Global Health Action 
INDEPTH NETWORK CAUSE-SPECIFIC MORTALITY
Malaria mortality in Africa and Asia: evidence from
INDEPTH health and demographic surveillance
system sites
P. Kim Streatfield1,2,3, Wasif A. Khan2,3,4, Abbas Bhuiya3,5,6,
Syed M.A. Hanifi3,5,6, Nurul Alam3,7,8, Eric Diboulo3,9,10, Ali Sié3,9,10,
Maurice Yé3,9,10, Yacouba Compaoré3,11,12, Abdramane B. Soura3,11,12,
Bassirou Bonfoh3,13,14, Fabienne Jaeger3,13,15, Eliezer K. Ngoran3,13,16,
Juerg Utzinger3,13,15, Yohannes A. Melaku3,17,18, Afework Mulugeta3,17,18,
Berhe Weldearegawi3,17,18, Pierre Gomez3,19,20, Momodou Jasseh3,19,20,
Abraham Hodgson3,21,22, Abraham Oduro3,21,22, Paul Welaga3,21,22,
John Williams3,21,22, Elizabeth Awini3,23,24,25, Fred N. Binka3,23,25,
Margaret Gyapong3,23,25, Shashi Kant3,26,27, Puneet Misra3,26,27,
Rahul Srivastava3,26,27, Bharat Chaudhary3,28,29, Sanjay Juvekar3,28,29,
Abdul Wahab3,30,31, Siswanto Wilopo3,30,31, Evasius Bauni3,32,33,
George Mochamah3,32,33, Carolyne Ndila3,32,33, Thomas N. Williams3,32,33,34,
Mary J. Hamel3,35,36, Kim A. Lindblade3,35,36, Frank O. Odhiambo3,35,36,
Laurence Slutsker3,35,36, Alex Ezeh3,37,38, Catherine Kyobutungi3,37,38,
Marylene Wamukoya3,37,38, Valérie Delaunay3,39,40, Aldiouma Diallo3,39,40,
Laetitia Douillot3,39,40, Cheikh Sokhna3,39,40, F. Xavier Gómez-Olivé3,41,42,
Chodziwadziwa W. Kabudula3,41,42, Paul Mee3,41,42, Kobus Herbst3,43,44,
Joël Mossong3,43,44,45, Nguyen T.K. Chuc3,46,47, Samuelina S. Arthur3,
Osman A. Sankoh3,48,49*, Marcel Tanner50 and Peter Byass51
1Matlab HDSS, Bangladesh; 2International Centre for Diarrhoeal Disease Research, Bangladesh;
3INDEPTH Network, Accra, Ghana; 4Bandarban HDSS, Bangladesh; 5Chakaria HDSS, Bangladesh;
6Centre for Equity and Health Systems, International Centre for Diarrhoeal Disease Research,
Bangladesh; 7AMK HDSS, Bangladesh; 8Centre for Population, Urbanisation and Climate Change,
International Centre for Diarrhoeal Disease Research, Bangladesh; 9Nouna HDSS, Burkina Faso;
10Nouna Health Research Centre, Nouna, Burkina Faso; 11Ouagadougou HDSS, Burkina Faso;
12Institut Supérieur des Sciences de la Population, Université de Ouagadougou, Burkina Faso;
13Taabo HDSS, Côte d’Ivoire; 14Centre Suisse de Recherches Scientifiques en Côte d’Ivoire, Abidjan,
Côte d’Ivoire; 15Swiss Tropical and Public Health Institute, Basel, Switzerland; 16Université Félix
Houphoët-Boigny, Abidjan, Côte d’Ivoire; 17Kilite-Awlaelo HDSS, Ethiopia; 18Department of Public
Health, College of Health Sciences, Mekelle University, Mekelle, Ethiopia; 19Farafenni HDSS, The
Gambia; 20Medical Research Council, The Gambia Unit, Fajara, The Gambia; 21Navrongo HDSS,
Ghana; 22Navrongo Health Research Centre, Navrongo, Ghana; 23Dodowa HDSS, Ghana; 24Dodowa
Health Research Centre, Dodowa, Ghana; 25School of Public Health, University of Ghana, Legon,
Ghana; 26Ballabgarh HDSS, India; 27All India Institute of Medical Sciences, New Delhi, India; 28Vadu
HDSS, India; 29Vadu Rural Health Program, KEM Hospital Research Centre, Pune, India; 30Purworejo
HDSS, Indonesia; 31Department of Public Health, Universitas Gadjah Mada, Yogyakarta, Indonesia;
32Kilifi HDSS, Kenya; 33KEMRI-Wellcome Trust Research Programme, Kilifi, Kenya; 34Department of
Medicine, Imperial College, St. Mary’s Hospital, London; 35Kisumu HDSS, Kenya; 36KEMRI/CDC
Research and Public Health Collaboration and KEMRI Center for Global Health Research, Kisumu,
Kenya; 37Nairobi HDSS, Kenya; 38African Population and Health Research Center, Nairobi, Kenya;
39Niakhar HDSS, Senegal; 40Institut de Recherche pour le Developpement (IRD), Dakar, Sénégal;
41Agincourt HDSS, South Africa; 42MRC/Wits Rural Public Health and Health Transitions Research
Unit (Agincourt), School of Public Health, Faculty of Health Sciences, University of the Witwatersrand,
Johannesburg, South Africa; 43Africa Centre HDSS, South Africa; 44Africa Centre for Health and
Authors are listed arbitrarily in order of their site code, and alphabetically within each site.
Global Health Action 2014. # 2014 INDEPTH Network. This is an Open Access article distributed under the terms of the Creative Commons CC-BY 4.0 1
License (http://creativecommons.org/licenses/by/4.0/), allowing third parties to copy and redistribute the material in any medium or format and to remix,
transform, and build upon the material for any purpose, even commercially, provided the original work is properly cited and states its license.
Citation: Glob Health Action 2014, 7: 25369 - http://dx.doi.org/10.3402/gha.v7.25369
(page number not for citation purpose)
INDEPTH Network
Population Studies, University of KwaZulu-Natal, Somkhele, KwaZulu-Natal, South Africa; 45National
Health Laboratory, Surveillance & Epidemiology of Infectious Diseases, Dudelange, Luxembourg;
46FilaBavi HDSS, Vietnam; 47Health System Research, Hanoi Medical University, Hanoi, Vietnam;
48School of Public Health, Faculty of Health Sciences, University of the Witwatersrand,
Johannesburg, South Africa; 49Hanoi Medical University, Hanoi, Vietnam; 50WHO Collaborating
Centre for Verbal Autopsy, Umeå Centre for Global Health Research, Umeå University, Umeå,
Sweden; 51MRC/Wits Rural Public Health and Health Transitions Research Unit (Agincourt),
School of Public Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg,
South Africa
Background: Malaria continues to be a major cause of infectious disease mortality in tropical regions.
However, deaths from malaria are most often not individually documented, and as a result overall
understanding of malaria epidemiology is inadequate. INDEPTH Network members maintain population
surveillance in Health and Demographic Surveillance System sites across Africa and Asia, in which individual
deaths are followed up with verbal autopsies.
Objective: To present patterns of malaria mortality determined by verbal autopsy from INDEPTH sites across
Africa and Asia, comparing these findings with other relevant information on malaria in the same regions.
Design: From a database covering 111,910 deaths over 12,204,043 person-years in 22 sites, in which verbal
autopsy data were handled according to the WHO 2012 standard and processed using the InterVA-4 model,
over 6,000 deaths were attributed to malaria. The overall period covered was 19922012, but two-thirds of the
observations related to 20062012. These deaths were analysed by site, time period, age group and sex to
investigate epidemiological differences in malaria mortality.
Results: Rates of malaria mortality varied by 1:10,000 across the sites, with generally low rates in Asia (one
site recording no malaria deaths over 0.5 million person-years) and some of the highest rates in West Africa
(Nouna, Burkina Faso: 2.47 per 1,000 person-years). Childhood malaria mortality rates were strongly
correlated with Malaria Atlas Project estimates of Plasmodium falciparum parasite rates for the same
locations. Adult malaria mortality rates, while lower than corresponding childhood rates, were strongly
correlated with childhood rates at the site level.
Conclusions: The wide variations observed in malaria mortality, which were nevertheless consistent with
various other estimates, suggest that population-based registration of deaths using verbal autopsy is a useful
approach to understanding the details of malaria epidemiology.
Keywords: malaria; Africa; Asia; mortality; INDEPTH Network; verbal autopsy; InterVA
Responsible Editors: Heiko Becher, University of Hamburg, Germany; Nawi Ng, Umeå University, Sweden.
*Correspondence to: Osman A. Sankoh, INDEPTH Network, PO Box KD213, Kanda, Accra, Ghana,
Email: osman.sankoh@indepth-network.org
A Corrigendum has been published for this paper. Please see http://www.globalhealthaction.net/index.php/
gha/article/view/27833
This paper is part of the Special Issue: INDEPTH Network Cause-Specific Mortality. More papers from
this issue can be found at http://www.globalhealthaction.net
Received: 3 July 2014; Revised: 6 September 2014; Accepted: 6 September 2014; Published: 29 October 2014
he epidemiology of malaria in Africa and Asia has and sought to generalise patterns of malaria burden using
Tbeen extensively, but not always systematically, sophisticated modelling techniques (5). Nevertheless,investigated. Many studies have focused on young malaria remains as an important cause of infectious dis-
children’s exposure to the disease (1), and to some extent ease mortality in many parts of Africa, and some areas in
the effects on pregnant women (2), without evaluating Asia and Latin America. WHO’s World Malaria Report
the malaria status of other population sub-groups. Few 2013 suggests that malaria mortality rates fell by more
studies have looked specifically at the impact of malaria than 40% from 2000 to 2012, a period during which there
on older people (3). Many data have been taken from was substantial international investment in malaria con-
heath facilities at various levels and may be influenced by trol (6). However, although malaria transmission has
patterns of health services utilisation rather than clearly successfully been reduced in many former high-incidence
representing malaria patterns within communities (4). settings, few areas have become malaria-free. The need for
Some work has taken whatever data may be available adequate, reliable evidence on malaria mortality in various
2 Citation: Glob Health Action 2014, 7: 25369 - http://dx.doi.org/10.3402/gha.v7.25369
(page number not for citation purpose)
Malaria mortality in Africa and Asia
Ballabgarh, India: Matlab, Bangladesh:
Nouna, Burkina Faso: CSMF 1.74% CSMF 0.005% AMK, Bangladesh:
Niakhar, Senegal: CSMF 25.53% 0.13/1,000 py 0.0003/1,000 py none recorded
CSMF 10.59% 2.47/1,000 py Kilite-Awlaelo, Ethiopia:
0.86/1,000 py FilaBavi, Vietnam:CSMF 2.11%
CSMF 0.37%
Ouagadougou, 0.09/1,000 py Vadu, India: 0.015/1,000 py
Burkina Faso: CSMF 0.06%
CSMF 9.46% 0.003/1,000 py Bandarban, Bangladesh:
0.56/1,000 py Chakaria, Bangladesh: CSMF 4.25%
Kisumu, Kenya: CSMF 0.25% 0.24/1,000 py
Navrongo, Ghana: CSMF 11.61% 0.02/1,000 py
CSMF 4.90% 2.15/1,000 py
0.48/1,000 py
Farafenni, Nairobi, Kenya:
The Gambia: Dodowa, Ghana: CSMF 1.07%
CSMF 8.63% CSMF 10.24% 0.11/1,000 py Purworejo, Indonesia:
0.91/1,000 py 0.78/1,000 py CSMF 2.15%
0.13/1,000 py
Kilifi, Kenya:
Taabo, CSMF 2.87%
Côte d'Ivoire: 0.17/1,000 py
CSMF 12.61%
1.21/1,000 py  
Agincourt, South Africa: Africa Centre, South Africa:
CSMF 1.02% CSMF 0.40%
0.09/1,000 py 0.06/1,000 py
Fig. 1. Map showing participating sites, with agesextime standardised cause-specific mortality fractions and mortality rates
for malaria.
populations therefore remains as important as ever, and data are not designed to form a representative network, each one
at the population level are crucially needed to validate and follows a geographically defined population longitudinally,
understand top-down estimates. systematically recording all death events and undertaking
As is the case for deaths from all diseases, malaria deaths VAs on deaths that occur. Sites with longer time-series may
are generally poorly verified and documented in Africa therefore be able to measure changes over time effectively.
and some parts of Asia. Attributing a death to malaria Our aim is to present the malaria mortality patterns at each
after the event is not easy  in highly endemic areas, acute site, comparing these community-level findings with other
febrile deaths may be likely to be described as malaria and information on malaria in Africa and Asia.
lead to over-attribution, whereas the converse may apply in
settings where malaria is uncommon. It has been suggested Methods
that over-attribution of malaria as a clinical diagnosis in The overall public-domain INDEPTH dataset (10) from
endemic areas may even be dangerous (7). Because most which these malaria-specific analyses are drawn is described
malaria deaths occur in areas not covered by routine death in detail elsewhere (11), with full details of methods used,
certification, verbal autopsy (VA) methods have been used which are also summarised here in Box 1. Briefly, the dataset
in many settings as the only available source of cause of documents 111,910 deaths in 12,204,043 person-years of
death data, but their validity in absolute terms for assign- observation across 22 sites, all processed in a standardised
ing malaria as a cause of death remains open to question. manner. The Karonga site in Malawi did not contribute VAs
Rapid diagnostic tests (RDTs) are becoming increasingly for children, and for that reason is excluded from further
widely used as a basis for malaria treatment decisions, and, analyses here. The InterVA-4 ‘high’ malaria setting was used
where RDT results are known from an illness leading to for all the West African sites, plus the East African sites
death, either positive or negative RDTresults may increase (with the exceptions, on the grounds of high altitude, of
the available VA information and hence the accuracy of Nairobi, Kenya and Kilite-Awlaelo, Ethiopia), on the basis
cause of death attribution. Consequently in the WHO 2012 of local experience. All other sites used the ‘low’ setting; the
VA standard, specific items on a recent positive or negative ‘very low’ setting was not used. The InterVA-4 guideline is
test result were introduced (8). However, it will be some that the ‘high’ setting is appropriate for an expected malaria
time before sufficient VAs are collected which include those cause-specific mortality fraction (CSMF) higher than about
data items to assess their utility as part of the VA process. 1%, though the setting chosen does not result in any great
In this paper, we present malaria-specific mortality dichotomisation of outputs; the clinical equivalent would be
rates derived from standardised VA data in 22 INDEPTH a physician’s knowledge that his/her current case comes from
Network Health and Demographic Surveillance Sites a setting where malaria is more or less likely, irrespective of
(HDSS) across Africa and Asia (9). Although these HDSSs particular symptoms.
Citation: Glob Health Action 2014, 7: 25369 - http://dx.doi.org/10.3402/gha.v7.25369 3
(page number not for citation purpose)
INDEPTH Network
Box 1. Summary of methodology based on the detailed was covered by site-level ethical approvals relating to on-
description in the introductory paper (11) going health and demographic surveillance in those
specific locations. No individual identity or household
Agesextime standardisation location data were included in the secondary data and no
To avoid effects of differences and changes in age specific ethical approvals were required for these pooled
sex structures of populations, mortality fractions analyses.
and rates have been adjusted using the INDEPTH
2013 population standard (12). A weighting factor Results
was calculated for each site, age group, sex and The CSMFs for malaria at each site are shown, together
year category in relation to the standard for the with the population-based malaria-specific mortality
corresponding age group and sex, and incorpo- rate per 1,000 person-years, in Fig. 1. In West African
rated into the overall dataset. This is referred to in sites, malaria CSMF ranged from 4.90% to 25.53%, with
this paper as agesextime standardisation in the malaria-specific standardised mortality rates ranging
contexts where it is used. from 0.48 to 2.47 per 1,000 person-years. In Eastern and
Cause of death assignment Southern Africa, CSMFs were 0.4011.61%, with rates
The InterVA-4 (version 4.02) probabilistic model from 0.06 to 2.15 per 1,000 person-years. In Asia, CSMFs
was used for all the cause of death assignments in were 04.25%, with rates from 0 to 0.24 per 1,000 person-
the overall dataset (13). InterVA-4 is fully compli- years. One site, AMK in Bangladesh, recorded no malaria
ant with the WHO 2012 Verbal Autopsy Standards deaths in over 0.5 million person-years of observation.
and generates causes of death categorised by Table 1 breaks down malaria-specific mortality rates by
ICD-10 groups (14). The data reported here were age group and site. Malaria mortality rates among infants
collected before the WHO 2012 VA standard was varied considerably, from 0 to 1.4 per 1,000 person-years,
available, but were transformed into the WHO with the highest rates not necessarily being in the locations
2012 and InterVA-4 format to optimise cross-site with highest overall malaria mortality. The largest num-
standardisation in cause of death attribution. For a bers of malaria deaths at most sites occurred in the 14
small proportion of deaths VA interviews were year age group, though the highest malaria mortality rate
not successfully completed; a few others contained in that age group was 0.43 per 1,000 person-years at Taabo,
Côte d’Ivoire. Malaria mortality rates in the 514 year age
inadequate information to arrive at a cause of
group were generally lower than the rates for younger
death. InterVA-4 assigns causes of death (max-
children. Similarly, malaria mortality rates among adults
imum 3) with associated likelihoods; thus cases for
were generally lower than those for children, although they
which likely causes did not total 100% were also
tended to increase among the elderly. Figure 2 shows
assigned a residual indeterminate component. This
malaria-specific mortality rates for each site by age group,
served as a means of encapsulating uncertainty in
split into time periods (19921999; 20002005 and 2006
cause of death at the individual level within the
2012), depending on periods when individual sites operated.
overall dataset, as well as accounting for 100% of
Logarithmic scales have been used to visualise both high
every death.
and low levels of malaria mortality while using the same
Overall dataset
scale for each site. For most sites and most periods there
The overall public-domain dataset (10) thus con- were generally U-shaped relationships between malaria
tains between one and four records for each death, mortality rates and age; naturally more random variation
with the sum of likelihoods for each individual was evident in sites with generally low malaria mortality
being unity. Each record includes a specific cause of because of relatively small numbers of cases.
death, its likelihood and its age-sex-time weighting. We undertook a sensitivity analysis to examine the
effects of the ‘high’ and ‘low’ InterVA-4 malaria settings
Deaths assigned to malaria were extracted from the across this large and diverse dataset. Re-running the
overall data set together with data on person-time exposed InterVA-4 model with the ‘high’ and ‘low’settings reversed
by site, year, age and sex. Overall malaria mortality as re- at site level gave the results shown in Fig. 3. Incorrect use of
flected here amounted to a total of 6,330.8 agesextime the ‘high’ setting in low malaria populations appeared to
standardised deaths, to which 8,076 individually recorded result in high relative rates of falsely attributed malaria,
deaths contributed some component of probable malaria although the numbers involved would still be relatively
mortality. As each HDSS covers a total population, rather small at the lowest endemicities. Conversely using the ‘low’
than a sample, uncertainty intervals are not shown. setting in high malaria populations reduced the number of
In this context, all of these data are secondary datasets malaria assignments. Although the rate ratios changed less
derived from primary data collected separately by each in high endemicity settings, the numbers of cases involved
participating site. In all cases the primary data collection would be important with increasing rates.
4 Citation: Glob Health Action 2014, 7: 25369 - http://dx.doi.org/10.3402/gha.v7.25369
(page number not for citation purpose)
Malaria mortality in Africa and Asia
Table 1. Malaria-specific deaths and mortality rates per 1,000 person-years, by age group and site
Age group at death
Country: Site Infant 14 years 514 years 1549 years 5064 years 65 years
Bangladesh: Matlab
Deaths 0.00 0.41 0.00 0.00 0.00 0.00
Rate/1,000 py 0.00 0.00 0.00 0.00 0.00 0.00
Bangladesh: Bandarban
Deaths 0.98 1.00 2.46 3.76 1.47 3.25
Rate/1,000 py 0.79 0.17 0.18 0.11 0.25 1.03
Bangladesh: Chakaria
Deaths 0.43 1.23 1.99 0.00 0.00 0.28
Rate/1,000 py 0.08 0.06 0.03 0.00 0.00 0.03
Bangladesh: AMK
Deaths 0.00 0.00 0.00 0.00 0.00 0.00
Rate/1,000 py 0.00 0.00 0.00 0.00 0.00 0.00
Burkina Faso: Nouna
Deaths 507.76 859.38 140.73 108.93 76.24 287.96
Rate/1,000 py 0.75 0.20 0.11 0.07 0.42 0.70
Burkina Faso: Ouagadougou
Deaths 19.48 68.03 17.90 8.56 2.72 4.43
Rate/1,000 py 0.72 0.19 0.10 0.04 0.24 0.90
Côte d’Ivoire: Taabo
Deaths 22.74 63.22 8.24 22.79 2.99 8.56
Rate/1,000 py 1.42 0.43 0.14 0.11 0.43 1.35
Ethiopia: Kilite-Awlaelo
Deaths 1.83 2.22 1.22 1.00 0.70 4.93
Rate/1,000 py 0.57 0.13 0.03 0.02 0.06 0.41
The Gambia: Farafenni
Deaths 35.28 113.11 38.72 43.35 19.85 43.46
Rate/1,000 py 1.06 0.33 0.15 0.09 0.55 1.15
Ghana: Navrongo
Deaths 121.42 283.42 39.50 12.34 9.45 32.61
Rate/1,000 py 0.42 0.10 0.04 0.02 0.06 0.14
Ghana: Dodowa
Deaths 4.74 49.53 28.83 154.67 45.91 138.68
Rate/1,000 py 0.28 0.14 0.06 0.03 0.21 0.26
India: Ballabgarh
Deaths 5.41 17.89 3.64 4.25 0.00 5.38
Rate/1,000 py 0.45 0.20 0.04 0.02 0.00 0.26
India: Vadu
Deaths 0.00 0.00 0.00 0.91 0.00 0.00
Rate/1,000 py 0.00 0.00 0.00 0.01 0.00 0.00
Indonesia: Purworejo
Deaths 2.42 3.13 2.00 4.34 5.64 13.50
Rate/1,000 py 0.85 0.19 0.05 0.02 0.14 0.19
Kenya: Kilifi
Deaths 38.53 90.21 36.03 14.84 3.97 12.72
Rate/1,000 py 0.17 0.04 0.02 0.01 0.05 0.18
Kenya: Kisumu
Deaths 672.20 1177.46 177.79 321.30 99.16 181.89
Rate/1,000 py 0.38 0.10 0.04 0.03 0.14 0.17
Kenya: Nairobi
Deaths 16.42 16.50 4.59 7.23 3.91 0.26
Rate/1,000 py 0.80 0.18 0.04 0.02 0.15 0.05
Citation: Glob Health Action 2014, 7: 25369 - http://dx.doi.org/10.3402/gha.v7.25369 5
(page number not for citation purpose)
INDEPTH Network
Table 1 (Continued )
Age group at death
Country: Site Infant 14 years 514 years 1549 years 5064 years 65 years
Senegal: Niakhar
Deaths 23.25 126.45 21.32 16.31 4.04 28.49
Rate/1,000 py 1.05 0.33 0.15 0.09 0.22 0.68
South Africa: Africa Centre
Deaths 8.67 13.84 7.37 9.44 1.53 7.22
Rate/1,000 py 0.33 0.12 0.03 0.02 0.03 0.17
South Africa: Agincourt
Deaths 12.45 29.39 19.45 54.40 7.56 4.93
Rate/1,000 py 0.28 0.14 0.05 0.03 0.08 0.08
Vietnam: FilaBavi
Deaths 0.00 0.00 0.00 0.55 0.00 2.46
Rate/1,000 py 0.00 0.00 0.00 0.01 0.00 0.14
Table 2 shows estimates of malaria-specific mortality years in 2010 (15). Since all the INDEPTH HDSSs cover
rates for the countries with INDEPTH sites reporting defined small areas, it was possible to extract a PfPR
here, for the under-5 and 5-plus age groups for compar- value for each endemic site from the MAP data. Where
ison with other sources of malaria mortality estimates. sites covered more than one cell of the MAP surface, all
INDEPTH estimates for countries with multiple sites the cells relating to the site were averaged. Data were
were derived as population-weighted average rates. available for 14 sites with childhood malaria mortality
The Malaria Atlas Project (MAP) produced geo- data; data were not available for seven sites in Vietnam,
referenced estimates of Plasmodium falciparum parasite India, Bangladesh and Ethiopia, presumably because of
rates (PfPR) across endemic areas for children aged 210 very low or uncertain endemicity. Figure 4 shows the
Bangladesh: Matlab Bangladesh: Bandarban Bangladesh: Chakaria Burkina Faso: Nouna Burkina Faso: Ouagadougou
10
1
.1
.01
Cote d'Ivoire: Taabo Ethiopia: Kilite Awlaelo Gambia: Farafenni Ghana: Navrongo Ghana: Dodowa
10
1
.1
.01
India: Ballabgarh India: Vadu Indonesia: Purworejo Kenya: Kilifi Kenya: Kisumu
10
1
.1
.01
Kenya: Nairobi Senegal: Niakhar South Africa: Agincourt South Africa: Africa Centre Vietnam: FilaBavi
10
1
.1
.01
1992-99 2000-05 2006-12
Fig. 2. Malaria mortality rates by site, age group and period at 20 INDEPTH Network sites.
6 Citation: Glob Health Action 2014, 7: 25369 - http://dx.doi.org/10.3402/gha.v7.25369
(page number not for citation purpose)
malaria mortality rate /1,000 p-y
infants
1-4 y
5-14 y
15-49 y
50-64 y
65+ y
infants
1-4 y
5-14 y
15-49 y
50-64 y
65+ y
infants
1-4 y
5-14 y
15-49 y
50-64 y
65+ y
infants
1-4 y
5-14 y
15-49 y
50-64 y
65+ y
infants
1-4 y
5-14 y
15-49 y
50-64 y
65+ y
Malaria mortality in Africa and Asia
0.018 to 2.47 per 1,000 person-years. Figure 5 shows the
500
low malaria site high malaria site
correlation between adult and child malaria rates for
these 17 sites, shown on logarithmic scales for clarity. As
100
50 expected, the sites from West Africa dominate the top-
right quadrant, together with Kisumu, on the shores of
10 Lake Victoria in Kenya. Other African and Asian sites
5 largely occupy the lower-left quadrant, with the Chakaria
site in Bangladesh showing very low malaria mortality
1 rates for both adults and children. The per-site correlation
.5 (represented by the line in Fig. 5) between agesextime
standardised adult and child malaria mortality rates
.1 was highly significant (R20.65, p0.0001), fitting the
.001 .005 .01 .05 .1 .5 1 5 10 50 relationship:
malaria cause-specific mortality fraction (%)
Adult malaria mortality rate
Fig. 3. Sensitivity analysis showing the effect of choosing ¼ e½ðchild malaria mortality rate1:002Þ1:157
the ‘wrong’ malaria endemicity setting (‘high’ and ‘low’
reversed) in processing VA data using the InterVA-4 model,
by site. Discussion
correlation between per-site malaria mortality rates for These results represent widely-based evidence on malaria
the 114 year age group as determined by InterVA-4 and mortality, which has not previously been documented at
the MAP PfPR values for the same geographic locations. the population level on this scale, using standardised
The line in Fig. 4 represents a highly significant correla- methods. The interpretation of findings at individual sites
tion (R20.69, p0.002), fitting the relationship: depends on local characteristics (1636). Two sites,
Ouagadougou in Burkina Faso and Nairobi in Kenya,
Malaria mortality rate ¼ e½ðPfPR0:6274Þþ0:7023 followed urban populations and recorded lower levels
of malaria than some of their rural neighbours. Bandarban
An important area of uncertainty in malaria epide- in Bangladesh is located in a frontier zone close to the
miology is the ratio of malaria-specific mortality rates Myanmar border, which may explain higher rates of
between children and adults. Seventeen sites recorded malaria compared with other sites in Bangladesh; this is
malaria deaths in both under-15 and over-15 year age consistent with WHO malaria mapping for Bangladesh
categories. Apart from one outlier (Dodowa, Ghana, (37). The very low overall levels of malaria mortality in
where the malaria-specific mortality rate ratio for over-15: Bangladesh are not only consistent with expectations, but
under-15 age categories was 2.5), in the remaining 16 form an important part of these analyses in that they
sites the malaria-specific mortality rate ratios for over- suggest our methods are capable of assigning malaria as
15:under-15 age categories were in the range 0.05 to 0.82, a cause of death with high specificity. Kisumu in Kenya is
while overall malaria-specific mortality rates ranged from located on the shores of Lake Victoria, in an area known to
Table 2. Within-country estimates of malaria-specific mortality rates derived from WHO/CHERG (42, 43), IHME (44)
compared with population-weighted average country rates from INDEPTH sites
WHO/CHERG IHME INDEPTH
Country Under 5 years 5 years and over Under 5 years 5 years and over Under 5 years 5 years and over
Bangladesh 0.05 0.004 0.05 0.02 0.02 0.006
Burkina Faso 9.94 0.15 8.34 1.19 6.08 1.00
Côte d’Ivoire 6.92 0.13 5.49 0.92 5.04 0.57
Ethiopia 0.38 ? 1.86 0.36 0.32 0.06
Ghana 2.90 0.11 2.99 0.58 2.40 0.30
India 0.06 0.02 0.04 0.04 0.53 0.03
Indonesia 0.11 0.03 0.80 0.04 0.74 0.08
Kenya 0.47 ? 1.86 0.44 3.35 0.31
Senegal 2.39 0.05 1.96 0.59 2.95 0.39
The Gambia 4.31 0.14 5.55 0.46 2.34 0.61
Vietnam 0.004 0.000 0.003 0.013 0 0.015
Citation: Glob Health Action 2014, 7: 25369 - http://dx.doi.org/10.3402/gha.v7.25369 7
(page number not for citation purpose)
malaria mortality ratio for ‘wrong’ category
INDEPTH Network
(basically diagnoses based on parasitaemia and fever)
13 14 (43, 44). Unfortunately however there were no data on
12 the presence or absence of malaria parasitaemia in cases
11
9 10 attributed to other causes, nor on parasite species for
8
the malaria cases. Most (64%) of the adult malaria deaths
.5 7 in this series came from hospitals in India, while the
childhood cases were mainly from Dar-es-Salaam city
5 6
3 4 (88%), though it should be noted that this study did not
2 aim to represent any particular population. Only 25% of
1 the malaria deaths mentioned the word ‘malaria’ in the
.05 open-ended part of the subsequent VA interview (which
did not contain any specific question on malaria), while
.005 .05 .5 69% of malaria case VAs for adults and 54% for children
Plasmodium falciparum parasite rate 2–10 years
reported severe respiratory symptoms. This may partly
reflect the tertiary facility settings of these cases, where
Fig. 4. Scatter plot of agesextime standardised InterVA some cases may have progressed to respiratory complica-
malaria mortality rates per 1,000 person-years for children
tions of malaria (45), or VA respondents may simply
aged 114 years versus Plasmodium falciparum parasite rate
 have noted hospital treatment for breathing difficultiesdata for children aged 2 10 years, for 14 INDEPTH HDSS
sites reporting malaria mortality which also had geo- in the trajectory towards death (46). Consequently, the
referenced parasite rate data for 2010 in the Malaria Atlas PHMRC dataset is not particularly useful in terms of
Project (15). Line shows correlation, R20.56. (1. Africa validating VA in general for malaria.
Centre, South Africa; 2. Agincourt, South Africa; 3. Nairobi, The WHO 2012 VA standard (8) includes indicators
Kenya; 4. Purworejo, Indonesia; 5. Bandarban, Bangladesh; relating to positive or negative malaria test results during
6. Kilifi, Kenya; 7. Dodowa, Ghana; 8. Navrongo, Ghana; 9. the final illness, as well as other relevant symptomatic
Farafenni, The Gambia; 10. Ouagadougou, Burkina Faso; parameters. However, because these data were collected
11. Niakhar, Senegal; 12. Taabo, Côte d’Ivoire; 13. Kisumu, before the WHO 2012 standard was directly implemented
Kenya; 14. Nouna, Burkina Faso).
for data capture, specific responses for these indicators
have higher malaria transmission than most other parts of were missing in over 90% of cases. However, a previous
the country, such as the coastal area around Kilifi (38). sensitivity analysis showed that InterVA-4 was generally
Kilite-Awlaelo is located in the Ethiopian highlands, at an relatively robust in relation to missing data items (46).
altitude around 2,000 m above sea level, at which malaria Nevertheless, the malaria-specific outputs here, using the
is typically unstable and epidemic in nature. The two WHO 2012 standard and the corresponding InterVA-4
model, show huge differences between locations and age
South African sites are located on the margins of malaria
groups, as might be expected. These plausible patterns
transmission, and some of the relatively few cases that
suggest that there may be at least a reasonable degree of
occurred may reflect travel, for example to neighbouring
validity in terms of InterVA-4’s assignment of malaria
Mozambique (39).
deaths. The application of a standard probabilistic model
The validity of VA cause of death assignment specifi-
such as InterVA-4 at least guarantees that inter-site
cally for malaria is difficult to determine precisely. The
differences are reflections of variations in the VA source
InterVA model has previously been used in a WHO study
data (13). If, alternatively, physicians at each site were used
of malaria treatment, showing a significant difference in
to assign cause of death, it would be easy for inter- and
malaria-specific mortality following a treatment delivery
intra-physician variations to contribute to apparent differ-
intervention (40). A review of VA methodological valida-
ences between sites and over time. This is the first time
tions in relation to hospital data found some examples such a large VA dataset relating to malaria has been
relating to malaria, but a generalisable formal validation compiled that spans complete populations in Africa and
for malaria mortality remains elusive (41). In principle Asia, covers a wide spectrum of endemicity, and uses
validity of VA methods for malaria as a cause of death standardised cause of death attribution. The sensitivity
could be established in a large VA dataset from an en- analysis reported here is important in justifying the design
demic area which included systematic parasitaemia test- assumptions in InterVA-4 that require local settings for
ing across all age groups. Operationally this could be malaria (and HIV) endemicity. The crossover region
incorporated in a minimally-invasive autopsy approach between the ‘high’ and ‘low’ settings, recommended at
(42). The Population Health Metrics Research Consor- 1%, has been seen as a difficult concept by some InterVA-4
tium (PHMRC) collected a ‘gold standard’ VA dataset users. However, the sensitivity analysis shown in Fig. 3
of 12,530 tertiary facility cases, which contained 216 cases suggests that this setting is both important and appro-
meeting the PHMRC definitions of a malaria death priate, and analogous to a clinician’s local knowledge
8 Citation: Glob Health Action 2014, 7: 25369 - http://dx.doi.org/10.3402/gha.v7.25369
(page number not for citation purpose)
malaria mortality rate 1–14 years (per 1,000 py)
Malaria mortality in Africa and Asia
available, modelled to represent the national situation as
far as is possible, and may include facility and community
sources, as well as diverse methods of cause of death
.5
assignment. The INDEPTH numbers come from the
specific HDSS populations as described above, which are
not intended to be nationally representative, but which
.05
are collected and processed in a standardised way across
the various countries represented. In the case of Kenya,
.005 for example, the higher INDEPTH rate for under-5s
reflects high malaria mortality in the Kisumu area. While
.05 .5 5 it would be inappropriate to over-interpret comparisons
malaria mortality rate <15 years (per 1,000 py)
of the rates presented in Table 2, it is clear that there are
western Africa eastern and southern Africa Asia substantial similarities between all three sources. IHME
and INDEPTH figures tend towards higher rates for the
Fig. 5. Scatter plot of agesextime standardised malaria 5-plus age group, though the reasons for this are not
mortality rates per 1,000 person-years for adults (15 years clear. In INDEPTH’s case, InterVA-4 appears to be
and over) and children (under 15 years), for 17 INDEPTH detecting a number of acute febrile illnesses among older
HDSS sites reporting malaria mortality among adults and people and attributing them as malaria; but there is
children. Line shows correlation, R20.65.
absolutely no associated biomedical evidence that these
deaths are indeed directly due to malaria.
of malaria endemicity, irrespective of the history and
However, Fig. 4 showed a strong correlation between
symptomatology of the next patient.
InterVA-4 estimates of childhood malaria mortality and
There are other major pieces of work describing ma-
geo-referenced parasite prevalence estimates from MAP
laria mortality in Africa and Asia, using totally different
(15). There are three possible consequences to consider.
methods, with which these findings can be compared and
Firstly, if one accepts the validity of the parasite prevalence
contrasted. The WHO World Malaria Report 2013 (6) sets
estimates, then the observed correlation suggests that for
out WHO’s most recent compilation of malaria reports children (notwithstanding the slightly different age groups
from its member countries, together with associated data of 114 years for mortality and 210 years for parasite
estimates in the WHO Global Health Observatory (47) prevalence), InterVA-4 is capturing a directly related
and, for children, from the Child Epidemiology Reference pattern of malaria mortality, across a 100-fold range of
Group (CHERG) (48). The Institute of Health Metrics endemicity. The second option is to accept the validity of
and Evaluation (IHME) has also published global and the InterVA-4 malaria mortality findings reported here,
country estimates of malaria mortality covering a similar in which case they add veracity to the parasite prevalence
time period, based on mathematical modelling of available estimates. Thirdly, if both the InterVA-4 and MAP findings
data (49). Both of these sources take the approach of are considered to be reasonably valid, then this correlation
gathering whatever malaria mortality data may be avail- establishes an interesting relationship between childhood
able across all endemic areas (to which this dataset now parasite prevalence and malaria mortality. This relation-
adds), and then making best estimates to fill in the con- ship seems to hold over a wide range of sites, even though
siderable gaps in the available data. it might be reasonable to presume that local factors such
Table 2 enables comparisons of malaria-specific mor- as the effectiveness of treatment and control programmes
tality rates for the countries with INDEPTH sites re- could also play a part. Previous work (among hospital-
porting here, for the under-5 and 5-plus age groups, with ised cases) in Tanzania showed relationships between
other sources of estimates. South Africa is not included age, transmission intensity and malaria mortality (50).
because the majority of the country is malaria-free, while Another modelling study sought to establish relationships
the two INDEPTH sites represent marginal transmission between malaria transmission and mortality, though
areas, making national estimates difficult to interpret. starting from a rather disparate group of datasets (51).
WHO and CHERG publish separate data estimates Figure 5 showed a strong correlation between InterVA-4
for all-age malaria deaths and under-5 malaria deaths, adult and childhood malaria mortality rates at the site
respectively; while these are largely congruent, allowing level. If InterVA-4 were generally misclassifying a wide
the calculation of 5-plus deaths, for Kenya and Ethiopia range of acute adult febrile illnesses as malaria, this would
the number of under-5 deaths exceeded total deaths, so not be the expected pattern. If there were appreciable
that no rate could be calculated for the 5-plus age group. misclassification, the so-called ‘malaria’ deaths in adults
Comparisons between these three data sources have to be might be expected to occur at a rate largely independent
interpreted with care. The WHO/CHERG and IHME of childhood malaria mortality, in the absence of any
numbers come from estimates based on such data as are hypothesis as to other causes of acute adult febrile
Citation: Glob Health Action 2014, 7: 25369 - http://dx.doi.org/10.3402/gha.v7.25369 9
(page number not for citation purpose)
malaria mortality rate 15+ years (per 1,000 py)
INDEPTH Network
mortality that happened to correlate geographically with the Governments of Australia, Bangladesh, Canada, Sweden and the
childhood malaria. However, there were clearly much UK for providing core/unrestricted support. The Ouagadougou site
acknowledges the Wellcome Trust for its financial support to the
higher rates of what InterVA-4 was calling ‘malaria’
Ouagadougou HDSS (grant number WT081993MA). The Kilite
among adults in West Africa, where malaria transmission Awlaelo HDSS is supported by the US Centers for Disease Control
is known to be the highest in the world. A more detailed and Prevention (CDC) and the Ethiopian Public Health Association
analysis of malaria mortality by age from the Kisumu site (EPHA), in accordance with the EPHA-CDC Cooperative Agree-
in Kenya showed complex and changing relationships ment No.5U22/PS022179_10 and Mekelle University, though these
findings do not necessarily represent the funders’ official views. The
between malaria mortality and age (52). Because malaria
Farafenni HDSS is supported by the UK Medical Research Council.
surveillance among older people has generally not been The Kilifi HDSS is supported through core support to the KEMRI-
given high priority, there appears to be a need for further Wellcome Trust Major Overseas Programme from the Wellcome
population-based research to further resolve this question. Trust. TNW is supported by a Senior Fellowship (091758) and CN
The public availability of these malaria mortality through a Strategic Award (084538) from the Wellcome Trust. This
data creates interesting opportunities for further analyses. paper is published with permission from the Director of KEMRI.
The Kisumu site wishes to acknowledge the contribution of the late
Apart from contributing to the overall body of malaria
Dr. Kubaje Adazu to the development of KEMRI/CDC HDSS,
mortality data, there are several other ways in which they which was implemented and continues to be supported through a
may be specifically useful. While one can debate the cooperative agreement between KEMRI and CDC. The Nairobi
generalisability of HDSS sites (53), the cross-site relation- Urban Health and Demographic Surveillance System (NUHDSS),
ships established here between gridded parasite preva- Kenya, since its inception has received support from the Rockefeller
Foundation (USA), the Welcome Trust (UK), the William and
lence data and childhood malaria mortality, and between
Flora Hewlett Foundation (USA), Comic Relief (UK), the Swedish
child and adult malaria mortality rates, could well be International Development Cooperation Agency (SIDA) and the
incorporated into wider estimations of malaria mortality. Bill and Melinda Gates Foundation (USA). The Agincourt site
acknowledges that the School of Public Health and Faculty of
Conclusions Health Sciences, University of the Witwatersrand, and the Medical
Research Council, South Africa, have provided vital support since
Measuring malaria mortality effectively continues to be a inception of the Agincourt HDSS. Core funding has been provided
global problem. As remarked in the context of malaria by the Wellcome Trust, UK (Grants 058893/Z/99/A; 069683/Z/02/Z;
transmission modelling (54), malaria mortality events 085477/Z/08/Z) with contributions from the National Institute on
frequently fall under the radar of health information Aging of the NIH, William and Flora Hewlett Foundation, and
systems. The data presented here, from a wide range of Andrew W Mellon Foundation, USA.
INDEPTH HDSSs across Africa and Asia, demonstrate
the value of detailed longitudinal surveillance in defined Conflict of interest and funding
populations, rather than relying on more disparate
The authors have not received any funding or benefits from
sources. VA may not be an ideal tool for tracking malaria,
industry or elsewhere to conduct this study.
but nevertheless the malaria-specific mortality rate esti-
mates obtained here using the WHO 2012 standard and the
InterVA-4 model closely correspond to other sources of References
estimates, despite the 1:10,000 range in the magnitude of
rates measured using the same methods in different settings. 1. Fink G, Olgiati A, Hawela M, Miller JM, Matafwali B.
Association between early childhood exposure to malaria and
More widespread use of these population-based approaches
children’s pre-school development: evidence from the Zambia
would add considerably to global understanding of malaria, early childhood development project. Malar J 2013; 12: 12.
and thereby inform control and elimination programmes. 2. Amegah AK, Damptey OK, Sarpong GA, Duah E,
Vervoorn DJ, Jaakkola JJK. Malaria infection, poor nutrition
Acknowledgements and indoor air pollution mediate socioeconomic differences in
adverse pregnancy outcome in Cape Coast, Ghana. PLoS ONE
We are grateful to all the residents of INDEPTH HDSS sites who 2013; 8: e69181.
have contributed personal information to this mortality dataset, and 3. Sankoh OA, Kynast-Wolf G, Kouyate B, Becher H. Patterns of
to the field staff who undertook so many verbal autopsy interviews adult and old-age mortality in rural Burkina Faso. J Public
and data management staff who handled the data at every partici- Health Med 2003; 25: 3726.
pating site. INDEPTH acknowledges all the site scientists who have 4. Okiro EA, Kazembe LN, Kabaria CW, Ligomeka J, Noor AM,
participated in bringing this work together, and who participated Ali D, et al. Childhood malaria admission rates to four
in analysis workshops in Ghana, Belgium, Thailand and the United hospitals in Malawi between 2000 and 2010. PLoS ONE 2013;
Kingdom. The INDEPTH Network is grateful for core funding 8: e62214.
from Sida, the Wellcome Trust, and the William & Flora Hewlett 5. Griffin JT, Ferguson NM, Ghani AC. Estimates of the changing
Foundation. The Umeå Centre for Global Health Research is core age-burden of Plasmodium falciparum malaria disease in
funded by Forte, the Swedish Research Council for Health, Working sub-Saharan Africa. Nat Commun 2014; 5: 3136.
Life and Welfare (grant 2006-1512). PB’s residency at the University 6. World Health Organization. World Malaria Report 2013. Geneva:
of the Witwatersrand Rural Knowledge Hub to analyse and draft World Health Organization; 2013. Available from: http://www.
these results was supported by the European Community Marie who.int/iris/bitstream/10665/97008/1/9789241564694_eng.pdf [cited
Curie Actions IPHTRE project (no. 295168). icddr,b is thankful to 25 February 2014].
10 Citation: Glob Health Action 2014, 7: 25369 - http://dx.doi.org/10.3402/gha.v7.25369
(page number not for citation purpose)
Malaria mortality in Africa and Asia
7. Reyburn H, Mbatia R, Drakeley C, Carneiro I, Mwakasungula using population-based verbal autopsy method in Kilite Awlaelo
E, Mwerinde O, et al. Overdiagnosis of malaria in patients with surveillance site. Health Policy Plan 2013; 28: 8918.
severe febrile illness in Tanzania: a prospective study. BMJ 2004; 24. Oduro AR, Wak G, Azongo D, Debpuur C, Wontuo P,
329: 1212. Kondayire F, et al. Profile: the Navrongo Health and Demo-
8. Leitao J, Chandramohan D, Byass P, Jakob R, Bundhamcharoen graphic Surveillance System. Int J Epidemiol 2012; 41:
K, Choprapawon C, et al. Revising the WHO verbal autopsy 96876.
instrument to facilitate routine cause-of-death monitoring. Glob 25. Gyapong M, Sarpong D, Awini E, Manyeh AK, Tei D,
Health Action 2013; 6: 21518, http://dx.doi.org/10.3402/gha.v6i0. Odonkor G, et al. Profile: the Dodowa Health and Demo-
21518 graphic Surveillance System. Int J Epidemiol 2013; 42: 168696.
9. Sankoh O, Byass P. The INDEPTH Network: filling vital gaps 26. Jasseh M, Webb EL, Jaffar S, Howie S, Townend J, Smith PG,
in global epidemiology. Int J Epidemiol 2012; 41: 57988. et al. Reaching Millennium Development Goal 4  the Gambia.
10. INDEPTH Network. INDEPTH Network Cause-Specific Trop Med Int Health 2011; 16: 131425.
Mortality  Release 2014. Oct 2014. Provided by the INDEPTH 27. Ng N, Hakimi M, Santosa A, Byass P, Wilopo SA, Wall S. Is
Network Data Repository. www.indepth-network.org. doi: 10. self-rated health an independent index for mortality among
7796/INDEPTH.GH003.COD2014.v1. older people in Indonesia? PLoS One 2012; 7: e35308.
11. Streatfield PK, Khan WA, Bhuiya A, Alam N, Sie A, Soura 28. Kant S, Misra P, Gupta S, Goswami K, Krishnan A,
AB, et al. Cause-specific mortality in Africa and Asia: evidence Nongkynrih B, et al. Profile: the Ballabgarh Health and
from INDEPTH Health and Demographic Surveillance System Demographic Surveillance System (CRHSP-AIIMS). Int J
sites. Glob Health Action 2014; 7: 25362, http://dx.doi.org/10. Epidemiol 2013; 42: 37644.
3402/gha.v7.25362 29. Hirve S, Juvekar S, Sambhudas S, Lele P, Blomstedt Y, Wall S,
12. Sankoh O, Sharrow D, Herbst K, Kabudula CW, Alam N, et al. Does self-rated health predict death in adults aged 50 years
Kant S, et al. The INDEPTH standard population for low- and and above in India? Evidence from a rural population under
middle-income countries, 2013. Glob Health Action 2014; 7: health and demographic surveillance. Int J Epidemiol 2012; 41:
23286, http://dx.doi.org/10.3402/gha.v7.23286 171927.
13. Byass P, Chandramohan D, Clark SJ, D’Ambruoso L, 30. Scott JA, Bauni E, Moisi JC, Ojal J, Gatakaa H, Nyundo C,
et al. Profile: The Kilifi Health and Demographic Surveillance
Fottrell E, Graham WJ, et al. Strengthening standardised
System (KHDSS). Int J Epidemiol 2012; 41: 6507.
interpretation of verbal autopsy data: the new InterVA-4 tool.
31. Odhiambo FO, Laserson KF, Sewe M, Hamel MJ, Feikin DR,
Glob Health Action 2012; 5: 19281, http://dx.doi.org/10.3402/
Adazu K, et al. Profile: the KEMRI/CDC Health and Demo-
gha.v5i0.19281
graphic Surveillance System  Western Kenya. Int J Epidemiol
14. World Health Organization. Verbal Autopsy Standards: the
2012; 41: 97787.
2012 WHO Verbal Autopsy Instrument. Geneva: WHO; 2012.
32. Oti SO, Mutua M, Mgomella GS, Egondi T, Ezeh A,
Available from: http://www.who.int/healthinfo/statistics/WHO_
Kyobutungi C. HIV mortality in urban slums of Nairobi,
VA_2012_RC1_Instrument.pdf [cited 25 February 2014].
Kenya 20032010: a period effect analysis. BMC Public Health
15. Gething PW, Patil AP, Smith DL, Guerra CA, Elyazar IR,
2013; 13: 588.
Johnston GL, et al. A new world malaria map: Plasmodium
33. Delaunay V, Douillot L, Diallo A, Dione D, Trape JF,
falciparum endemicity in 2010. Malar J 2011; 10: 378.
Medianikov O, et al. Profile: the Niakhar Health and Demo-
16. Razzaque A, Nahar L, Akter Khanam M, Streatfield PK.
graphic Surveillance System. Int J Epidemiol 2013; 42: 100211.
Socio-demographic differentials of adult health indicators in
34. Kahn K, Collinson MA, Gómez-Olivé FX, Mokoena O,
Matlab, Bangladesh: self-rated health, health state, quality of
Twine R, Mee P, et al. Profile: Agincourt health and socio-
life and disability level. Glob Health Action 2010; 3: 4618, demographic surveillance system. Int J Epidemiol 2012; 41:
http://dx.doi.org/10.3402/gha.v3i0.4618 9881001.
17. INDEPTH Network. Bandarban HDSS. Available from: http:// 35. Herbst AJ, Mafojane T, Newell ML. Verbal autopsy-based
www.indepth-network.org/Profiles/Bandarban HDSS.pdf. cause-specific mortality trends in rural KwaZulu-Natal, South
18. Hanifi MA, Mamun AA, Paul A, Hasan SA, Hoque S, Sharmin Africa, 20002009. Popul Health Metr 2011; 9: 47.
S, et al. Profile: the Chakaria Health and Demographic 36. Huong DL, Minh HV, Vos T, Janlert U, Van DD, Byass P.
Surveillance System. Int J Epidemiol 2012; 41: 66775. Burden of premature mortality in rural Vietnam from 19992003:
19. Lindeboom W, Das SC, Ashraf A. Health and Demographic analyses from a Demographic Surveillance Site. Popul Health
Surveillance Report 2009  Abhoynagar and Mirsarai. Dhaka, Metr 2006; 4: 9.
Bangladesh: ICDDR,B; 2011. 37. World Health Organization. Malaria country profiles: Bangladesh.
20. Sié A, Louis VR, Gbangou A, Müller O, Niamba L, Stieglbauer Geneva: WHO; 2013. Available from: http://www.who.int/
G, et al. The Health and Demographic Surveillance System malaria/publications/country-profiles/profile_bgd_en.pdf [cited
(HDSS) in Nouna, Burkina Faso, 19932007. Glob Health 25 February 2014].
Action 2010; 3: 5284, http://dx.doi.org/10.3402/gha.v3i0.5284 38. World Health Organization. Malaria country profiles: Kenya.
21. Rossier C, Soura A, Baya B, Compaoré G, Dabiré B, Geneva: WHO; 2013. Available from: http://www.who.int/
Dos Santos S, et al. Profile: the Ouagadougou Health and malaria/publications/country-profiles/profile_ken_en.pdf [cited
Demographic Surveillance System. Int J Epidemiol 2012; 41: 25 February 2014].
65866. 39. World Health Organization. Malaria country profiles: South
22. Kouadio MK, Righetti AA, Abé NN, Wegmüller R, Weiss MG, Africa. Geneva: WHO; 2013. Available from: http://www.who.int/
N’goran EK, et al. Local concepts of anemia-related ill- malaria/publications/country-profiles/profile_zaf_en.pdf [cited 25
nesses and public health implications in the Taabo health February 2014].
demographic surveillance system, Côte d’Ivoire. BMC Hematol 40. Lemma H, Byass P, Desta A, Bosman A, Constanzo G, Toma
2013; 13: 5. L, et al. Deploying artemether-lumefantrine with rapid testing
23. Weldearegawi B, Ashebir Y, Gebeye E, Gebregziabiher T, in Ethiopian communities: impact on malaria morbidity,
Yohannes M, Mussa S, et al. Emerging chronic non-communicable mortality and healthcare resources. Trop Med Int Health
diseases in rural communities of Northern Ethiopia: evidence 2010; 15: 24150.
Citation: Glob Health Action 2014, 7: 25369 - http://dx.doi.org/10.3402/gha.v7.25369 11
(page number not for citation purpose)
INDEPTH Network
41. Leitao J, Desai N, Aleksandrowicz L, Byass P, Miasnikof P, 48. Child Epidemiology Reference Group. Estimated numbers of
Tollman S, et al. Comparison of physician-certified verbal deaths by cause in children younger than 5 years by WHO
autopsy with computer-coded verbal autopsy for cause of death region and country, 2010. Available from: http://cherg.org/
assignment in hospitalized patients in low-and middle-income publications/Country-level-child-deaths-2000-2010.xls [cited 25
countries: systematic review. BMC Med 2014; 12: 22. February 2014].
42. Mudenda V, Lucas S, Shibemba A, O’Grady J, Bates M, Kapata 49. Murray CJL, Rosenfeld LC, Lim SS, Andrews KG,
N, et al. Tuberculosis and tuberculosis/HIV/AIDS-associated Foreman KJ, Haring D, et al. Global malaria mortality be-
mortality in Africa: the urgent need to expand and invest in tween 1980 and 2010: a systematic analysis. Lancet 2012; 379:
routine and research autopsies. J Infect Dis 2012; 205(Suppl 2): 41331.
S3406. 50. Reyburn H, Mbatia R, Drakeley C, Bruce J, Carniero I, Olomi
43. Murray CJ, Lopez AD, Black R, Ahuja R, Ali SM, Baqui A, R, et al. Association of transmission intensity and age with
et al. Population Health Metrics Research Consortium gold clinical manifestations and case fatality of severe Plasmodium
standard verbal autopsy validation study: design, implementa- falciparum malaria. JAMA 2005; 293: 146170.
tion, and development of analysis datasets. Popul Health Metr 51. Ross A, Maire N, Molineaux L, Smith T. An epidemio-
2011; 9: 27. logic model of severe morbidity and mortality caused by
44. Population Health Metrics Research Consortium. PHMRC Plasmodium falciparum. Am J Trop Med Hyg 2006; 75(Suppl 2):
Gold Standard Verbal Autopsy Data 20052011. Available 6373.
from: http://ghdx.healthmetricsandevaluation.org/search/site/ 52. Desai M, Buff AM, Khagayi S, Byass P, Amek N, van Eijk A,
phmrc [cited 25 February 2014]. et al. Age-specific malaria mortality rates in the KEMRI/CDC
45. Byass P, Campbell H, O’Dempsey TJ, Greenwood BM. Coin- Health and Demographic Surveillance System in Western
cidence of malaria parasitaemia and abnormal chest X-ray Kenya, 20032010. PLoS ONE 2014; 9: e106197.
findings in young Gambian children. J Trop Med Hyg 1991; 94: 53. Byass P, Sankoh O, Tollman SM, Högberg U, Wall S. Lessons
223. from history for designing and validating epidemiological sur-
46. Byass P. Usefulness of the Population Health Metrics Research veillance in uncounted populations. PLoS ONE 2011; 6: e22897.
Consortium gold standard verbal autopsy data for general 54. Noor AM, Kinyoki DK, Mundia CW, Kabaria CW, Mutua
verbal autopsy methods. BMC Medicine 2014; 12: 23. JW, Alegana VA, et al. The changing risk of Plasmodium
47. World Health Organization. Global Health Observatory: Number falciparum malaria infection in Africa: 20002010: a spatial and
of malaria deaths. Available from: www.who.int/gho/malaria/ temporal analysis of transmission intensity. Lancet 2014; 383:
epidemic/deaths/en/ [cited 25 February 2014]. 173947.
12 Citation: Glob Health Action 2014, 7: 25369 - http://dx.doi.org/10.3402/gha.v7.25369
(page number not for citation purpose)