The Use of Mosquito Nets and the Prevalence of
Plasmodium falciparum Infection in Rural South Central
Abdisalan M. Noor1,2*, Grainne Moloney3, Mohamed Borle3, Greg W. Fegan1,4, Tanya Shewchuk5,
Robert W. Snow1,2
1Malaria Public Health & Epidemiology Group, Centre for Geographic Medicine Research-Coast, Kenya Medical Research Institute/Wellcome Trust Research Programme,
Nairobi, Kenya, 2Centre for Tropical Medicine, University of Oxford, John Radcliffe Hospital, Headington, Oxford, United Kingdom, 3Food Security Analysis Unit- Somalia,
Parklands, Nairobi, Kenya, 4Infectious Diseases Epidemiology Unit, Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine,
London, United Kingdom, 5United Nations Children’s Fund, Somalia Support Centre, Nairobi, Kenya
Background: There have been resurgent efforts in Africa to estimate the public health impact of malaria control
interventions such as insecticide treated nets (ITNs) following substantial investments in scaling-up coverage in the last five
years. Little is known, however, on the effectiveness of ITN in areas of Africa that support low transmission. This hinders the
accurate estimation of impact of ITN use on disease burden and its cost-effectiveness in low transmission settings.
Methods and Principal Findings: Using a stratified two-stage cluster sample design, four cross-sectional studies were
undertaken between March-June 2007 across three livelihood groups in an area of low intensity malaria transmission in
South Central Somalia. Information on bed net use; age; and sex of all participants were recorded. A finger prick blood
sample was taken from participants to examine for parasitaemia. Mantel-Haenzel methods were used to measure the effect
of net use on parasitaemia adjusting for livelihood; age; and sex. A total of 10,587 individuals of all ages were seen of which
10,359 provided full information. Overall net use and parasite prevalence were 12.4% and 15.7% respectively. Age-specific
protective effectiveness (PE) of bed net ranged from 39% among ,5 years to 72% among 5–14 years old. Overall PE of bed
nets was 54% (95% confidence interval 44%–63%) after adjusting for livelihood; sex; and age.
Conclusions and Significance: Bed nets confer high protection against parasite infection in South Central Somalia. In such
areas where baseline transmission is low, however, the absolute reductions in parasitaemia due to wide-scale net use will be
relatively small raising questions on the cost-effectiveness of covering millions of people living in such settings in Africa with
nets. Further understanding of the progress of disease upon infection against the cost of averting its consequent burden in
low transmission areas of Africa is therefore required.
Citation: Noor AM, Moloney G, Borle M, Fegan GW, Shewchuk T, et al. (2008) The Use of Mosquito Nets and the Prevalence of Plasmodium falciparum Infection in
Rural South Central Somalia. PLoS ONE 3(5): e2081. doi:10.1371/journal.pone.0002081
Editor: Don Husereau, Canadian Agency for Drugs and Technologies in Health, Canada
Received January 16, 2008; Accepted March 20, 2008; Published May 7, 2008
Copyright: ? 2008 Noor 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: FFSAU funded training of assessment teams, data collection, paid enumerators and data entry clerks. FSAU nutrition surveillance project is funded
primarily by OFDA-USAID and receives support from UNICEF, SIDA and EC for conducting nutrition assessments in Somalia. RDTs and anti-malarial treatment were
provided by UNICEF through GFATM funding (SOM-202-G01-M-00). AMN is supported by the Wellcome Trust as a Research Training Fellow (#081829). RWS is
supported by the Wellcome Trust as Principal Research Fellow (#079081). Both AMN and RWS acknowledge the support of the Kenyan Medical Research Institute.
The funders did not have a role in study design, data collection and analysis, decision to publish, or preparation of manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: firstname.lastname@example.org
The evidence on the public health impact supporting the wide-
scale use of insecticide treated nets (ITNs) in Africa is drawn from
areas of stable malaria transmission where Plasmodium falciparum
infection prevalence in the community is often over 40% [1,2].
There is a paucity of parasitological or health impact data on the
benefits of net/ITN in areas of Africa that support low stable or
Across the horn of Africa the dominant vector is Anopheles
arabiensis , a less efficient vector compared to its sibling species
across the central belt of Africa, An. gambiae ss. The semi-arid
conditions of large parts of northern Sudan, Ethiopia, Eritrea,
Djibouti and Somalia support malaria transmission conditions that
result in low P. falciparum parasite prevalence among resident
communities . Surprisingly little is known about the malaria
infection and disease epidemiology in the semi-arid settings of the
East and Horn of Africa, with the exception of studies in Eastern
Sudan [5–8]. Areas of low, stable or unstable malaria transmission
provide different challenges for prevention and control strategies
to those prescribed for more stable transmission areas. Where
parasite exposure is infrequent the clinical consequences of P.
falciparum infection are more likely to directly relate to the risk of
infection compared to areas of high intensity transmission where
the cumulative effects of repeated infection on the development of
clinical immunity are more pronounced . The impact and
PLoS ONE | www.plosone.org1May 2008 | Volume 3 | Issue 5 | e2081
recommendations for the deployment of ITN in these areas of
Africa remains unclear.
With the fall of the government in 1989, Somalia has been
without a central authority and has suffered the ravages of civil
war. In this fragile setting several international relief agencies and
non-governmental organizations currently support the national
ministries of health of the three self-declared states of South-
Central, Puntland and Somaliland in the delivery of preventative
and curative services . In 2004, the Global Fund for Aids, TB
and Malaria (GFATM), awarded Somalia USD 12.8 million to
support the new national malaria control strategy [11,12] With
these funds the United Nations Children’s Fund (UNICEF)
coordinated various partners and the ministries of health to
provide over 700,000 free and/or highly subsidised ITNs by mid
2007 [13,14]. Here we report the parasitological impact of net use
in 2007 from a series of community-based surveys undertaken in
an area of low intensity malaria transmission in South Central
The objective of this study was to examine the effectiveness of
mosquito bednets delivered under routine operational conditions
in South-Central Somalia, an area of generally low malaria
The Food and Agriculture Organization-Food Security Analysis
Unit (FAO-FSAU) has undertaken regular surveys since 1995 in
all regions of Somalia to monitor the nutritional status of children
less than 5 years of age and internally displaced groups . In
2007, four cross-sectional nutritional survey rounds were under-
taken by FAO-FSAU in the regions of Bay (March, April-May)
and Gedo (June), Middle Shabelle and Lower Shabelle (June) of
South Central Somalia (Figure 1). UNICEF and the World Health
Organization (WHO) requested that in addition to routine
nutrition data collection FAO-FSAU undertook investigations of
malaria prevalence and bed net use among individuals of all age
The four survey regions (Figure 1) are densely populated with
an estimated 1.6 million people located between the two major
rivers (the Shabelle and Juba) that transect the southern part of
Somalia. The only published malaria surveys reported from this
area derive from work by Warsame and colleagues in the late
1980’s where P. falciparum infection prevalence among children
aged 1–9 years was 18% in Lower Shabelle . Anopheles arabiensis
is the only reported malaria vector in the area . A UNICEF
multiple indicators cluster survey (MICs) showed that bed net
coverage among children aged less than 5 years in August 2006 in
the four study regions was 8.3% with 84.7% of all nets being ITN
of which 88.1% were long-lasting insecticidal nets . Southern
Somalia has a savannah climate with the long rains beginning in
March-April. The flat irrigated land between the two perennial
rivers, however, provides fertile land for crops (maize and
sorghum). Floods in this area are a relatively common occurrence.
Description of procedures or Investigations undertaken
A stratified two-stage cluster sampling design was used whereby
for each survey sampling was stratified by livelihood zone:
pastoral; agro-pastoral; and riverine. A livelihood zone refers to
a specific food economy group in which communities share similar
types and levels of assets; similar options for production and
income generation; and often with similar socio-cultural and
environment profile and are therefore vulnerable to similar risks
. Communities were defined as pastoral if they engaged
primarily in livestock production and were nomadic (moved with
their livestock from place to place in search of water and pasture);
agro-pastoral if they practiced mixed crop and livestock produc-
tion; and riverine if they lived along the river and were primarily
involved in crop production and river-based economy. Lower and
Middle Shabelle regions were combined and samples derived for
agro-pastoralist and riverine zones; Bay had pastoralist and agro-
pastoralist zones; and Gedo had all three zones resulting in a total
of seven livelihood zones. In each livelihood zone, 30 rural clusters
were first randomly sampled and within each cluster, 30
households were further randomly sampled. The 210 sampled
clusters were derived from a sampling frame of settlements and
populations for each livelihood developed by the United Nations
Development (UNDP) project and UNICEF/WHO for polio
vaccination campaigns colla `ted by the FAO Somali Water and
Land Information Management (SWALIM) and supplemented
with additional information from field visits . Survey tools,
which were based on the standard MICs bednet and parasitolog-
ical questions , were pre-tested for internal consistency and
validity through dummy interviews and pilot field surveys. During
field work each household head was approached by an
interviewer, trained in survey procedures and extraction of blood
samples, who explained the purpose of the surveys and obtained a
household verbal consent.
Following the completion of the nutritional assessment in each
cluster, community members were further investigated as part of
the malaria assessment survey. Sampling continued from one
randomly selected household to the next across the cluster until
approximately fifty people had been examined. Respondents were
asked about the use of a mosquito net during the night prior to the
survey day and provided a finger prick blood sample that was
examined for the presence of P. falciparum infection using a rapid
diagnostic test (RDT) (Paracheck Pf
Systems, Goa, India). Unfortunately no information was collected
on whether the net had been treated with an insecticide nor where
the nets were obtained from, we revisit this is the discussion and
interpretation of the results. Additional information was recorded
on the age and sex of survey participants. All Individuals found to
have a positive RDT result were treated with nationally
recommended age-weight specific sulphadoxine-pyrimethamine
artesunate combination therapy.
TM, Orchid Biomedical
Ethical approval was provided through permission by the
Ministry of Health Somalia, Transitional Federal Government of
Somalia Republic, Ref: MOH/WC/XA/146./07, dated 02/02/
07. Informed verbal consent was sought from all participating
households and individuals.
Data entry and storage was undertaken using EPI Info (Epi
InfoTMVersion 6, CDC, USA), through customized data entry
screens with in-built range and consistency checks. Each
household member was assigned a unique identifier. Descriptive
summaries of infection prevalence by bed net use and community-
level covariates were generated using STATA version 9.2
(Statacorp 2003, College Station, USA) and MS Excel 2007
(Microsoft, Redmond, USA). To account for the clustered nature
of the data, the svy command in STATA was used with the cluster
as the primary sampling unit (psu) stratified by livelihood zones. All
results were weighted (weight=1/probability of selection) to
account for unequal probabilities of selection of clusters across
Net Use and Parasite Infection
PLoS ONE | www.plosone.org2May 2008 | Volume 3 | Issue 5 | e2081
livelihood zones. To test for differences in proportions in net use or
infection prevalence a Pearson x2test accounting for survey design
(clustering and stratification) was used and the test statistic
converted to an F-statistic using the second-order Rao and Scott
(1984) correction yielding wider confidence intervals and conser-
vative P-values compared to the uncorrected x2test.
For each livelihood data the mhodds command in STATA 9.2
was used to examine the association between parasite infection
and sleeping under a bed net the night before survey, adjusting for
the separate confounding effects of livelihoods, age and sex using
the Mantel-Haenszel method. Odds ratios; their 95% confidence
intervals (CI); and associated P-values for the Mantel-Haenszel x2
test are reported.
Of the sampled 210 clusters, 201 were surveyed and 9
pastoralist clusters could not be located during survey. A total
10,587 individuals were seen during the four surveys in 201
clusters (Figure 1). A total of 132 (1.2%) RDT results were invalid
and recorded as un-interpretable. A further 96 (0.9%) respondents
had no information recorded on net use the previous night. These
228 (2.1%) individuals from 4 clusters have been excluded from
the subsequent analysis. The distribution of the remaining 10,359
respondents from 197 clusters who provided a valid RDT and net
use result by livelihood, age and sex are shown in Table 1.
The overall reported use of nets during the night prior to the
survey was 12.4% (95% CI: 8.4–16.5) (Table 1). Net use was
significantly higher among communities classified as riverine
(20.9%, 13.1–28.7) compared to pastoralists (8.6%; F1, 96=9.0,
P=0.004) or agro-pastoralists although the difference was not
significant for the latter (12.6%; F1, 134=2.7, P=0.104). Children
under the age of five years had similar net use compared to
children 5–14 years of age (F1, 193=1.9, P=0.174) and adults (F1,
194=0.6, P=0.438) (Table 1). However evidence of a difference in
net use by children 5–14 years of age was very weak compared to
adults (Table 1: F1, 194=3.5, P=0.061). There was no significant
difference in reported net use by sex (Table 1: F1, 194=0.1,
Overall P. falciparum infection prevalence was 15.7% (95% CI:
12.8–18.6) (Table 1). The highest prevalence was recorded among
the pastoralist communities (24.1%) and the lowest prevalence
among riverine communities (7.2%; F1,
Infection prevalence was significantly lower among adults (9.8%)
compared to children under five years of age (19.6%; F1,
194=56.0, P,0.001) and children 5–14 years (20.5%; F1,
194=55.7, P,0.001). There was no significant difference in
infection prevalence among children aged below five years
compared to those 5–14 years (F1, 193=0.4, P=0.518). Infection
prevalence was higher among males (17.0%) compared to females
(14.5%; F1, 194=9.3, P=0.003).
Table 1 also summarizes the prevalence of infection by use of
bed nets. Overall infection prevalence was significantly lower
among net users: 6.9% (95% CI: 2.7–11.1) compared to non-net
users 17.0% (95% CI: 13.9–20.0); (F1, 194=9.9, P,0.002). Across
all livelihoods, age and sex categories prevalence was significantly
lower among net users although among the riverine communities
the evidence for this was very weak (F1, 36=3.7, P=0.062).
To examine in more detail the effects of net use on P. falciparum
infection prevalence Mantel-Haenszel odds ratios were calculated
that adjust for the effects of age and sex within each livelihood
grouping (Table 2). Within each livelihood group the use of nets
Figure 1. Map of South-Central Somalia regions showing the distribution of clusters where surveys were undertaken by livelihood
zone. 6/197 clusters could not be geo-located and therefore not shown in the figure.
Net Use and Parasite Infection
PLoS ONE | www.plosone.org3May 2008 | Volume 3 | Issue 5 | e2081
significantly reduced the chances of being infected with P.
falciparum with an adjusted protective effectiveness of between
60% (95% CI: 42–72, P,0.001) among pastoralist communities to
48% (95% CI: 22–65, P=0.001) among riverine communities
(Table 2). When all data were combined and adjusted for
livelihood and sex, bed nets reduced significantly the probability
of infection in each age group, with a protective effectiveness
among children 5–14 years of age of 71% (95% CI: 57–80,
P,0.001); followed by adults (48%, 95% CI: 24–64, P,0.001);
and children under five years of age (39%, 95% CI: 17–55,
P=0.002). Overall, after adjusting for livelihood, sex and age the
use of bed nets had a protective effectiveness against parasite
infection of (54%, 95% CI: 44–63, P,0.001).
There are remarkably few studies on the impact of ITN on
infection prevalence among African communities living under
conditions of low transmission  and none we can identify
undertaken in the semi-arid areas of Africa where transmission is
maintained by An. arabiensis. After a limited period of national net
distribution in Somalia we have shown that among the varied
communities of South Central Somalia net use remains low with
only 12.4% of residents reporting using a net (Table 1). Among
children aged below 5 years 12.4% (Table 1) were reported using a
net compared to only 8.7% reported during a cluster sample
survey in the same regions 8–10 months earlier . Among non-
net users P. falciparum infection prevalence was overall 17.0%
ranging from 7.9% among the riverine communities to 25.3%
among the pastoralists. We cannot explain the paradoxical
observation that communities closer to rivers had lower infection
prevalence and would require a more comprehensive vector biting
and larval entomological survey to explore this further. Similar
findings, however, have been reported in The Gambia where
communities living near mosquito breeding sites had proportion-
ately lower parasite rates [22,23]. Nevertheless the ranges of
reported prevalence support the view that transmission intensity in
this area of Somalia is between hypo- and mesoendemic. The
overall age, sex and livelihood adjusted protective effect of nets
among these communities was 54% ranging from 60% among
pastoralist communities to 48% among communities located along
the rivers (Table 2). The largest effects were seen among pastoralist
and agro-pastoralist children aged 5–14 years. We suspect that
most of these nets were treated with insecticide and the majority of
those treated were LLIN (see methods), however, the specifics of
net treatment were not recorded.
The protective effect of nets on infection prevalence among the
sampled populations in South Central Somalia was consistent
across age-groups (Table 2). Given the relatively higher prevalence
of infection through older childhood and into adulthood it is
important to recognize the need to provide ITN to all members of
a community and not focus only on young children in areas of low
transmission. This resonates with recent calls for high coverage
among all community members across the range of transmission
settings  where it is also recognized that individuals older that
five years contribute to transmission.
Several studies in stable, hyper- holoendemic areas of Africa
have examined the impact of nets and ITN on parasite
prevalence. The meta-analysis of randomized controlled trials
(RCT) suggests that the median protective effect of ITN on
infection prevalence in children aged less than 15 years is only
13% . Including an examination of non-RCT studies it is clear
that the ranges of reported protection vary widely largely a result
of different studies examining different age groups and using
different designs. This makes comparisons difficult and presents
several interpretation problems. First, investigations that have
been part of randomized controlled trials, where efforts are made
to maximize coverage of net or ITN use among communities
neighboring ‘‘control’’ communities may have under-estimated
the effects upon infection prevalence [25–30] due to a shadowed,
wide-area protection provided by the intervention areas across
control areas [25–32]. Second, the combined active clinical
Table 1. Characteristics of 10,359 survey participants who provided a valid RDT result and whose bed net use was known in 197
clusters of South Central Somalia
Sleeping under a bed
net, % (95% CI) [n]
Total Pf positive %
(95% CI) [n]
Pf positive sleeping
under a bed net, %
(95% CI) [n]
Pf positive not
sleeping under a bed
net, % (95% CI) [n]
Overall 10,35912.4 (8.4–16.5) [1,418] 15.7 (12.8–18.6) [1,603] 6.9 (2.7–11.1) 17.0 (13.9–20.0) [1,488]
Type of dominant type
Pastoralists 30.1 (3,113)8.6 (4.5–12.6)  24.1 (17.8–30.4) 11.6 (5.3–17.9)  25.3 (18.7–31.8) 
Agro-pastoralist49.9 (5,165)12.6 (6.4–18.8) 13.4 (10.1–16.7) 6.3 (0.0–12.6) 14.4 (11.0–17.9) 
Riverine20.1 (2,081)20.9 (13.1–28.7)  7.2 (3.2–11.2) 4.2 (0.1–7.6) 7.9 (3.4–12.5) 
Age category (years)*
,5 30.9 (3,202)12.4 (8.4–16.3)  19.6 (15.6–23.5) 10.0 (2.8–17.3) 20.9 (16.8–25.0) 
5–1427.0 (2,799)14.0 (8.7–19.2)  20.5 (16.7–24.3) 6.2 (1.8–10.7)  22.8 (18.8–26.8) 
.1442.1 (4,358)11.5 (7.8–15.3) 9.8 (7.4–12.2)  5.0 (2.3–7.6) 10.5 (7.9–13.0) 
Male48.2 (4,997)12.3 (8.1–16.5)  17.0 (13.9–20.0) 7.7 (2.6–12.8) 18.3(15.0–21.5) 
Female51.8 (5,362)12.6 (8.6–16.5)  14.5 (11.6–17.5) 6.1 (2.4–9.8) 15.7 (12.6–18.9) 
*age of 47 individuals was missing
*Proportions and their precisions have been adjusted for clustering and stratification and are weighted using the inverse probability of selection of a cluster within a
Net Use and Parasite Infection
PLoS ONE | www.plosone.org4May 2008 | Volume 3 | Issue 5 | e2081
detection surveillance and treatment among the same children
examined for the prevalence of infection will bias the results away
from significance as more control children will receive effective
treatment during the surveillance period if nets protect against
clinical events [27,28,30,33–35]. These combined effects might
explain why several investigations using single cross-sectional
studies of infection prevalence against the reported use of nets
and ITN under routine operational delivery conditions have
tended to provide higher estimates of protection (51–63%) than
those described during an RCT design [21,36]. Finally, studies
that have examined reductions in parasite prevalence due to nets
or ITN among older children in high transmission areas will have
been unable to identify the true impact on the incidence of
infection through a simple cross-sectional estimate of prevalence
[27,28,33,35]. The duration of sub-patent or un-treated infec-
tions in an individual host may be many months making the
distinction between old and new infections difficult due to
saturation of multiple infections in high transmission areas.
Conversely, in very young, immunologically semi-naive infants at
Kilifi, on the Kenyan coast, comparisons of those sleeping under
ITN with community randomized control infants showed that
infection rates were reduced by 50% among those using ITN
. Similarly an operational effectiveness study of ITN
protection against infection prevalence in children aged less than
24 months in Tanzania showed a 62% protection against
infection prevalence .
Our findings among a wider age group living in an area with a
low rate of parasite exposure might therefore be expected to
correspond to levels of protection described among young infants
living in much higher transmission settings in Kenya  and
Tanzania . Our findings are also consistent with observations
made during non-RCT investigations of net/ITN use in The
Gambia  and the unstable, highland areas of Kenya . The
review by Lengeler (2004) might have under-estimated the impact
on the prevalence of infection consequent upon ITN use. We
would argue that the incidence of new infections might be reduced
by over 50% in all transmission settings but measurement, through
prevalence surveys, is critically dependent upon the selected age
range in different areas of differing intensities of malaria
Halving the risks of P. falciparum infection among all age groups
through the use of nets in an area of low transmission intensity has
not been previously described among the semi-arid areas of Africa.
These findings appear at first glance to support the wide-scale use
of ITN in these areas. However, coverage remains poor and many
more people would need to be reached to achieve a significant
population-attributable impact. This then leaves an issue hard to
resolve with the current data. Where infection prevalence is very
low, such as the communities in the South Central region of
Somalia, when does it become cost-ineffective to deliver ITN to
the entire population? Covering 1.6 million people with an ITN to
reduce infection risks from 17% to 7% might be seen as an
Table 2. Results of the Mantel-Haenszel adjusted odds ratios of using bed nets on Pf infection by livelihood adjusting for the
confounding effect of age and sex.
% sleeping under
a bed net the night
before survey who
were positive for
% NOT sleeping under a
bed net the night before
survey who were positive
for parasite infection
Adjusted* Odds Ratio,
(95% CI), x2(df), P-value
Overall adjusted** Odds Ratio,
(95% CI), x2(df), P-value
,5 19.2 (15/78) 29.0 (228/787) 0.58 (0.33–1.04)
3.4 (1), P=0.068
25.0 (5), P,0.001
5–149.0 (9/100)29.3 (227/774)0.24 (0.12–0.49)
18.4 (1), P,0.001
.147.1 (10/141)14.4 (178/1,233)0.45 (0.23–0.88)
5.8 (1), P=0.06
,5 12.2 (24/197)18.9 (257/1363)0.60 (0.38 –0.94)
5.2 (1), P=0.023
26.4 (5), (P,0.001)
5–14 6.2 (14/225)19.0 (221/1,162)0.28 (0.16–0.48)
21.9 (1), P,0.001
.145.4 (14/257) 8.6 (168/1,961) 0.61 (0.36–1.10)
,5 7.1 (12/169) 10.2 (62/608)0.68 (0.36–1.28)
1.5 (1), (P=0.223)
10.5 (5), (P=0.001)
5–147.4 (8/108)15.8 (68/430)0.42 (0.20–0.92)
5.1 (1), P=0.024
.146.3 (9/143)12.7 (79/623)0.48 (0.23–0.97)
4.3 (1), P=0.037
,5 11.5 (51/444)19.8 (547/2,758) 0.61 (0.45–0.83)
10.0 (1), P,0.001
61.2 (17), P,0.001
5–147.2 (31/433)21.8 (516/2,366)0.29 (0.20–0.43)
44.6 (1), P,0.001
.146.1 (33/541)11.1 (425/3,817)0.52 (0.36–0.75)
12.3 (1), P,0.001
*The age-specific Odds Ratios within each livelihood were adjusted for sex
**Within each livelihood the overall Odds were adjusted for both age and sex. For the combined data the odds were adjusted for livelihood, age and sex
Net Use and Parasite Infection
PLoS ONE | www.plosone.org5May 2008 | Volume 3 | Issue 5 | e2081
expensive way to tackle the problem of malaria in this region. The
answers to this dilemma would require a better understanding of
the health risks consequent upon infection to compute a cost-per-
disability-adjusted-life-year averted and a comparison with other
approaches to malaria infection reduction in semi-arid, semi-
nomadic and conflict areas of Somalia.
A particular challenge in Somalia, due to the large pastoral
community, is maintaining a consistent survey sampling frame and
in some cases there are difficulties in locating nomadic pastoralist
clusters as seen in this study. To minimize the effect of this, a large
sample size was therefore selected. In addition, our estimates of
protective effectiveness on infection prevalence might have
increased if it were possible to separate out ITN from untreated
nets and estimate the impact of ITN alone. We present these
estimates of impact under routine delivery conditions where we
had no opportunity for strict randomization criteria nor, because
of the opportunistic nature of the surveys, have we been able to
adjust for the many other potential confounders of infection and
net use for example socio-economic status or treatment seeking
behaviours. However, the results were internally consistent
between groups sharing similar livelihoods and economic activities
adding confidence to the observed protective effects.
The authors are grateful to the FAO-FSAU survey team for their
invaluable supervision and support during the field surveys and subsequent
data entry and cleaning and Bruno Moonen of MERLIN specifically for
helping with training for the parasite survey. We thank Dr Awdal, the
Minister for Health, Transitional National Government of Somalia for
providing permission and support for the survey. We thank Priscilla
Gikandi and Victor Alegana for additional data cleaning and geo-
referencing. We are also grateful to Dave Smith and Emelda Okiro for
their comments on earlier versions of the manuscript.
Conceived and designed the experiments: GM MB TS. Performed the
experiments: GM MB. Analyzed the data: AN RS GF GM. Contributed
reagents/materials/analysis tools: TS. Wrote the paper: AN RS GF TS.
Other: Data cleaning: AN.
1. Lengeler C (2004) Insecticide-treated bed nets and curtains for preventing
malaria. The Cochrane Database of Systematic Reviews 2004, Issue 2. Art. No.:
CD000363.pub2. DOI: 10.1002/14651858.CD000363.pub2.
2. Phillips-Howard PA, Nahlen BL, Kolczak MS, Hightower AW, Ter Kuile FO,
et al. (2003) Efficacy of permethrin-treated bed nets in the prevention of
mortality in young children in an area of high perennial malaria transmission in
western Kenya. Am J Trop Med Hyg 68: 23–29.
3. Mouchet J, Carnevale P, Coosemans M, Julvez J, Manguin S, et al. (2004)
Biodiversite du paludisme dans le monde. Eurotext, UK: John Libbey. 428 p.
4. Guerra CA, Gikandi PW, Tatem AJ, Noor AM, Smith DL, et al. (2008) The
limits and intensity of Plasmodium falciparum: implications for malaria control
and elimination worldwide. PLoS Med 77(Suppl 6): 79–87.
5. Babiker HA (1998) Unstable malaria in Sudan: the influence of the dry season.
Plasmodium falciparum population in the unstable malaria area of eastern
Sudan is stable and genetically complex. Trans R Soc Trop Med Hyg 92:
6. Roper C, Richardson W, Elhassan IM, Giha H, Hviid L, et al. (1998) Seasonal
changes in the Plasmodium falciparum population in individuals and their
relationship to clinical malaria: a longitudinal study in a Sudanese village.
Parasitology 116: 501–510.
7. Hamad AA, El Hassan IM, El Khalifa AA, Ahmed GI, Abdelrahim SA, et al.
(2000) Chronic Plasmodium falciparum infections in an area of low intensity
malaria transmission in the Sudan. Parasitology 120: 447–456.
8. Giha HA, Rosthoj S, Dodoo D, Hviid L, Satti GMH, et al. (2000) The
epidemiology of febrile malaria episodes in an area of unstable and seasonal
transmission. Trans R Soc Trop Med Hyg 94: 645–651.
9. Snow RW, Marsh K (2002) The consequences of reducing Plasmodium
falciparum transmission in Africa. Advances in Parasitology 52: 235–264.
10. SACB (2000) Strategic Framework in support of the Health Sector in Somalia.
(Volume 1). Developed at the SACB Health Strategy Development Workshop
Nakuru, Kenya, 24–28 May 1999.
11. GFATM (2002) Round 2 Somalia project proposal. http://www.theglobalfund.
org/search/docs/2SOMM_134_0_full.pdf. Accessed 21stNovember 2007.
12. Capobianco E (2005) Somalia National Malaria Control Strategy 2005–2010.
13. GFATM (2006) Round 6 Somalia project proposal. http://www.theglobalfund.
org/search/docs/6SOMM_1418_0_full.pdf. Accessed 21stNovember 2007.
14. Global Fund Malaria Grant SOM-202-G01-M-00 Performance Report,
Somalia. 23rdAugust 2007 print version. http://www.theglobalfund.org/
search/docs/2SOMM_134_246_gpr.pdf. Accessed 5thJanuary 2008.
15. FSAU (2006) Project background. http://www.fsausomali.org/index.php@
id=3.html. Accessed 21stNovember 2007.
16. Warsame M, Perlmann H, Ali S, Hagi H, Farah S, et al. (1989) The sero-
reactivity against Pf155 (RESA) antigen in villagers from a meso-endemic area in
Somalia. Trop Med Parasitol 40: 412–414.
17. Kamal M (2007) Entomological surveillance in Somalia. Consultancy report for
WHO Somalia. October 25–November 26, 2007.
18. UNICEF (2006) Somalia Multiple Indicator Cluster Survey, Somalia. August-
19. FSAU (2005) Analytical systems. http://www.fsausomali.org/200510112504_
baselines.php. Accessed 11thNovember 2007.
20. FAO-SWALIM (2007) Databases. http://geonetwork.faoswalim.org:8080/
geonetwork. Accessed 11thNovember 2007.
21. Guyatt HL, Corlett S, Robinson T, Ochola SA, Snow RW (2002) Malaria
prevention in highland Kenya: indoor residual house-spraying vs. insecticide-
treated bednets. Trop Med Int Health 7: 298–303.
22. Thomson MC, D’Alessandro U, Bennett S, Connor SJ, Langerock P, et al.
(1994) Malaria prevalence is inversely related to vector density in the Gambia,
West Africa. Trans Roy Soc Trop Med Hyg 88: 638–643.
23. Thomson M, Connor S, Bennett S, D’Alessandro U, Milligan P, et al. (1996)
Geographical perspectives on bednet use and malaria transmission in the
Gambia, west Africa. Soc Sci Med 43: 101–112.
24. Killeen GF, Smith TA, Fergusin HM, Mshinda H, Abdulla S, et al. (2007)
Preventing childhood malaria in Africa by protecting adults from mosquitoes
with Insecticide-treated nets. PLoS Med 4: e229.
25. Ter Kuile F, Terlouw D, Phillips-Howard P, Hawley W, Friedman J, et al.
(2003) Impact of permethrin-treated bed nets on malaria and all cause morbidity
in young children in an area of intense perennial malaria transmission in western
Kenya: cross-sectional survey. Am J Trop Med Hyg 68: 100–107.
26. Snow RW, Rowan KM, Greenwood BM (1987) A trial of permethrin-treated
bed nets in the prevention of malaria in Gambian children. Trans R Soc Trop
Med Hyg 81: 563–567.
27. Snow RW, Rowan KM, Lindsay SW, Greenwood BM (1988a) A trial of bed
nets (mosquito nets) as a malaria control strategy in a rural area of The Gambia,
West Africa. Trans R Soc Trop Med Hyg 82: 212–215.
28. Snow RW, Lindsay SW, Hayes RJ, Greenwood BM (1988b) Permethrin-treated
bed nets (mosquito nets) prevent malaria in Gambian children. Trans R Soc
Trop Med Hyg 82: 838–842.
29. Habluetzel A, Cuzin N, Diallo DA, Belem S, Cousens S, et al. (1999) Insecticide-
treated curtains reduce the prevalence and intensity of malaria infection in
Burkina Faso. Trop Med Int Health 8: 557–564.
30. Henry MC, Assi SB, Rogier C, Dossou-Yovo J, Chandre F, et al. (2005)
Protective efficacy of Lambda-Cyhalothrin treated nets in Anopheles gambiae
pyrethroid resistance areas of Cote D’Ivoire. Am J Trop Med Hyg 73: 859–864.
31. Howard S, Omumbo JA, Some ES, Donelly CA, Snow RW (2000) Evidence for
a mass community effect of insecticide treated bed nets on the incidence of
malaria on the Kenyan Coast. Trans R Soc Trop Med Hyg 94: 357–360.
32. Hawley WA, Philips-Howard PA, ter Kuile FO, Terlouw DJ, Vulule JM, et al.
(2003) Community-wide effects of permethrin-treated bed nets on child
mortality and malaria morbidity in Western Kenya. Am J Trop Med Hyg 68:
33. Sexton JD, Ruebush II TK, Brandling-Bennet DA, Breman JG, Roberts JM, et
al. (1990) Permethrin-impregnated curtains and bed nets prevent malaria in
Western Kenya. Am J Trop Med Hyg 43: 11–18.
34. D’Alessandro U, Olayeye BO, McGuire W, Langerock P, Bennett S, et al.
(1995) Mortality and morbidity from malaria in Gambian children after
introduction of an impregnated bednet programme. Lancet 345: 479–483.
35. Fraser-Hurt N, Felger I, Edoh D, Steiger S, Mashaka M, et al. (1999) Effect of
insecticide-treated bed nets on haemoglobin values, prevalence and multiplicity
of infection with Plasmodium falciparum in a randomized controlled trial in
Tanzania. Trans R Soc Trop Med Hyg 93: 47–51.
36. Clarke S, Bogh C, Brown R, Pinder M, Walraven G, et al. (2001) Do untreated
bednets protect against malaria? Trans R Soc Trop Med Hyg 95: 457–462.
37. Snow RW, Molyneux CS, Warn PA, Omumbo J, Nevill CG, et al. (1996) Infant
parasite rates and IgM seroprevalence as a measure of exposure to Plasmodium
Net Use and Parasite Infection
PLoS ONE | www.plosone.org6 May 2008 | Volume 3 | Issue 5 | e2081
falciparum during a randomized controlled trial of insecticide-treated bed nets Download full-text
on the Kenyan Coast. Am J Trop Med Hyg 55: 144–149.
38. Abdulla S, Schellenberg J, Nathan R, Mukasa O, Marchant T, et al. (2001)
Impact on malaria morbidity of a programme supplying insecticide treated nets
in children aged under 2 years in Tanzania: community cross sectional study.
British Medical Journal 322: 270–273.
Net Use and Parasite Infection
PLoS ONE | www.plosone.org7May 2008 | Volume 3 | Issue 5 | e2081