Malaria Journal
BioMed Central
Open Access
Research
Spatial risk profiling of Plasmodium falciparum parasitaemia in a high
endemicity area in Côte d'Ivoire
Giovanna Raso*†1,2,3, Kigbafori D Silué†1,4, Penelope Vounatsou3,
Burton H Singer5, Ahoua Yapi4, Marcel Tanner3, Jürg Utzinger3 and
Eliézer K N'Goran1,4
Address: 1Département Environnement et Santé, Centre Suisse de Recherches Scientifiques, Abidjan, Côte d'Ivoire, 2School of Population Health,
University of Queensland, Brisbane, Australia, 3Department of Public Health and Epidemiology, Swiss Tropical Institute, Basel, Switzerland, 4UFR
Biosciences, Université de Cocody-Abidjan, Abidjan, Côte d'Ivoire and 5Office of Population Research, Princeton University, Princeton, USA
Email: Giovanna Raso* - giovanna.raso@gmail.com; Kigbafori D Silué - kigbafori.silue@csrs.ci;
Penelope Vounatsou - penelope.vounatsou@unibas.ch; Burton H Singer - singer@princeton.edu; Ahoua Yapi - yapiah@yahoo.fr;
Marcel Tanner - marcel.tanner@unibas.ch; Jürg Utzinger - juerg.utzinger@unibas.ch; Eliézer K N'Goran - eliezerngoran@yahoo.fr
* Corresponding author †Equal contributors
Published: 11 November 2009
Malaria Journal 2009, 8:252
doi:10.1186/1475-2875-8-252
Received: 3 August 2009
Accepted: 11 November 2009
This article is available from: http://www.malariajournal.com/content/8/1/252
© 2009 Raso et al; licensee BioMed Central Ltd.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0),
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
Background: The objective of this study was to identify demographic, environmental and
socioeconomic risk factors and spatial patterns of Plasmodium falciparum parasitaemia in a high
endemicity area of Africa, and to specify how this information can facilitate improved malaria
control at the district level.
Methods: A questionnaire was administered to about 4,000 schoolchildren in 55 schools in
western Côte d'Ivoire to determine children's socioeconomic status and their habit of sleeping
under bed nets. Environmental data were obtained from satellite images, digitized ground maps and
a second questionnaire addressed to school directors. Finger prick blood samples were collected
and P. falciparum parasitaemia determined under a microscope using standardized, qualitycontrolled methods. Bayesian variogram models were utilized for spatial risk modelling and
mapping of P. falciparum parasitaemia at non-sampled locations, assuming stationary and nonstationary underlying spatial dependence.
Results: Two-thirds of the schoolchildren were infected with P. falciparum and the mean
parasitaemia among infected children was 959 parasites/μl of blood. Age, socioeconomic status, not
sleeping under a bed net, coverage rate with bed nets and environmental factors (e.g., normalized
difference vegetation index, rainfall, land surface temperature and living in close proximity to
standing water) were significantly associated with the risk of P. falciparum parasitaemia. After
accounting for spatial correlation, age, bed net coverage, rainfall during the main malaria
transmission season and distance to rivers remained significant covariates.
Conclusion: It is argued that a massive increase in bed net coverage, particularly in villages in close
proximity to rivers, in concert with other control measures, is necessary to bring malaria
endemicity down to intermediate or low levels.
Page 1 of 16
(page number not for citation purposes)
Malaria Journal 2009, 8:252
Background
Malaria remains one of the most pressing public health
and poverty-related issue in the developing world, particularly in sub-Saharan Africa [1]. Each year, malaria might
claim the lives of >1 million individuals. There are >500
million episodes of clinical Plasmodium falciparum malaria
and the global burden might exceed 40 million disabilityadjusted life years (DALYs) [2-4]. Mortality, morbidity
and economic losses due to malaria could be reduced significantly if effective measures, such as sleeping under
long-lasting insecticidal nets (LLINs) and access to
prompt diagnosis and effective treatment using artemisinin-based combination therapy (ACT) were made available to all those in need [5]. Interventions aiming at the
control and local elimination of malaria require reliable
risk maps in order to enhance the efficacy and cost-effectiveness of control measures. Since parasitaemia is correlated with clinical manifestations of malaria [6],
parasitaemia risk maps are a useful tool for the spatial targeting of control interventions. Ongoing blood sampling
at the household level on a broad scale is expensive and
not practical for surveillance purposes. District-level planning and targeting would be greatly facilitated by rapid
and non-invasive identification of high-risk zones.
Over the past decade, geographical information system
(GIS) and remote sensing technologies have been widely
used for mapping malaria [1,7,8]. However, purely GIS
and remote sensing approaches have a number of shortcomings, due to their inability to quantify the relation
between environmental factors and malaria risk and, consequently, infer predictions from statistical models [9].
Furthermore, classical statistical models have been widely
employed to evaluate the relationship between disease
risk and demographic, environmental and socioeconomic
factors, assuming independence of spatially-explicit data
[10-12]. Since disease data cluster in space, the assumption of independence is violated, and hence the statistical
significance of the model covariates often overestimated
[13]. It follows that predictive risk models lack accuracy.
In recent work by the authors, Bayesian non-stationary
geostatistical models were employed for spatial risk profiling of malaria [9,14,15]. The strengths of these models are
their accountancy for spatial dependence in the data, and
the assumption of non-stationary spatial processes. The
use of non-stationary models is further justified on the
ground that local characteristics related to human behaviour and environment, including vector ecology, depend
on location. Consequently, assuming stationarity may
provide unreliable results when analyzing spatiallyexplicit disease data.
Here, risk factors and spatial patterns of P. falciparum parasitaemia among school-aged children in a high endemic-
http://www.malariajournal.com/content/8/1/252
ity setting of western Côte d'Ivoire are elucidated. An
integrated approach, using GIS and remotely-sensed environmental data, questionnaire and parasitological survey
data and Bayesian geostatistical models was employed.
Finally, the use of non-stationary models for risk profiling
of P. falciparum parasitaemia at a regional scale was
explored. The identified risk factors can help district
health planners to implement malaria control interventions in a spatially-explicit manner, followed by monitoring and surveillance so that control tools can be finetuned over time to enhance their performance [16].
Methods
Study area and population
This study was carried out in the region of Man, a mountainous area in the western part of Côte d'Ivoire, which is
highly endemic for P. falciparum malaria, as well as
helminth infections [17-21]. Climate conditions are tropical with rains occurring from March to October with
highest precipitation observed in July and August. The dry
season extends from November to February.
The present study was carried out between October 2001
and February 2002. Schoolchildren from 57 rural schools
attending grades 3-5 were invited for finger prick blood
samples and two questionnaires were administered, one
addressed to schoolchildren and the second one to school
directors.
Ethical clearance
The study protocol was approved by the institutional
research commissions of the Swiss Tropical Institute
(Basel, Switzerland) and the Centre Suisse de Recherches
Scientifiques (Abidjan, Côte d'Ivoire). Ethical clearance
was obtained by the Ministry of Health in Côte d'Ivoire.
Cross-sectional surveys
Thin and thick blood films were prepared from finger
prick blood samples on microscope slides, air-dried and
transferred to a laboratory in the town of Man. Slides were
stained with 10% Giemsa and examined under a light
microscope by experienced laboratory technicians. The
number of Plasmodium spp. parasites was counted by
assuming a standard white blood cell count of 8,000/μl of
blood.
The schoolchildren questionnaire was used to obtain
information about assets on ownership and household
characteristics (total of 12 indicators), and perceived
symptoms and diseases (total of 17 morbidity indicators).
In addition, children were asked whether they slept under
a bed net and whether they were living in the village of the
school or in a nearby village or hamlet. An asset-based
approach was used to stratify schoolchildren into five
socio-economic groups [19].
Page 2 of 16
(page number not for citation purposes)
Malaria Journal 2009, 8:252
http://www.malariajournal.com/content/8/1/252
The questionnaire addressed to school directors included
three main topics, i.e., (i) village demographics, (ii)
health issues and (iii) local environment (e.g., presence of
swamps, irrigation fields and pasture nearby the village
and the estimated distances). In case the school directors
felt they were not sufficiently acquainted to respond to
these questions, they were invited to consult with other
village authorities.
were utilized to generate a smooth map of P. falciparum
parasitaemia using Bayesian kriging [25].
Environmental data
Geographical coordinates for each school were collected
using a hand-held Magellan 320 global positioning system (GPS; Thales Navigation, Santa Clara, CA, USA). Distance to rivers was calculated from digitized ground maps.
Normalized difference vegetation index (NDVI) and land
surface temperature (LST) were downloaded at 1 × 1 km
spatial resolution from Moderate Resolution Imaging
Spectroradiometer (MODIS) from USGS EROS Data Centre. Rainfall estimate (RFE) data with an 8 × 8 km spatial
resolution from Meteosat 7 satellite were obtained from
the Africa Data Dissemination Service (ADDS). NDVI, LST
and RFE were downloaded for the period of September
2001 to August 2002 and processed as detailed elsewhere
[22]. A digital elevation model (DEM) was employed originating from the Shuttle Radar Topography Mission
(SRTM) to delineate watersheds [23].
over-dispersion (extra variation) r. The covariates Xij and
Data management and analysis
Data were entered twice and cross-checked. Geographical
data were displayed in ArcView GIS version 3.2 (Environmental Systems Research Institute, Inc., Redlands, CA,
USA). Schoolchildren were subdivided into two age
groups: (i) 6-10 years and (ii) 11-16 years. Bed net coverage was calculated as the percentage of schoolchildren
who reported sleeping under a bed net at the unit of the
school.
All demographic, environmental and socioeconomic covariates were fitted into negative binomial regression models on the P. falciparum parasitaemia data, using STATA
version 9.0 (Stata Corporation, College Station, TX, USA).
Covariates with a significance level <0.15 were built into
three different spatial models for P. falciparum parasitaemia using WinBUGS version 1.4 (Imperial College &
Medical Research Council, London, UK). The models
were (i) a stationary Bayesian negative binomial regression model, and (ii) two non-stationary Bayesian negative
binomial regression models. To take into account the spatial heterogeneity, location-specific random effects were
integrated in the logistic models, assuming that they are
distributed according to a multivariate normal distribution with variance-covariance matrix related to the variogram of the spatial process. Markov chain Monte Carlo
(MCMC) simulation was employed to estimate the model
parameters [24]. Model covariates from the final model
Model specification
To model P. falciparum parasitaemia, let Zij be the P. falci-
parum parasite count in blood films of schoolchild j in village i. It was assumed that Zij arises from a negative
binomial distribution, Zij~Nb(μij, r) with mean μij and
village-specific random effect ϕi were modeled with
log(μij) as the outcome, that is log(μij) = X ijT β + ϕi, where
β is the vector of regression coefficients. The spatial correlation was introduced on the ϕi's by assuming that ϕ =
(ϕ1, ϕ2, ..., ϕN)T has a multivariate normal distribution, β
~MVN(0, Σ), with variance-covariance matrix Σ. Moreover, an isotropic spatial process was assumed, i.e., Σmn = σ2
exp(-udmn), where dmn is the Euclidean distance between
village m and village n, σ2 is the geographic variability
known as the sill, and u is a smoothing parameter that
controls the rate of correlation decay with increasing distance. To take into account non-stationarity, the study
area was partitioned in K subregions and a local stationary
spatial process ω k was assumed in each subregion k = 1,
..., K. One type of model included ecological subregions,
i.e., watersheds of rivers, whereas the other type included
fixed subregions, i.e., the study area was subdivided into
two subregions on a diagonal from the north-western corner to the south-eastern corner of the study area. Spatial
correlation in the study area was viewed as a mixture of
the different spatial processes. The spatial random effect ϕi
at location i was modeled as a weighted average of the
subregion-specific (independent) stationary processes as
follows: φi =
K
∑1 aikω ki ,
with weights aik, which are
k=
decreasing functions of the distance between location i
and the centroids of the subregions k [26]. Assuming ωk
~MVN(0, Σk), (∑ k )ij = σ k2corr(d ij ; u k ) , it follows that
K
φ = N(0, ∑ A kT ∑ k A k ) , with Ak = diag{a1k, a2k, ..., ank}.
k =1
The range is defined as the minimum distance at which
spatial correlation between locations is below 5%. It can
be calculated as 3 u and is expressed in meters.
k
Page 3 of 16
(page number not for citation purposes)
Malaria Journal 2009, 8:252
Model implementation
Following a Bayesian model specification, prior distributions for the model parameters were adopted. Vague Nor-
mal distributions for the β parameters with large variances
(i.e., 10,000), gamma prior for r with large variance,
inverse gamma priors for σ k2 and uniform priors for uk, k
= 1, ..., K were chosen. MCMC simulation was employed
to estimate the model parameters [24]. A single chain
sampler with a burn-in of 5,000 iterations was run. Convergence was assessed by inspection of ergodic averages of
selected model parameters. Covariates from the binomial
regression models were used to generate a smooth risk
map for P. falciparum parasitaemia using Bayesian kriging
[25].
Model performance and predictive ability
The deviance information criterion (DIC) was utilized to
assess the model performance [27]. Additionally, a twostage approach was adapted for assessment of model performance based on the predictive ability. First, a training
sample from the current database was utilized by fitting
individual-level data from 43 randomly selected schools
into the negative binomial regression models. The individual-level data from the remaining schools were utilized
for prediction purposes. 95%, 75%, 50%, 25% and 1-5%
Bayesian credible intervals (BCIs) of the posterior predictive distribution of test individuals were calculated. The
model with the highest percentage of correctly predicted
individual parasitaemia within the interval with the
smallest coverage was considered as the best predicting
one. Second, the predictive ability of the models was
assessed using a Bayesian p-value analogue calculated
from the predictive posterior distribution, recently presented by Gosoniu and colleagues [9]. The Bayesian pvalue
is
calculated
as
1000
1 / 1000∑ j =1 min(I(p irep( j) > p iobs ), I(p irep( j) < p iobs )) . I(·)
denotes the number of points fulfilling the specific condition in the argument, piobs is the observed parasitaemia of
an individual and pirep = pirep(1),..., pirep(1000) are 1,000 replicated data from the predictive distribution for a test individual. When the median of the predictive posterior
distribution is close to 0.5, the model predicts the
observed data well. The model with median p-values closest to 0.5 is considered the best performing one.
Results
Study cohort
A total of 3,962 schoolchildren had complete data
records, i.e., were individually interviewed and had P. fal-
http://www.malariajournal.com/content/8/1/252
ciparum parasitaemia results from blood film examination. There were 2,340 boys (59.1%) and 1,622 girls
(40.9%). With regard to age, 1,684 children (42.5%) were
between 6 and 10 years, whereas 2,278 children (57.5%)
were aged 11-16 years.
Plasmodium falciparum parasitaemia
Almost two out of three children were infected with P. falciparum (64.9%). Other Plasmodium species were rare:
Plasmodium malariae and Plasmodium ovale infections were
found in 117 (3.0%) and 7 children (0.2%), respectively.
All subsequent analyses focus on P. falciparum. At the unit
of the school, the prevalence of P. falciparum ranged from
34.0% to 91.9%.
Among P. falciparum-infected children, the mean parasitaemia was 959 parasites/μl of blood. Whilst approximately a third of the children had no P. falciparum
infection as determined by light microscopy, a third of the
children had a P. falciparum parasitaemia <500 parasites/
μl of blood (37.9%), and one-fourth had a parasitaemia
ranging between 500 and 5,000 parasites/μl of blood
(25.1%). Only 72 (1.8%) of the children had a parasite
count >5,000 parasites/μl of blood. At the unit of the
school, the mean P. falciparum parasitaemia ranged from
63 to 2,178 parasites/μl of blood.
Risk profiling and spatial patterns
Results of the bivariate non-spatial analyses are shown in
Table 1. Children aged 6-10 years were at a significantly
higher risk of having a high P. falciparum parasitaemia
than their older peers. Sex was not significantly associated
with P. falciparum parasitaemia. Children from the fourth
and fifth quintile (the less poor and least poor) were at a
higher risk of having higher parasitaemia levels compared
to the poorest schoolchildren. Other significant risk factors included not sleeping under a bed net, bed net coverage at the unit of school, NDVI, RFE, LST, close proximity
to standing water (rivers, swamps and irrigated fields) and
absence of pasture near villages. There was no significant
association between P. falciparum parasitaemia and distance to the closest health care facility. Finally, no significant association was found between P. falciparum
parasitaemia and the children's place of residence (living
in the same village as the school or in a nearby village or
hamlet).
The mean P. falciparum parasitaemia at the unit of the
school is shown in Figure 1. Three schools in the northeastern part of the study area and one in the central part
had a mean parasitaemia >1,500 parasites/μl of blood.
Spatial analyses and model performance
Results of the spatial analyses are summarized in Table 2.
Children's age, bed net coverage and mean RFE during the
Page 4 of 16
(page number not for citation purposes)
Malaria Journal 2009, 8:252
http://www.malariajournal.com/content/8/1/252
Table 1: Results of the bivariate negative binomial regression models for P. falciparum parasitaemia among 3,962 schoolchildren from
55 rural schools of western Côte d'Ivoire.
Source of data
School registry
Questionnaire addressed to schoolchildren
Health district registry
Questionnaire addressed to school directors
Satellite images
Digitized ground maps
Indicator
Plasmodium falciparum parasitaemia
Age
6-10 years
11-16 years
Socioeconomic status
Most poor
Very poor
Poor
Less poor
Least poor
Sleeping under a bed net
Bed net coverage
<25%
≥ 25%
Living in the same village as school
Distance to health care facility
Presence of swamps for rice cultivation
Distance to irrigated field
<500 m
500-999 m
≥ 1000 m
Presence of pasture
NDVI
Mean I§
Mean II¶
Mean III||
Annual mean
Mean of the transmission season
Annual mean NDVI (categorized)
<0.65
0.65-0.70
>0.70
RFE
Mean Id
Mean IIe
Mean IIIf
Sum of annual rainfall
Mean of the transmission seasong
Maximum LST
Distance to rivers
Distance to rivers (categorized)
<500 m
500-999 m
≥ 1000 m
IRRa
95% CIb
P-value (AICc)
1.00
0.70
0.60, 0.83
<0.001
1.00
1.02
0.93
1.34
1.34
0.75
0.79, 1.32
0.72, 1.20
1.04, 1.73
1.04, 1.73
0.58, 0.98
0.005
0.040
1.00
0.38
0.86
1.00
0.85
0.28, 0.53
0.70, 1.06
0.93, 1.08
0.72, 1.01
<0.001
0.162
0.916
0.124
1.00
0.55
0.86
0.77
0.32, 0.96
0.59, 1.22
0.62, 0.96
0.117
0.022
1.03
1.06
0.95
1.00
1.14
0.94, 1.12
0.97, 1.15
0.88, 1.03
0.92, 1.08
1.05, 1.23
0.532 (46,255)
0.177 (46,253)
0.205 (46,254)
0.950 (46,255)
0.001 (46,245)
1.00
1.66
1.31
1.30, 2.12
1.05, 1.64
<0.001 (46,240)
1.06
1.01
1.11
1.17
1.22
1.09
0.85
0.98, 1.15
0.94, 1.09
1.02, 1.20
1.08, 1.26
1.13, 1.32
1.00, 1.18
0.78, 0.92
0.122 (46,253)
0.742 (46,255)
0.015 (46,249)
<0.001 (46,238)
<0.001 (46,230)
0.048
<0.001 (46,241)
1.00
0.98
0.62
0.80, 1.21
0.51, 0.76
< 0.001 (46,236)
aIRR: incidence-rate ratio; bCI: confidence interval; cAIC: Akaike information criterion; dMean I: mean value during the month prior to blood sample
collection; eMean II: mean value during the month of collection and the previous month; fMean III: mean value during the month of collection and
the two previous months; gThe transmission season is during June to August
main malaria transmission season (June to August) were
significant covariates in the stationary negative binomial
regression model. The three covariates were also found
significant in the non-stationary model with ecological
subregions. In contrast, mean RFE during the transmission season was not significant in the negative binomial
regression model with fixed subregions. In the latter
model, age, bed net coverage and distance to rivers were
significant covariates. There was a clear over-dispersion (r
= 0.16) of the data. The range where spatial correlation is
below 5% was 1.9 km in the stationary model. For the
non-stationary model with ecological subregions, the
ranges were 2.3 km, 1.9 km and 2.1 km, respectively. For
the non-stationary model with fixed subregions, the
Page 5 of 16
(page number not for citation purposes)
Malaria Journal 2009, 8:252
http://www.malariajournal.com/content/8/1/252
Figure
Mean
d'Ivoire
P. 1
during
falciparum
the parasitaemia
school year 2001/2002
among 3,962 schoolchildren from 55 sampled schools in the region of Man, western Côte
Mean P. falciparum parasitaemia among 3,962 schoolchildren from 55 sampled schools in the region of Man,
western Côte d'Ivoire during the school year 2001/2002. The normalized difference vegetation index (NDVI) is displayed in the background.
ranges were 1.9 km and 2.3 km. Geographical variability
differed depending on the subregion in the non-stationary model with ecological subregions.
For the assessment of the model performance, the spatial
models without the covariates bed net coverage, presence
of swamps, distance to irrigated fields and presence of pasture were used, since no information was available for prediction. Results of the spatial analyses of those models are
shown in Table 3. The differences between DICs for the
three models were only marginal, and hence the results
suggest that the stationary and the non-stationary models
performed similarly. Table 4 summarizes the results of the
models' predictive ability using different BCIs. Virtually
no difference was found between the stationary and the
non-stationary models, although the latter type of models
seemed to perform slightly better at the smallest BCIs. The
p-values calculated from the predictive distribution of the
1,034 selected individuals for model validation revealed
similar distributions for all three models, including medi-
ans, suggesting that the models had the same predictive
ability.
Risk mapping
Figures 2, 3, 4, 5, 6 and 7 display the results from the three
stationary and non-stationary P. falciparum parasitaemia
models. The maps were based on models without the covariates bed net coverage, presence of swamps, distance to
irrigated fields and presence of pasture, as this information was missing for prediction. There is a clear difference
in the parasitaemia predictions between stationary and
non-stationary models. In the non-stationary map
inferred from the non-stationary model with ecological
subregions, the predicted parasitaemia was considerably
higher in the north-eastern part of the study area compared to the maps derived from the stationary model and
the non-stationary model with fixed subregions. However, the standard deviations of the predicted parasitaemia inferred from the non-stationary model with
ecological subregions show that in this area the prediction
Page 6 of 16
(page number not for citation purposes)
Malaria Journal 2009, 8:252
http://www.malariajournal.com/content/8/1/252
Table 2: Multivariate stationary and non-stationary spatial analyses results for P. falciparum parasitaemia for the region of Man,
western Côte d'Ivoire.
Indicator
Bayesian negative binomial regression models
Stationary
Age (years)
6-10
11-16
Socioeconomic status
Most poor
Very poor
Poor
Less poor
Least poor
Sleeping under a bed
net
Bed net coverage (%)
< 25
≥ 25
Presence of swamps
for rice cultivation
Distance to irrigated
field (m)
< 500
500-999
≥ 1000
Presence of pasture
NDVI (categorized)
< 0.65
0.65-0.70
> 0.70
Mean rainfall during
transmission season
Maximum LST
Distance to rivers (m)
< 500
500-999
≥ 1000
rc
ρ1d
ρ2
ρ3
σ12e
σ 22
σ 32
DICf
Non-stationary with ecological
subregions
Non-stationary with fixed subregions
IRRa
95% BCIb
IRRa
95% BCIb
IRRa
95% BCIb
1.00
0.72
0.60, 0.85
1.00
0.71
0.60, 0.83
1.00
0.71
0.59, 0.83
1.00
1.02
0.97
1.14
1.08
0.86
0.77, 1.33
0.74, 1.26
0.85, 1.51
0.79, 1.44
0.63, 1.14
1.00
1.00
0.96
1.09
1.04
0.88
0.75, 1.29
0.73, 1.24
0.82, 1.43
0.76, 1.39
0.65, 1.17
1.00
1.01
0.97
1.10
1.03
0.87
0.77, 1.30
0.74, 1.25
0.83, 1.43
0.76, 1.37
0.65, 1.16
1.00
0.51
1.11
0.24, 0.98
0.71, 1.63
1.00
0.49
1.23
0.27, 0.82
0.80, 1.77
1.00
0.50
1.13
0.26, 0.90
0.73, 1.67
1.00
0.74
0.92
0.99
0.17, 2.23
0.35, 2.14
0.57, 1.59
1.00
0.66
0.95
0.88
0.23, 1.48
0.47, 1.74
0.57, 1.32
1.00
0.64
0.89
0.92
0.21, 1.59
0.39, 1.76
0.58, 1.43
1.00
1.42
0.76
1.28
0.81, 2.43
0.42, 1.33
1.00, 1.61
1.00
1.57
0.78
1.24
0.95, 2.40
0.46, 1.24
1.01, 1.51
1.00
1.61
0.78
1.20
0.97, 2.49
0.46, 1.24
0.95 1.54
0.93
0.75, 1.15
0.95
0.76, 1.15
0.93
0.72, 1.16
1.00
1.13
0.72
0.16
0.0016
0.68, 1.74
0.44, 1.09
0.16, 0.17
0.0005, 0.003
0.66, 1.56
0.42, 0.94
0.16, 0.17
0.0002, 0.002
0.0005, 0.003
0.17, 0.57
0.73, 1.56
0.46, 1.08
0.16, 0.17
0.0002, 0.002
0.0005, 0.003
0.0002, 0.002
0.02, 0.39
0.40, 3.08
0.03, 0.60
1.00
1.03
0.63
0.16
0.0013
0.0016
0.33
1.00
1.08
0.71
0.16
0.0014
0.0016
0.0013
0.14
1.20
0.21
46,044.9
0.13
0.58
0.02, 0.39
0.26, 1.16
46,047.4
46,045.3
aIRR: incidence-rate ratio; bBCI: Bayesian credible interval; cr: over-dispersion parameter; dρ: scalar parameter representing the rate of decline of
correlation with distance between points; eσ2: estimate of geographic variability; fDIC: deviance information criterion; a composite measure of how
well the model does, i.e. a compromise between fit and complexity, with smaller DICs indicating better performance of the model.
error is highest. Nonetheless, all three standard deviation
maps show increased standard errors in the north-eastern
part of the Man area in western Côte d'Ivoire.
Discussion
Current anti-malarial prophylaxis and treatment, and vector control using insecticides are susceptible to the emergence of resistant malarial parasites and vectors. Hence,
there is a pressing need for other interventions incorporated into the programme that can delay the onset of
Page 7 of 16
(page number not for citation purposes)
Malaria Journal 2009, 8:252
http://www.malariajournal.com/content/8/1/252
Table 3: Multivariate stationary and non-stationary spatial analyses results for P. falciparum parasitaemia for the region of Man,
western Côte d'Ivoire.
Indicator
Bayesian negative binomial regression models
Stationary
Age (years)
6-10
11-16
Socioeconomic status
Most poor
Very poor
Poor
Less poor
Least poor
Sleeping under a bed
net
NDVI (categorized)
<0.65
0.65-0.70
>0.70
Mean rainfall during
transmission season
Maximum LST
Distance to rivers (m)
<500
500-999
≥ 1000
rc
ρ1d
ρ2
ρ3
σ12e
σ 22
σ 32
DICf
Non-stationary with ecological subregions
Non-stationary with fixed subregions
IRRa
95% BCIb
IRRa
95% BCIb
IRRa
95% BCIb
1.00
0.72
0.60, 0.85
1.00
0.71
0.60, 0.83
1.00
0.71
0.59, 0.83
1.00
1.02
0.98
1.15
1.09
0.81
0.77, 1.33
0.74, 1.27
0.87, 1.52
0.80, 1.46
0.60, 1.07
1.00
1.00
0.96
1.11
1.07
0.81
0.75, 1.30
0.73, 1.25
0.84, 1.46
0.79, 1.42
0.61, 1.07
1.00
1.00
0.96
1.11
1.05
0.82
0.76, 1.23
0.73, 1.22
0.83, 1.44
0.77, 1.39
0.61, 1.08
1.00
1.44
0.77
1.27
0.84, 2.35
0.43, 1.29
1.03, 1.56
1.00
1.54
0.76
1.25
0.94, 2.36
0.46, 1.19
1.00, 1.54
1.00
1.69
0.78
1.20
1.04, 2.54
0.49, 1.21
0.96, 1.51
0.92
0.75, 1.12
0.91
0.75, 1.10
0.89
0.72, 1.09
1.00
1.19
0.73
0.16
0.0016
0.76, 1.80
0.46, 1.10
0.15, 0.17
0.0005, 0.002
0.98, 1.55
0.39, 0.92
0.15, 0.17
0.0003, 0.003
0.0006, 0.003
0.18, 0.54
0.73, 1.56
0.46, 1.08
0.15, 0.17
0.0002, 0.002
0.0005, 0.002
0.0002, 0.002
0.03, 0.45
0.37, 2.73
0.06, 0.68
1.00
1.01
0.61
0.16
0.0014
0.0016
0.33
1.00
1.08
0.71
0.16
0.0013
0.0016
0.0014
0.18
1.08
0.26
46,044.6
0.14
0.60
0.02, 0.37
0.27, 1.19
46,046.4
46,045
These models were used for calculation of model performance and prediction.
aIRR: incidence-rate ratio; bBCI: Bayesian credible interval; cr: over-dispersion parameter; dρ: scalar parameter representing the rate of decline of
correlation with distance between points; eσ2: estimate of geographic variability; fDIC: deviance information criterion; a composite measure of how
well the model does, i.e. a compromise between fit and complexity, with smaller DICs indicating better performance of the model.
resistance. There is also a need for new drugs and insecticides and a malaria vaccine, coupled with improved monitoring and surveillance [28]. Mapping areas where people
are at an elevated risk of infection and P. falciparum parasitaemia is important for the design and implementation
of district-based malaria control interventions.
Here an integrated approach for spatial risk profiling of P.
falciparum parasitaemia was used, building on previous
research pertaining to the mapping and prediction of
helminth infections in the Man region, western Côte
d'Ivoire [29]. Reasons why this approach is termed 'integrated' are as follows. First, a diversity of data (demographic, environmental and socioeconomic) was
obtained from different sources, including cross-sectional
questionnaire and epidemiological surveys and remote
sensing. Second, data covered different spatial scales. For
example, RFE, LST and NDVI data were collected by
remote sensing at a large spatial scale. At a small spatial
scale, data on proximity to standing water (e.g., swamps
and irrigated agricultural fields) were obtained from questionnaires addressed to school directors and from digitized maps. Third, the data were collated, stored and
managed using a GIS. Finally, Bayesian geostatistical
models were employed to produce smoothed risk maps of
P. falciparum parasitaemia, and to compare model outcomes assuming either stationary or non-stationary
dependence. Age, socioeconomic status, sleeping under a
bed net, bed net coverage and different environmental
factors - both small-scale (e.g., close proximity to standing
Page 8 of 16
(page number not for citation purposes)
Malaria Journal 2009, 8:252
http://www.malariajournal.com/content/8/1/252
Table 4: Percentage of test individuals with P. falciparum parasitaemia falling within selected Bayesian credible intervals (BCIs).
BCIs
95%
75%
50%
25%
5%
4%
3%
2%
1%
Bayesian negative binomial regression model
Stationary
Non-stationary (ecological subregions)
Non-stationary (fixed subregions)
99%
94%
63%
14%
2%
1%
1%
1%
0%
99%
93%
63%
13%
1%
1%
1%
1%
0%
99%
93%
64%
13%
2%
1%
1%
1%
0%
water) and large-scale (e.g., LST, NDVI and RFE) - were
significant risk factors for P. falciparum parasitaemia.
Interestingly, after introducing spatial correlation into the
regression analyses, age, bed net coverage and - depending
on the type of the model - mean RFE over the malaria
transmission season, and distance to rivers appeared to be
significant risk factors for P. falciparum parasitaemia.
Appraisal of model performance revealed no difference
when comparing stationary with non-stationary models.
However, the non-stationary model with ecological subregions showed that the geographical variability is different
between subregions.
Two shortcomings of the present study should be noted.
First, school-aged children are usually not the most
severely affected group with malaria in highly endemic
Figure
Smoothed
region of2 Man,
map western
of P. falciparum
Côte d'Ivoire
parasitaemia derived from a stationary negative binomial regression model using Bayesian in the
Smoothed map of P. falciparum parasitaemia derived from a stationary negative binomial regression model
using Bayesian in the region of Man, western Côte d'Ivoire.
Page 9 of 16
(page number not for citation purposes)
Malaria Journal 2009, 8:252
http://www.malariajournal.com/content/8/1/252
Figure
Standard
model using
3deviation
Bayesian
map
kriging
of theinpredicted
the regionP.of
falciparum
Man, western
parasitaemia
Côte d'Ivoire
derived from a stationary negative binomial regression
Standard deviation map of the predicted P. falciparum parasitaemia derived from a stationary negative binomial regression model using Bayesian kriging in the region of Man, western Côte d'Ivoire.
areas. Since the western part of Côte d'Ivoire is holoendemic for malaria [15,17,18,21], it is likely that schoolaged children have acquired some kind of immunity to
malarial parasites [30,31]. However, parasitaemia levels
in school-aged children might be higher than in younger
children. Underlying reasons are that school-aged children in high endemicity areas are mainly asymptomatic
carriers, they might be more exposed to mosquito bites
due to their behaviour, they are less likely to be treated
because of a lower incidence of clinical malaria, and
hence they might harbour considerably more parasites
than preschool-aged children. Second, due to the possibility of sequestration mechanisms of infected erythrocytes
from peripheral blood, as well as partially acquired
immunity, microscopic examination of only a single finger prick blood sample might have underestimated the
true prevalence of infection, and P. falciparum parasitaemia might have been slightly different [32-34].
sleeping under a bed net and bed net coverage) and a host
of environmental factors. As expected, children who
reported sleeping under a bed net were less likely to have
a high malaria parasitaemia as were children from schools
with a bed net coverage >25%. A study from rural Tanzania revealed that people from poorer households were less
likely to access preventive measures [35]. A similar result
has been reported for the population under study here
[19]. Based on these observations and the common belief
that the poorest population segments would share the
highest burden of malaria, the current results surprisingly
point in the opposite direction: schoolchildren from better-off households were more likely to have a higher parasitaemia than their poorer peers. This result is in
accordance with previous work focusing on spatial risk
profiles of P. falciparum prevalence in the same group of
children [15] and consequently warrants further investigation.
Notwithstanding these shortcomings, several risk factors
were found to be associated with P. falciparum parasitaemia, including demographic factors (e.g., age), socioeconomic factors, personal preventive measures (e.g.,
For the current mapping of P. falciparum parasitaemia, a
similar geostatistical approach was used as before when
modeling P. falciparum prevalence data [15] and common
helminth infections [22,23,36]. Importantly, the statisti-
Page 10 of 16
(page number not for citation purposes)
Malaria Journal 2009, 8:252
http://www.malariajournal.com/content/8/1/252
subregions
Figure
Smoothed
4 map
usingofBayesian
P. falciparum
kriging
parasitaemia
in the region
derived
of Man,
from
western
a non-stationary
Côte d'Ivoire
negative binomial regression model with ecologic
Smoothed map of P. falciparum parasitaemia derived from a non-stationary negative binomial regression
model with ecologic subregions using Bayesian kriging in the region of Man, western Côte d'Ivoire.
cal significance of several covariates changed once spatial
correlation had been taken into account. For example,
children's socioeconomic status, sleeping under a bed net
and several environmental factors - most notably LST,
NDVI, close proximity to standing water and presence of
pasture - were not significant anymore in the spatial models. This issue might be explained because omission of
spatial correlation, when analysing spatially-explicit data,
overestimates the significance of the regression coefficients [13]. In contrast to previous spatial analyses of P.
falciparum prevalence data, it was found that environmental factors such as rainfall during the main malaria transmission season and distance to the nearest permanent
river were significant predictors for P. falciparum parasitaemia. These environmental covariates are related to the
presence and abundance of malaria vectors, including
Anopheles gambiae and Anopheles funestus, which are the
key vector species as found in previous work in the nearby
forest and wet Savannah zones of Côte d'Ivoire [37,38]
and the medium-sized town of Man located in the centre
of the current study area [20]. As shown in a study from
Burkina Faso these vectors breed in small pools (An. gambiae) and larger semi-permanent water bodies (An. funes-
tus) [39]. In previous research pertaining to P. falciparum
prevalence data, most of the environmental factors
included had a large spatial scale and none of the environmental covariates was found significant [15]. Hence, it
was concluded that environmental data at a small spatial
scale are necessary for more precise spatial risk profiling at
the district level where decisions are usually made for the
control of malaria and other infectious diseases. Indeed,
including information obtained from interviewing the
directors of schools about the proximity of residential
houses to standing water revealed a number of significant
environmental covariates in the non-spatial analyses,
although there was a lack of statistical significance in the
spatial models. At a more local or regional scale, only distance to rivers, which was used as a proxy for standing
water, was significant in one of the spatially-explicit models. Further ground-based investigations are required,
since only data derived from questionnaires and digitized
maps were used rather than ecological surveys to explore
small-scale environmental features. It will also be interesting to determine the use of topography-derived wetness
indices, which have been linked to household malaria risk
at small spatial scale in two communities in the Kenyan
Page 11 of 16
(page number not for citation purposes)
Malaria Journal 2009, 8:252
http://www.malariajournal.com/content/8/1/252
Figure
Standard
model with
5deviation
ecologicmap
subregions
of the predicted
using Bayesian
P. falciparum
krigingparasitaemia
in the regionderived
of Man,from
western
a non-stationary
Côte d'Ivoirenegative binomial regression
Standard deviation map of the predicted P. falciparum parasitaemia derived from a non-stationary negative
binomial regression model with ecologic subregions using Bayesian kriging in the region of Man, western Côte
d'Ivoire.
highlands [40]. Perhaps somewhat surprising at first, the
present spatial analyses showed that RFE during the main
malaria transmission season, which is rather a broad scale
indicator, indicated the spatial heterogeneity of parasitaemia in the study area. This observation might be
explained by the distinct climatic conditions, i.e., higher
precipitation in the mountainous northern part of the
study area.
Comparing the performance of different models did not
reveal any significant difference in the predictive ability
between stationary and non-stationary models, and hence
the predicted parasitaemia risk maps were similar. Interestingly though, the non-stationary model with ecological
subregions predicted a slightly larger area with high parasitaemia in the north-eastern part of the study area. The
corresponding standard deviations of the map showed
that uncertainty was particularly high in this subregion. A
likely explanation of this observation is that there were
fewer sampled locations in that specific subregion (Figure
1). However, uncertainty in the north-eastern part of the
study area was also elevated (though to a lesser extent)
when employing a stationary and a non-stationary model
with fixed subregions. Of note, the spatial parameters in
the non-stationary model with ecologic subregions
revealed that geographic variability differed between subregions. Consequently, this would rule in favour of using
non-stationary models for predicting P. falciparum parasitaemia. Previous spatial analyses of P. falciparum prevalence in the same area revealed that non-stationary
models performed somewhat better than stationary models [15].
An important aspect of the current study is that the statistical model approach influences not only the spatial
parameter estimates, including the prediction maps and
standard deviations of the prediction, but also the significance of malaria risk indicators. Depending on the statistical model chosen, i.e., stationary or non-stationary, the
significance of several environmental factors changed. For
example, in the stationary and the non-stationary models
with ecological subregions, mean RFE during the main
malaria transmission season was significantly explaining
the geographical heterogeneity, whereas in the non-sta-
Page 12 of 16
(page number not for citation purposes)
Malaria Journal 2009, 8:252
http://www.malariajournal.com/content/8/1/252
Figure
regions
Smoothed
using
6 map
Bayesian
of P. falciparum
kriging inparasitaemia
the region ofderived
Man, western
from a non-stationary
Côte d'Ivoire negative binomial regression model with fixed subSmoothed map of P. falciparum parasitaemia derived from a non-stationary negative binomial regression
model with fixed subregions using Bayesian kriging in the region of Man, western Côte d'Ivoire.
tionary model with fixed subregions, this covariate was
not significant. Instead, distance to rivers appeared as a
significant covariate in the non-stationary model with
fixed subregions. Such differing results have also been
reported by others when comparing stationary and nonstationary models for the risk of malaria across Mali [9].
The covariate mean RFE during the main malaria transmission season had the lower BCIs near 1 in both stationary and non-stationary models with ecological
subregions, and the increase in odds due to increased rainfall was only 0.28 and 0.24, respectively. In contrast, the
non-stationary model with fixed subregions seems particularly promising, as the upper BCI for the covariate distance to rivers was not close to 1 and the parasitaemia risk
decreased by over a third with increasing distance from
rivers.
Employing a spatially-explicit risk profiling approach,
demographic, environmental and socioeconomic risk factors were identified that govern the geographic distribution of P. falciparum parasitaemia in a high endemicity
area at the district level. This information can be utilized
for designing and implementing malaria control interventions. In particular, at the time of the study in 2001/2002,
virtually no malaria control interventions were carried out
in the region of Man. The very low frequency of schoolchildren reported sleeping under a bed net (< 10%) documents this issue [19]. Although bed nets were available for
purchase from local dispensaries and the district hospital
in the town of Man, the price was perceived as too high. It
is speculated that the malaria situation in this region has
not improved, partially explained by an armed conflict
starting in September 2002 that also hit the region of Man
and resulted in a collapse of the health care delivery systems [41,42]. Available information supports this claim;
coverage of bed nets (ITNs) was reported below 5% in
Côte d'Ivoire at a national scale [43] and in the Man
region in particular [44]. The results further suggest that
health-seeking regarding prevention and treatment of
malaria at dispensaries was weak, as no statistical significance was found with regards to distance to a health post.
Conclusion
A massive scale-up of bed net coverage in the region of
Man, ideally promoting LLINs is indicated. Villages
located in the north-eastern part of the study area and
those in close proximity to rivers should be targeted first
to have the strongest impact. Once control interventions
Page 13 of 16
(page number not for citation purposes)
Malaria Journal 2009, 8:252
http://www.malariajournal.com/content/8/1/252
Figure
Standard
model with
7deviation
fixed subregions
of the predicted
using Bayesian
P. falciparum
kriging
parasitaemia
in the region
derived
of Man,
from
western
a non-stationary
Côte d'Ivoire
negative binomial regression
Standard deviation of the predicted P. falciparum parasitaemia derived from a non-stationary negative binomial regression model with fixed subregions using Bayesian kriging in the region of Man, western Côte d'Ivoire.
will start to take off, it is conceivable that the malaria situation will become more heterogeneous across the Man
region, and hence stationarity in modeling prevalence and
parasitaemia will no longer be justified, as control interventions are likely to vary depending on location. Future
field studies will elucidate whether the presented integrated risk profiling and control approach can also be
employed for rigorous monitoring and performance evaluations of the district-level malaria control programme.
List of abbreviations
ACT: Artemisinin-based Combination Therapy; ADDS:
Africa Data Dissemination Service; AIC: Akaike Information Criterion; BCI: Bayesian Credible Interval; CI: Confidence Interval; DALYs: Disability-Adjusted Life Years;
DEM: Digital Elevation Model; DIC: Deviance Information Criterion; GIS: Geographical Information system;
GPS: Global Positioning System; ITN: Insecticide-Treated
Nets; LLINs: Long-Lasting Insecticidal Nets; LST: Land Surface Temperature; MCMC: Markov Chain Monte Carlo;
MODIS: Moderate Resolution Imaging Spectroradiometer; NDVI: Normalized Difference Vegetation Index; RFE:
Rainfall Estimates; SRTM: Shuttle Radar Topography Mission.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
GR contributed to the conception and design, participated
in the data collection, carried out the spatial analyses and
interpretation of the data and drafted the manuscript.
KDS was involved in the data collection, quality control,
data analyses and drafting of the manuscript. PV contributed to the analysis of the data and drafting of the manuscript. BHS was involved in the interpretation of the data
and critical revision of the manuscript. AY was involved in
the acquisition of data. MT contributed to the conception
and design. JU contributed to the conception and design,
interpretation of the data and drafting of the manuscript.
EKN was involved in the conception and design as well as
the critical revision of the manuscript. All authors read
and approved the initial submission and the revised version of the manuscript.
Page 14 of 16
(page number not for citation purposes)
Malaria Journal 2009, 8:252
http://www.malariajournal.com/content/8/1/252
Acknowledgements
We thank the staff members of the health and sanitation district of the
region of Man, the education officers, directors, teachers, schoolchildren
and the field and laboratory technicians (M. Traoré, K.L. Lohourignon, B.A.
Sosthène, A. Allangba and S. Diabaté). This investigation received financial
support from the Swiss National Science Foundation to P. Vounatsou
(project no. 3252B0-102136/1) and J. Utzinger (project no. PPOOB-102883
and PPOOB-119129), and the University of Queensland to G. Raso through
a Postdoctoral Research Fellowship and Early Career Research Grant
(project no. 2007002086). E.K. N'Goran is grateful to Fairmed for financial
support.
18.
19.
20.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Hay SI, Guerra CA, Gething PW, Patil AP, Tatem AJ, Noor AM,
Kabaria CW, Manh BH, Elyazar IRF, Brooker S, Smith DL, Moyeed
RA, Snow RW: A world malaria map: Plasmodium falciparum
endemicity in 2007. PLoS Med 2009, 6:e1000048.
Sachs J, Malaney P: The economic and social burden of malaria.
Nature 2002, 415:680-685.
Snow RW, Guerra CA, Noor AM, Myint HY, Hay SI: The global distribution of clinical episodes of Plasmodium falciparum
malaria. Nature 2005, 434:214-217.
Lopez AD, Mathers CD, Ezzati M, Jamison DT, Murray CJL: Global
and regional burden of disease and risk factors, 2001: systematic analysis of population health data. Lancet 2006,
367:1747-1757.
Gomez Sambo L: Defining and defeating the intolerable burden of malaria III. Progress and perspectives. Am J Trop Med
Hyg 2007, 77(Suppl 6):iii.
Bouvier P, Rougemont A, Breslow N, Doumbo O, Delley V, Dicko A,
Diakite M, Mauris A, Robert CF: Seasonality and malaria in a
West African village: does high parasite density predict fever
incidence? Am J Epidemiol 1997, 145:850-857.
Craig MH, Snow RW, le Sueur D: A climate-based distribution
model of malaria transmission in sub-Saharan Africa. Parasitol
Today 1999, 15:105-111.
Graves PM, Osgood DE, Thomson MC, Sereke K, Araia A, Zerom M,
Ceccato P, Bell M, Del Corral J, Ghebreselassie S, Brantly EP, Ghebremeskel T: Effectiveness of malaria control during changing
climate conditions in Eritrea, 1998-2003. Trop Med Int Health
2008, 13:218-228.
Gosoniu L, Vounatsou P, Sogoba N, Smith T: Bayesian modelling
of geostatistical malaria risk data. Geospat Health 2006,
1:127-139.
Patz JA, Strzepek K, Lele S, Hedden M, Greene S, Noden B, Hay SI,
Kalkstein L, Beier JC: Predicting key malaria transmission factors, biting and entomological inoculation rates, using modelled soil moisture in Kenya. Trop Med Int Health 1998,
3:818-827.
Snow RW, Gouws E, Omumbo J, Rapuoda B, Craig MH, Tanser FC,
le Sueur D, Ouma J: Models to predict the intensity of Plasmodium falciparum transmission: applications to the burden of
disease in Kenya. Trans R Soc Trop Med Hyg 1998, 92:601-606.
Brooker S, Leslie T, Kolaczinski K, Mohsen E, Mehboob N, Saleheen
S, Khudonazarov J, Freeman T, Clements A, Rowland M, Kolaczinski
J: Spatial epidemiology of Plasmodium vivax, Afghanistan.
Emerg Infect Dis 2006, 12:1600-1602.
Cressie N: Regional mapping of incidence rates using spatial
Bayesian models. Med Care 1993, 31(Suppl 5):60-65.
Gosoniu L, Vounatsou P, Sogoba N, Maire N, Smith T: Mapping
malaria risk in West Africa using a Bayesian nonparametric
non-stationary model. Comput Stat Data Anal 2009, 53:3358-3371.
Silué KD, Raso G, Yapi A, Vounatsou P, Tanner M, N'Goran EK,
Utzinger J: Spatially-explicit risk profiling of Plasmodium falciparum infections at a small scale: a geostatistical modelling
approach. Malar J 2008, 7:111.
Keiser J, Singer BH, Utzinger J: Reducing the burden of malaria
in different eco-epidemiological settings with environmental
management: a systematic review. Lancet Infect Dis 2005,
5:695-708.
Utzinger J, N'Goran EK, N'Dri A, Lengeler C, Xiao SH, Tanner M:
Oral artemether for prevention of Schistosoma mansoni
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
infection: randomised controlled trial.
Lancet 2000,
355:1320-1325.
Raso G, Luginbühl A, Adjoua CA, Tian-Bi NT, Silue KD, Matthys B,
Vounatsou P, Wang Y, Dumas ME, Holmes E, Singer BH, Tanner M,
N'Goran EK, Utzinger J: Multiple parasite infections and their
relationship to self-reported morbidity indicators in a community of rural Côte d'Ivoire. Int J Epidemiol 2004, 33:1092-1102.
Raso G, Utzinger J, Silué KD, Ouattara M, Yapi A, Toty A, Matthys B,
Vounatsou P, Tanner M, N'Goran EK: Disparities in parasitic
infections, perceived ill health and access to health care
among poorer and less poor schoolchildren of rural Côte
d'Ivoire. Trop Med Int Health 2005, 10:42-57.
Matthys B, N'Goran EK, Kone M, Koudou BG, Vounatsou P, Cisse G,
Tschannen AB, Tanner M, Utzinger J: Urban agricultural land use
and characterization of mosquito larval habitats in a
medium-sized town of Côte d'Ivoire. J Vector Ecol 2006,
31:319-333.
Matthys B, Vounatsou P, Raso G, Tschannen AB, Becket EG, Gosoniu
L, Cisse G, Tanner M, N'Goran EK, Utzinger J: Urban farming and
malaria risk factors in a medium-sized town in Côte d'Ivoire.
Am J Trop Med Hyg 2006, 75:1223-1231.
Raso G, Matthys B, N'Goran EK, Tanner M, Vounatsou P, Utzinger J:
Spatial risk prediction and mapping of Schistosoma mansoni
infections among schoolchildren living in western Côte
d'Ivoire. Parasitology 2005, 131:97-108.
Beck-Wörner C, Raso G, Vounatsou P, N'Goran EK, Rigo G, Parlow
E, Utzinger J: Bayesian spatial risk prediction of Schistosoma
mansoni infection in western Côte d'Ivoire using a remotelysensed digital elevation model. Am J Trop Med Hyg 2007,
76:956-963.
Gelfand AE, Smith AFM: Sampling-based approaches to calculating marginal densities. J Am Stat Assoc 1990, 85:398-410.
Diggle PJ, Tawn JA, Moyeed RA: Model-based geostatistics. J Roy
Stat Soc C-App 1998, 47:299-326.
Banerjee S, Gelfand AE, Knight JR, Sirmans CF: Spatial modeling of
house prices using normalized distance-weighted sum of stationary processes. J Bus Econ Stat 2004, 22:206-213.
Spiegelhalter DJ, Best N, Charlin BP, Linde A van der: Bayesian
measures of model complexity and fit. J Roy Stat Soc, Ser B 2002,
64:583-639.
Greenwood BM, Fidock DA, Kyle DE, Kappe SH, Alonso PL, Collins
FH, Duffy PE: Malaria: progress, perils, and prospects for eradication. J Clin Invest 2008, 118:1266-1276.
Raso G, Vounatsou P, Singer BH, N'Goran EK, Tanner M, Utzinger J:
An integrated approach for risk profiling and spatial prediction of Schistosoma mansoni-hookworm coinfection. Proc Natl
Acad Sci USA 2006, 103:6934-6939.
Gupta S, Snow RW, Donnelly CA, Marsh K, Newbold C: Immunity
to non-cerebral severe malaria is acquired after one or two
infections. Nat Med 1999, 5:340-343.
Grobusch MP, Kremsner PG: Uncomplicated malaria. Curr Top
Microbiol Immunol 2005, 295:83-104.
Delley V, Bouvier P, Breslow N, Doumbo O, Sagara I, Diakite M, Mauris A, Dolo A, Rougemont A: What does a single determination
of malaria parasite density mean? A longitudinal survey in
Mali. Trop Med Int Health 2000, 5:404-412.
Coleman RE, Sattabongkot J, Promstaporm S, Maneechai N, Tippayachai B, Kengluecha A, Rachapaew N, Zollner G, Miller RS, Vaughan
JA, Thimasarn K, Khuntirat B: Comparison of PCR and microscopy for the detection of asymptomatic malaria in a Plasmodium falciparum/vivax endemic area in Thailand. Malar J 2006,
5:121.
Menge DM, Ernst KC, Vulule JM, Zimmerman PA, Guo H, John CC:
Microscopy underestimates the frequency of Plasmodium falciparum infection in symptomatic individuals in a low transmission highland area. Am J Trop Med Hyg 2008, 79:173-177.
Armstrong Schellenberg J, Victora CG, Mushi A, de Savigny D, Schellenberg D, Mshinda H, Bryce J: Inequities among the very poor:
health care for children in rural southern Tanzania. Lancet
2003, 361:561-566.
Raso G, Vounatsou P, Gosoniu L, Tanner M, N'Goran EK, Utzinger J:
Risk factors and spatial patterns of hookworm infection
among schoolchildren in a rural area of western Côte
d'Ivoire. Int J Parasitol 2006, 36:201-210.
Briet OJ, Dossou-Yovo J, Akodo E, van de Giesen N, Teuscher TM:
The relationship between Anopheles gambiae density and
Page 15 of 16
(page number not for citation purposes)
Malaria Journal 2009, 8:252
38.
39.
40.
41.
42.
43.
44.
http://www.malariajournal.com/content/8/1/252
rice cultivation in the savannah zone and forest zone of Côte
d'Ivoire. Trop Med Int Health 2003, 8:439-448.
Koudou BG, Tano Y, Doumbia M, Nsanzabana C, Cisse G, Girardin
O, Dao D, N'Goran EK, Vounatsou P, Bordmann G, Keiser J, Tanner
M, Utzinger J: Malaria transmission dynamics in central Côte
d'Ivoire: the influence of changing patterns of irrigated rice
agriculture. Med Vet Entomol 2005, 19:27-37.
Costantini C, Li SG, Della Torre A, Sagnon N, Coluzzi M, Taylor CE:
Density, survival and dispersal of Anopheles gambiae complex
mosquitoes in a West African Sudan savanna village. Med Vet
Entomol 1996, 10:203-219.
Cohen JM, Ernst KC, Lindblade KA, Vulule JM, John CC, Wilson ML:
Topography-derived wetness indices are associated with
household-level malaria risk in two communities in the western Kenyan highlands. Malar J 2008, 7:40.
Betsi NA, Koudou BG, Cisse G, Tschannen AB, Pignol AM, Ouattara
Y, Madougou Z, Tanner M, Utzinger J: Effect of an armed conflict
on human resources and health systems in Côte d'Ivoire:
prevention of and care for people with HIV/AIDS. AIDS Care
2006, 18:356-365.
Fürst T, Raso G, Acka CA, Tschannen AB, N'Goran EK, Utzinger J:
Dynamics of socioeconomic risk factors for neglected tropical diseases and malaria in an armed conflict. PLoS NTDs 2009,
3:e513.
Noor AM, Mutheu JJ, Tatem AJ, Hay SI, Snow RW: Insecticidetreated net coverage in Africa: mapping progress in 2000-07.
Lancet 2009, 373:58-67.
UNICEF: The State of the World's Children 2007. In The double
dividend of gender equality New York: The United Nations Children's
Fund (UNICEF); 2007:160.
Publish with Bio Med Central and every
scientist can read your work free of charge
"BioMed Central will be the most significant development for
disseminating the results of biomedical researc h in our lifetime."
Sir Paul Nurse, Cancer Research UK
Your research papers will be:
available free of charge to the entire biomedical community
peer reviewed and published immediately upon acceptance
cited in PubMed and archived on PubMed Central
yours — you keep the copyright
BioMedcentral
Submit your manuscript here:
http://www.biomedcentral.com/info/publishing_adv.asp
Page 16 of 16
(page number not for citation purposes)