Sub-National Targeting of Seasonal Malaria Chemoprevention in the Sahelian Countries of the Nouakchott Initiative

Abstract

Background: Seasonal malaria chemoprevention (SMC) has been shown to be highly efficacious against clinical malaria in areas where transmission is acutely seasonal. SMC targeting depends on a complex interplay of climate, malaria transmission and population distribution. In this study a spatial decision support framework was developed to identify health districts suitable for the targeting of SMC across seven Sahelian countries and northern states of Nigeria that are members of the Nouakchott Initiative.

Methods: A spatially explicit decision support framework that links information on seasonality, age-structured population, urbanization, malaria endemicity and the length of transmission season was developed to inform SMC targeting in health districts. Thresholds of seasonality, population and receptive risks were defined to delineate SMC suitable health districts and define the age range of children for targeting. Numbers of children were then computed for the period 2015-2020 in SMC districts. For 2015, this was combined with maps of length of malaria transmission seasons and WHO recommended treatment regimen to quantify the number of tablets required across the SMC health districts.

Results: A total of 597 Sahelian health districts were mapped, out of which 478 (80.1%) were considered suitable for SMC based on seasonality and endemicity thresholds. These districts had an estimated 119.8 million (85%) of the total population in 2015. In the six years from 2015-2020, it is estimated that a total of 158 million children 3m to <5 years, 121 million of whom were in rural areas, will need SMC to achieve universal coverage in the Sahel. If the upper age limit of SMC targeted children was increased to <10 years in low transmission districts, a total 177 million overall, of whom 135 million were rural children, will require chemoprevention in 2015-2020. In 2015 alone, an estimated 49-72 million SP tablets and 148-217 million AQ tablets will be needed to cover all or rural children respectively under the different scenarios of upper age limits.

Conclusions: Our proposed framework provides a standardised approach to support targeting and scale up of SMC by the countries of the Nouakchott Initiative. Our analysis suggests that the vast majority of the population in this region are likely to benefit from SMC and substantial resources will be required to reach universal coverage each year.

Fig 1. A spatial decision support framework for identifying areas suitable for seasonal chemoprevention and quantifying the size of the population of target children and the amount of the required antimalarial tablets.

For additional details of the definition of inputs, processes and outputs see the S1 File.

Fig 2. Map of Sahelian health districts (n = 597) showing A) the distribution of population in 2015 at 1 x 1 km spatial resolution [source: www.worldpop.org]; B) areas (orange) where 60% or more of the annual total rainfall occurs in any three consecutive months; C) the median number of malaria transmission months in an average year in seasonal health districts [source: Grover-Kopec et al 2006; D) population adjusted estimates of P. falciparum parasite rate standardized to the ages 2 to 10 years (PAPfPR_2-10) for the year 2000 [source: Noor et al 2014].

In Nigeria, health districts from the states of Bauchi, Borno, Jigawa, Kano, Katsina, Kebi, Sokoto, Yobe and Zamfara were included in the analysis.

Fig 3. Map of the Sahel showing the health districts that are not suitable for SMC targeting (grey n = 119) and SMC suitable health districts (light to dark brown, n = 478) classified by age class of target children.

The SMC suitable districts were those where PAPfPR_2-10 in 2000 was ≥5% and 80% of the population lived in areas where ≥60% or more of the annual total rainfall occurs in any three consecutive months. In SMC health districts where PAPfPR_2-10 in 2010 was ≥5% to 10% (n = 123) children 3 months to <10 years of age were targeted for SMC while those where it was >10% (n = 355) children 3 months to <5 years of age were targeted. *All inputs are either generated at or resampled to surfaces of 1 x 1 km spatial resolutions. A) Monthly Africa Rainfall Estimates version 2 (RFE 2.0) data from 2002–2009 at 10 × 10 km spatial resolution [NOAA 2013] were used to generate average long term monthly rainfall which are then used to define average seasonality (Section D in S1 File); B) Maps of total population are disaggregated by age structure (3 months to below 5 years; 5 years to below 10 years) using data from census and household surveys and by urban and rural using population density, night time lights and other land cover classifications (Section C in S1 File). Countries should use most recent census and survey data for population projections and age categorisations; C) For all countries except Niger and Mauritania PfPR_2-10 data from the period 1980–2012 were used to estimate endemicity from 2000 and 2010 (Section F in S1 File). D) A map based on the presumed relationship between P. falciparum transmission, temperature and rainfall to define the length of transmission seasons was downloaded as a grid surface from International Research Institute for Climate and Society website [IRI URL]. The map was at spatial resolution of approximately 50 x 50 km and was resampled to 1 x 1 km (Section E in S1 File); E) The approach by Cairns et al (2012) that identified acute malaria seasonality as areas where 60% or more of the annual total rainfall occurred in three consecutive months was used. This approach had a high sensitivity (95%) of areas where over 60% of malaria cases occurred in 4 consecutive months (Section D in S1 File,); F) Data from a variety of international and national sources were used to develop the most recent boundaries of health districts (Section B in S1 File). Due to population growth and changes in governance health districts change frequently and countries should continuously update these boundary changes. G-J) Health districts where ≥80% of population lived in areas of acute malaria seasonality and had 2000 PAPfPR_2-10 ≥ 5% were considered suitable for SMC (Section F in S1 File). This endemicity threshold allowed for the inclusion of areas where current risk is low but where receptive risk is still high. Population estimates by age class, urban and rural were extracted to each health district (S1 File); K) In districts where 2010 PfPR_2-10 was 5% to ≤10%, children aged 3 months to <10 years were targeted for SMC and 3 to months to <5 years in higher transmission districts (Section F in S1 File). Countries can update the contemporary description of risk using most recent survey data. A decision also needs to be made on whether or not to include urban areas. L) The median number of transmission months was extracted for each health district from the climate based map of length of transmission (Section E in S1 File) and was multiplied by the estimated number of SMC targeted children and the 1 SP and 3 AQ tablets per child per month (Section F in S1 File).