LLIN Evaluation in Uganda Project (LLINEUP): factors associated with ownership and use of long-lasting insecticidal nets in Uganda: a cross-sectional survey of 48 districts

Over the past decade, impressive reductions in malaria cases and deaths have been documented worldwide, following substantial investment in malaria control [1]. Between 2000 and 2015, an estimated 663 million clinical cases of malaria were averted by malaria control interventions; nearly 70% of these were attributed to use of long-lasting insecticidal nets (LLINs) [2]. However, malaria control achievements have not been uniform, and recent evidence suggests that the global decline in malaria burden may have stalled [1]. In Uganda, expanded coverage of malaria control interventions has been associated with reduced malaria incidence and prevalence [3–5], but malaria control gains have been fragile in high transmission areas and the burden of malaria in Uganda remains high [1, 6, 7].

Long-lasting insecticidal nets have been shown to reduce malaria morbidity and mortality across a range of epidemiological settings, and are the most widely used vector control tool [8, 9]. To achieve and maintain universal coverage with LLINs, defined as universal access to and use of LLINs by populations at risk of malaria, the World Health Organization (WHO) recommends the distribution of one LLIN for every two persons at risk of malaria through mass campaigns, conducted every 3 years [10]. Between May 2013 and August 2014, Uganda conducted its first Universal Coverage Campaign (UCC), to distribute LLINs free-of-charge nationwide. The aim of the UCC was to deliver at least one LLIN for every 2 residents to over 90% of households [11]. This campaign was led by a national coordination committee chaired by the Uganda Ministry of Health supported by key implementing partners. LLINs were distributed locally by village health teams (VHTs). Uganda’s last Malaria Indicator Survey, conducted in 2014–15 following the UCC campaign, found that 90% of households owned at least one LLIN, while 69% of residents had slept under an LLIN the previous night [3]. However, more recent data indicate that Uganda suffers from a LLIN ownership gap of over 40%, representing the proportion of households that own at least one net, but not enough nets for every two occupants [1]. Furthermore, results from a comprehensive surveillance programme suggest that although the UCC successfully increased LLIN coverage levels in Uganda, the effect on clinical malaria indicators was limited [4]. These findings raise concerns about attrition and use of LLINs [12, 13], and the potential impact of pyrethroid resistance [4, 14, 15]. To better understand patterns of LLIN ownership, coverage, and use, a cross-sectional community survey in 48 districts in Eastern and Western Uganda was conducted. This is the first large-scale survey of LLIN coverage in Uganda since the 2014–15 Malaria Indicator Survey, and will serve as the baseline for an ongoing cluster-randomized trial to evaluate the impact of LLINs with, and without, piperonyl butoxide (PBO) on parasite prevalence in community children aged 2–10 years (ISRCTN 17516395).

LLINs are the cornerstone of vector control, but maintaining high coverage is a challenge. To better understand factors associated with LLIN ownership, coverage, and use, a cross-sectional survey in 48 districts across Uganda was conducted 29–53 months (2.4–4.4 years) after a national campaign aiming to achieve universal coverage. LLIN ownership, adequate coverage, and use of LLINs have all reduced markedly in Uganda since the 2014–15 Malaria Indicator Survey, suggesting that the Ugandan population is no longer adequately protected by LLINs. Time since the last UCC and household wealth were strongly associated with all outcome measures. Households that had received LLINs via the last UCC more recently, and wealthier households, were more likely to own and use LLINs, and to be adequately covered, as were smaller households with fewer residents. School-aged children were less likely to use nets, particularly in households without adequate coverage, as were residents who were more distantly related to the head of household. These findings highlight the important issue of net attrition. Strategies to minimize attrition and to target at-risk groups through continuous net distribution should be considered. National mass distribution campaigns may also need to be carried out more frequently.

To achieve and maintain universal LLIN coverage, the WHO recommends that countries distribute nets free-of-charge through universal coverage campaigns, supplemented by continuous distribution through different channels, including antenatal clinics [23]. LLINs are assumed to remain effective for at least 3 years under field conditions [24, 25], and the WHO recommends that mass campaigns be repeated every 3 years [10]. However, a growing body of evidence calls the 3-year lifespan of LLINs into question [26–30]. Although the WHO recognizes that ‘coverage gaps can start to appear almost immediately post-campaign due to net deterioration, loss of nets, and population growth’, the most recent guidelines on achieving and maintaining universal coverage with LLINs do not address strategies to minimize attrition [10]. Instead, emphasis is placed on the need to monitor national coverage and durability of LLINs [10]. Studies of net durability focus on survivorship (the proportion of nets distributed which remain in the household), physical integrity (presence of net damage, including the number and size of holes), and bio-efficacy (tested using WHO cone bioassays to assess mosquito knock-down and mortality after exposure). Survivorship of LLINs has been shown to range from 75% at 24 months after distribution in Rwanda [27], to 65% at 2–4 years of follow-up in Tanzania [28], to only 35% at 12 months post-distribution in Zambia [29]. Studies from Ethiopia and Madagascar suggest that physical deterioration of nets begins soon after distribution, with approximately half of LLINs assessed showing some damage by 6 months [26, 30]. In Ethiopia, 30% of nets were classified as unusable due to poor condition at 26–30 months [26], and in Tanzania, 39% of nets were considered unserviceable at 2–4 years [28]. In Zambia, only 56% of nets met criteria for functional survival (nets present and classified as ‘good’ or ‘damaged’) at 30 months giving an estimated mean survival time of 2.5–3 years [29], while in Rwanda, considering survivorship plus integrity, the lifespan of LLINs was estimated to be closer to 2 years [27]. Bio-efficacy results are mixed, but suggest that insecticidal effectiveness of LLINs may also be less than 3 years. In Tanzania, 63.6% of nets tested met the WHO cone assay criteria (caused more than 95% knock-down and/or > 80% mortality in pyrethroid susceptible Anopheles gambiae sensu stricto) 2–4 years after distribution [28], while in Zambia only 34.2% of LLINs met these criteria at 24 months (tested using susceptible An. gambiae) [29], and in Madagascar, average mortality was low at 12 months (ranging from 6.9 to 25.9% in susceptible Anopheles arabiensis) [30]. The net attrition documented in this study adds to this evidence, highlighting the need to re-evaluate the effective lifespan of LLINs under field conditions and to develop strategies to minimize attrition. Recognizing the operational challenges of delivering nets nationwide, the frequency of mass distribution campaigns and the approach to replacing nets through continuous distribution channels should also be reconsidered.

In this study, poorer households were less likely to own and use LLINs. The link between malaria and poverty is well-described [31–34], and multiple studies have demonstrated an association between poverty and poor LLIN coverage and use. Mass distribution of free LLINs has been strongly advocated to maximize coverage and ensure equitable access to LLINs [35]. However, evidence on whether distribution of free nets reduces inequity in protection by malaria control interventions, and risk of malaria infection, is limited. A recent study evaluated the change in equity in ownership of LLINs in 19 sub-Saharan African countries between 2003 and 2014, by comparing baseline and endline Demographic and Health Survey and Malaria Indicator Survey data. This study concluded that equity of net ownership had improved in 13 countries (including Uganda), was unchanged in 2 countries (Mali and Mozambique), and had worsened in 4 countries (favouring poorer households in Madagascar and Senegal; favouring wealthier households in Angola and Niger) [36]. Pooled multi-country analyses suggested that the significant increase in LLIN ownership favoured the poorest households in most countries. These results are encouraging, suggesting that mass distribution campaigns can reach poorer households. However, whether increased LLIN ownership translates into higher net use, and lower malaria risk, in poorer households is less clear. An analysis of Malaria Indictor Survey data collected in 2006–2009 in Angola, Tanzania and Uganda, aiming to evaluate the effectiveness of targeted distribution of free bed nets in achieving equity, found mixed results [37]. Wealth was strongly associated with household net ownership, as well as net use and malaria test positivity in children under-five. Distribution of free nets narrowed the gap in net ownership between poorer and wealthier households, but wealthier households were still more likely to own nets in all three countries, and use of nets by children in poorer households was not improved by distribution of free nets. In all three countries, children from poorer households were more likely to test positive for malaria than their wealthier counterparts [37]. These results suggest that household wealth is strongly associated with net ownership and use in Uganda, and that poorer households might be at risk for higher net attrition. Effective strategies to improve LLIN coverage and use among the poor, especially in between national distribution campaigns need to be identified.

In this study, school-aged children were less likely to use LLINs than children under-five or older residents. Typically, malaria control efforts have focused on children under-five (and pregnant women) because they bear the brunt of malaria morbidity and mortality. In endemic areas, risk of clinical disease and death declines throughout childhood due to the gradual acquisition of immunity gained through repeated infection [38]. By adolescence, most malaria infections are asymptomatic, although pregnancy again places women at increased risk. While older children generally experience fewer clinical malaria episodes in high transmission areas, the burden of malaria in schoolchildren is not insignificant, and has been associated with reduced school attendance, cognition, learning and school performance [39, 40]. Moreover, parasite prevalence is often highest in school-aged children, who serve as reservoirs of infection for the onward transmission of malaria [41–43]. Children in this ‘neglected’ age group are often less well-covered by LLINs [11, 42, 44], a finding supported by the results of this study. School-aged children suffer consequences of malaria and pose an important public health risk for transmission to other community members. This age group should be targeted with interventions aiming to improve LLIN coverage and use. School interventions are an attractive option given that schools are organized and well-distributed geographically, and thus provide access to a large proportion of this target group. Furthermore, children are considered change agents; targeting them can potentially lead to improved LLIN use within the household [45].

This study had several limitations. First, self-report was used to measure net use, which could have under-estimated or over-estimated the actual use of LLINs. Although nets were observed by field team members during the day-time surveys, night-time net use was not observed. Although this is the standard approach to measuring LLIN use [46], self-reported measures have been shown to overestimate LLIN adherence by 13.6% as compared to objective measures [47], suggesting that the true proportion of residents who slept under a LLIN the previous night could be lower than our estimates. Second, the ability to understand why individuals choose to use nets or not, is limited by the quantitative nature of the questionnaire. Further exploration using qualitative research methods would be required to better understand local perceptions and motivations for net use [12]. Third, although net attrition is an important issue, no data were collected on net quality or durability in this survey. However, as part of the LLINEUP project, net durability and bio-efficacy of LLINs will be assessed approximately 12 months after distribution via the 2017–18 national campaign in Uganda.

In Uganda, LLIN ownership, coverage, and use were all well-below desired targets 2.5–4.5 years after LLINs were distributed through a national campaign. Net attrition may compromise the effectiveness of LLINs, leaving a substantial proportion of the population unprotected. The lifespan of LLINs under field conditions, and the optimum frequency of national LLIN mass distribution campaigns, should be reassessed. Moreover, strategies for continuous distribution of LLINs to schoolchildren and communities should be further explored to increase LLIN coverage of vulnerable groups. The results of this study will serve as a baseline for the ongoing Uganda PBO Net study, a cluster-randomised trial designed to evaluate the impact of LLINs, including conventional LLINs, and new generation LLINs combining a pyrethroid insecticide with a synergist (piperonyl butoxide), distributed by the Ugandan Ministry of Health in 2017–2018.