uva-dare (digital academic repository) a critical enquiry ... · charlotte gryseels,1* melanie...

UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

A Critical Enquiry into Variability of Insecticidal Net Use in Cambodia:Implications for Assessing Appropriateness of Malaria Elimination Interventions

Gryseels, C.; Bannister-Tyrrell, M.; Uk, S.; Set, S.; Suon, S.; Gerrets, R.; Grietens, K.P.DOI10.4269/ajtmh.18-0730Publication date2019Document VersionFinal published versionPublished inAmerican Journal of Tropical Medicine and HygieneLicenseCC BY

Link to publication

Citation for published version (APA):Gryseels, C., Bannister-Tyrrell, M., Uk, S., Set, S., Suon, S., Gerrets, R., & Grietens, K. P.(2019). A Critical Enquiry into Variability of Insecticidal Net Use in Cambodia: Implications forAssessing Appropriateness of Malaria Elimination Interventions. American Journal of TropicalMedicine and Hygiene, 100(6), 1424–1432. https://doi.org/10.4269/ajtmh.18-0730

General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s)and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an opencontent license (like Creative Commons).

Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, pleaselet the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the materialinaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letterto: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. Youwill be contacted as soon as possible.

Download date:06 Apr 2021

https://doi.org/10.4269/ajtmh.18-0730https://dare.uva.nl/personal/pure/en/publications/a-critical-enquiry-into-variability-of-insecticidal-net-use-in-cambodia-implications-for-assessing-appropriateness-of-malaria-elimination-interventions(f7515dfe-b92f-4575-ac6a-ae55ff36c28b).htmlhttps://doi.org/10.4269/ajtmh.18-0730

Am. J. Trop. Med. Hyg., 100(6), 2019, pp. 1424–1432doi:10.4269/ajtmh.18-0730Copyright © 2019 by The American Society of Tropical Medicine and Hygiene

A Critical Enquiry into Variability of Insecticidal Net Use in Cambodia: Implications for AssessingAppropriateness of Malaria Elimination Interventions

Charlotte Gryseels,1* Melanie Bannister-Tyrrell,1 Sambunny Uk,2 Srun Set,2 Sokha Suon,2 René Gerrets,3 andKoen Peeters Grietens1

1Institute of Tropical Medicine, Antwerp, Belgium; 2National Center for Parasitology Entomology and Malaria Control, Phnom Penh, Cambodia;3Amsterdam Institute for Social Science Research, University of Amsterdam, Amsterdam, The Netherlands

Abstract. Distributing long-lasting insecticidal nets (LLINs) to individuals living in malaria-endemic regions is a cor-nerstone of global malaria control. National malaria control programs aim to achieve “universal coverage” of at-riskpopulations to reach LLINs’ full potential to reducemalaria, progress ofwhich is thenmeasured by indicators constructedfrom standardized questionnaires. Through an exploration of variability in LLIN use in Cambodia, we argue that indicatorsof universal coverage of LLINs are not sufficiently commensurate with the realities they are intended to measure, limitingthe suitability of the data to serve program and policy purposes in a malaria elimination era. Reflecting on the varioussources of variability in LLIN use, we apply and extend the concept of “appropriateness” as a third prong to the widelyused “efficacy” and “effectiveness” criteria for evaluating LLINs as a tool for malaria prevention. Describing first thedifferent dimensions of the intervention and the sociocultural context separately, we will further show how the variabilityunderlying both is affected and induced by inappropriate aspects of the intervention and themeasurements of its impact.We consider the gap between “net use” and the numerical representations of such local net use justifies further explo-ration of potential strategies to improve LLIN use in subgroups where persisting malaria transmission clusters.

INTRODUCTION

The distribution of long-lasting insecticidal nets (LLINs) tocommunities living in malaria-endemic countries is still acornerstone of malaria control today. Because of a suc-cessful scaling-up of control measures, including LLINs andinsecticide-treated nets (ITNs), and an overall decliningmalaria incidence, malaria elimination in many countries hasonce again been declared feasible since 2007. In the sameyear, the WHO began recommending that all malaria controlprograms ensure “universal coverage” of freely distributed orhighly subsidized LLINs in populations at risk.1 Universalcoverage is currently defined by two indicators2: 1) the per-centage of people who have access to LLINs in the householdand 2) the percentage of people reporting having slept underan LLIN the previous night. The WHO defines “operationalsuccess (as opposed to a target of 100% coverage) as theobservation in surveys of at least 80% coverage in termsof these indicators.”2 To measure whether countries haveachieved “operational success,” The Roll Back Malaria Part-nership provides standardized household questionnairesfor use in malaria indicator surveys. There are indicators forhousehold ownership of (sufficient) LLINs and an indicatorfor individual LLIN use, operationalized as the respondentreporting to have slept under a LLIN the night before thesurvey.3

Whenassessingwhether universal coveragewith LLINs hasbeen achieved and the extent to which ongoing malariatransmission is attributable to “residual” transmission,4 indi-cators of LLIN use are necessarily abstracted from the so-ciocultural context to allow comparison between settingsand the generation of generalizable strategies for malariacontrol and elimination. However, variability, often at fine

socio-geographical scales, exists in the extent to whichcommunities and/or individuals “comply” to interventionssuch as LLINs. Long-lasting insecticidal nets are objects thatcanaffect andare affectedby sleepingarrangements, housingstructure, the spatial organization within a house, and be-havior from visiting friends and relatives.5–9 Acceptabilitystudies across the malaria-endemic regions of the world havelong since shown a discrepancy between the designs ofLLINs and targeted users’ social requirements of LLINs andother bed nets.6,9–14 These important findings have seldomprompted the malaria community to revise LLIN programs orimpacted on how the induced variability in LLIN use is mea-sured through indicators of “ownership,” “coverage,” and“use” when evaluating the average effectiveness of LLINdistribution programs. The impact of numerous social, cul-tural, economic, and environmental factors5,7,9,15,16 is aver-aged at the level of the lowest administrative unit (e.g., districtor country) required for methodological and reporting pur-poses. Although stabilizing variability in LLIN use across timeand place renders it measurable (i.e., quantifiable), this re-ification also obscures the phenomenon that is being mea-sured, which is the extent to which people are effectivelyprotected from biting vectors while sleeping at night, in (all)places where they are exposed. Although gaps betweensimple indicators and their measurement targets are ubiqui-tous, in a malaria elimination context, failure to recognize andcapture variability in LLIN use leads to assumptions aboutcauses of ongoing malaria transmission once universal cov-erage is assumed, and accordingly, to the prioritization ofspecific types of public health interventions and research.In addition to the indicators, the intervention itself is also at

best minimally adapted to the diverse contexts in which LLINsare intended to be used. Failure to localize LLINs can be un-derstood as relating to the (in)appropriateness of the in-tervention. Although this concept has been formally applied toclinical care,17,18 there has been minimal extension of theconcept of appropriateness in global health, especially forpreventive public health interventions, which are instead often

*Address correspondence to Charlotte Gryseels, Medical Anthropol-ogy Unit, Department of Public Health, Institute of Tropical Medicine,Nationalestraat 155, Antwerp 2000, Belgium. E-mail: [email protected]

1424

mailto:[email protected]:[email protected]

assessed in terms of the efficacy–effectiveness paradigm.Unlike concepts such as effectiveness, which estimate aver-age net intervention effect under real-life conditions, theconcept of appropriateness was introduced in clinical care toexpand the evidence base underlying clinical decisions aboutwhich procedures and tests are appropriate for patients in acertain setting or context, as well as to focus attention onreducing unwarranted variation in medical care practices.17 Inthis article, we aim to adapt and expand the concept of ap-propriateness to the context ofmalaria control and eliminationinterventions and go beyond “acceptability” studies. We de-fine “appropriateness” as 1) the extent to which LLINs as anintervention (both their design and distribution) align with theconditions and associated requirements of the contexts inwhich they are intended to be used and 2) how well the indi-cators that aim to capture its effect align with the real-lifevariability in LLIN use.

METHODS

Study design. Ethnographic research was embedded in acommunity-randomized trial investigating the impact of massdistribution of personal topical repellents in addition to thedistribution of LLINs on malaria prevalence in Ratanakiriprovince, Cambodia, between 2012 and 2013. The ethno-graphic research was conducted ancillary to the trial. Thesocial science field team worked independently of the trialteam and developed relationships of trust with the studypopulation during extended stays in the study villages. Vari-ability in LLINs was an emerging theme during ethnographicfieldwork and in an additional structured observation survey ofpersonal topical repellent and bed net use (including LLIN,ITN, and non-treated nets), we identified population sub-groups who did not use LLINs or other bed nets, or used bednets less frequently because they were using the providedrepellent (for more details, see Ref. 19). After the conclusion ofthe trial, the social science team carried out an explanatoryqualitative study in 2014, aiming to further characterize thesecontexts and conditions of variability in bed net use andnonuse, the results of which are presented in this article. Thisstudy made use of in-depth (n = 130) and informal interviews(n = 21) with both the identified subgroups described pre-viously, as well as with other relevant key informants such ashealth center staff, village malaria workers (VMWs), villagechiefs, and other community members.Study site. Research was conducted in Ratanakiri Prov-

ince, a remote region of northeastern Cambodia borderingVietnam. Several indigenous populations inhabit this region,including Jarai, Tompuon, Kreung, Lon, Prov, Kavet, Kachok,and Lao peoples, as well as a rapidly growing Khmer migrantpopulation, the predominant ethnic group in Cambodia. Mostof our ethnographic investigations during the trial took place intwo predominantly Jarai villages in Oyadao district, twoTompuon villages along the Tonle San River in Voen Sai dis-trict, and one Tompuon village located further south in theprovince near the Srepok River in Lumphat district. We con-ducted additional observations and informal conversations inthe district town centers and the district health centers, andprivate practices in the district centers and in the provincialcapital Banlung. After the end of the trial, the explanatoryqualitative studywasconducted in the villages selected for thestructured observation survey, which included three Jarai

villages in Oyadao district, three Kreung villages in Ochumdistrict, three Tompuon villages in Lumphat district, and oneTompuon and one Kreung village in Voen Sai district.Malaria is endemic at low prevalence in Ratanakiri, primarily

caused byPlasmodium vivax andPlasmodium falciparum, butPlasmodium malariae and Plasmodium ovale infections alsooccur.20 As measured by passive case detection, overall an-nual incidence rates per 1,000 inhabitants of all Plasmodiumspecies decreased from 85.9 in 2010 to 30.4 in 2014 inRatanakiri.21

The Cambodian National Malaria (CNM) Control Programimplements the WHO-recommended malaria control strate-gies throughprovincial health departments, operational healthdistricts, local health centers at district and commune levels,and at community level through VMWs, who are trained todiagnose and treat malaria with rapid diagnostic tests andfirst-line artemisinin-combination therapies.22 In collaborationwith the health centers, VMWs alsomanage the distribution ofLLINs, one for every person registered in a malaria-endemicvillage. The national malaria control guidelines stipulate thisdistribution should occur every 2 years. Beyond these officialcontrol measures, villagers can obtain non-treated bed nets(“market nets”), non-treated hammock nets, personal topicalrepellents, and mosquito coils at local markets, sometimes atprices subsidized by international non-governmental organi-zations. A range of unlicensed therapeutics, including phar-maceutical drug “cocktails” and traditional remedies, arereadily accessible in the pluralistic health-care system.Data collection. Ethnographic research was carried out

using interviewing and participant observation as the maindata collection techniques. All interviews were either carriedout in English and translated to Khmer or were conducted inKhmer. When informants did not speak Khmer fluently, villageguides or other key informants of the same village assisted inconducting interviews and were appropriately reimbursed fortheir time.Informal conversations with people in the villages and at the

health centers during our stays in the villages—during the trial,four visits of 2 months each, and, for the explanatory quali-tative study, one visit of 2 months—constituted an importantpart of our data collection process. Participant observationwas used as an ethnographic technique that enabled the re-searchers to contrast people’s actual behavior to people’sreported behavior. It included both continuous informal con-versations and observations of people’s practices in theirhomes, in village, and farm settings, and during social orceremonial events taking place in the village. Both informalconversations and observations were transcribed in fieldnotes, and each single observation and conversation consti-tuted a separate data entry in the final database.Sampling. Participants for in-depth and informal inter-

viewing were theoretically selected and gradually includedfrom the village populations (including self-reported users andnonusers of both LLIN and non-treated market nets), publichealth center staff, private health-care providers, and pro-vincial health department staff, usually through snowballsampling techniques. Other respondents were identified froma quantitative structured observation study (methods de-scribed in detail in Ref. 19) and included everyone from thefollowing subgroups: household leaders of households 1)whowere observed to not use any bed nets (n = 29) or 2) who self-reported using bed nets less frequently because of the

VARIABILITY OF INSECTICIDAL NET USE IN CAMBODIA 1425

availability of the trial repellents (n = 60). Informants were ad-ditionally classified according to variables, such as gender,age, subsistence strategies, locality, ethnicity, and occupa-tion, to allow for internal variation and comparison.Data analysis. Preliminary data were frequently reviewed

and analyzed in the field, concurrent with data collection. Al-though we made use of question guides, interviews werenevertheless open to allow for emergent information. In theinitial phase of research, raw data were coded inductively inMicrosoft Word. When preliminary coding was performed,new hypotheses often emerged and the question guides wereadapted to reflect the new hypotheses. These were furthertested in the field until theoretical saturation was reached. Afinal coding tree was developed based on the results of theanalytic process during fieldwork and applied to the data.NVivo 9 Qualitative Analysis software (QSR International,Doncaster, Victoria, Australia) was used for all final datamanagement and deductive analysis. Coding queries wereused to test relationships between codes or between codesand attributes of respondents.Ethical approvals. The Institutional Review Board of the

Institute of Tropical Medicine, Antwerp, and the CambodianNational Ethics Committee for Health Research granted ethi-cal approval. All participants were explained the objectives ofthe study, the risks and benefits involved, the type of ques-tions that were going to be asked, the use of the results forpublications, and their right to stop the interview andwithdrawparticipation at any time. They were then free to provide theiroral consent to participate in the study. All interviewers fol-lowed the Code of Ethics of the American AnthropologicalAssociation.

RESULTS

The concept of appropriateness is intricately tied to thevariability in net use we introduced earlier. Defining first thedifferent dimensions of the intervention and the socioculturalcontext separately, we will go on to show how the variabilityunderlying both is affected and induced by inappropriate as-pects of both the intervention and the measurements of itsimpact.The sociocultural context. The indigenous ethnic groups

of Ratanakiri originate from different geographic and linguisticlineages. Nevertheless, they are often closely related, and asneighbors, they interact and intermarry to the extent that itwould bewrong to view each ethnic group as a separate entityfrom the larger social system and ecological niche theyshare.23 Most indigenous communities in Ratanakiri practiceslash-and-burn agriculture, including cultivation of dry rice(“upland rice”) and various vegetables such as eggplant,gourd, corn, cucumber, sesame, and yam at farms cleared inthe forest.23 Each village has its own forest territory, wheremostly nuclear families work farm plots.24 Entire families fre-quently stay in bamboo houses constructed at their farmsbecause of the distance to the village house, spiritual relationswith the cleared lands, and the necessity to protect crops fromwild animals and roaming cattle at night and during the day.Livelihood practices are therefore intimately interwoven withsleeping arrangements. Household composition at farm andvillage houses is fluid: teenagers often move between villagesand fields for social and economic activities25; children some-times remain in the villageswith grandmothers to attend school

when the Khmer teachers are present; and adult men comeback for a night or two to participate in social events such asgeneral drinking parties, funerals, weddings, or other ceremo-nies that are mostly characterized by drinking rice wine.People frequently engage in forest activities suchashunting

and logging, which primarily occur in the evening or nighttime,and overnight stays are common. Such activities are usuallyperformedbygroupsofmen in theassociated forest spacesofa particular village. Other forest activities, such as collectingwild roots, vines, and fruits, are carried out by larger groups ofmen, women, and children during certain times of the year.Sleeping spaces are flexible and not reducible to a predictablerotation of mobile individuals in one sleeping space.The intervention. There are three distinct dimensions of

LLINs as an intervention: the product, procurement, and thedistribution strategy for the product.The product. Long-lasting insecticidal nets are designed to

reduce the incidence of malaria by offering personal pro-tection to users from exposure to infectious biting vectors.This protection is assured by three distinct functions: 1)forming a protective physical barrier between the mosquitoand its blood meal, 2) killing the mosquitoes that come intocontact with the insecticide with which the nets are impreg-nated, and 3) the repellency effects of the insecticide that re-duces mosquito entry into the house where LLIN users sleep.The insecticidal properties of nets also offer household-leveland community-level protection by reducing the overall pop-ulation of infected vectors, if used sufficiently at individuallevel. Implicitly, therefore, LLINs are intended for use indoorsand/or in locations where they can be hung, when people aresleeping, at hours of the day (usually nighttime) that coincidewith vector biting peaks. At the time of our fieldwork, Olysetnets (Olyset® Net; Sumitomo Chemical UK PLC, London,United Kingdom) were distributed in Ratanakiri. The companypresents its product as a quality LLIN because its main char-acteristics include the follows: 1) “hybrid polymer and con-trolled insecticide release technology to repel, kill, andpreventmosquitos from biting for up to 5 years”; 2) being “washable,”meaning they have “passed the > 20 washes test required bythe WHO to be designated an LLIN”; and 3) being very strongbecause of their thick fibers (180 deniers).26 In addition, Olysetnet mesh sizes are relatively large (4 mm) than many otherbrands of LLINs (usually 1.6 mm) to increase ventilation.26

The appropriateness of the product in the socioculturalcontext. The appropriateness of the nets available in Rata-nakiri (both Olyset LLIN and non-treated market nets) in re-lation to varying housing and environmental factors directeduse: small farmhouses are usually only able to fit the smallersized distributed LLIN and larger village houses are able to fitthe larger market nets, which despite indicating variable LLINuse, may nonetheless be a suitable arrangement for malariaprevention as sleeping at farmhouses has been shown to in-crease malaria risk.20

Our informants considered the fabric of the LLINs distrib-utedby theNationalMalariaControl Programat the timeof thisstudy simultaneously too hard and too fragile, and especially,the mesh sizes too large. Mesh size is an important charac-teristic of nets that induces the perception of protection andprivacy: large mesh sizes are reported to let in “all sorts ofinsects,” including mosquitos, despite the insecticide; theyform less of a protective barrier between the safe and do-mesticated “indoor” and the untamed “outdoor” when

1426 GRYSEELS AND OTHERS

sleeping in farmhouses located next to forests that harborwild animals and vindictive spirits; and they do not offer anyprivacy in homes where newlyweds are sleeping next to theirparents. Contrary to the LLINs, untreated nets from localmarkets or traveling salesmen are soft and accommodatelarge families, and are perceived to have smaller mesh.Moreover, their mesh sizes are smaller than LLINs, whichleads to the empirically soundperceptions of sufficient insectprotection as well as greater privacy. This difference in ac-ceptability between the distributed LLINs and the marketnets leads to differences in actual use; those families that canafford them and have the relatively large space that is re-quired to hang market nets are more likely to be using themarket nets on a regular basis.The procurement processes. Variability in LLIN use can be

related to priorities and decisions at national and internationallevels concerning procurement and mass distribution ofstandardized, quality-assured (i.e., WHOPesticide EvaluationScheme [WHOPES] approved) LLINs for implementationacross the world’s highly diverse malaria-endemic regions.The malaria-epistemic community,27–29 which is global inoutlook and with rapidly expanding membership in the mod-ernmalaria eliminationera, is aproduct of current global healthhegemonies that emphasize rapid up-scaling of interventionsglobally, innovation, and setting ambitious targets to driveglobal progress toward the ultimate goal of malaria eradica-tion.30 Accordingly, malaria control programs are advised toestablish single national policies for achieving universal LLINcoverage that ensure cost-effective and equitable provision ofLLINs to all populations at risk of malaria. Long-lasting in-secticidal net distributions are therefore preceded by clearlydefined procurement criteria. First, the WHO provides rec-ommendations to national malaria control programs forselecting approved LLINs, based on bioassay tests and netdurability tests performedby theWHOPES. Second, basedonthese recommendations, donors such as the Global Fundconstruct a list of authorized suppliers of WHOPES-approvednets that national malaria control programs can direct donorfunding toward. Procurement is envisioned as an open tenderprocess to acquire contracts for the cheapest possible in-ternationally standardized LLINs.The appropriateness of the procurement in the sociocultural

context. Although procurement is envisioned as an opentender process, in Cambodia, it has been reported to involvecorrupt bidding practices.31 As a consequence, donor controland financial auditing of the execution of international policiesand programs have tightened, which appears to have had theeffect of inducing delays in the implementation of LLIN dis-tributions. Accordingly, there has been relatively little em-phasis in these procurement processes toward selectingLLINs with characteristics that are most appropriate for thespecific target populations.The distribution. The LLINs purchased at the national level

then need to be distributed to populations at risk of malaria.There are several distribution strategies possible. Almostuniversally applied is the free-of-charge distribution at thelocal level of a predefined number or nets per person (usuallyone net per two household members). Additional strategiesexist, such as socially promoted and subsidized nets for saleat markets; and hang-up keep-up strategies, to tackle theproblem of new nets being saved for the future instead of asreplacement of the current nets.32–35

Cambodian national malaria policy calls for distribution ofLLINs to all people living in malaria-endemic regions. Theprogramaims to ensure that the knownbarriers to theeffectivescaling-up of insecticidal nets are adequately addressed: 1)LLINSaredistributed locally and freeof charge, so there are nofinancial barriers; 2) LLINS are distributed one per one personper village unit, so every single household member shouldhave access to an LLIN; and 3) distribution rounds occur onceevery 2 years, assuring continued access to effective LLINs.The appropriateness of the distribution in the sociocultural

context. This also translates to distribution practices at districtandvillage level, where several barriers emerge to assuring theLLIN distribution runs as envisioned by international donors.Local distribution rounds are organized by district healthcenter staff, with the support of provincial health departments,in collaborationwith village authorities suchas village chiefs orVMWs. On the announced day of the LLIN distribution, reg-istered residentsmustpresent their “bednet booklet”—that is,the administrative paperwork that registered residents receiveto receive LLINs—to the health center staff and have itstamped at the central location in the village where the dis-tribution is organized. In addition to themany Khmer migrantsthat are not registered in Ratanakiri because they have re-cently migrated from another province where they are regis-tered, there are also registered indigenous inhabitants who donot receive the “bed net booklet” as they were away from thevillage center at the time the booklet was distributed. Difficultsocial relations may exist with key community members suchas the village chief, VMW, or health center staff organizing thedistribution rounds, which causes some villagers to be“omitted” from the village list. Key informants from villagesand the health system also report that somewhere on the hi-erarchical line (from government official, provincial healthdepartment, and health center staff to village chief or VMW)LLINs are frequently withheld from distribution because of asystem of social privilege in Cambodia that allows for andenables government officials to keep items such as LLINs forother family members or other purposes.36 Related distribu-tional practices therefore also reflect and reproduce the socialinequalities inherent in a system of patronage.36

Variations in LLIN use in time and place. Assumptions ofuniversal coverage are further challenged by place-based andtemporal variations in LLIN use.Place-based variation. Intended distribution of one LLIN per

one person aims to cover individuals in one sleeping spaceand does not consider the varying sleeping spaces individualsuse.When families are divided throughout several residences,it is not straightforward for the mobile household member tocarry a LLIN from sleeping space to sleeping space andleaving other household members exposed. As there arevarying needs in relation to the social events in the village or faraway fields that need attending, and in relation to child care inthe village in absence of most family members working atfields, there is always some proportion of the population re-quiring flexible and multiple sleeping spaces, and thereforeless likely to be using a LLIN. Net priorities (for both LLIN andmarket nets) also varied and related to the multiple residencesystem; families in Ratanakiri often preferred to keep theirmarket nets in the villagehousesimply because thehousewasstructurally better suited to contain the large-sized marketnets needed to accommodate large families. Long-lastinginsecticidal netswere often preferred for farmhouses because


insect nuisance was perceived to be higher at farms andpeople reported not being able to sleep there without a netseparating them from nuisance insects. Especially in suchhigh nuisance settings, the mesh sizes were considered im-portant and large mesh sizes were considered inappropriate,given the insects thatwere perceived tobeable to enter the net.Temporal variation. Long-lasting insecticidal net use varies

in accordance with agricultural calendars, with time since thelast LLINdistribution, and over the life course. Indigenousmenand women engage frequently in cultural events and cere-monies such as funerals, weddings, harvest, or planting cer-emonies and other ceremonies to satisfy spirits, which occurboth variably and at fixed times of the agricultural cycle. Theseevents are always linked to the consumption of rice wine.Although now drinkingmay bemore often linked to “parties,”these events lead to most of a village population or a par-ticular social subgroup going to bed late at night and oftentoo intoxicated to find, or care about finding, a bed net tosleep in. Such events are frequent inRatanakiri, and thereforecould skew standardized questions related to LLIN use thatexpect a straightforward answer. People who report to al-ways sleep in a LLIN in surveys may in fact usually do so butnot when they return home drunk from such (frequent) ricewine ceremonies or parties.In relation to the temporal decline in net quality, both LLINs

and market nets tend to tear quickly in local households inRatanakiri.

Yeswe liked to sleep in it [distributed LLIN] because it hadinsecticide to protect from mosquitoes. But my wifewashed it often because my children always pee on thebednet. So itworeout of insecticide andbecame fragile. Ithas big holes so after it wore out of insecticide, themosquitoes and other small insects could enter it.(Kachok farmer)

As nets are commonly urinated on by children and picked atby chickens andmice, frequent washing of nets is common ascleanliness is considered a social virtue. Although people areaware that frequent washing of LLINs reduces the effect of theinsecticide and makes the nets tear more quickly, theynonetheless have to wash the nets weekly to avoid the smellbecoming unbearable, which highlights the contrast betweenthe required durability and stated durability of the distributedLLINs as exceeding (only) 20washes. Thewear and tear LLINsincur under these conditions led many of our informants tostop using the LLINs relatively quickly after distribution. Thegap between projected LLIN durability and their actual life-span is another example of temporal variability that cannot bemeasured with indicators that assume a temporally stablereality. Indeed, the most persistent comment was that theinherent low quality of the distributed LLINs was thought tomake the nets tear “by themselves.” Some informants told usthey perceive the low quality of LLINs to reflect the low socialstatus the national malaria control program affords them as“ethnic minorities.” They did not think the LLINs lasted longerthan a year after distribution, and therefore many experienceda gap in protection between distributions during which theysimply did not have nets to use. Thus, if they report in surveysthat theydonot useanLLIN, thismeasure of “noncompliance”in part refers to a lack of access to a good and durable net for

the local context. Although thosewho canwill buymarket netsto use in the time between distributions, most indigenousfarmers cannot afford (enough) market nets to cover allhousehold members.Another temporal dimension of the variability in LLIN use re-

lates to different age groups and life stages among indigenouspeoples in Ratanakiri. Families will prioritize the protection ofsmall children who sleep together with their mothers.

There are many mosquitoes until 8 to 9 PM. I worry aboutsmall children; I let them sleep in the bed net. As I am anadult, there is less of a problem. When my children getfever, they get convulsions. I spent 100,000 riels [∼ $25USD] for treatment formy children at the health center. Hehad an infusion there for three days. (Tompuon farmer).

Children are perceived to be more vulnerable to illness andthe biomedical treatments required to cure them are thoughtto be more expensive. However, in this setting, malaria riskclusters in adolescent and adult men, particularly whensleeping and working in or near the forest.20,37–39 Men andoldermale childrenwere given least priority to sleep protectedand revert to sleeping on mats with blankets or hammockswithout nets when distributed LLINs are no longer fit for use.Men particularly go through different life stages that can belinked to different uses of bed nets. Groups of young adoles-cent boys sleep at friends’ houses, in empty houses, or incommunal village houses with hammocks (mostly withoutnets), avoiding the fixed sleeping spaces that more readilyenable use of LLINs within their family home(s). Likewise, thesociality enabled by hammocks is in linewith the social eventsadolescent boys prefer to participate in, and the sleeping sitesthat are available to them during such social events are simplymore hammock friendly.

Likeme,mybednet hasbeenhung in the village house, butI sometimes went to sleep at the farm with friends. Forexample, there is amanwhohas abig house, his househasenough spaces for boys to sleep together, around 16–17boys sleeping with hanging hammocks, but some peopleprefer to sleep with bed net. (Tompuon teenage boy)

Social custom further affords adult bachelors the freedomto participate in various evening social activities and sleep invarious locations as preferred; a practice also facilitated bylightweight and easy-to-carry hammocks. When they getmarried, men start sleeping together with their wives in a bednet, offering additional privacy in shared houses with ex-tended families. However, when a man loses his wife, eitherbecause of untimely passing or to divorce, or remains abachelor even in later years, it is common for him to stop usingbed nets, stressing again the importance of privacy and therequired size of the net and mesh that affords a sense of pri-vacywhencommitting to consistent net use.Menwho state toalways sleep in a bed netwhen asked in a surveywill likely onlydo so while they (still) require that privacy.

DISCUSSION

Although data derived from malaria indicator and othersurveys estimate a high level of bed net use in Ratanakiri and


implicitly assume this use is continuous and consistent,40,41

our research demonstrates that there is considerable vari-ability in local sleeping patterns and net use that resists beingcaptured by quantitative indicators derived from questionsthat assume a stable pattern or straightforward response. Inboth policy and research circles, the default, or implicit, as-sumptions about the generic social context in which LLINs areused include that sleeping arrangements are stable, each netwill be used by up to two adults, and that distributing sufficientLLINs based on household size to every household at-risk,coupled with behavioral change communication interven-tions, should achieve universal coverage. Although it has beenrecognized that the net use gap, referring to the difference innumbers between household ownership of at least one LLINand household use of LLINs, is not simply due to a failure ofbehavioral change communication but to a lack of intra-household access,42 our results show that deriving “use” fromthequestions that construct these indicators further overlooksthe social processes that shape the variation in net use inpractice.5 These indicators conceal certain windows in spaceand time inwhich socially clustered groups of people at-risk ofmalaria may be exposed to biting vectors because LLINs arenot fully aligned to use in the local sociocultural and environ-mental conditions. As such, nonuse of LLINs is not randomlydistributed within a population, but clustered by certain socialcharacteristics, potentially creating sufficient opportunities formalaria transmission to persist in this pre-elimination setting.For policy-makers and implementers, the relevance of

contextualized findings such as ours is frequently counteredwith the argument that the minimal increase in LLIN useexpected by taking “user preferences” into account is notworth the investment that countries have to make to accom-modate those preferences.43 Narrowly conceived “user pref-erences” (e.g., color and shape of nets) alone may indeed notimpact on LLIN use substantially in nationally aggregateddata, and this logic holds true in settings that are still focusedon controlling malaria and where further upscaling of LLINscan substantially reduce transmission. Existing malaria indi-catorsmaybeappropriate for rapidly assessing the rollout anduptakeof LLINs inmalaria control settings, but they should notbe uncritically extended to address important knowledgegaps in the new malaria elimination era.In elimination settings,with increasing heterogeneity,where

persisting transmission clusters in specific populations andlocations despite assumed universal coverage of LLINs, in-dicators at population level are less useful for assessing crit-ical gaps in coverage. Furthermore, when universal coverageis assumed, persisting malaria transmission is frequently at-tributed to “residual transmission,” axiomatically defined as alltransmission that continues after universal coverage with ef-fective LLINs and IRS has been achieved.4 In reality, ongoingtransmission may be due to truly “residual” transmission,which in Ratanakiri province is not only caused by early andoutdoor biting vectors and outdoor evening and night-timeactivities7,44 but also due to variability in LLIN use partly in-duced by a lack of appropriateness of the intervention. Itremains unclear to what extent “indoor” night-time bitingcontributes to transmission in Ratanakiri, although our pre-vious research suggests this could still account for almost50% of the total exposure.7 Furthermore, in Ratanakiri, even“early” mosquito biting can often be categorized under “so-cial” nighttime because sleeping times are much earlier than

frequently assumed,7 as also shown for Vietnam,45 andtherefore can be addressed with LLINs.In addition to the commonly used criteria of “efficacy” and

“effectiveness” when assessing LLINs as a tool for malariaprevention,46–50 we have proposed “appropriateness” as athird key concept by which to evaluate LLINs and other globalhealth interventions.Wehaveargued that appropriateness is aconcept that applies equally to themeasurement of the impactof the intervention as to the intervention itself. Although theconcept of appropriateness will require substantial develop-ment to be fully realized, we propose a heuristic tool (Figure 1)to aid researchers and program implementers to reflect onhow the (in)appropriateness of interventions relates to the (in)appropriateness of indicators. For example, indicators thatassume stable behaviors and consistency in LLIN use cannotaddress variability in LLIN use, in itself partly induced by a lackof appropriateness of the distributed LLINs in the local con-text. And yet, without recognizing ormeasuring this variability,programmers lack an “evidence base” to guide optimizing theappropriateness of LLIN interventions in different contexts.More crucially, optimizing the appropriateness of LLIN inter-ventions could meaningfully reduce malaria incidence, in-cluding in settings where universal coverage is assumedto have been achieved or nearly achieved. As of yet, othercontrol measures such as (spatial) repellents or insecticide-treated hammocks have not proven their efficacy or cost-effectiveness in interrupting transmission in low transmissionsettings. Trialing novel technologies should not fully divertattention and resources from theopportunity to reconsider theappropriateness of and increase LLIN use, one of the mosteffective tools for reducing malaria throughout the malaria-endemic world.The inappropriateness of measurements used for the

evaluation of an intervention cycles back to mischaracteriza-tions of the appropriateness of the intervention, as evident inguidance issuedby theWHO for addressingwhat is implied assuboptimal (and decontextualized) LLIN use.

Where there is evidence that LLINs are not being ade-quately cared for or used, programmes should design andimplement behaviour change communication activitiesaimed at improving these behaviours (p. 3).1

By contrast, optimizing appropriateness could lead to re-duced variability in LLIN use, which would in turn render in-dicators of LLIN use more meaningful. For example, the studypopulation in Ratanakiri required multiple functions fromLLINs in addition to protection frombitingmosquitoes. Similarto results shown in Peru,6 where the authors called for anincreased understanding of the “adequacy” of bed nets in thecontext of implementation, nets created a physical barrierproviding users with privacy and prevented dirt from fallingonto sleeping spaces from thatched roofs. Also shown inSenegal and Tanzania,51,52 nets were mainly valued for theirability to protect from general insect nuisance, including in-sects smaller than local species of Anopheles. By directlyengaging with the fact that the most vulnerable householdsfrequently require or adapt household items to have multiplefunctions, and not dismissing untreated market net use or nonet use as misinformed or incorrect, LLINs could be de-veloped to fulfill these additional requirements to increase


effective use.6 Moreover, the LLINs distributed in Ratanakiriwere not thought to be of the quality their specifications claim:in local conditions, these LLIN are reported to deterioratequickly andare therefore perceived tobe inadequate. This lackof appropriateness for local social and environmental condi-tions has been shown to limit LLIN use in other malaria-endemic settings.5–7,9,17

CONCLUSION

Failure to measure and account for variability in LLIN useleads researchers and programmers to underestimate thecontinuing potential of LLINs for reducing and interruptingmalaria transmission, including the opportunity to improveLLIN appropriateness as a path toward increasing effectiveuse. The importance of these arguments should be recog-nized in consideration of the shift in research priorities towardinterventions for targeting residual transmission, aswell as thejustification of mass drug administration programs and otherelimination interventions, all of which are intended for imple-mentation after malaria control programs assume that uni-versal LLIN coverage has been achieved.

Received September 6, 2018. Accepted for publication March 16,2019.

Published online April 15, 2019.

Acknowledgments: We would like to thank the epidemiological re-search team coordinated by Somony Heng in Cambodia, and ourcolleagues at ITM Marc Coosemans, Vincent Sluydts, and LiesDurnez for all the hard work and their important conceptual andpractical contributions during both fieldwork and analysis.Wewouldalso like to thank Tho Sochantha and Mao Sokny at CNM for their

logistical support, and all the malaria staff at the provincial healthdepartment of Ratanakiri Province, including the Village MalariaWorkers, for facilitating the research. Finally and most importantly,we thank all the indigenous families in Ratanakiri that opened theirhomes to us.

Financial support: This studywas financed by theBelgianDirectorate forDevelopment Cooperation within the framework of an institutional col-laborationbetweenCNM,PhnomPenh, and ITM,Antwerp.This researchbuilt further on data collected during a community-randomized trialfunded by Bill & Melinda Gates Foundation (#OPP1032354).

Authors’ addresses: Charlotte Gryseels, Melanie Bannister-Tyrrell,and Koen Peeters Grietens, Institute of Tropical Medicine, Antwerp,Belgium, E-mails: [email protected], [email protected],and [email protected]. Sambunny Uk, Srun Set, and Suon Sokha,National Center for Parasitology Entomology and Malaria Control,PhnomPenh, Cambodia, E-mails: [email protected], [email protected], and [email protected]. René Gerrets, Amsterdam In-stitute for Social Science Research, Amsterdam, The Netherlands,E-mail: [email protected].

This is an open-access article distributed under the terms of theCreative Commons Attribution License, which permits unrestricteduse, distribution, and reproduction in any medium, provided theoriginal author and source are credited.

REFERENCES

1. WHO Global Malaria Programme, 2007. Insecticide-TreatedNets: AWHOPosition Statement. Geneva, Switzerland:WHO.Available at: http://www.ivcc.com/sites/ivcc.mrmdev.co.uk/files/content/itnspospaperfinal.pdf. Accessed October 6,2015.

2. WHO, 2013. Vector Control Technical Expert Group Report toMPAC: Methods for Maintaining Coverage with Long-LastingInsecticidal Nets (LLINs). Geneva, Switzerland: WHO. Availableat: https://www.who.int/malaria/mpac/mpac_sep13_vcteg_llin_survival_report.pdf?ua=1. Accessed October 6, 2015.

FIGURE 1. Framework for assessing the appropriateness of the intervention.


mailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]://creativecommons.org/licenses/by/4.0/http://www.ivcc.com/sites/ivcc.mrmdev.co.uk/files/content/itnspospaperfinal.pdfhttp://www.ivcc.com/sites/ivcc.mrmdev.co.uk/files/content/itnspospaperfinal.pdfhttps://www.who.int/malaria/mpac/mpac_sep13_vcteg_llin_survival_report.pdf?ua=1https://www.who.int/malaria/mpac/mpac_sep13_vcteg_llin_survival_report.pdf?ua=1

3. Roll Back Malaria Partnership to End Malaria, 2013. Malaria In-dicator Survey Toolkit. Available at: https://malariasurveys.org/toolkit.cfm. Accessed October 6, 2015.

4. Killeen GF, 2014. Characterizing, controlling and eliminating re-sidual malaria transmission.Malar J 13: 330.

5. Dunn CE, Le Mare A, Makungu C, 2011. Malaria risk behaviours,socio-cultural practices and rural livelihoods in southern Tan-zania: implications for bednet usage.SocSciMed72:408–417.

6. Peeters Grietens K et al., 2013. Traditional nets interfere with theuptakeof long-lasting insecticidal nets in thePeruvianAmazon:the relevanceof net preference for achieving high coverage anduse. PLoS One 8: e50294.

7. Gryseels C et al., 2015. Re-imagining malaria: heterogeneity ofhuman and mosquito behaviour in relation to residual malariatransmission in Cambodia.Malar J 14: 165.

8. Grietens KP, Xuan XN, Ribera J, Duc TN, Bortel Wv, Ba NT, VanKP, Xuan HL, D’Alessandro U, Erhart A, 2012. Social determi-nants of long lasting insecticidal hammock use among the Ra-glai ethnic minority in Vietnam: implications for forest malariacontrol. PLoS One 7: e29991.

9. MonroeA,HarveySA, LamY,MuhangiD, Loll D, Kabali AT,WeberR, 2014. “People will say that I am proud”: a qualitative study ofbarriers to bed net use away from home in four Ugandan dis-tricts.Malar J 13: 82.

10. Atkinson J-A, Bobogare A, Fitzgerald L, Boaz L, Appleyard B,Toaliu H, 2009. A qualitative study on the acceptability andpreference of three types of long-lasting insecticide-treatedbed nets in Solomon Islands: implications for malaria elimina-tion.Malar J 8: 119.

11. Gore-Langton GR et al., 2015. Investigating the acceptability ofnon-mesh, long-lasting insecticidal nets amongst nomadiccommunities in Garissa County, Kenya using a prospective,longitudinal study design and cross-sectional household sur-veys.Malar J 14: 52.

12. Binka FN, Adongo P, 1997. Acceptability and use of insecticideimpregnated bednets in northern Ghana. Trop Med Int Health2: 499–507.

13. Gunasekaran K, Sahu SS, Vijayakumar KN, Jambulingam P,2009. Acceptability, willing to purchase and use long lastinginsecticide treated mosquito nets in Orissa state, India. ActaTrop 112: 149–155.

14. Ingabire CM, Rulisa A, van Kempen L, Muvunyi C, KoenraadtCJM, Van Vugt M, Mutesa L, Van Den Borne B, Alaii J, 2015.Factors impeding the acceptability and use of malaria pre-ventive measures: implications for malaria elimination in east-ern Rwanda.Malar J 14: 136.

15. PeetersGrietensK et al., 2015.Characterizing types ofmobility toinform differential and targetedmalaria elimination strategies innortheast Cambodia. Sci Rep 5: 16837.

16. Monroe A, Asamoah O, Lam Y, Koenker H, Psychas P, Lynch M,Ricotta E, Hornston S, Berman A, Harvey SA, 2015. Outdoor-sleeping and other night-time activities in northern Ghana: im-plications for residual transmission and malaria prevention.Malar J 14: 35.

17. Brook RH, 2009. Assessing the appropriateness of care–its timehas come. JAMA 302: 997–998.

18. Sanmartin C et al., 2008. Appropriateness of healthcare inter-ventions: concepts and scopingof the published literature. Int JTechnol Assess Health Care 24: 342–349.

19. Gryseels C et al., 2015. Factors influencing the use of topicalrepellents: implications for the effectiveness of malaria elimi-nation strategies. Sci Rep 5: 16847.

20. Sluydts V et al., 2014. Spatial clustering and risk factors of malariainfections in Ratanakiri Province, Cambodia.Malar J 13: 387.

21. Heng S, Durnez L, Mao S, Siv S, Tho S, Mean V, Sluydts V,Coosemans M, 2017. Passive case detection of malaria inRatanakiri Province (Cambodia) to detect villages at higher riskfor malaria.Malar J 16: 104.

22. Canavati SE, Lawpoolsri S, Quintero CE, Nguon C, Ly P,PukrittayakameeS,SintasathD,SinghasivanonP,GrietensKP,Whittaker MA, 2016. Villagemalaria worker performance key tothe elimination of artemisinin-resistant malaria: a WesternCambodia health system assessment.Malar J 15: 282.

23. Bourdier F, 2006.TheMountain of PreciousStones. PhnomPenh,Cambodia: The Center for Khmer Studies.

24. Bourdier F, 1998. Health, women and environment in a marginalregion of north-eastern Cambodia. GeoJournal 44: 141–150.

25. Gryseels C et al., 2015. High mobility and low use of malaria pre-ventive measures among the Jarai male youth along theCambodian-Vietnameseborder.AmJTropMedHyg93:810–818.

26. Tami A, Mubyazi G, Talbert A, Mshinda H, Duchon S, Lengeler C,2004. Evaluation of Olyset insecticide-treated nets distributedseven years previously in Tanzania.Malar J 3: 19.

27. Haas PM, 2011. Introduction: epistemic communities and in-ternational policy coordination. Int Organ 46: 1–35.

28. KnorrCetinaK,1999.EpistemicCultures:HowtheSciencesMakeKnowledge. Cambridge, MA: Harvard University Press.

29. Dalglish SL, George A, Shearer JC, Bennett S, 2018. Epistemiccommunities in global health and the development of childsurvival policy: a case study of iCCM. Health Policy Plan30 (Suppl 2): ii12–ii25.

30. Kelly AH, Beisel U, 2011. Neglected malarias: the frontlines andback alleys of global health. BioSocieties 6: 71–87.

31. The Global Fund: Office of the Inspector General, 2013. In-vestigation Report of Global Fund Grants to CambodiaPrincipal Recipients CNM, NCHADS and MoH and NCHADSSub-Recipient MEDiCAM. Geneva, Switzerland: The GlobalFund. Available at: https://www.theglobalfund.org/media/2774/oig_gfoig13050investigationcambodia_report_en.pdf?u=636784021740000000. Accessed October 6, 2015.

32. Aleme A, Girma E, Fentahun N, 2014. Willingness to pay forinsecticide-treated nets in Berehet District, Amhara Region,Northern Ethiopia: implication of social marketing. Ethiop JHealth Sci 24: 75–84.

33. NathanR,MasanjaH,MshindaH, Schellenberg JA, deSavignyD,Lengeler C, TannerM, VictoraCG, 2004.Mosquito nets and thepoor: can social marketing redress inequities in access? TropMed Int Health 9: 1121–1126.

34. Grabowsky M, Nobiya T, Selanikio J, 2007. Sustained high cov-erage of insecticide-treated bednets through combined catch-up and keep-up strategies. Trop Med Int Health 12: 815–822.

35. Willey BA, Paintain S, Mangham L, Armstrong J, Willey BA, 2012.Strategies for delivering insecticide-treated nets at scale formalaria control: a systematic review. Bull World Health Organ90: 672–684.

36. Brinkley J, 2012. Cambodia’s Curse: The Modern History of aTroubled Land. New York, NY: Public Affairs.

37. Cui L et al., 2012. Malaria in the Greater Mekong Subregion:heterogeneity and complexity. Acta Trop 121: 227–239.

38. Erhart A, ThangND, HungNQ, Toi le V, Hung le X, Tuy TQ,Cong leD,SpeybroeckN,CoosemansM,D’AlessandroU, 2004.Forestmalaria in Vietnam: a challenge for control. Am J TropMed Hyg70: 110–118.

39. Erhart A et al., 2005. Epidemiology of forest malaria in centralVietnam: a large scale cross-sectional survey.Malar J 4: 58.

40. Sluydts V et al., 2016. Efficacy of topical mosquito repellent(picaridin) plus long-lasting insecticidal nets versus long-lastinginsecticidal nets alone for control of malaria: a cluster rando-mised controlled trial. Lancet Infect Dis 16: 1169–1177.

41. Huy R, Sovannaroth S, Lek D, Chan V, Chea S, Yeang O, BouKheng T, Montha M, Yom Y, Theassy HC, 2013. CambodiaMalaria Survey. Phnom Penh, Cambodia: Malaria Consortium.

42. Koenker H, Kilian A, 2014. Recalculating the net use gap: a multi-country comparisonof ITNuse versus ITNaccess.PLoSOne9:e97496.

43. Koenker H, Yukich JO, 2017. Effect of user preferences on ITNuse: a review of literature and data.Malar J 16: 233.

44. Durnez L, Mao S, Denis L, Roelants P, Sochantha T, CoosemansM, 2013. Outdoor malaria transmission in forested villages ofCambodia.Malar J 12: 329.

45. Peeters Grietens K, Xuan XN, Van Bortel W, Duc TN, Ribera JM,Ba Nhat T, Van KP, Le Xuan H, D’Alessandro U, Erhart A, 2010.Low perception of malaria risk among the Ra-glai ethnic mi-nority in south-central Vietnam: implications for forest malariacontrol.Malar J 9: 23.

46. Lengeler C, Snow RW, 1996. From efficacy to effectiveness:insecticide-treated bednets in Africa. Bull World Health Organ74: 325–332.

47. Nordon C, Karcher H, Groenwold RHH, Ankarfeldt MZ, Pichler F,Chevrou-severac H, Rossignol M, Abbe A, Abenhaim L;


https://malariasurveys.org/toolkit.cfmhttps://malariasurveys.org/toolkit.cfmhttps://www.theglobalfund.org/media/2774/oig_gfoig13050investigationcambodia_report_en.pdf?u=636784021740000000https://www.theglobalfund.org/media/2774/oig_gfoig13050investigationcambodia_report_en.pdf?u=636784021740000000https://www.theglobalfund.org/media/2774/oig_gfoig13050investigationcambodia_report_en.pdf?u=636784021740000000

GetReal Consortium, 2016. The “efficacy-effectiveness gap”:historical background and current conceptualization. ValueHealth 19: 75–81.

48. Beer N, Ali AS, de SavignyD, Al-Mafazy A-WH,RamsanM, AbassAK, Omari RS, Björkman A, Källander K, 2010. System effec-tiveness of a targeted free mass distribution of long lasting in-secticidal nets in Zanzibar, Tanzania. Malar J 9: 173.

49. Killeen GF, Kihonda J, Lyimo E, Oketch FR, Kotas ME, MathengeE, Schellenberg JA, Lengeler C, Smith TA, Drakeley CJ, 2006.Quantifying behavioural interactions between humans andmosquitoes: evaluating the protective efficacy of insecticidalnets against malaria transmission in rural Tanzania.BMC InfectDis 6: 161.

50. WHO Global Malaria Programme, 2017. Achieving and Main-taining Universal Coverage with Long-Lasting Insecticidal Netsfor Malaria Control. Available at: http://apps.who.int/iris/bitstream/handle/10665/259478/WHO-HTM-GMP-2017.20-eng.pdf?sequence=1. Accessed May 5 2018.

51. Berthe S, Loll D, Faye SL, Wone I, Koenker H, Arnold B, Weber R,2014. “When I sleep under the net, nothing bothers me; I sleepwell and I’m happy”: Senegal’s culture of net use and how in-conveniences to net use do not translate to net abandonment.Malar J 13: 357.

52. Koenker HM, Loll D, Rweyemamu D, Ali AS, 2013. A good night’ssleep and the habit of net use: perceptions of risk and reasonsfor bed net use in Bukoba and Zanzibar.Malar J 12: 203.

http://apps.who.int/iris/bitstream/handle/10665/259478/WHO-HTM-GMP-2017.20-eng.pdf?sequence=1http://apps.who.int/iris/bitstream/handle/10665/259478/WHO-HTM-GMP-2017.20-eng.pdf?sequence=1http://apps.who.int/iris/bitstream/handle/10665/259478/WHO-HTM-GMP-2017.20-eng.pdf?sequence=1

uva-dare (digital academic repository) a critical enquiry ... · charlotte gryseels,1* melanie...

Documents