Skip to main content

Table 3 Summery of GIS applications and geographic factors associated with malaria and anaemia among children 0-5y of age living in low- or middle-income countries

From: A systematic review of the application and utility of geographical information systems for exploring disease-disease relationships in paediatric global health research: the case of anaemia and malaria

 Author (year) GIS application Geographic factors of malaria or anaemia
Anthony et al., (1992) [37] Dot map of malaria incidence in one of the study villages. Malaria point prevalence varied within and across villages in Indonesia. Incidence of malaria infections greatest in Yapimakot (39.1%), followed by Dabolding (34.95), Kabiding (31.9%) and Kutdol (28.6%). Prevalence of malaria ~50% lower among populations living in areas of forest-covered mountain slopes above the valley compared to villagers.
Giardina et al. (2012) Giardina et al. [28] Geospatial analysis, remote sensing data, and choropleth maps used to estimate environmental/climatic predictors of malaria. Prevalence of malaria varied across survey locations in Senegal (lowest in northern regions, highest in the sourthern regions). High geographical variation in parasitaemia prevalence, including urban (1.3%) vs. rural (8.47%) differences (reduced odds for urban areas by 81%, 95% BCI: 55%-93%).
Gordon (2004) [38] Choropleth and symbology maps used to depict worldwide prevalence estimates and other related geographic features such as climate suitability for vector transmission. Annual deaths from malaria in 2002 by WHO region highest in Africa (978,661) and lowest in Europe (44).
Hightower et al., (1998) [27] GIS used to perform spatial analyses and link location information to parasitology and entomology databases. Prevalence of parasitemia tended to decrease with increasing household distance from larval habitat (p = 0.3437) except during the dry month of September. Average number of trapped An. gambiae mosquitoes was related to the distance of the household to the nearest breeding site for September (p = 0.0039), but not wet month of June (p = 0.1530). Opposite relationship was found for An.
funestus(June p = 0.0191, September p = 0.6608).
MARA/ARMA (1998) MARA/ARMA [31] Various thematic maps used to depict relevant environmental (e.g. climatic) and population characteristics (e.g. density), and disease prevalence/incidence data. Childhood (0-4y) population exposed to malaria mortality risk was higher in areas with 50% malaria transmission stability than areas with 90%. In Kenya the number of children < 5y who die or develop clinical malaria varies across areas of high, medium, low, or unstable malaria endemicity. In Mali an inverse U-shaped association found between malaria prevalence and distance to a water source (total population estimate).
Mbogo (1993) [25] Vector map of study area. Prevalence of asymptomatic infections (with or without parasitaemia concentration ≥ 5000/uL) was higher in rural area of Sokoke compared to Kalifi town, Kenya. Higher proportion of children recruited from Sokoke reported to the District Hopsital with febrile illness and high parasitemia.
Mbogo (1995) [36] Vector map of study area. Spatial patterns of severe disease varied across study sites indpendently of transmission intensity and entomological innoculation rate (EIR).
Root (1999) Root [34] Choropleth maps to depict spatial patterns of <5 mortality in 20 sub-Saharan African countries. High mortality rates in East/South Africa and in vicinity of Lake Victoria represented heterogeneity in disease environments, indicating spatial impact and correlation between intensity of malaria transmission and observed mortality patterns.
Schellenberg et al., (1998) [26] Choropleth maps used to depict quintiles of severe malaria presenting to District Hospital and layout of all-weather roads. Admission rates significantly higher in children living within 5 km from hospital (31.6/1000 child-years at risk) compared to those > 25 km away (5.0 per 1000 child-years at risk). Children living > 2.5 km away from nearest road were significantly less likely to be admitted compared to those living < 0.5 km (Adj RR = 0.47, 95%CI: 0.3-0.9).
Snow (1998a) [32] Dot density map of projected population distribution according to modelled predictions of regions of stable malaria endemicity. High transmission intensity conditions identified around Lake Victoria (affecting 677,000 children < 5y). Largest number of children 0-4y exposed to areas of moderate stable malaria endemicity. Highest risk of malaria mortality and hospital admission in areas of high and moderate stable malaria endemicity, respectively.
Snow (1999a) [35] [30] Dot density map of population distribution from communities exposed to at least 50% probability of malaria transmission according to a fuzzy logic climate model. Wide geographical variation in estimates of malaria mortality in childhood. Deaths in hospital due to malaria per 1000 catchment childhood population highest in Sukutu, The Gambia (range 0.33-2.8) compared to other sites.
Snow (1999b) [30] [30] Thematic maps of climate suitability for stable transmission, interpolated population density, and zones of malaria risk in Africa. Higher median mortality and morbidity rates in areas of stable transmission with ≥ 0.2 climate suitability than malaria risk area in South Africa with ≥ 0.5 climate suitability.
WHO (2008b) [7] Choropleth maps of global incidence of malaria (and malaria related deaths) in 2006. Variation in estimated burden of malaria (cases and deaths) in 2006 among children < 5y within and across 30 high burden countries.
WHO (2010) [33] Choropleth maps of geographical distribution of confirmed malaria cases/1000 population. Variation in estimated malaria cases among children < 5y across 24 selected countries between 2000–2009.
Mainardi ( 2012) [21] Spatial distribution of anaemia prevalence, including comparison between countries, and association with urabanizaiton. Geographical variation in average proporiton of children with moderate or severe anaemia. Localization/urbanization was inversely associated with moderate and severe anaemia (OLS). Increased median time to a water source was significantly associated with lower prevalence of moderate (p < 0.01), but not severe anaemia (GWR). Widespread anaemia prevalence observed in mainly inland regions in West Africa, and a few specific areas in Eastern and central Africa.
Snow 1994 [22] Vector map of study areas in Kenya and Tanzania. Higher prevalence of parasitaemia among children 0-4y in Ifakara compared to Kilifi. Higher prevalence of severe anaemia among children 0-4y in Kilifi than Ifakara.
WHO (2008a) [2] Choropleth maps of global anaemia prevalence and public health significance by country. Prevalence of anaemia among pre-school aged children (0.5-4.99y) highest in Africa (global range 23.1 to 67.6%).
Tanzanian NBS and ICF International 2012 [39] Choropleth maps of anaemia prevalence by region. Anaemia prevalence ranged from 42% among two in-land regions (Rukwa and Kilimanjaro) to 78% in the northern island region of Unguja.
Greenwell 2006 [29] Choropleth maps of anaemia or malaria prevalence, as well as malaria transmision by country (vector) or overall (raster). Overlayed dot density maps were used to show cluster locations. Children in areas of moderate malaria prevalence were at highest risk of severe anaemia. The validity of haemoglobin measurements was dependent on whether the assessment was conducted during a high malaria transmission season.
Magalhaes 2011 [24] Dot density map of anaemia prevalence by DHS location. Choropleth maps of predictive geogrpahical risk or variation of anaemia or Hb concentration. Mean haemoglobin was lowest in Burkina Faso, and a large spatial cluster of low mean haemoglobin and high anaemia risk was predicted for an area shared by Burkina Faso and Mali.