Progress in coverage of bed net ownership and use in Burkina Faso 2003–2014: evidence from population-based surveys

Background Use of insecticide-treated bed nets (ITNs) is the cornerstone of malaria prevention. In 2010 and 2013, the Burkina Faso Government launched mass distribution campaigns of ITNs to increase coverage of ownership and use in the country. This study assessed the progress towards universal bed net coverage in Burkina Faso. Methods The authors used data from the Burkina Faso 2003 and 2010 Demographic and Health Surveys (DHS), the 2006 Multiple Indicator Cluster Surveys (MICS) and the 2014 Malaria Indicator Survey (MIS). For each survey, the authors computed key malaria prevention indicators in line with recommendations from the Survey and Indicator Task Force of the Roll Back Malaria Monitoring and Evaluation Reference Group. The trends over a decade was assessed by calculating percentage point change between 2003 and 2014. Results At national level, the proportion of households owning at least one ITN increased substantially from 5.6, 95% CI (4.7, 6.5%) in 2003 to 89.9% (88.5, 91.2%) in 2014, with low heterogeneity between regions. The proportion of households owning at least one ITN per two people increased significantly from 1.8% (1.4, 2.3%) in 2003 to 49.2% (47.3, 51.0%) in 2014. ITN use in the general population increased from 2.0% (1.6, 2.3%) in 2003, to 67.0% (65.3, 68.7%) in 2014. A similar trend was observed among children under the age of five years, increasing from 1.9% (1.5, 2.4%) in 2003 to 75.2% (73.2, 77.3%) in 2014, and among pregnant women, increasing from 3.0% (1.9, 4.2%) in 2003 to 77.1% (72.9, 81.3%) in 2014. The intra-household ownership gap was 67.0% (61.5, 72.4%) in 2003, but decreased significantly to 45.3% (43.6, 47.1%) in 2014. The behavioural gap, which was relatively low in 2013 with only 20.0% of people who had access to an ITN but were not using it, further decreased to 5.9% in 2014. Conclusion Burkina Faso made considerable progress in coverage of ITN ownership, access and use between 2003 and 2014, as a result of the two free mass distribution campaigns in 2010 and 2013. However, ITN coverage remains below the national targets of 100% for ownership and 80% for use. The results of 90% of ownership and 67% of use confirm that free mass distribution campaigns of ITNs are effective; however, there is room for improvement to reach and maintain optimal coverage of ITN ownership and use.


Background
Insecticide-treated bed nets (ITNs) are effective tools for malaria control [1]. Meta-analyses have shown that ITNs are associated with an 18% reduction in child mortality [2], 51% decrease in uncomplicated malaria incidence and 17% reduction in parasite prevalence in children [3]. In the past decade, the rapid scale-up of bed nets in sub-Saharan Africa (SSA) contributed to the significant decline of malaria burden in the region [4,5]. Sustaining high coverage of this intervention is critical to decrease further the burden of the disease and reach the long term-goal of malaria elimination. It is estimated that  30.13, 1200 Bruxelles, Belgium Full list of author information is available at the end of the article a minimum of 150 million ITNs per year are needed to maintain a constant pool of 450 million functioning ITNs to protect individuals at risk in SSA [6].
In 2001, a nationwide survey in Burkina Faso estimated that only 12.4% of children under the age of five were sleeping under a net, compared to 23.2% in 2005. Among pregnant women this proportion was 10.0% in 2001 and 27.5% in 2005 [20]. To rapidly increase coverage of ITN ownership and use, particularly among vulnerable groups, the Government of Burkina Faso initiated a first national-scale, free distribution campaign of ITNs in 2010. The aim of the campaign was to ensure that households had access to at least one ITN for every two people through the distribution of about eight million long-lasting insecticidal nets (LLINs). Moreover, in 2013, the national malaria control programme (NMCP) launched the second free LLIN distribution campaign to scale-up the coverage of ITNs in the country. This campaign aimed to ensure that 100% of households owned at least one ITN, and reach 80% ITN use by 2015. In 2014, the Burkina Faso Government decided to conduct the first Malaria Indicator Surveys (MISs) to assess coverage and impact of scaled-up malaria interventions. MISs were developed by the Roll Back Malaria (RBM) Monitoring and Evaluation Reference Group (MERG) with the aim to help national ministries of health collect key and timely information on malaria control at national level [21]. As Burkina Faso aims to achieve universal coverage with LLINs, this paper assessed the progress and gaps in coverage of bed net ownership and use based on RBM/ MERG-recommended indicators [21].

Methods
The authors analysed regional trends of ITNs ownership, access and use indicators in Burkina Faso over 11 years. These indicators were computed using data from the 2003 and 2010 Demographic and Health Surveys (DHS) [22,23], 2006 Multiple Indicators Cluster Survey (MICS) [24] and the first national MIS 2014 [25]. At the time of these surveys, Burkina Faso was divided into 13 administrative regions.

Data from Demographic and Health Survey 2003 and 2010
DHS 2003 (between June and December 2003) and DHS 2010 (between May 2010 and January 2011) were conducted during the high transmission season. DHS was designed to obtain national and regional estimates for malaria indicators. The DHS surveys followed a twostage selection process in which a random sample of clusters was first selected from the most recent national sample frame. In the second stage, all households were listed and the final list of households selected by systematic random sample. In the Burkina Faso DHS, the sample was selected in two stages, stratified by place of residence (urban and rural) with enumeration areas (EAs) as the first-stage sampling units, and households as the secondstage sampling units. Further details are provided in the DHS reports [22,23].

Data from MIS 2014
The MIS data were collected from October to November 2014 (at the end of the high transmission season), using the standard malaria indicator questionnaires developed by the RBM and the DHS Program. The dataset consists of malariometric information, demographic characteristics and socio-economic status on a nationally representative sample of 6448 households from 252 clusters, of which 52 are in the urban areas. These clusters were derived from a stratified two-stage cluster design. A detailed description of the sampling strategies is documented in the final report of the 2014 Burkina Faso MIS [25].

Data from MICS 2006
Multiple Indicator Cluster Surveys are typically carried out by government organizations with the support and assistance of UNICEF to fill data gaps for monitoring the children and women wellbeing.  [24].

Indicators
Ownership, access, use, and gap indicators were calculated from the datasets of households and individual household members, as recommended by MERG [21] ( Table 1).

Statistical methods
Data was analysed using Stata version 14 software and the maps were made using the R software. Point estimates (in percentage) and 95% confidence intervals were computed for each indicator and data point. In addition the percentage point changes between the baseline (2003) and endline (2014) were computed to assess change in the indicator and statistical significance assess at 5% level. Change by region and socio-demographical factor of each indicator between 2003 and 2014 were explored using the difference between weighted proportions (with svy prop command for survey data analysis) in Stata version 14 followed by a Lincom command (Linear combination of estimators). The survey mean command followed by Lincom (to compute two-sample t-test for difference in means with sampling weights) was used for the continuous variable access. Using this approach, we were directly testing whether the observed difference was significantly superior to zero.  (Table 2).

ITN
Insecticide-treated bed nets coverage also increased significantly in different regions from 2003 to 2014. The percentage point increases were consistently high across regions (from 76.9 to 94.5). Compared to 2003 and 2010, ITN ownership was rather stable across the country and displayed limited geographic heterogeneity in 2014 (Fig. 2).

ITN ownership at household level (P2: households with at least one ITN for every two people)
The proportion of households with enough ITNs for every household member, i.e., at least one ITN for every two people, was 1.  Table 3). In these two regions, 62.5 and 60.4% households, respectively, owned at least one ITN for every two members in 2014 (Fig. 2).

Use of ITN at individual level (P4)
In 2003, 2.0% (1.6, 2.3%) slept under a net. In 2010, the proportion of people who slept under an ITN was 31.5% (29.8, 33.2%) compared to 67.0% (65.3, 68.7%) in 2014, suggesting a considerable increase of 65.0% points from 2003 to 2014 (p < 0.001, Fig. 1). In urban areas, 5.6% of individuals used ITNs in 2003, a proportion that increased to 61.8% by 2014 (p < 0.001). In rural areas, a significant increase was also observed, with 1.2% of people who used an ITN in 2003 and 68.8% in 2014 (p < 0.001). The substantial increase in the proportion of people who slept under ITNs was observed across all quintiles of wealth. Use of ITNs increased significantly from 0.8% in 2003 to 63.7% in 2014 (p < 0.001) in the poorest wealth quintile. In the second poorest wealth quintile, ITN use increased from 0.8% in 2003 to 69.1% in 2014 (p < 0.001), compared to an increase of 54.9% points in the richest quintile (6.1-61.0%) ( Table 5).
In terms of regions, the largest increase occurred in the Central-East and Central-South regions. The absolute increase in ITN use in the Central-East region was 75.9% points (p < 0.001), increasing from 2.6% in 2003 to 78.6% in 2014. The corresponding estimates for the Central-South region were 1.4% in 2003 and 75.4% in 2014 (p < 0.001) (Fig. 3).

Use of ITN among children under 5 years of age (P5)
In 2003, 1.9% (1.5, 2.4%) of children under 5 years of age were sleeping under an ITN, compared to 9.6% (7.6,  (Table 5). Overall, the use of ITNs among children under five years has increased significantly from 2003 to 2014 (p < 0.001) (Fig. 1). Analysis of ITN use by age band showed a significant increase from 2003 to 2014. In children younger than 12 months, use of ITNs increased from 1.9% in 2003 to 77.0% in 2014 (p < 0.001). Among children aged 12-23 months, the proportion that used ITNs increased from 2.2% in 2003 to 76.7% in 2014, suggesting an absolute increase of 74.5% points between the two periods (p < 0.001). Substantial increases in the use of ITNs also were observed in older children (ages 24, 36 and 48 months) from 2003 to 2014 (Table 6).
In urban settings, 6. In wealth quintiles, the smallest increases were observed in children under 5 years from the richest wealth quintile, with an increase from 7.0% in 2003 to 69.7% in 2014 (Table 6).
Marked increases in ITN use were also achieved in all regions over the specified period; however, the Centre-East and Centre-Nord regions displayed the greatest increase in ITN use compared to the other regions with an increase from 2.5 to 86.5% and from 0.5 to 82.1% (Table 6, Fig. 3).

Use of ITN among pregnant women (P6)
The use of ITNs by pregnant women in Burkina Faso  Table 7). The corresponding estimates in rural areas were 2.1% in 2003 and 78.8% in 2014, a significant improvement in ITN use among pregnant women between these periods (p < 0.001). A trend similar to that of ITN use in children under five years was found when analyses were performed by wealth quintile.

ITN ownership and use gaps
In 2003, 94.4% (93.5, 95.2%) of the study households did not possess an ITN (Fig. 1). Among those who owned at least one ITN, 67.0% (61.5, 72.4%) did not have sufficient bed nets to protect all members (intra-household net ownership gap). However, 19.4% (n = 94) of these households had excess ITNs (i.e., more than one ITN for every two people). A significant proportion (21.9%, n = 316) of the population with sufficient access to ITNs did not actually use them the night before the survey. In 2010, 43.0% (41.0, 45.2%) of the study households did not have an ITN. The intra-household net ownership gap was 67.6% (65.8, 69.3%), indicating that about two-thirds of the households with at least one ITN did not have sufficient ITNs to protect all members. This gap is presented in Table 8 by background characteristics and shows that the gap was very high in large household size (93.4%) and rural areas (70.4%). However, 18.3% (n = 1475) of these households had excess ITNs (i.e., more than one ITN for every two people). A small proportion (12.7%, n = 3700) of the population with sufficient access to ITNs did not actually use them.
In 2014, only 10.6% (9.2, 12.0%) of the study households did not have an ITN. The intra-household net ownership gap was 45.3% (43.6, 47.1%), indicating that about half of the households with at least one ITN did not have sufficient ITNs to protect all members. This gap was 80.7% in large household size and well above the national average (Table 8). However, 34.5% (n = 1926) of these households had excess ITNs (i.e., more than one ITN for every two people). A small proportion (5.9%, n = 1562) of the population with sufficient access to ITNs did not actually use them. In contrast, this proportion was 13.1% in urban areas and only 3.4% in rural areas (Table 9).

Discussion
The Government of Burkina Faso set a national goal to increase ITN ownership, access and use. These data provide evidence of the remarkable increase in the coverage of ITN ownership, particularly in 2014 after the second free distribution campaign. Indeed, ownership, access and use indicators calculated following MERG's recommendations [21] dramatically increased between 2003 and 2014 and was particularly successful at reaching the poorest populations. The increasing trend in ITN ownership described here, is consistent with data from 19 SSA countries during a similar time period [26]. The data show that the two free distribution campaigns substantially increased ITN ownership and reduced inequity among populations in Burkina Faso. These findings are consistent with other free mass distribution campaigns that have been carried out in SSA [9,27], demonstrating that this strategy can be used to rapidely scale-up ITN coverage in areas with low coverage and reduce social inequity. However, despite the significant progress, less than 50% of households own enough ITNs to protect every household members (Fig. 1). These campaigns should not represent the only mechanism by which ITNs are distributed to poorest communities and vulnerable populations [9]. In Burkina Faso, ITNs were provided for free to pregnant women and children under five years of age through routine channels, such as antenatal care and immunization campaigns. ITNs were also available for purchase in retail shops and stores. This could explain both the slight increase in bed net ownership (Fig. 1) and use (Fig. 3). However, the relative contribution of these distribution channels remains, to date, very limited. More than 90% of the ITNs were obtained during the free distribution campaign [25]. To reach and maintain high ITN coverage in Burkina Faso, there is a need to improve the contribution of the routine distribution through ANC and vaccination programs and develop alternative strategies, such as the continuous distribution of ITN in schools and by community health workers for replacement [28]). Ownership and behavioural gap analyses provide complementary information regarding ITN ownership and use. The results reveal geographical and sociogeographic  Remarkably, the ownership gap increased in Centre-Nord. This increase can be attributed to the 2003 gap value, which is clearly an outlier: 35.8% gap, while the gap in all other regions fell between 59 and 90 (Fig. 4). By contrast with ownership gap, behavioural gap remained stable across the country. However, significant decrease was observed in Boucles de Mouhoun, Centre, Centre-Sud, and Centre-Est. All four regions displayed higher-than-average behavioural gap in 2003. In this respect, Centre consistently displayed higher behavioural

Table 3 Proportion of households owning at least one insecticide-treated bed net for every two members
All estimates take into account sample weights

Table 4 Proportion of population having access to an insecticide-treated bed net
All estimates take into account sample weights  gap values for all years studied, most likely because the population in this region is concentrated in urban habitat (Ouagadougou). This result could be explained by higher population dynamics in the capital region. Interestingly also, the change in behaviour is very recent in this region (in 2010, behavioural gap was 40%). In 2014, the behavioural gap was uniformly low across the country (0-15%). This reduction is probably a result of the health promotion programmes initiated by the Government of Burkina Faso to improve awareness concerning malaria prevention methods [25]. Household size is the main factor associated with ownership gap in this study. Large households with at least one bed net lacked additional bed nets to protect all family members (vs 25.5% for small households). Conversely, households in urban settings and from the richest Table 5 Proportion of population who slept under an insecticide-treated bed net the night before the survey All estimates take into account sample weights

years old who slept under an insecticide-treated bed net the night before the survey
All estimates take into account sample weights  quintile of the population more frequently owned enough bed nets than households located in rural settings or with a lower wealth index. This result is consistent with the findings of other studies showing that ITN coverage is lower in urban areas because mass distribution campaigns usually focus on rural communities [29,30].
Therefore, future strategies for ITN distribution should pay partiuclar attention to urban areas. Overall, the behavioural gap was very low in 2014. However, households located in urban settings and from the richest quintile of wealth index have higher gaps, because they might have other alternative prevention Table 8 Proportion of households owning at least one insecticide-treated bed net but with fewer than one net for every two household members (intra-ownership gap) All   This study has a few limitations, however, they do not affect the validity of the results. This study was based on exiting data, and was limited by available data. Survey data were collected during different seasons of the year. MIS data were collected during the high transmission period while DHS data were collected during the end of the transmission period. This difference could potentially affect the trends analysis and may have under-or overestimated the effect size, as ITN use can be seasonal depending on the perceived nuisance of mosquitoes [31]. Furthermore, the measures presented in this paper were self-reported and therefore susceptible to social desirability biases.

Conclusion
Following the two free mass distribution campaigns in 2010 and 2013, Burkina Faso has made considerable progress in coverage of ITN ownership, access and use between 2003 and 2014. However, bed net coverage remains below national targets of 100% for ownership and 80% for use. To reduce significantly the malaria burden in Burkina Faso, the NMCP needs to increase further and sustained ITN ownership and use in the general population. The free mass distribution campaigns contributed effectively to increase INT ownership and use in Burkina Faso. The NMCP should continue implementing these campaigns to reach the universal coverage target.
In addition, these campaigns should be complemented by other bed net distribution mechanisms (through antenatal care, immunization) to identify and replace nets that are worn, damaged or lost between free mass distribution campaigns. Furthermore, NMCP should have an effective behaviour change communication component in all distribution mechanisms to ensure that the population use bed nets consistently.

Authors' contributions
FKS and SS conceived the study. SS and MP performed the statistical analysis. SS, FKS and MP drafted the manuscript. FKS and YY contributed to the manuscript by giving substantial intellectual inputs. All authors read and approved the final manuscript.

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