This study was conducted in September 2010 prior to planned mass distribution campaigns in Abia and Plateau States, which took place in August 2012 and December 2010, respectively. The results document high levels of Plasmodium infection and anemia in both states, extremely low (<1%) IRS coverage and low bed net ownership and use. Low IRS coverage across the sampled population is not unexpected, as IRS in Nigeria is limited to ‘target’ areas including: densely populated municipalities, areas with short malaria transmission seasons, areas where LLINs are difficult to implement, and institutional locations [5].
Around the same time (October 2010), the first national malaria indicator survey (MIS) was conducted throughout Nigeria in order to evaluate the scale-up of malaria prevention and control measures [1]. While the MIS provides national and zone-level estimates for interventions, state-level evaluation is also critical as mass net distribution campaigns are done on a state-by-state basis. In addition, since state campaigns have been conducted in different years, many of the 2010 MIS aggregate zonal estimates combine data from states that had already completed mass LLIN distribution with others that had not. As far as we are aware, this study is the first to report baseline estimates of malaria prevention measures, malaria prevalence and anemia in individual Nigerian states prior to scaled–up mass distribution campaigns targeting universal coverage.
Overall age-adjusted Plasmodium prevalence by microscopy (all ages) was similar between Abia (36.1%) and Plateau (36.6%) with almost one third of infections in Abia state being P.malariae. These represent some of the only modern estimates of Plasmodium prevalence across all age groups in Nigeria, as recent surveys including 2010 MIS tend to focus on specific sub-populations like children [1, 19, 20], neonates [21, 22], pregnant women [23], or those infected with HIV [24–26]. Prevalence estimates for children under five years by microscopy were similar for Plateau (43.5%, 95% CI: 36.6%–50.7%) compared to the 2010 MIS estimate for the larger area of the North Central zone in which it is located (49.4%, CI not reported), but likely different in Abia (42.0%, 95% CI: 35.7%–48.6%) compared to the 2010 MIS South East zone estimate (27.6%, CI not reported) [1]. One recent study conducted among all aged individuals during the dry season in Lagos State, South West zone, estimated an overall prevalence of 14.7%, with prevalence highest in the 5–14 year age group [27]. This trend with age is consistent with our results, where prevalence was highest in the 5–9 and 10–14 age groups in both Abia and Plateau.
A high level of concordance was observed in Plasmodium prevalence between microscopy and RDT, though RDT-estimates were slightly lower than microscopy. This contrasts with observations from other large surveys that consistently observe higher RDT-prevalence attributed to antigen persistence following treatment or submicroscopic infections [28]. Differences between this study and MIS 2010 results in RDT-prevalence estimates for children under five years were less pronounced than for microscopy results. However, unlike our study, which utilized a Pf/Pan combination RDT, the MIS 2010 utilized Paracheck PF®, an RDT that only detects P. falciparum-specific histidine-rich protein-2. MIS RDT results thus likely underestimated the overall Plasmodium prevalence in some areas through undiagnosed non-falciparum infections, as nearly one third of malaria infections in Abia in the present study were identified as P. malariae by microscopy. A significant proportion of non-falciparum infections were also identified by microscopy in South East and North Central zones in the MIS 2010 [1]. Historically, a significant proportion of P. malariae and P. ovale infections were also reported by The Garki Project, in Kano State, North West zone, from 1969 to 1976 [29]. Taken together, these results demonstrate that non-falciparum infections are prevalent in parts of Nigeria and highlight the importance of utilizing multi-species RDTs to monitor trends of all Plasmodium parasites. In addition to variation in prevalence between species, our study highlights large heterogeneity in prevalence between clusters within states that deserves further investigation to improve malaria risk stratification of all species in Nigeria.
In this study, more than half of children less than 11 years in both states were found to be anemic (mild, moderate or severe), with prevalence higher in Abia than in Plateau and also higher among children less than 5 years. Our results are consistent with WHO’s estimate that two-thirds of preschool-age children in Africa are anemic [30], and within Nigeria, are similar to those from the 2010 MIS, which found 71.7% anemia prevalence in South East zone and 56.0% in North Central zone among children under five [1]. Malaria is a major cause of childhood anemia in malaria endemic areas where it accounts for approximately half of pediatric admissions for severe anemia [31, 32]. Given the similar malaria prevalence between the two states, it is not immediately clear why anemia was significantly higher in Abia. Perhaps the higher prevalence of P. malariae or the slightly longer malaria transmission season may contribute. Other causes of anemia include iron and other nutritional deficiencies, blood disorders, inflammation and other acute and chronic diseases [16]. Thus differences in diet and genetic composition may also contribute to higher anemia in Abia. However, we hypothesize that repeated statewide MDA in Plateau, but not Abia, of deworming drugs from 2003–2012 for the elimination of lymphatic filariasis, as well as since 2008 for treatment of schistosomiasis in school-age children [33, 34], may have reduced the prevalence of helminth infections that have been shown to interact with malaria infection to worsen anemia [35]. Indeed, a recent survey of school-aged children has confirmed a higher prevalence of hookworm infection in Abia compared to Plateau (D. Evans, personal communication).
Prior to 2009, Nigeria’s policy was to provide free net distribution to children under five and pregnant women (vulnerable groups) only. As part of Nigeria’s aim to reduce by 50% malaria-related morbidity and mortality by 2013, the country embarked in 2009 on a strategy of scaled-up mass distribution of universal coverage with free long-lasting insecticidal net (LLINs) across the 36 states and Federal Capital Territory. The new policy goal is to reach at least 80% of households with an average of two nets per household but the delivery of nets for these scaled-up distributions, supported by The Global Fund and other donors, took place over a five year period (2009-2013) on a state-by-state basis.
The net ownership figures estimated here for both states in 2010 are much lower than the current ministry target, reflecting previous policy. In order to place our state-level results in the context of previous net distribution strategy and coverage estimates, we reviewed results from the DHS 2003 [9] the study of Oresanya et al[36], the DHS 2008 [3], and the MIS 2010 [1] that reported zonal level estimates. In the South East zone (a group of five states including Abia, Figure 1), household net ownership of at least one net of any type was 5.8% in 2003, reportedly increased after scale up to 36.5% in 2005, decreased to 13.4% in 2008 and was up to 35% in MIS 2010. Our estimate of ownership for Abia state only in 2010 (10.1%) was surprisingly low, given that many more than 10% of households would have a vulnerable group member. It could be explained by 1) lack of net replacement since 2005, although we note that half of the nets in Abia in the present study were less than one year old; or 2) inter-state differences between Abia and other states within the South East zone, perhaps mainly reflecting the mass distribution in Anambra State that took place in 2009. Similar review of net ownership in North Central zone (a group of six states including Plateau, Figure 1), showed it to be 14.9% in 2003 [9], 19.0% in 2005 [36], 15.9% in 2008 [3] and 32.7% in MIS 2010 [1]. The latter estimate was similar to results of the current study in Plateau State only (35.1%) and suggests a large increase in net ownership from 2008 to 2010. Despite these similar estimates for state and zone, intra-zonal differences between states also likely exist in North Central zone; for example scale up to universal coverage occurred in 2009 in Niger State and likely biased the zone estimate upwards. The higher ownership overall in Plateau may be partly due to efforts by The Carter Center to increase and maintain net ownership by integrating distribution with MDA for onchocerciasis and lymphatic filariasis [10]. As in Abia, approximately half of nets observed in Plateau were less than one year old. The wide variation between states in both baseline coverage and in past and future timing of scale up distribution highlights the importance of state-level surveys in evaluating the impact of Nigeria’s mass net distribution strategy.
In both Abia and Plateau, household members had taken the initiative to purchase about half of the nets currently owned, despite differences in wealth profiles between the two states. Around one-third of nets in both Abia (37.0%) and Plateau (30.3%) had been obtained through health facilities, although not all such nets were provided free-of-charge: 9.8% and 19.4% of nets obtained from health facilities in Abia and Plateau, respectively, were reportedly purchased. One third of all nets were obtained from markets or shops, indicating significant existing demand for nets prior to statewide mass distribution, as was also observed in Enugu State, South East zone [37]. Unlike the present results, other studies of net ownership in Nigeria have observed inequity prior to mass distribution campaigns, though with conflicting trends--some report highest ITN ownership among wealthiest households [3, 38], while an earlier report found inverse associations with wealth [9]. It will be important to document whether demand for nets translates into sustained net use in Nigeria once the access to free nets increases, as studies from other African countries have revealed declines in net use among households owning nets following mass distribution campaigns [39, 40].
Net use estimates follow similar trends to ownership. Overall net use in 2010 estimated in this study for children under five years and pregnant women in Abia (6.0%; 5.7%, respectively) and Plateau (19.1%; 21.0%) was far below ministry target of 80% for both populations. Past trends in the South East zone for net use by children under five in all households show it was 4.4% in 2003 [9], 16.0% in 2005 [36], 14.3% in 2008 [3] and 17.4% in MIS 2010 [1]. For pregnant women in South East (not assessed in 2005) the corresponding figures were 2% in 2003, 10.2% in 2008 and 12% in MIS 2010. These indicate substantial heterogeneity in net use within the South East zone and that Abia was lower than its zone average in 2010, although this is to be expected given the low net ownership. In North Central zone, trends in net use by under fives were fairly stable at 8.9% in 2003 [9], 7.3% in 2005 [36] and 9.7% in 2008 [3] but doubled to 18.9% by MIS 2010 [1]. Pregnant women showed a similar trend at 9.2% in 2003 and 9.4% in 2008 but greater increase by MIS 2010 to 36.7%. Thus Plateau state was above average in net use by children under five (19.1%) and below average for pregnant women (21.0%) compared to its surrounding zone.
Among households that owned nets, net use by children under five and pregnant women was five-fold higher in Abia and 2.5- to 3-fold higher in Plateau compared to all households; however, only about one third (in Abia) and one half (in Plateau) of vulnerable groups reported sleeping under a net the previous night. Yet 61.3% of nets in Abia and 80.4% of nets in Plateau were reportedly used by a household member last night, indicating that nets are being used by persons other than children under five and pregnant women within households, especially in Abia. Use by other members is not surprising given that the mean number of individuals per household in each state exceeds by a factor of 3.7 the number of nets currently available per household. Analysis of the early scale-up of malaria prevention measures across sub-Saharan Africa has shown that the primary driver of net use is the relative availability of nets within households [41], and recent application of additional MERG-recommended net indicators to 2010 MIS data demonstrates that 61% and 71% of households with an ITN in South East and North Central zones, respectively, did not have enough nets for each household member (defined as one ITN per two persons) [18]. Using the same indicator, we found that 86% and 82% of households with a net (of any type) in Abia and Plateau, respectively, did not have sufficient number of nets.
Previous studies [38, 41] have observed that net use among children is not significantly associated with household wealth after net distributions. In the present study, which was conducted prior to mass distributions, net use was positively associated with wealth in Plateau, but not in Abia. Interestingly, DHS 2008 and MIS 2010 both reported inverse associations between wealth and net use at the national level [1, 3]. Also in contrast to other studies, which reveal a female bias in net use in Nigeria [1, 42], significant differences between the sexes were not observed in either state in the present study.
Age was significantly associated with net use among all households in Abia and Plateau, which was highest in children under five and those over 20 years. This is in agreement with other surveys from Nigeria [1, 38, 42], which consistently observe that net usage is lowest among older children and young adults. This finding is very important given that older children are the group with highest prevalence of Plasmodium infection. In addition to improving access to nets, this points to a significant need for education regarding malaria prevention and net use, including by children over five years old. That this is needed even among those who own nets is illustrated by the finding that ‘not wanting to use net’ was the most common reason for nets not being hung last night and that 25% of nets in Abia had never been used.
In an effort to address these gaps, The Carter Center has developed behavior change communication (BCC) materials that emphasize strategies for increasing net use that were identified among consistent net users during focus group discussions conducted in Plateau State. In addition, BCC materials incorporate health education about lymphatic filariasis and malaria. This innovative, integrated health messaging approach was driven by the fact that both diseases share the same Anopheles vector and the belief that heightened awareness of LF-associated sequelae, which include swelling of the limbs (lymphedema, elephantiasis) and genital organs (hydrocele), is likely to promote increased net usage, particularly among adolescents and males.
As with any survey, there are limitations to note. Results from this study represent a single cross-sectional sample, which was collected during peak malaria season. We compared results with the 2010 MIS survey, which was conducted approximately one month after our survey. However, the DHS surveys of 2003 and 2008 were conducted during the months of March to August and June to October, respectively, which overlap periods of typically lower malaria transmission. Care should thus be taken when comparing our results with the DHS, particularly malaria parasite prevalence estimates, as well as utilization of malaria prevention measures, since net use has been observed to decline during dry seasons [43–45]. Studies of this type are also reliant upon self-reported data for many questions. In an effort to verify net ownership and ever-use of nets, survey teams visually inspected nets within households and observed whether the net was still sealed in its original packaging. However, it was not possible to verify use of net the previous night or other self-reported data. The survey was also conducted by independent groups of survey teams in each state, and unidentified sources of systematic error between teams may have biased state level estimates and the inferred differences between states. Likewise, slides from Abia and Plateau were read in separate laboratories. Although quality control was conducted by the same individual for slides from both states, systematic differences in initial slide reading between states could have occurred. Nonetheless, RDT data closely matched the overall microscopy prevalence estimates for each state, suggesting that gross errors between states, and overall, did not exist.