ArticlePDF Available

Can Human Movements Explain Heterogeneous Propagation of Dengue Fever in Cambodia?


Abstract and Figures

Determining the factors underlying the long-range spatial spread of infectious diseases is a key issue regarding their control. Dengue is the most important arboviral disease worldwide and a major public health problem in tropical areas. However the determinants shaping its dynamics at a national scale remain poorly understood. Here we describe the spatial-temporal pattern of propagation of annual epidemics in Cambodia and discuss the role that human movements play in the observed pattern. We used wavelet phase analysis to analyse time-series data of 105,598 hospitalized cases reported between 2002 and 2008 in the 135 (/180) most populous districts in Cambodia. We reveal spatial heterogeneity in the propagation of the annual epidemic. Each year, epidemics are highly synchronous over a large geographic area along the busiest national road of the country whereas travelling waves emanate from a few rural areas and move slowly along the Mekong River at a speed of ∼11 km per week (95% confidence interval 3-18 km per week) towards the capital, Phnom Penh. We suggest human movements - using roads as a surrogate - play a major role in the spread of dengue fever at a national scale. These findings constitute a new starting point in the understanding of the processes driving dengue spread.
Content may be subject to copyright.
Can Human Movements Explain Heterogeneous
Propagation of Dengue Fever in Cambodia?
Magali Teurlai
, Rekol Huy
, Bernard Cazelles
, Raphae
¨l Duboz
, Christophe Baehr
, Sirenda Vong
1Epidemiology and Public Health Unit, Institut Pasteur du Cambodge, Phnom Penh, Cambodia, 2IRD UMR LOCEAN, UMR ESPACE-DEV, New-Caledonia, France, 3National
Dengue Control Program, National Centre for Parasitology, Entomology and Malaria Control, Ministry of Health, Phnom Penh, Cambodia, 4Ecologie & Evolution, UMR
7625, CNRS-UPMC-ENS, Paris, France, 5UMMISCO UMI 209 IRD - UPMC, Bondy, France, 6CIRAD UPR Agirs, Montpellier, France, 7Me
´o France, CNRM, Toulouse, France,
8CNRS, GAME URA 1357, Toulouse, France
Determining the factors underlying the long-range spatial spread of infectious diseases is a key issue
regarding their control. Dengue is the most important arboviral disease worldwide and a major public health problem in
tropical areas. However the determinants shaping its dynamics at a national scale remain poorly understood. Here we
describe the spatial-temporal pattern of propagation of annual epidemics in Cambodia and discuss the role that human
movements play in the observed pattern.
Methods and Findings:
We used wavelet phase analysis to analyse time-series data of 105,598 hospitalized cases reported
between 2002 and 2008 in the 135 (/180) most populous districts in Cambodia. We reveal spatial heterogeneity in the
propagation of the annual epidemic. Each year, epidemics are highly synchronous over a large geographic area along the
busiest national road of the country whereas travelling waves emanate from a few rural areas and move slowly along the
Mekong River at a speed of ,11 km per week (95% confidence interval 3–18 km per week) towards the capital, Phnom
We suggest human movements – using roads as a surrogate – play a major role in the spread of dengue fever
at a national scale. These findings constitute a new starting point in the understanding of the processes driving dengue
Citation: Teurlai M, Huy R, Cazelles B, Duboz R, Baehr C, et al. (2012) Can Human Movements Explain Heterogeneous Propagation of Dengue Fever in
Cambodia? PLoS Negl Trop Dis 6(12): e1957. doi:10.1371/journal.pntd.0001957
Editor: Duane J. Gubler, Duke University-National University of Singapore, Singapore
Received May 13, 2012; Accepted October 29, 2012; Published December 6, 2012
Copyright: ß2012 Teurlai et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: MT was supported financially by the Fondation Pierre Ledoux-Jeunesse Internationale and the International Division of Institut Pasteur. BC is supported
by the FP7 Cooperation Programme from the European Community, DenFREE (FP7-HEALTH-2011-282378). The funders had no role in study design, data
collection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail:
Cambodia is a low-income tropical country, hyper-endemic for
all four serotypes of dengue infection. As such, dengue epidemics
occur every year during the rainy season and result in considerable
morbidity and economic burden. The basis for these recurrent
epidemics is an increased vector activity during the rainy season
and complex interactions between hosts and viruses with short
lived cross-protective immunity [1–3]. In the absence of a vaccine,
control is limited to vector control measures. Regarding dengue
dynamics, locally, dengue outbreaks are explosive and tend to be
focal, perhaps reflecting the limited dispersal of the vector, which
visits few houses in a life-time, and have a limited flight range
[4,5]. At an international scale, human movement is known to be a
major factor responsible for the virus transportation among big
urban centres [6]. At a national scale, little is known about the
spatial propagation of the disease. In many endemic countries,
urban centres are thought to act as a reservoir of the virus from
where it can spread to the rest of the country [6–9]. In Thailand
[7] and more recently in Southern Vietnam [8], researchers
demonstrated the existence of a travelling wave either within the 3-
year periodic mode or the annual mode of oscillation. However,
the underlying factors responsible for these waves remain
unknown [1,7–10]. Hypotheses include immunological host-virus
interactions, differences in virus virulence, or heterogeneity of the
spatial distribution of the host population [1]. Synchronisation of
cases has also been observed in the 3 year periodic band in
Thailand [7]. Mechanisms responsible for higher synchronicity
could include climatic forcing [8,10].
The objective of this study was to determine whether there was
a spatial-temporal pattern of dengue propagation in Cambodia
repeating year after year. The characterisation of such patterns is
important to understand the forces driving dengue spatial spread
and aid better control and logical allocation of public health
Study area
Cambodia is a low-income country located in South-East Asia,
divided into 24 provinces and 180 districts covering 181,035 km
Out of the 13.4 million people in 2008, more than 80% live in
PLOS Neglected Tropical Diseases | 1 December 2012 | Volume 6 | Issue 12 | e1957
rural areas; 1.3 million people (9.9%) live in Phnom Penh, the
capital city. Cambodia is lagging behind other countries in South-
East Asia in terms of economic or demographic development. For
example, unlike Thailand, the demographic transition has not
occurred yet. The South-West and the North of the country (see
grey districts in Figure 1) are mountainous regions with low
population density (mean district density of 18 people per km
The rest of the country is composed of flat plains with few cities
(mainly Phnom Penh, Kampong Cham, Siem Reap and
Battambang, see Figure 1) scattered among rural areas. The
weather is warm all year long, and the climate is dominated by the
annual monsoon cycle, with a dry season (December to April)
alternating with a wet season (May to November). Climatic
variation from one area to another is limited. Temperature is
homogeneous across the country and ranges annually from 21uC
to 35uC. As a result, mosquitoes can be active all year long when
considering only temperature. Rainfall or water availability are
more likely to be the factors limiting the vector’s activity.
The data
Cambodian National surveillance recorded 109,332 dengue
cases during 2002–2008, a period over which the reporting process
was stable. Cases were reported passively from public hospitals
and actively from 5 major sentinel hospitals located in the cities of
Siem Reap (1 hospital), Kampong Cham (1 hospital) and Phnom
Penh (3 hospitals). Cases were clinically diagnosed using the 1997
World Health Organization (WHO) case definitions, allowing
clinical and paraclinical (haematocrit and platelets count) distinc-
tion between classic dengue fever, dengue haemorrhagic fever and
dengue shock syndrome [2]. Of note, the new WHO case
definition was only introduced in 2009 [11]. To increase
specificity, only severe cases (e.g. dengue haemorrhagic fever)
affecting children less than 16 years old were recorded in the
database. Assuming that epidemic patterns of dengue would be
stochastic in low population density areas, we excluded the 45 (/
180) districts with less than 20 people per km
from the analysis,
dismissing 3298 declared cases (3.03% of the total declared cases).
Since patients’ districts of residence were recorded, we calculated
weekly incidence rates for each of the 135 remaining districts.
Denominators for incidence rates were interpolated linearly using
the 1998 and 2008 national censuses. Based on age distribution
similarities between provinces, and on similarities between 1998
and 2008 age structures, we assumed that age structure was
homogeneous over the country and have not standardised
incidence rates according to age. As national surveillance data
were made available to be utilised for such temporal and spatial
analyses and have been routinely published nationally and once
internationally [2] no specific approval was requested from the
Ministry of Health’s National Ethics Committee. Moreover, data
that were provided by the National Programme were anonymised
prior to transfer to the Institut Pasteur du Cambodge. We
subsequently randomly assigned new codes to each record and
deleted the previous ones to unlink the present dataset to the
national database.
Temporal analysis
Time series of dengue incidence were square-root transformed
to stabilise the variance, subsequently centred to zero-mean and
normalised to unit variance. We then performed a wavelet
transform and used wavelet phase analysis to describe dynamic
patterns of dengue. This spectral method is well-suited for the
analysis of non-stationary time-series such as epidemic curves. It is
particularly useful to filter the data in any given frequency band
and extract the phase of any given periodic component [10,12–14]
(see Figure 2 for an illustration of the method).
The wavelet transform was done using R software [15] and
functions translated from Cazelles’ Matlab toolbox. We used a
Morlet wavelet as the mother wavelet. All equations used and
vocabulary relative to wavelet analysis are detailed in [10,12–14].
The wavelet coefficients corresponding to a period ranging from
0.8 to 1.2 year were used to reconstruct filtered time series
corresponding to the annual epidemic in each district. These
filtered time series, called ‘‘annual component of incidence’’, are
illustrated in Figure 2B.
The phases of the epidemic have been computed in the periodic
band 0.8–1.2 year (see equation (5) in [10]), thus obtaining, for
Figure 1. Map of mean annual dengue fever incidence rates in
districts of Cambodia. Mean annual incidence rates (in number of
cases declared per 100,000 people per year) are calculated over 2002 to
2008, for districts with more than 20 people per km
. Cambodia is
surrounded by the Indian Ocean (bottom left), Thailand (West), Lao
(North) and Vietnam (East and South-East). Phnom Penh, the capital, is
represented by a circle, Siem Reap by a triangle, Kampong Cham by a
square and Battambang by a lozenge. Blue lines represent the Mekong
River, going north to south, and the Tonle Sap River linking the Tonle
Sap central Lake to the Mekong River. Green lines represent national
roads. Grey districts have less than 20 people per km
Author Summary
Dengue fever is a mosquito borne viral infection. It has
become a major public health problem during the past
decades: only 9 countries were affected in the 1970s;
dengue is now endemic in more than 100 countries. In the
absence of any vaccine or specific treatment, control of
dengue fever is currently limited to vector control
measures, which are difficult to implement and hardly
sustainable, especially in low income countries. To
implement efficient control measures, it is crucial to
understand the dynamics of propagation of the disease
and the key factors underlying these dynamics. In this
study, data from 8-year national surveillance in Cambodia
were analysed. Dengue fever follows a recurrent pattern of
propagation at the national scale. The annual epidemics
originate from a few rural areas identified in this work. This
study also suggests additional evidence for the role of
human movement in the spatial dynamics of the disease,
which should be accounted for in control measures. These
results differ from the current knowledge about dengue
dynamics and are therefore of interest for future research.
Spatial Dynamics of Dengue Fever in Cambodia
PLOS Neglected Tropical Diseases | 2 December 2012 | Volume 6 | Issue 12 | e1957
each district, a single time series of the phase of the cyclic annual
component of the epidemic (Figure 2C). This phase can be viewed
as the timing of the epidemics, and is almost not influenced by the
intrinsic value of incidence. For a given week, if the seasonal
epidemic occurs at the same time in two districts, the two annual
components have the same phase. Another example in Figure 2,
district #306 (a rural district located around Kampong Cham) has
an advanced phase compared to Phnom Penh: the epidemic in
Phnom Penh is lagging behind the one in district #306.
By calculating phase differences, one can then determine in
which order districts are affected by the annual epidemic. This
ranking allowed us to identify districts hit early by the seasonal
epidemic. Time series of the temporal lag between seasonal
epidemics at different locations were thus estimated from phase
differences, according to equations (6)–(8) in [10]. This temporal
lag is proportional to the phase difference.
In some districts identified as early districts, there were not
enough cases reported to be confident that there was significant
dengue transmission ongoing, given that cases are declared on a
clinical basis. Therefore, for each district, we identified epidemic
years using the national epidemic threshold [2,16]. The national
weekly threshold [2] for an epidemic was calculated as two
standard deviations above a three weeks sliding mean computed
over the weekly national incidence rates of the five past available
non epidemic years [16] (from 2002 to 2006 in our study). We
excluded from the analysis time periods from districts with low
incidence, only including epidemic years defined as districts with a
weekly incidence rate from January to December above the
national threshold for an epidemic during two consecutive weeks.
Subsequent analyses were restricted to the years with detected
Spatial analysis of dengue propagation
For a given year, to evaluate the speed at which dengue
epidemic propagates along a given geographical axis comprised of
I districts, we performed a regression analysis: Y
with i in [1, I], Y
the annual mean of the temporal lag between
district i and a district located along the axis and selected as a time
reference to compute temporal lags, and X1
the corresponding
distance as the crow flies, in km, between the centres of district i
and the reference district. The speed of dengue propagation along
the axis was estimated as the inverse of the regression slope b1.
To compare the dynamic pattern along J different axes, we
performed, each year, an analysis of covariance [17]: Y
, with i in [1, I
], j in [1, J], Y
the annual mean of the
temporal lag between district i of axis j and a district located along
the axis j and chosen as a time reference, X1
the corresponding
distance separating the centres of districts i and the reference
district on axis j, and b0
and b1
the intercept and slope of the
regression line of axis j. We will call ‘‘p-value of interaction’’ the p-
value of the hypothesis that the model slopes b1
are all equal, or in
other words, that the speed of propagation is the same along
different axes. If the model gave a p-value of interaction below
0.05, the propagation was considered heterogeneous along the
different axes, and separate regressions were performed for each
As we wanted to know whether the results were consistent from
year to year, we performed this analysis each year.
Figure 2. Weekly raw incidence rates, filtered incidence and phase of four districts in Cambodia. The four districts are: district #306
(black), a rural district located around Kampong Cham; Phnom Penh (red); District #307 (green), a rural district located mid-way between #306 and
Phnom Penh; District #104 (blue). (A) Weekly raw incidence rates (in number of cases declared per 100,000 people per week). (B) Annual component
of incidence, obtained by filtering raw weekly incidence in the 0.8–1.2 year periodic band using wavelet analysis. (C) Phase of the annual component
of incidence, computed in the 0.8–1.2 periodic band using wavelet analysis (see Methods).
Spatial Dynamics of Dengue Fever in Cambodia
PLOS Neglected Tropical Diseases | 3 December 2012 | Volume 6 | Issue 12 | e1957
All analyses and figures were performed using R software [15].
All confidence intervals (C.I.) provided were calculated using
classical methods for calculating a confidence interval around a
mean, unless stated otherwise.
Figure 3 shows the weekly incidence in the 135 (/180) most
populous districts (average district population of 78,000), where
105,598 of the 109,332 reported patients reside. Annual epidemics
do not appear synchronous, peaking at different times of the year
in different districts (between May and October). Unexpectedly, in
Phnom Penh, the capital, where the virus circulates during the dry
season [2], the annual epidemic lags behind the one in other
districts indicating that Phnom Penh is not the starting point.
Inspection of weekly and mean annual incidence maps over the
whole period (not shown here, but partly visible in Figure 1, Movie
S1, and Figure S1) showed two possible geographic axes seminal in
the propagation of dengue: the national road between Kampong
Cham and Siem Reap – the busiest one in Cambodia –, and the
Mekong River.
To characterise and compare the spatio-temporal pattern of
dengue incidence along those two axes and determine the focal
starting areas of the annual epidemic, we performed a phase
analysis using a wavelet approach (see Methods).
The examination of wavelet power spectra (Figure S2) revealed
the annual seasonal component of incidence is the most powerful
in 117/135 districts, meaning that in Cambodia, the seasonal cycle
of dengue incidence time series has more power than the inter-
annual cycle. A significant 2–3 years periodic component was
detected in some districts, but our time series were too short in
time to study it.
Figure 4 represents, in a time-space domain, the phase of the
annual component of incidence filtered in the 0.8–1.2 year
periodic band for each district of the two geographic features
identified previously (see Movie S2 for maps of these phases in all
135 districts). White parts represent years when no epidemic was
detected in the district (see Methods). The phase represents the
timing of the epidemic in each district. For a given week, if all
districts have the same phase (same colour), the epidemics occur at
the same time.
First, the pattern of spatial synchronicity observed is consistent
from year to year. To test this result, each year, we ranked the 135
districts according to their phase during the epidemic period
(weeks 13 to 39), from the district where the epidemic has the most
advanced phase to the one with the highest phase delay. We then
performed a Spearman correlation test on these ranks, comparing
ranks from year n with ranks of years n+1. Except in 2007,
correlation coefficients were all higher than 0.38 and significant
(all p-values,0.03), inferring that districts are affected in a similar
chronological order year after year. In 2007, the order in which
districts were affected by the epidemic was not significantly
correlated to the order of the previous year.
This ranking also allowed us, each year, to identify districts
where the annual epidemic appears early. We have identified three
starting points, all located in rural areas: district #306 and a few
rural districts around (district #306 being the most early of them),
district #104 and 2 other districts around, and, some years, district
#805, located along the Mekong River, South to Phnom Penh,
along the Vietnamese border (see Figure 4). These districts are
very similar to other rural districts included in the analysis,
Figure 3. Apparent dengue haemorrhagic fever weekly inci-
dence rates in the 135 most populous districts of Cambodia.
Weekly incidence rates (cases per 100,000 people per week) were
computed in each of the 135 districts where population density is
higher than 20 people per km
in Cambodia. Districts are ranked by
increasing distance to Phnom Penh from bottom to top.
Figure 4. Phases of the annual component of incidence for
districts located along two geographic axes. Phases are
computed in the 0.8–1.2 year periodic band. (A) Map of the two
geographic areas chosen: the national road in blue, and the Mekong
River in orange. (B) Phase of districts along the Mekong River (orange in
Figure 4A), presented from the most southerly to the most northerly
from bottom to top. (C) Phase of districts along the national road (blue
in Figure 4A), presented from West to East from bottom to top. The
arrows indicate districts: 1, #306; 2, Phnom Penh; 3, #805 (Figure 4B); 4,
#104 (Figure 4C).
Spatial Dynamics of Dengue Fever in Cambodia
PLOS Neglected Tropical Diseases | 4 December 2012 | Volume 6 | Issue 12 | e1957
composed of a flat flood plain, with a mean population density of
155 people per km
. They are all located away from the three
urban centres where sentinel hospitals involved in active surveil-
lance are. They are consistently the same, year after year.
Secondly, Figure 4 demonstrates that the propagation of dengue
is heterogeneous. Annual epidemics are highly synchronous along
the national road linking Kampong Cham to Siem Reap whereas
an oriented propagation emanates from a rural area located
around Kampong Cham (district #306) and ends in Phnom Penh
11 weeks later (95% C.I. 4–18 weeks). This was obscured when
districts were classified independently of the geographic area,
taking into account only the distance to Phnom Penh (Figures S3,
S4 and S5). Some years, another oriented propagation is seen from
a rural area near the Vietnamese border (district #805) towards
Phnom Penh (Figure 4B).
To evaluate the speed of propagation along each axis, we
calculated the temporal lag of the seasonal pattern in each district
relative to the district #306 [10] (see Methods). This district was
chosen as a time reference point to calculate time lags because it is
located at the intersection of the Mekong River and the national
road areas, which was convenient to present results of both
geographic areas on the same figure (Figure 5). Moreover, based
on the ranking of districts using their phase, district #306 was the
one with the most number of districts lagging behind over the
study period (132, 134, 131, 118, 124 and 104 districts/134
respectively from 2002 to 2007). The results of the analysis of
covariance performed on the mean annual lag to compare the
speed of propagation of the epidemic in each geographic axis
(Figure 4A) are presented in Table 1. Results show that regardless
of the year, the evolution of the temporal lag between epidemics
according to the distance differs significantly (a-level of 0.05) from
one area to the other. We therefore performed separate regressions
for each of the two axes (Figure 5). Propagation is always faster
along the national road than along the Mekong River. The speed
of propagation of the travelling wave along the Mekong River,
evaluated by the inverse of the regression slope, is estimated to be
11 km per week (95% C.I. 3–18 km per week) over the study
period. Along the national road, the speed is quasi-instantaneous
with respect to our time domain sampling rate.
To test the role Phnom Penh plays in this heterogeneity, we ran
the same analysis, but excluding Phnom Penh districts. The
propagation remained significantly heterogeneous at an alpha-
level of 5%, except in 2002 and 2007, and remained significantly
heterogeneous at an alpha-level of 10% in 2002 and 2007 (results
not shown).
The removal of districts with low population density has no
effect at all on results shown. The exclusion of low incidence years
did not modify the two conclusions arrived at in the paper (onset of
the seasonal epidemic in highly localised rural areas, and
heterogeneous propagation in the country), but removed the
potential bias linked with the very low number of cases during the
non epidemic years (see Figures S6 and S7 that show the same as
Figures 4 and 5 when all years are included).
To sum up, our results show that the seasonal epidemic
consistently starts in the same three rural districts in Cambodia.
Then the propagation is not homogeneous in the country. In
districts located along the busiest road, dengue epidemics appear
simultaneously and early (with all districts being hit in less than a
month), whereas districts located along the Mekong River get hit
by the seasonal epidemic later, with the delay being proportional
to the distance to district #306.
In 2007 an exceptional epidemic occurred in Cambodia with
DENV-3 as the dominant etiological agent, and a four fold
increase in weekly incidence rate compared to the four previous
years (Figures 2 and 3). During this epidemic, the wave in the one-
year periodic mode travelled from Kampong Cham to Phnom
Penh in only 7 weeks, at a speed of 17 km per week, a higher
synchronicity reflecting a more rapid propagation during this
peculiar event. The sequence in which districts were affected by
the seasonal variation was also modified, as shown by non-
significant Spearman rank coefficients when correlations were
calculated between any year and the year 2007.
Figure 5. Scatterplot of mean annual temporal lags between epidemics against distances between districts. Temporal lags between
epidemics and distances are computed relative to district #306. The lines show the linear regressions between the mean annual temporal lag of the
annual epidemic in each district and the distance for 2002 (A), 2003 (B), 2004 (C), 2005 (D), 2006 (E) and 2007 (F). Colours represent the geographic
localisation of each district, according to Figure 4A. The number of districts included in the analysis changes every year, according to whether an
epidemic occurred in the district (Table 1). Error bars represent the 95% C.I. associated with the mean. Normality and homoscedasticity of residuals
were confirmed using the Shapiro-Wilks and the Bartlett tests respectively (alpha level of 0.05).
Spatial Dynamics of Dengue Fever in Cambodia
PLOS Neglected Tropical Diseases | 5 December 2012 | Volume 6 | Issue 12 | e1957
A rural onset
Regarding the onset of the national epidemic, there is a
tendency for the dengue season to start in few rural districts more
often than in any other district moving from rural areas towards
urban centres, with, for example, annual epidemics in districts
#306 and #104 (Figure 4A) leading epidemics in the rest of the
country by 3 weeks on average over the study period (95% C.I. 2–
4 weeks). In addition, recent prospective cohort data showed that
rural areas were affected by dengue to the same degree as urban
areas or, as during the 2007 epidemic, at even higher incidence
rates [18]. This finding is not consistent with the common thought
that the most populated areas spread the disease [7–9] or known
mechanisms underlying travelling waves [7,13]. One plausible
explanation for a rural origin of the spread, also supported by a
recent study in Vietnam [19], is that more than 80% of rural
Cambodians do not have access to public water supply, and store
their drinking water in big jars that have been identified as major
breeding sites for Aedes mosquitoes in Cambodia [2]. This result
(onset in rural areas) has strong implications regarding the control
of dengue in low-income countries.
Factors that could explain the heterogeneous pattern of
To our knowledge, it is the first time that a recurrent
heterogeneous pattern of propagation of dengue is revealed in
surveillance data at a national scale. This pattern of propagation
could come from actual transmission of dengue viruses between
districts, via infected mosquitoes or humans, or correspond merely
to spatial differences in the emergence of the epidemic, differences
due either to spatial differences in local (within-district) dynamics,
or to forcing by extrinsic factors such as climate (Moran effect).
Given the limited dispersal range of Aedes spp. vectors and the high
synchronicity of the epidemics over a 400 km wide area along the
national road, it is unlikely that viral transmission from one place
to another via infected mosquitoes can account for the pattern
Climate is quite homogeneous in Cambodia, with a narrow
temperature range across the country, and the monsoon striking
the country from South to North within a month only, around
April-May. Climate, by influencing the vector’s life cycle, is clearly
driving the dengue seasonality observed in Cambodia. It could
easily explain the high synchronicity of epidemics observed along
the national road if this synchronisation was observed country-
wide. However, two of our observations preclude climatic forcing
to be the underlying factor for the spatial-temporal pattern
observed. First, climate is more homogeneous in the country
compared to the observed spatial-temporal pattern of dengue
fever. The existence of a very slow oriented dengue propagation
along the Mekong River, going North to South, with the epidemic
in some districts lagging more than 3 months behind the one in
districts that are only less than 200 kilometres in distance cannot
be accounted for by climatic differences between districts.
Secondly, the epidemic starts as early as in March in the three
districts identified as starting points, whereas the monsoon only
begins end of April/beginning of May.
We believe the heterogeneous propagation observed is related
to the heterogeneity of traffic on the roads of the country, where
traffic allows the movement of human carriers or transported
mosquitoes infected with dengue. The national road –the busiest
country road connecting the two main economic cities within 4–
5 hours – probably explains synchronicity. By contrast, the
existence of dirt roads along the entire Mekong River, on which
traffic and population movement between smaller villages are
slower and more difficult than along the national road supports
our hypothesis. Despite its epidemiological relevance, our under-
standing of the relationship between human movement and
pathogen transmission remains limited [20–22]. The importance
of the relationship between human movement and pathogen
transmission in explaining the spatio-temporal dynamics of dengue
incidence or other diseases has been increasingly pointed-out
during the last decade [5,8,9,23–25] and explored mainly through
a theoretical modelling approach [22,26,27].
2007, a peculiar year
The higher synchronisation and higher incidence levels across
the country in 2007 argues for an association with higher net
reproduction ratios of infection due to lower herd immunity when
a new serotype is introduced [25,28]. This hypothesis is supported
by the fact that serotype 3 invaded Cambodia at the end of year
2006, replacing serotype 2, in place since 1999, as the major
circulating virus [2]. The fact that heterogeneity of propagation
remained despite this serotype change supports the hypothesis that
another factor – such as human movement – plays an important
role in the dynamics.
Major limitations of our analysis include underreporting
inherent to passive surveillance data and potential selection bias
Table 1. Results of the analysis of covariance and linear regressions (see methods).
Covariance analysis Regression (‘‘Mekong River’’) Regression (‘‘national road’’)
year n
p-value of interaction n
1/slope estimate
2002 40 0.005 22 0.02 11 18 0.70
2003 44 ,0.001 22 0.001 9 22 0.22
2004 31 ,0.001 22 ,0.001 8 9 0.07
2005 38 ,0.001 19 ,0.001 8 19 0.09
2006 40 ,0.001 22 0.004 13 18 0.22
2007 52 0.001 26 0.006 17 26 0.378
Number of districts where an annual epidemic occurred according to the national threshold, along the Mekong River, the national road (Figure 4A), or, for the
covariance analysis, both.
P-value of the regression slope estimate.
Inverse of the estimate of the regression slope b1, in km per week (see methods).
Spatial Dynamics of Dengue Fever in Cambodia
PLOS Neglected Tropical Diseases | 6 December 2012 | Volume 6 | Issue 12 | e1957
leaning towards underreporting from urban centres compared
with rural areas. However, underreporting would only affect the
amplitude of the epidemics in each district and therefore have little
effects on the study of synchronism when using the wavelet phase
analysis approach [12]. Secondly, it is unlikely that rural
Cambodians were over-represented as most hospitals and those
that recruit dengue patients through active surveillance are free of
charge and located in urban areas (three cities). One could also
assume that collecting data from both an active and a passive
surveillance system could affect the timing of detection of the
epidemics, with active surveillance sites more likely to reveal a
small increase in incidence levels earlier than passive surveillance
sites. This would lead to a bias in the analysis. However, after a
thorough analysis of the surveillance system (Institut Pasteur
Cambodia’s not published report), we believe that severe cases are
scarcely missed even by the passive surveillance system; we also
believe that our results are not affected by this bias, given the fact
that districts affected early in the epidemic are located in rural
areas, kilometres away from the urban sentinel sites. Another
common limitation when analysing surveillance data is the
introduction of spatial or temporal biases due to difficulties in
standardising surveillance systems in time and space, especially in
low income countries. We thus, on purpose, excluded any results
or comments that would rely on spatial differences in incidence
levels only. We found the wavelet approach very robust to these
biases when exploring spatial-temporal patterns: unlike many
other temporal methods, the level of incidence in a given district
does not influence the calculation of time lags between epidemics.
One could think that not standardising incidence according to
age could impair the results obtained. But age standardisation only
affects the results by modifying incidence levels, and as our results
do not rely on differences in incidence levels, this is not a real
Our findings and speculations require further research for
additional evidence. Firstly, reasons as to why the three areas
identified as being hit early by the epidemic repeatedly year after
year (districts #306, #104 and #805) are unclear. This warrants
further filed investigations to identify specific factors that trigger
the epidemic in these settings.
Secondly, the data we analysed here can only reveal the spatio-
temporal dynamics and help make hypotheses on underlying
factors. Given the results of this study, and particularly the
heterogeneity of dengue spatial dynamics, surveillance data on the
spatial distribution of the serotypes (and genotypes if possible) of
the co-circulating dengue viruses would help validate (or not) our
hypotheses on dengue dynamics in Cambodia, using independent
Lastly, given the potential benefit in term of disease control
from demonstrating the efficient role of humans’ movement in
dengue spatial transmission, one might consider further investiga-
tions in this direction, collecting data and using dynamic models
for instance.
Supporting Information
Figure S1 Maps of weekly incidence rates in Cambodia.
Maps show weeks 5, 9, 11, 14, 18, 22, 25, 28, 31, 35, 38, 42 and 50
of year 2002 in Cambodian districts (in number of cases declared
per 100,000 people per week).
Figure S2 Map of Cambodian districts with their
identifying numbers and wavelet power spectra of the
weekly incidence rates of districts with more than 20
people per km
.Power is colour coded, blue indicating low
power. Dashed lines represent the limit of the cone of influence
where data are modified by edge effects. The black lines show the
0.05 alpha-level of significance computed based on 500 boot-
strapped series [10].
Figure S3 Phases of the annual component of incidence
for districts located along the ‘‘Mekong’’ axis (orange in
figure 4A). Phases are computed in the 0.8–1.2 year periodic
band. Districts are ranked by increasing distance to Phnom Penh,
from bottom to top. The arrows indicate: 1, district #306; 2,
Phnom Penh; 3, district #805.
Figure S4 Phases of the annual component of incidence
for districts located along the ‘‘national road’’ axis (blue
in figure 4A). Phases are computed in the 0.8–1.2 year periodic
band. Districts are ranked by increasing distance to Phnom Penh,
from bottom to top. The arrow indicates district #104.
Figure S5 Phases of the annual component of incidence
for districts with more than 20 people per km
.Phases are
computed in the 0.8–1.2 year periodic band. Districts are ranked
by increasing distance to Phnom Penh, from bottom to top. The
arrows indicate: 1, Phnom Penh; 2, District #306; 3, district
#805; 4, district #104 (see figure 4A).
Figure S6 Phases of the annual component of incidence
for districts located along two geographic axes (see
Figure 4A for a map of the two geographic areas chosen).
Phases are computed in the 0.8–1.2 year periodic band. Figure
shows the same as Figure 4B and 4C, but with all years included.
(A) Phase of districts along the Mekong River (orange in
Figure 4A), presented from the most southerly to the most
northerly from bottom to top. (B) Phase of districts along the
national road (blue in Figure 4A), presented from West to East
from bottom to top. The arrows indicate districts: 1, #306; 2,
Phnom Penh (Figure S6A) and #104 (Figure S6B).
Figure S7 Scatterplot of mean annual temporal lags
between epidemics against distances between districts.
Temporal lags between epidemics and distances are computed
relative to district #306. The lines show the linear regressions
between the mean annual temporal lag of the annual epidemic in
each district and the distance for 2002 (A), 2003 (B), 2004 (C),
2005 (D), 2006 (E) and 2007 (F). Colours represent the geographic
localisation of each district, according to Figure 4A. Error bars
represent the 95% C.I. associated with the mean. Figure shows the
same as Figure 5, but with all years included. Each year, the
number of units included in the analysis is 27 for the Mekong axis,
and 26 for the national road axis.
Movie S1 Evolution of weekly incidence rates in Cam-
bodia. Incidence rates are shown on maps of Cambodian districts
from January 2002 to December 2008.
Movie S2 Evolution the weekly phases of the annual
component of incidence. Phases are computed in the 0.8–1.2
year periodic band and shown on maps of Cambodian districts
from January 2002 to December 2008.
Spatial Dynamics of Dengue Fever in Cambodia
PLOS Neglected Tropical Diseases | 7 December 2012 | Volume 6 | Issue 12 | e1957
We thank Tristan Rouyer for providing an R toolbox (translated from
Cazelles’ Matlab toolbox) for performing wavelet analysis, Ke´vin Cazelles,
Christophe Menke`s and Morgan Mangeas for their help with parts of the
analysis, and Yann Que´au for useful comments.
Author Contributions
Conceived and designed the experiments: MT BC SV. Analyzed the data:
MT RH BC RD CB SV. Wrote the paper: MT RH BC RD CB SV.
Provided the data: RH. Designed the toolbox to perform wavelet analysis:
1. Wearing HJ, Rohani P (2006) Ecological and immunological determinants of
dengue epidemics. Proc Natl Acad Sci U S A 103: 11802–11807.
2. Huy R, Buchy P, Conan A, Ngan C, Ong S, et al. (2010) National dengue
surveillance in Cambodia 1980–2008: epidemiological and virological trends
and the impact of vector control. Bull World Health Organ doi:10.2471/
3. Hay SI, Myers MF, Burke DS, Vaughn DW, Endy T, et al. (2000) Etiology of
interepidemic periods of mosquito-borne disease. Proc Natl Acad Sci U S A 97:
4. Mammen MP Jr, Pimgate C, Koenraadt CJM, Rothman AL, Aldstadt J, et al.
(2008) Spatial and temporal clustering of dengue virus transmission in Thai
villages. PLoS Med doi:10.1371/journal.pmed.0050205.
5. Harrington LC, Scott TW, Lerdthusnee K, Coleman RC, Costero A, et al.
(2005) Dispersal of the dengue vector Aedes aegypti within and between rural
communities. Am J Trop Med Hyg 72: 209–220.
6. Chan KL (1985) Singapore’s dengue hemorrhagic fever control programme: a
case study on the successful control of Aedes aegypti and Aedes albopictus using
mainly environmental measures as a part of integrated vector control. Tokyo:
southeast Asian Medical Information Centre.
7. Cummings DAT, Irizarry RA, Huang NE, Endy TP, Nisalak A, et al. (2004)
Travelling waves in the occurrence of dengue haemorrhagic fever in Thailand.
Nature 427: 344–347.
8. Thai KTD, Cazelles B, Nguyen NV, Vo LT, Boni MF, et al. (2010) Dengue
dynamics in Binh Thuan province, southern Vietnam: periodicity, synchronicity
and climate variability. PLoS Negl Trop Dis doi:10.1371/journal.pntd.0000747.
9. Rabaa MA, Hang VTT, Wills B, Farrar J, Simmons CP, et al. (2010)
Phylogeography of recently emerged DENV-2 in southern Viet Nam. PLoS
Negl Trop Dis doi:10.1371/journal.pntd.0000766.
10. Cazelles B, Chavez M, McMichael AJ, Hales S (2005) Nonstationary influence
of el Nin˜o on the synchronous dengue epidemics in Thailand. PLoS Med
11. WHO (2009) Dengue guidelines for diagnosis, treatment, prevention and
control. Available:
9789241547871_eng.pdf. Accessed 2012 Apr 15.
12. Torrence C, Compo GP (1998) A practical guide to wavelet analysis. Bull Am
Meteorol Soc 79: 61–78.
13. Grenfell BT, Bjornstad ON, Kappey J (2001) Travelling waves and spatial
hierarchies in measles epidemics. Nature 414: 716–723.
14. Cazelles B, Chavez M, Constantin de Magny G, Gue´gan JF, Hales S (2007)
Time-dependent spectral analysis of epidemiological time-series with wavelets.
J R Soc Interface 4: 625–636.
15. R Development Core Team (2011) R: A language and environment for
statistical computing. R Foundation for Statistical Computing, Vienna, Austria.
ISBN 3-900051-07-0. Available: Accessed 2012 Nov
16. Stroup DF, Williamson GD, Hernon JL (1989) Detection of aberrations in the
occurrence of notifiable diseases surveillance data. Stat Med 8: 323–329.
17. Zuur AF, Ieno EN, Smith GM (2007) Analysing Ecological Data. New York:
Springer. 685p.
18. Vong S, Khieu V, Glass O, Ly S, Duong V, et al. (2010) Dengue incidence in
urban and rural Cambodia: results from population-based active fever
surveillance, 2006–2008. PLoS Negl Trop Dis doi:10.1371/jour-
19. Schmidt WP, Suzuki M, Thiem VD, White RG, Tsuzuki A, et al. (2011)
Population density, water supply and the risk of dengue fever in Vietnam: cohort
study and spatial analysis. PLoS Med doi:10.1371/journal.pmed.1001082.
20. Vazquez-Prokopec GM, Stoddard ST, Paz-Soldan V, Morrison AC, Elder JP, et
al. (2009) Usefulness of commercially available GPS data-loggers for tracking
human movement and exposure to dengue virus. International Journal of Health
Geographics 8 doi:10.1186/1476-072X-8-68.
21. Gonzales MC, Hidalgo CA, Barabasi AL (2008) Understanding individual
human mobility patterns. Nature 453: 779–782.
22. Stoddard ST, Morrison AC, Vazquez-Prokopec GM, Paz-Soldan V, Kochel TJ,
et al. (2009) The role of human movement in the transmission of vector-borne
pathogens. PLoS Negl Trop Dis doi:10.1371/journal.pntd.0000481.
23. Morrison AC, Getis A, Santiago M, Rigau-Perez JG, Reiter P (1998)
Exploratory space-time analysis of reported dengue cases during an outbreak
in Florida, Puerto Rico, 1991–1992. Am J Trop Med Hyg 58: 287–298.
24. Honorio NA, Ribeiro-Nogueira RM, Torres Codec¸o C, Carvalho MS,
Gonc¸alves-Cruz O, et al. (2009) Spatial evaluation and modeling of dengue
seroprevalence and vector density in Rio de Janeiro, Brazil. PLoS Negl Trop Dis
25. Viboud C, Bjornstad ON, Smith DL, Simonsen L, Miller MA, et al. (2006)
Synchrony, waves, and spatial hierarchies in the spread of influenza. Science
312: 447–451.
26. Adams B, Kapan DD (2009) Man bites mosquito: understanding the
contribution of human movement to vector-borne disease dynamics. PLoS
One doi:10.1371/journal.pone.0006763.
27. Barmak DH, Dorso CO, Otero M, Solari HG (2011) Dengue epidemics and
human mobility. Phys Rev E, 84: 011901 doi: 10.1103/PhysRevE.84.011901.
28. Longini IM, Fine PEM, Thacker SB (1986) Predicting the global spread of new
infectious agents. Am J Epidemiol 123: 383–391.
Spatial Dynamics of Dengue Fever in Cambodia
PLOS Neglected Tropical Diseases | 8 December 2012 | Volume 6 | Issue 12 | e1957
... Chareonsook et al. (1999) showed that DHF in Thailand, which was originally thought to be an urban disease, has spread to most areas of Thailand, and is now more common in rural than urban areas and studies suggest that rural dengue incidence can surpass urban and semi-urban communities within the same region (Reller et al., 2012;Vong et al., 2010). In addition, several studies have stressed that rural settings play an important role in the timing of dengue epidemics in Southeast Asia, with the seasonal dengue waves typically arriving later in major urban centers (Cazelles and Cazelles, 2014;Cuong et al., 2013;Teurlai et al., 2012). ...
... Because the DENFREE data covers only a relatively short period of time, surveillance data were added to improve the estimations. Surveillance of dengue is conducted at the national level in Cambodia, through the National Dengue Surveillance system (NDSS) Teurlai et al., 2012), involving the paediatric departments of several hospitals throughout the country. Since surveillance is hospital based, mostly severe cases are observed. ...
Full-text available
Dengue dynamics are shaped by the complex interplay between several factors, including vector seasonality, interaction between four virus serotypes, and inapparent infections. However, paucity or quality of data do not allow for all of these to be taken into account in mathematical models. In order to explore separately the importance of these factors in models, we combined surveillance data with a local-scale cluster study in the rural province of Kampong Cham (Cambodia), in which serotypes and asymptomatic infections were documented. We formulate several mechanistic models, each one relying on a different set of hypotheses, such as explicit vector dynamics, transmission via asymptomatic infections and coexistence of several virus serotypes. Models are confronted with the observed time series using Bayesian inference, through Markov chain Monte Carlo. Model selection is then performed using statistical information criteria, but also by studying the coherence of epidemiological characteristics (reproduction numbers, incidence proportion, dynamics of the susceptible class) in each model. Considering the available data, our analyses on transmission dynamics in a rural endemic setting highlight both the importance of using two-strain models with interacting effects and the lack of added value of incorporating vector and explicit asymptomatic components.
... This is true both for emerging infectious diseases whose spatial trajectory needs to be predicted [7][8][9][10][11] and those vaccine-preventable infectious diseases whose elimination is stymied by the stubborn mosaic of under-vaccination [12][13][14][15][16] or asynchronous metapopulation persistence [17,18]. Key to this effort has been attempting to understand the drivers of the observed spatial ecology of these systems, including the impact of host movement [5,11,19,20] and workflows [7], urban-rural gradients [1], regional variation in climate seasonality [4,[21][22][23] or population demography [3] and the impact of road networks [24]. These studies have identified insights that may, in principle, inform control policies including the implementation of spatially synchronised pulsed vaccination to increase the likelihood of elimination when epidemics are asynchronous [17,25,26], quantifying the impacts of border closures [9] and geographic barriers [27,28] on transmission. ...
Full-text available
Pertussis has resurfaced in the UK, with incidence levels not seen since the 1980s. While the fundamental causes of this resurgence remain the subject of much conjecture, the study of historical patterns of pathogen diffusion can be illuminating. Here, we examined time series of pertussis incidence in the boroughs of Greater London from 1982 to 2013 to document the spatial epidemiology of this bacterial infection and to identify the potential drivers of its percolation. The incidence of pertussis over this period is characterized by 3 distinct stages: a period exhibiting declining trends with 4-year inter-epidemic cycles from 1982 to 1994, followed by a deep trough until 2006 and the subsequent resurgence. We observed systematic temporal trends in the age distribution of cases and the fade-out profile of pertussis coincident with increasing national vaccine coverage from 1982 to 1990. To quantify the hierarchy of epidemic phases across the boroughs of London, we used the Hilbert transform. We report a consistent pattern of spatial organization from 1982 to the early 1990s, with some boroughs consistently leading epidemic waves and others routinely lagging. To determine the potential drivers of these geographic patterns, a comprehensive parallel database of borough-specific features was compiled, comprising of demographic, movement and socio-economic factors that were used in statistical analyses to predict epidemic phase relationships among boroughs. Specifically, we used a combination of a feed-forward neural network (FFNN), and SHapley Additive exPlanations (SHAP) values to quantify the contribution of each covariate to model predictions. Our analyses identified a number of predictors of a borough’s historical epidemic phase, specifically the age composition of households, the number of agricultural and skilled manual workers, latitude, the population of public transport commuters and high-occupancy households. Univariate regression analysis of the 2012 epidemic identified the ratio of cumulative unvaccinated children to the total population and population of Pakistan-born population to have moderate positive and negative association, respectively, with the timing of epidemic. In addition to providing a comprehensive overview of contemporary pertussis transmission in a large metropolitan population, this study has identified the characteristics that determine the spatial spread of this bacterium across the boroughs of London.
... Stringent international air travel bans during the COVID-19 pandemic implemented globally have resulted in sharp decreases in dengue cases in countries where dengue is almost predominantly imported, such as Australia, due to a precipitous decline in the importation of exotic pathogens by infected overseas travelers [10]. However, within dengue endemic countries, international air travel bans are unlikely to alter local dengue transmission patterns substantially; nationally, human movements along busiest national road of the country have been found to play a major role in the spread of dengue infection in dengue endemic countries [11]. Impacts of movement restrictions induced by lockdown measures on dengue transmission plausibly vary between countries, with differing dengue transmission dynamics and severities. ...
... Seasonal DENV transmission in non-endemic regions has received limited attention [19,20,24,[38][39][40][41][42]. A better understanding of transmission in these areas will shed light on several aspects of DENV transmission beyond non-endemic regions by enabling assessment of the initial phases of an outbreak when DENV begins circulating [17,43]. This is difficult in endemic areas where DENV has a virtually uninterrupted transmission. ...
Full-text available
Dengue is steadily increasing worldwide and expanding into higher latitudes. Current non-endemic areas are prone to become endemic soon. To improve understanding of dengue transmission in these settings, we assessed the spatiotemporal dynamics of the hitherto largest outbreak in the non-endemic metropolis of Buenos Aires, Argentina, based on detailed information on the 5,104 georeferenced cases registered during summer-autumn of 2016. The highly seasonal dengue transmission in Buenos Aires was modulated by temperature and triggered by imported cases coming from regions with ongoing outbreaks. However, local transmission was made possible and consolidated heterogeneously in the city due to housing and socioeconomic characteristics of the population, with 32.8% of autochthonous cases occurring in slums, which held only 6.4% of the city population. A hierarchical spatiotemporal model accounting for imperfect detection of cases showed that, outside slums, less-affluent neighborhoods of houses (vs. apartments) favored transmission. Global and local spatiotemporal point-pattern analyses demonstrated that most transmission occurred at or close to home. Additionally, based on these results, a point-pattern analysis was assessed for early identification of transmission foci during the outbreak while accounting for population spatial distribution. Altogether, our results reveal how social, physical, and biological processes shape dengue transmission in Buenos Aires and, likely, other non-endemic cities, and suggest multiple opportunities for control interventions.
... There, synchronous annual outbreaks originating in rural areas occurred along the main national highway in early phases, while spread to areas along the Mekong river occurred later. 59 S. Dass et al. ...
Background We studied the spatiotemporal spread of a chikungunya virus (CHIKV) outbreak in Sarawak state, Malaysia, during 2009–2010. Methods The residential addresses of 3054 notified CHIKV cases in 2009–2010 were georeferenced onto a base map of Sarawak with spatial data of rivers and roads using R software. The spatiotemporal spread was determined and clusters were detected using the space-time scan statistic with SaTScan. Results Overall CHIKV incidence was 127 per 100 000 population (range, 0–1125 within districts). The average speed of spread was 70.1 km/wk, with a peak of 228 cases/wk and the basic reproduction number (R0) was 3.1. The highest age-specific incidence rate was 228 per 100 000 in adults aged 50–54 y. Significantly more cases (79.4%) lived in rural areas compared with the general population (46.2%, p<0.0001). Five CHIKV clusters were detected. Likely spread was mostly by road, but a fifth of rural cases were spread by river travel. Conclusions CHIKV initially spread quickly in rural areas mainly via roads, with lesser involvement of urban areas. Delayed spread occurred via river networks to more isolated areas in the rural interior. Understanding the patterns and timings of arboviral outbreak spread may allow targeted vector control measures at key transport hubs or in large transport vehicles.
... Although it is possible for dengue infections to occur in workplaces, it was found in one study that 60% of dengue cases live less than 200m apart came from the same transmission chain, revealing that residential areas are a focal point of transmission [31]. Additionally, an increased frequency of movement within urban and predominantly residential neighbourhoods was found to increase the risk of dengue infection [66]. Patterns in urban area structure and population density could also influence the rate of dengue incidence, as agglomerations of high-rise buildings have a lower dengue incidence as compared to low-rise buildings [67]. ...
... Although it is possible for dengue infections to occur in workplaces, it was found in one study that 60% of dengue cases live less than 200m apart came from the same transmission chain, revealing that residential areas are a focal point of transmission [31]. Additionally, an increased frequency of movement within urban and predominantly residential neighbourhoods was found to increase the risk of dengue infection [66]. Patterns in urban area structure and population density could also influence the rate of dengue incidence, as agglomerations of high-rise buildings have a lower dengue incidence as compared to low-rise buildings [67]. ...
Full-text available
An estimated 105 million dengue infections occur per year across 120 countries, where traditional vector control is the primary control strategy to reduce contact between mosquito vectors and people. The ongoing sars-cov-2 pandemic has resulted in dramatic reductions in human mobility due to social distancing measures; the effects on vector-borne illnesses are not known. Here we examine the pre and post differences of dengue case counts in Malaysia, Singapore and Thailand, and estimate the effects of social distancing as a treatment effect whilst adjusting for temporal confounders. We found that social distancing is expected to lead to 4.32 additional cases per 100,000 individuals in Thailand per month, which equates to 170 more cases per month in the Bangkok province (95% CI: 100-242) and 2008 cases in the country as a whole (95% CI: 1170-2846). Social distancing policy estimates for Thailand were also found to be robust to model misspecification, and variable addition and omission. Conversely, no significant impact on dengue transmission was found in Singapore or Malaysia. Across country disparities in social distancing policy effects on reported dengue cases are reasoned to be driven by differences in workplace-residence structure, with an increase in transmission risk of arboviruses from social distancing primarily through heightened exposure to vectors in elevated time spent at residences, demonstrating the need to understand the effects of location on dengue transmission risk under novel population mixing conditions such as those under social distancing policies.
We developed a method to produce time-varying maps for dengue transmission risk, using the Ross-Macdonald framework and differential equations to estimate spatially the basic reproduction number (R0) of a vector-borne disease. The components of the R0 formula were derived partly from a mosquito population dynamics model integrating meteorological and environmental variables, and partly from temperature-dependent functions of vector competence and the extrinsic incubation period. The method was applied on Reunion Island, a tropical island located in the Indian Ocean, where the mosquito Aedes (Stegomyia) albopictus has been responsible for large and numerous outbreaks of dengue. As a validation, predicted maps and dynamic outputs were compared with the distribution of confirmed dengue cases registered during the year 2018 in Reunion Island. The results highlight strong agreements between the observed epidemiological patterns and predicted R0 distribution and temporal dynamics. This finding demonstrates the relevance and efficiency of the spatialised basic reproduction number (R0) to develop an operational dynamic mapping tool for dengue surveillance and control. The resulting method could be of great use in a health policy-making context, providing a time and space awareness to the dengue risk perception.
For the last many years, dengue has been reported to be one of the main causes of death in Malaysia. For more than 40 years, Malaysia is suffering from this endemic problem. The mortality and morbidity are reported for the dengue cases in a higher number of conformed cases in Malaysia. As per statistics, 136,992 cases were reported from 2008 to 2012, the highest in the record. As per the report from the Ministry of Health (MOH) of Malaysia, 77% of cases are reported from the urban area and 23% from the rural area. Since much research have been carried out in history, many researchers had concluded their novel research work, but still dengue cases are not controlled. Hence, this research suggests a novel way to visualize dengue cases and the occurrence of cases. This research uses machine learning technology combined with (geographical information system) GIS to predict dengue cases in Malaysia. The area of research is limited to the Selangor state of Malaysia as this is the most vulnerable area for dengue cases. This research focuses on unsupervised learning techniques to predict the density of cases. K-mean, KNN, and Expectation-Maximization (EM) algorithms are used to cluster the cases and visualize the pattern of dengue spread. In conclusion, all these information are mapped on dynamic mapping which will give the exact coordinates where dengue can occur. Based on this location, the fogging team can be informed and can target a specific area.
Full-text available
Buku monograf tentang Karakteristik dan Pola Penyebaran Demam Berdarah Dengue di Wilayah Endemis
Full-text available
Aedes aegypti, the major vector of dengue viruses, often breeds in water storage containers used by households without tap water supply, and occurs in high numbers even in dense urban areas. We analysed the interaction between human population density and lack of tap water as a cause of dengue fever outbreaks with the aim of identifying geographic areas at highest risk. We conducted an individual-level cohort study in a population of 75,000 geo-referenced households in Vietnam over the course of two epidemics, on the basis of dengue hospital admissions (n = 3,013). We applied space-time scan statistics and mathematical models to confirm the findings. We identified a surprisingly narrow range of critical human population densities between around 3,000 to 7,000 people/km² prone to dengue outbreaks. In the study area, this population density was typical of villages and some peri-urban areas. Scan statistics showed that areas with a high population density or adequate water supply did not experience severe outbreaks. The risk of dengue was higher in rural than in urban areas, largely explained by lack of piped water supply, and in human population densities more often falling within the critical range. Mathematical modeling suggests that simple assumptions regarding area-level vector/host ratios may explain the occurrence of outbreaks. Rural areas may contribute at least as much to the dissemination of dengue fever as cities. Improving water supply and vector control in areas with a human population density critical for dengue transmission could increase the efficiency of control efforts. Please see later in the article for the Editors' Summary.
Full-text available
In this work we explore the effects of human mobility on the dispersion of a vector borne disease. We combine an already presented stochastic model for dengue with a simple representation of the daily motion of humans on a schematic city of 20x20 blocks with 100 inhabitants in each block. The pattern of motion of the individuals is described in terms of complex networks in which links connect different blocks and the link length distribution is in accordance with recent findings on human mobility. It is shown that human mobility can turn out to be the main driving force of the disease dispersal.
Full-text available
Background: Dengue vaccines are now in late-stage development, and evaluation and robust estimates of dengue disease burden are needed to facilitate further development and introduction. In Cambodia, the national dengue case-definition only allows reporting of children less than 16 years of age, and little is known about dengue burden in rural areas and among older persons. To estimate the true burden of dengue in the largest province of Cambodia, Kampong Cham, we conducted community-based active dengue fever surveillance among the 0-to-19-year age group in rural villages and urban areas during 2006-2008. Methods and findings: Active surveillance for febrile illness was conducted in 32 villages and 10 urban areas by mothers trained to use digital thermometers combined with weekly home visits to identify persons with fever. An investigation team visited families with febrile persons to obtain informed consent for participation in the follow-up study, which included collection of personal data and blood specimens. Dengue-related febrile illness was defined using molecular and serological testing of paired acute and convalescent blood samples. Over the three years of surveillance, 6,121 fever episodes were identified with 736 laboratory-confirmed dengue virus (DENV) infections for incidences of 13.4-57.8/1,000 person-seasons. Average incidence was highest among children less than 7 years of age (41.1/1,000 person-seasons) and lowest among the 16-to-19-year age group (11.3/1,000 person-seasons). The distribution of dengue was highly focal, with incidence rates in villages and urban areas ranging from 1.5-211.5/1,000 person-seasons (median 36.5). During a DENV-3 outbreak in 2007, rural areas were affected more than urban areas (incidence 71 vs. 17/1,000 person-seasons, p<0.001). Conclusion: The large-scale active surveillance study for dengue fever in Cambodia found a higher disease incidence than reported to the national surveillance system, particularly in preschool children and that disease incidence was high in both rural and urban areas. It also confirmed the previously observed focal nature of dengue virus transmission.
Full-text available
Revealing the dispersal of dengue viruses (DENV) in time and space is central to understanding their epidemiology. However, the processes that shape DENV transmission patterns at the scale of local populations are not well understood, particularly the impact of such factors as human population movement and urbanization. Herein, we investigated trends in the spatial dynamics of DENV-2 transmission in the highly endemic setting of southern Viet Nam. Through a phylogeographic analysis of 168 full-length DENV-2 genome sequences obtained from hospitalized dengue cases from 10 provinces in southern Viet Nam, we reveal substantial genetic diversity in both urban and rural areas, with multiple lineages identified in individual provinces within a single season, and indicative of frequent viral migration among communities. Focusing on the recently introduced Asian I genotype, we observed particularly high rates of viral exchange between adjacent geographic areas, and between Ho Chi Minh City, the primary urban center of this region, and populations across southern Viet Nam. Within Ho Chi Minh City, patterns of DENV movement appear consistent with a gravity model of virus dispersal, with viruses traveling across a gradient of population density. Overall, our analysis suggests that Ho Chi Minh City may act as a source population for the dispersal of DENV across southern Viet Nam, and provides further evidence that urban areas of Southeast Asia play a primary role in DENV transmission. However, these data also indicate that more rural areas are also capable of maintaining virus populations and hence fueling DENV evolution over multiple seasons.
Full-text available
Dengue is a major global public health problem with increasing incidence and geographic spread. The epidemiology is complex with long inter-epidemic intervals and endemic with seasonal fluctuations. This study was initiated to investigate dengue transmission dynamics in Binh Thuan province, southern Vietnam. Wavelet analyses were performed on time series of monthly notified dengue cases from January 1994 to June 2009 (i) to detect and quantify dengue periodicity, (ii) to describe synchrony patterns in both time and space, (iii) to investigate the spatio-temporal waves and (iv) to associate the relationship between dengue incidence and El Niño-Southern Oscillation (ENSO) indices in Binh Thuan province, southern Vietnam. We demonstrate a continuous annual mode of oscillation and a multi-annual cycle of around 2-3-years was solely observed from 1996-2001. Synchrony in time and between districts was detected for both the annual and 2-3-year cycle. Phase differences used to describe the spatio-temporal patterns suggested that the seasonal wave of infection was either synchronous among all districts or moving away from Phan Thiet district. The 2-3-year periodic wave was moving towards, rather than away from Phan Thiet district. A strong non-stationary association between ENSO indices and climate variables with dengue incidence in the 2-3-year periodic band was found. A multi-annual mode of oscillation was observed and these 2-3-year waves of infection probably started outside Binh Thuan province. Associations with climatic variables were observed with dengue incidence. Here, we have provided insight in dengue population transmission dynamics over the past 14.5 years. Further studies on an extensive time series dataset are needed to test the hypothesis that epidemics emanate from larger cities in southern Vietnam.
A practical step-by-step guide to wavelet analysis is given, with examples taken from time series of the El NiñoSouthem Oscillation (ENSO). The guide includes a comparison to the windowed Fourier transform, the choice of an appropriate wavelet basis function, edge effects due to finite-length time series, and the relationship between wavelet scale and Fourier frequency. New statistical significance tests for wavelet power spectra are developed by deriving theoretical wavelet spectra for white and red noise processes and using these to establish significance levels and confidence intervals. It is shown that smoothing in time or scale can be used to increase the confidence of the wavelet spectrum. Empirical formulas are given for the effect of smoothing on significance levels and confidence intervals. Extensions to wavelet analysis such as filtering, the power Hovmöller, cross-wavelet spectra, and coherence are described. The statistical significance tests are used to give a quantitative measure of changes in ENSO variance on interdecadal timescales. Using new datasets that extend back to 1871, the Niño3 sea surface temperature and the Southern Oscillation index show significantly higher power during 1880-1920 and 1960-90, and lower power during 1920-60, as well as a possible 15-yr modulation of variance. The power Hovmöller of sea level pressure shows significant variations in 2-8-yr wavelet power in both longitude and time.
Introduction.- Data management and software.- Advice for teachers.- Exploration.- Linear regression.- Generalised linear modelling.- Additive and generalised additive modelling.- Introduction to mixed modelling.- Univariate tree models.- Measures of association.- Ordination--first encounter.- Principal component analysis and redundancy analysis.- Correspondence analysis and canonical correspondence analysis.- Introduction to discriminant analysis.- Principal coordinate analysis and non-metric multidimensional scaling.- Time series analysis--Introduction.- Common trends and sudden changes.- Analysis and modelling lattice data.- Spatially continuous data analysis and modelling.- Univariate methods to analyse abundance of decapod larvae.- Analysing presence and absence data for flatfish distribution in the Tagus estuary, Portugual.- Crop pollination by honeybees in an Argentinean pampas system using additive mixed modelling.- Investigating the effects of rice farming on aquatic birds with mixed modelling.- Classification trees and radar detection of birds for North Sea wind farms.- Fish stock identification through neural network analysis of parasite fauna.- Monitoring for change: using generalised least squares, nonmetric multidimensional scaling, and the Mantel test on western Montana grasslands.- Univariate and multivariate analysis applied on a Dutch sandy beach community.- Multivariate analyses of South-American zoobenthic species--spoilt for choice.- Principal component analysis applied to harbour porpoise fatty acid data.- Multivariate analysis of morphometric turtle data--size and shape.- Redundancy analysis and additive modelling applied on savanna tree data.- Canonical correspondence analysis of lowland pasture vegetation in the humid tropics of Mexico.- Estimating common trends in Portuguese fisheries landings.- Common trends in demersal communities on the Newfoundland-Labrador Shelf.- Sea level change and salt marshes in the Wadden Sea: a time series analysis.- Time series analysis of Hawaiian waterbirds.- Spatial modelling of forest community features in the Volzhsko-Kamsky reserve.
Dengue has been reportable in Cambodia since 1980. Virological surveillance began in 2000 and sentinel surveillance was established at six hospitals in 2001. Currently, national surveillance comprises passive and active data collection and reporting on hospitalized children aged 0-15 years. This report summarizes surveillance data collected since 1980. Crude data for 1980-2001 are presented, while data from 2002-2008 are used to describe disease trends and the effect of vector control interventions. Trends in dengue incidence were analysed using the Prais-Winsten generalized linear regression model for time series. During 1980-2001, epidemics occurred in cycles of 3-4 years, with the cycles subsequently becoming less prominent. For 2002-2008 data, linear regression analysis detected no significant trend in the annual reported age-adjusted incidence of dengue (incidence range: 0.7-3.0 per 1000 population). The incidence declined in 2.7% of the 185 districts studied, was unchanged in 86.2% and increased in 9.6%. The age-specific incidence was highest in infants aged < 1 year and children aged 4-6 years. The incidence was higher during rainy seasons. All four dengue virus (DENV) serotypes were permanently in circulation, though the predominant serotype has alternated between DENV-3 and DENV-2 since 2000. Although larvicide has been distributed in 94 districts since 2002, logistic regression analysis showed no association between the intervention and dengue incidence. The dengue burden remained high among young children in Cambodia, which reflects intense transmission. The national vector control programme appeared to have little impact on disease incidence.