15 years tremendous efforts have been made to attain universal coverage of ITN and gestational use of antimalarial drugs, the two most cost-effective tools of prevention of malaria. Several countries in SSA have received free ITN campaign and distribution, thereby succeeded in increasing household ownership of ITNs. However, there are evidences on the gap between ownership and optimum use of the preventive services [35], addressing which could help the programs attain the full potential of the services being provided [36]. Achieving the universal coverage and utilisation goal of ITN and IPTp, can be greatly enhanced by improving people knowledge and awareness about malaria and encouraging adoption of preventive behaviour among the population. Social media strategies have proven to be useful in promoting health knowledge among adult populations in all settings, and can be leveraged by malaria elimination programs in Africa as well. Health policy making should focus on empowering women with health knowledge as it can help women develop their own perceptions and knowledge base of the particular disease and enable them to better communicate their health issues with physicians.

For long-terms success, the broader macroeconomic conditions e.g. research and funding, should also be given special attention. As more countries are progressing towards malaria elimination, the donor financing malaria programs has also been declining since 2010 [37, 38]. Many malaria-eliminating countries have projected national declines in funding from the Global Fund to Fight AIDS, Tuberculosis, and Malaria (GFATM), which has been the largest financial supporter of malaria since 2002 [39]. The changing priority of donor agencies can put the countries in SSA in a difficult situation with respect to their pursuit of malaria elimination. Since these are the least wealthy of all countries who are experiencing the highest incidence, it is very unlikely that national funding alone will be able to keep up with intervention and elimination efforts. National leaders in Africa need to gather efforts to reduce dependency on foreign donors and develop regional partnership for addressing local challenges.

Strengths and limitations of the study

Among the strengths of the present study is the large sample size, generalizability of the findings, women aged 15–49 years, and comparative presentation of the data that allows understanding of the situation in a particular country relative to others. However, there are several limiting points including the secondary nature of the data which means that authors had no control over the measurement and selection of the variables. Although the incidence of malaria is perennial in some regions, for certain there is a factor of seasonality. Therefore, not being able to account for high or low transmission period of the surveys might have affected the prevalence of ITN use. For instance in Kenya, the prevalence of ITN use was found to be significantly higher during the rainy season compared to that in dry season [30]. Apart from the seasonal variation in ITN use, another potential confounder for the prevalence might be the use of unobserved protective behaviour or mechanisms e.g. better housing environment, insecticide sprays, and traditional methods. Information on antimalarial drug use for the last pregnancy is subject to some degree of recall error. We were also unable to deduce the variation in the predictability of different types of mass media (digital, paper, health worker) on ITN use and uptake of antimalarial drug. There was no information on the exact contents of the messages women received through the above mentioned sources. Lastly, data were cross-sectional and therefore no causal relationship can be made from our analysis.