An insecticide-treated bed-net campaign and childhood malaria in Burkina Faso

Abstract Objective To investigate if the first national insecticide-treated bed-net campaign in Burkina Faso, done in 2010, was followed by a decrease in childhood malaria in a district with high baseline transmission of the disease. Methods We obtained data on the prevalence of Plasmodium falciparum parasitaemia in children aged 2 weeks to 36 months from malaria surveys in 2009 and 2011. We assessed morbidity in children younger than 5 years by comparing data from the Nouna health district’s health management information system before and after the campaign in 2010. We analysed mortality data from 2008 to 2012 from Nouna’s health and demographic surveillance system. Findings The bed-net campaign was associated with an increase in the reported use of insecticide-treated nets. In 2009, 73% (630/869) of children reportedly slept under nets. In 2011, 92% (449/487) did. The campaign had no effect on the proportion of young children with P. falciparum parasitaemia after the rainy season; 52% (442/858) in 2009 and 53% (263/499) in 2011. Cases of malaria increased markedly after the campaign, as did the number of children presenting with other diseases. The campaign was not associated with any changes in child mortality. Conclusion The 2010 insecticide-treated net campaign in Burkina Faso was not associated with a decrease in care-seeking for malaria or all-cause mortality in children younger than 5 years. The most likely explanation is the high coverage of nets in the study area before the campaign which could have had an effect on mosquito vectors, limiting the campaign’s impact.


Introduction
Malaria remains a major threat to the health of some three billion people in 108 countries where the disease is endemic. 1 According to the World Health Organization (WHO), in 2012 malaria caused 207 million episodes of disease and 627 000 deaths. 2 About 85% of illness and 90% of deaths attributable to malaria occur in sub-Saharan Africa, with young children being most affected. 3 After a partially successful malaria eradication campaign in the 1950s and 1960s, the global burden of the disease increased between the 1970s and the 1990s due to a combination of fewer control measures and the development of resistance to antimalarial drugs and insecticides. 3,4 After the start of the Roll Back Malaria Initiative in 1998, the global malaria burden decreased again, thanks to the large-scale introduction of interventions such as artemisinin-based combination therapy and insecticide-treated bed nets. 2,3 This was associated with a reduction in malaria transmission intensity and the malaria burden in sub-Saharan Africa. 3,5 Although surveys indicate that insecticide-treated nets have reduced the malaria burden and all-cause child mortality in Africa, 6 most recent reports of success in the region have come from islands, the peripheries of endemic areas and smaller countries with substantial external support. 3,5,7 The reported malaria burden has changed little in highly endemic African countries despite increased coverage of insecticide-treated nets. [7][8][9][10][11][12][13][14][15][16] Moreover, few data are available from countries in sub-Saharan Africa in which the malaria burden is highest. 2 The aim of our study was to evaluate the effect of the first national campaign designed to achieve universal coverage of insecticide-treated bed nets, which took place in 2010, in a high malaria transmission area of Burkina Faso. In particular, we investigated the effect of the campaign on the prevalence of parasitaemia in young children, childhood morbidity and child mortality.

Methods
The study took place in Nouna health district in rural, northwestern Burkina Faso, where the annual rainfall varies between 500 and 1000 mm. 17,18 The district's population was 311 000 in 2010 and almost exclusively comprised subsistence farmers. By 2013, it was served by 33 first-line health facilities and one hospital in the district capital Nouna. The Centre de Recherche en Santé de Nouna runs a health and demographic surveillance system (HDSS) site that covers Nouna town and 58 of the 300 villages in Nouna health district (Fig. 1). 19 In the past, malaria was hyperendemic or holoendemic in the area and most cases occurred during and shortly after the rainy season, from July to November. 20 The disease was also the principle cause of death in young children in the district. [21][22][23] Malaria control interventions were introduced in 2002 and, by 2009, malaria was predominantly mesoendemic in the area. 18,24 By early 2010, 59% of households owned at least one insecticidetreated bed net, 67% of pregnant women received intermittent preventive treatment at least once and 34% of children with malaria younger than 5 years received an artemisinin-based combination therapy. 25 After the national, insecticide-treated net campaign in 2010, 99% of households had at least one net. 26 Moreover, 66% of young children in the Centre de Recherche Objective To investigate if the first national insecticide-treated bed-net campaign in Burkina Faso, done in 2010, was followed by a decrease in childhood malaria in a district with high baseline transmission of the disease. Methods We obtained data on the prevalence of Plasmodium falciparum parasitaemia in children aged 2 weeks to 36 months from malaria surveys in 2009 and 2011. We assessed morbidity in children younger than 5 years by comparing data from the Nouna health district's health management information system before and after the campaign in 2010. We analysed mortality data from 2008 to 2012 from Nouna's health and demographic surveillance system. Findings The bed-net campaign was associated with an increase in the reported use of insecticide-treated nets. In 2009, 73% (630/869) of children reportedly slept under nets. In 2011, 92% (449/487) did. The campaign had no effect on the proportion of young children with P. falciparum parasitaemia after the rainy season; 52% (442/858) in 2009 and 53% (263/499) in 2011. Cases of malaria increased markedly after the campaign, as did the number of children presenting with other diseases. The campaign was not associated with any changes in child mortality. Conclusion The 2010 insecticide-treated net campaign in Burkina Faso was not associated with a decrease in care-seeking for malaria or all-cause mortality in children younger than 5 years. The most likely explanation is the high coverage of nets in the study area before the campaign which could have had an effect on mosquito vectors, limiting the campaign's impact.

Study data
The effect of the insecticide-treated net campaign on the prevalence of Plasmodium falciparum parasitaemia in the blood of young children was assessed from malaria surveys: (i) two cross-sectional surveys of children aged 0.5 to 36 months in eight villages were conducted in June and December 2011, after the campaign; and (ii) two identical surveys were conducted in the two villages in the same months in 2009, before the campaign, as part of a research project comparing child health data from 1999 and 2009. 18 For both the 2009 and the 2011 survey, 460 children were randomly selected from the Nouna HDSS register for the June survey and followed up in the December survey. The sample size enabled a reduction in malaria parasitaemia prevalence of 15% to be detected with a power of 80% at a significance level of 0.05.
Associations of the campaign with changes in childhood morbidity were assessed by comparing data from Nouna health district's health management information system for the two years before the campaign (i.e. 2008 and 2009) with data for the two years after the campaign (i.e. 2011 and 2012). For the health management information system, health workers at local health centres sent paper versions of summary statistics on specific diseases to district health teams, who entered data into an electronic database, which were forwarded to regional and national health offices. In this study, we used data from Nouna health district's database on uncomplicated and severe cases of malaria and, for comparison, on cases of other childhood diseases in children younger than 5 years.
Changes in child mortality were assessed using data for the period 1993 to 2012 from the Nouna HDSS, whose data collection procedures have been described elsewhere. 19 In brief, fieldworkers visited every household about three times per year and recorded all new births, deaths and migrations and other demographic and health variables.

Results
Data on the prevalence of P. falciparum parasitaemia from the four crosssectional surveys conducted in children aged 0.5 to 36 months are shown in Table 1. Although the reported use of insecticide-treated nets during the night before the survey increased from 73% (630/869) in 2009 to 92% (449/487) in 2011, overall, the prevalence of parasitaemia was similar in the two years, at around 52% (442/858 and 263/499).
Data on morbidity in infants and in children younger than 5 years in Nouna health district are shown in Fig. 2 and Table 2. Over the period, 243 057 consultations were reported for all children: 82 116 for infants and 160 941 for children aged 1 to 4 years. Malaria was the leading diagnosis and accounted for more than half of all consultations (154 919); about 10% (13 618) of all malaria cases were categorized as severe. The next leading causes of illness were respiratory disease (59 070), diarrhoea (17 196) and malnutrition (11 684). Both the total number of consultations and the number of consultations for leading causes of illness at health facilities increased continuously from 2008 until 2012. During this period, the number of cases of malaria (both uncomplicated and severe), respiratory disease and diarrhoea roughly doubled in all age groups, whereas the number of malnutrition cases increased by a factor of about 10. The number of uncomplicated malaria cases increased more in older children than infants. The number of measles cases also increased over the study period, but not consistently. There was no clear trend in the low number of asthma cases.
Overall, there was a substantial decline in all-cause child mortality between 1993 and 2012, 21,23 but there was no decrease in the period of interest between 2008 and 2012 (Fig. 3). Moreover, between 2008 and 2012 mortality showed a characteristic seasonal pattern, with no change in the pattern after the insecticide-treated net campaign (Fig. 4). During the period, mortality remained relatively constant in children aged 1 to 4 years but, in infants, there was a slight increasing trend between 2009 and 2012 (Fig. 5). Mortality was significantly lower in Nouna town than the rural neighbouring villages and after 2009 the rural area showed an increasing trend in mortality (Fig. 6).  Notes: The graph shows cases in children younger than 5 years. Asthma and measles cases are not shown due to low numbers. Data source: Health management information system of Nouna health district, Burkina Faso.

Research
Malaria control, Burkina Faso Valérie R Louis et al.

Discussion
The recent scaling-up of malaria control interventions and increases in international funding have coincided with many reports and publications suggesting that the epidemiology of malaria in sub-Saharan Africa has undergone a major change. However, most of the success stories come from areas with low malaria transmission intensity. 3,5,7,28,29 Moreover, the relationship between cause and effect has not always been clear and substantial changes in malaria incidence and prevalence have often preceded expanded coverage by an intervention. 14,[30][31][32][33][34][35] The main finding of our study is that the national insecticide-treated net campaign, which resulted in very high household ownership and use of nets, had no clear effect on the burden of malaria in young children in a highly endemic area. We found that the campaign had not changed the prevalence of malaria parasitaemia; there was even an increase in the annual number of cases. Moreover, there was no substantial change in all-cause child mortality. Recent reports indicate that malaria control programmes in several countries in sub-Saharan Africa with a high malaria burden have also failed to observe any benefits. [7][8][9][10][11][12][13][14][15][16] In addition, our study found a significant difference in child mortality between urban and rural populations in Burkina Faso, which is in agreement with similar findings in other countries in sub-Saharan Africa. 3 A few studies have reported on changes in the prevalence of parasitaemia with insecticide-treated net campaigns. One systematic review found that net use was associated with only a small reduction of around 10%. 36 In highly endemic, mainland Equatorial Guinea, a moderate increase in net use by young children was followed by a slight reduction in the malaria parasite prevalence. 37 A large controlled study of insecticide-treated net campaigns in Nigeria found lower parasite prevalence in intervention areas, but the results were not conclusive. 38 In highly endemic areas of western Kenya, the introduction of nets was initially associated with a decrease in parasite prevalence, which increased rapidly thereafter. 39 In our study area in Nouna, surveys indicated that the prevalence of the malaria parasite in village children declined markedly from 1999 to 2009, when insecticide-treated net coverage  was increasing moderately, 18 whereas the prevalence was unaffected by the 2010 insecticide-treated net campaign. These findings were confirmed by an in-depth longitudinal study in one village. 40 Most studies on the effect of insecticide-treated net campaigns are based on health management information system data. In Kenya, one study of paediatric inpatients in 17 hospitals showed a mixed pattern of disease incidence over a 10-year period during which coverage with malaria control interventions, including insecticidetreated nets, increased. 8 Another study done in the same time period showed reductions in malaria admissions to six hospitals but increases in two. 14 A study from Malawi found some increase in paediatric malaria admissions to four hospitals between 2000 and 2010 despite moderately increased use of insecticide-treated nets. 12 Similarly, a study at the main hospital in the country was unable to show any change in severe malaria admissions over this period. 10 In Togo, a mass distribution campaign of insecticide-treated nets was associated with a significant reduction in anaemia in children but also with a 159% increase in malaria incidence in one district. 9,41 In addition, a study from Uganda demonstrated that paediatric malaria admissions to five hospitals increased significantly from 1999 to 2009 when there was a massive roll-out of malaria control interventions, including insecticide-treated nets. 15 Moreover, Ugandan national health statistics showed that the malaria case burden increased from 3.5 million in 2000 to 12.2 million in 2008. However, only 33% of children were sleeping under nets in 2009. 15 In Zambia, the incidence of malaria increased in one district between 2006 and 2012 despite greater use of insecticide-treated nets. 13 In contrast, the comprehensive roll-out of malaria control interventions in Rwanda -a hypoendemic country -including an insecticide-treated net campaign, from 2005 was associated with a large reduction in malaria cases. 28 In our study area, the number of cases of uncomplicated or severe malaria increased after the insecticide-treated net campaign, as did cases of other diseases in young children. The likely explanations are an increase in the population and in parents' healthcare-seeking behaviour, better reporting, service coverage and service quality.
Insecticide-treated nets have been found in a systematic review to reduce all-cause child mortality by about 20% under controlled conditions. 36 Similar findings were reported in Rwanda and Togo where insecticide-treated net programmes were associated with significant reductions in child mortality. 9,28 However, the first studies to demonstrate the efficacy of insecticide-treated nets in sub-Saharan Africa compared coverage of around 80% in intervention groups with 0% in  control groups. 36 Also, in the two studies in Rwanda and Togo, coverage increased from very low to very high levels. 9,28 In our study, in contrast, coverage by insecticidetreated nets was already high before the 2010 campaign. Moreover, a decrease in child mortality has been documented in the Nouna study area over the last two decades. 19,21,23 We s u g g e s t f o u r p o t e n t i a l explanations why the insecticide-treated net campaign in Burkina Faso had no overall beneficial association with malaria burden or mortality in young children. First, the insecticide-treated nets could have been ineffective because of faults in the nets themselves or because the Anopheles vector developed resistance to pyrethroid or changed its behaviour. Second, compliance with the insecticidetreated net intervention could have been lower than reported. Third, the efficacy of the insecticide-treated net intervention in an area of high malaria transmission intensity could have been reduced by a saturation effect. Fourth, an unknown biological or social factor could have decreased efficacy.
Currently, there is no evidence that pyrethroid-resistance is an important problem in the study area. 42 Moreover, studies in sub-Saharan Africa indicate that there is no association between increased resistance of malaria vectors to pyrethroids and decreased efficacy of bed nets. 3,43,44 The 2010 national insecticidetreated net campaign in Burkina Faso seems to have been well implemented and annual surveys in Nouna health district show that net use has remained high. 45 However, changes in vector behaviour, as reported in other endemic regions, cannot be excluded. 46 The most likely explanation for the study's findings is the possibility of a saturation effect with insecticide-treated nets. It has been shown in several sub-Saharan Africa countries that high coverage with nets has a mass effect on the mosquito vector, such that community members who are not sleeping under insecticide-treated nets are also protected from malaria. 3 At present, the level of coverage needed to provide this mass effect is unknown and it may be that the threshold is far below universal coverage. This hypothesis is supported by mathematical modelling which shows that, when insecticidetreated net use in a community reaches 35% to 65%, individuals not using nets are protected at a similar level as those using them. 47 More research is needed on the relationship between insecticide-treated net coverage and these saturation effects in areas with different degrees of malaria endemicity.
Our study has strengths and limitations. One strength is that our findings are based on data from a wellestablished research centre with quality-controlled data collection and management procedures. Moreover, we used data from three different sources which increases the validity of our results. 48 One limitation is that since the data came partly from the area covered by HDSS and partly from the whole of Nouna health district, the quality of data could have varied. Furthermore, as with any observational study, known or unknown confounders could have influenced the findings, particularly as they were based on health management information system data. For example, increased coverage by a community-based health insurance scheme in Nouna health district could have had an influence.
However, during the study period, only a small proportion of households participated in the scheme. 49 Although a few new health facilities have opened in the district in recent years, their presence does not explain the large increase in consultations over a short period of time. Moreover, there was no major variation in the rainfall pattern over the study period (range: 566-763 mm/year). The large increase in the reported number of malnutrition cases may have been due to continuous improvements in primary health-care services in the district in recent years, which have included the expansion of malnutrition treatment services. There were no changes to health management information system procedures during the study period.
In conclusion, the 2010 insecticidetreated net campaign in Burkina Faso was not associated with a decrease in malaria burden or all-cause mortality in children younger than 5 years. This was most likely due to the high baseline coverage of insecticide-treated nets in the study area. Additional tools are needed to reduce the burden of malaria in hightransmission settings. ■