Cdc protocol post-campaign survey senegal 2009111

PROJECT OVERVIEW
Title: Post-Campaign Survey to Evaluate Senegal’s 2009 Nationwide Long-lasting
Insecticide Treated Net (LLIN) Distribution

Protocol summary:
In 2009, Senegal conducted its first national LLIN distribution campaign, distributing in two
phases approximately 2.3 million LLINs nationwide. The first phase, that excluded 6 districts in
Dakar, included the distribution of mebendazole and vitamin A; the second phase in Dakar
included only the distribution of LLINs. While national LLIN distributions have become
increasingly common across sub-Saharan Africa, most are linked with vaccinations in addition to
the other interventions, and distribute all the elements of the campaign at distribution points.
Senegal used a door-to-door strategy, delivering a voucher for an LLIN, together with
mebendazole and vitamin A during the first phase, to each child aged 6-59 months during
household visits; vouchers were later redeemed at distribution points. The Senegal campaign also
included a number of communications strategies to advertise the campaign and communicate the
importance of using insecticide-treated nets (ITNs). Senegalese artist Youssou N’Dour partnered
with the Youssou Ndour Foundation and Malaria No More to develop a mass media campaign
involving television and radio spots and a popular song about malaria prevention with LLINs.
A Malaria Indicator Survey (MIS) was conducted in late 2008 and early 2009, and found that
approximately 22% of the population in general, 28% of children < 5 years and 27% of pregnant
women, had slept under an LLIN the previous night. While the goal for LLIN usage of the
National Malaria Control Program (NMCP) has shifted to universal coverage for the entire
population, the distribution was targeted to children < 5 years. In order to guide further
distribution efforts, it is important to determine to what extent the targeted campaign reached the
general population, and to evaluate the effectiveness of the communications strategies for
promoting campaign participation and LLIN use.
We propose to conduct a community-based, cross-sectional cluster survey with clusters selected
according to probability proportional to size sampling, according to standard MIS methodology.
Selected clusters will be mapped with GPS-assisted PDAs, and households selected by simple
random sample. The questionnaire will be administered on PDA, and will consist of questions
regarding bednet ownership and usage, campaign participation, socioeconomic status, and
household composition.
The results will be crucial as Senegal composes a new five year strategic plan in the upcoming
year, including LLIN strategy, and for planning of future LLIN distribution.
Investigators/collaborators/funding mechanism(s)/Federalwide Assurance numbers/
“engagement in research” status (se
The survey will be funded by the President’s Malaria Initiative (PMI). The Senegal PMI team and
NMCP will coordinate the study and primary CDC researchers will be Dr. Robert Perry, the CDC
PMI Resident Advisor in Senegal, and Dr. Julie Thwing. The implementing partner in Senegal will
be the Centre de Researche et Developpment Humain (CRDH, Center for Research and Human
Development), Director, Salif Ndiaye.
INTRODUCTION
Literature review/current state of knowledge about project topic:
LLINs are a major component of malaria prevention. An effective method to reach and maintain high coverage is to hold mass distribution campaigns, followed by a system of routine distribution to missed children and those born after the distribution. Governments and partner agencies closely scrutinize mass campaigns given their size, demands for resources and visibility. Key strategies for malaria prevention in the Senegal NMCP 2006-2010 Strategic Plan are the distribution of LLINs to pregnant women and children under five years of age and improving and reinforcing communication on the use of ITNs, with a goal that 80% of each group will sleep under an ITN by 2010. The NMCP also is following the “catch-up” and “keep-up” strategies endorsed by the RBM partnership to rapidly increase and maintain high coverage with LLINs. PMI also supports the “catch-up” and “keep-up” strategies and reinforcing communication programs to increase ITN use, with goals that 90% of households possess at least one LLIN and that 85% of pregnant women and children under five years of age sleep under an ITN by 2010. During the campaign, several partners have implemented communications strategies to improve rates of LLIN utilization. The Senegalese Red Cross Society (SRC) volunteers conducted door-to-door visits in 7 of 14 regions to provide communication messages and offer to hang LLINs still not in use. The Youssou Ndour Foundation and Malaria No More are beginning a two-year “Surround Sound” media campaign using Senegalese musicians and their music to boost LLIN usage. Most districts conducted post-campaign door-to-door activities to boost LLIN usage. Peace Corps volunteers also participated in communications activities throughout the country, and are organizing formal communications campaigns in Linguère District (Louga Region) and in Saraya District (Kédougou Region). Senegal ascribes to a universal coverage objective for LLINs. The WHO and other international organizations have also recommended that LLIN distributions target the entire population rather than the traditional vulnerable groups of children under five and pregnant women. The WHO has proposed a working definition of universal coverage as 1 LLIN per every 2 people. However, the Senegal NMCP campaign strategy remains one that targets children under five. In several regions of the country additional distributions have also targeted children under five. The World Bank Nutritional Enhancement Program in early 2008 sponsored the distribution of LLINs in 10 regions of the country to children living in communities served by the program. Shortly after that campaign the PMI distributed LLINs in July 2008 to children in communities not covered by the Nutrition Program in six regions. In November 2009 the Senegal River Basin Development Program will distribute additional LLINs to children living in the three regions covered by that program. The Peace Corps distributed additional LLINs in Kédougou Region in 2009 to cover all sleeping spaces remaining after the national campaign, with needs estimated based on 1 LLIN per 2 people, less the number of LLINs distributed in June. Future LLIN strategies and policy will require an evaluation of the progress to universal coverage achieved by repeated under five campaigns and the description of the remaining gaps. Coverage of interventions delivered during a campaign is also measured through administrative data on the number of interventions delivered, with administrative coverage being the number delivered divided by the estimated target population. Administrative data is readily available during the campaign and can also provide, soon after the campaign, coverage estimates for any level of the health care system, from national to local. Despite these advantages, administrative data only allows the calculation of coverage based on individual children and do not permit the calculation of standard indicators such as the proportion of households with at least one LLIN and the median number of LLINs per household. These data also leave no opportunity to ask additional questions to determine LLIN utilization rates, to evaluate the success of communication strategies, or to collect information from children who did not participate in the campaign. Given these problems and the need to assess the varied strategies and approaches to conducting integrated campaigns, high-quality post-campaign surveys are highly recommended by international partners such as the Alliance for Malaria Prevention, a workstream of the Harmonization Working Group of RBM. These surveys should evaluate all interventions given during the campaign and should use standardized questionnaires and statistically sound methodology such as what is used for a Demographic and Health Survey (DHS), Malaria Indicator Survey (MIS), or Multiple Indicator Cluster Survey (MICS).
Justification for study: The 2009 LLIN distribution was Senegal’s first nationwide free LLIN
distribution. It is critical to assess the progress towards reaching universal coverage achievable
through this strategy and to identify and describe the remaining gaps. After several mass
distributions it is also critical to evaluate communication strategies promoting LLIN use. This
information will guide the development of not only future campaigns but also a system for routine
distribution of LLINs. The results of this study will also form part of the evaluation of the current
NMCP strategic plan and help guide the development of the 2011-2015 plan.

Intended/potential use of study findings:
The findings from this study are intended to guide the planning and implementation of future
LLIN communication activities to increase utilization and LLIN distribution strategies by the
NMCP and its partners.

Study design/location(s):
The evaluation in Senegal will be a nationwide, community-based, cross-sectional cluster survey,
with clusters to be interviewed selected by two stage probability proportional to size sampling,
and with the final selection of households to be interviewed selected by simple random sample. It
will be held December 2009, five months after the June phase of the campaign and two months
after the October phase, and at the end of the rainy season.
Objectives:
1. Compare regions to assess ability of different communication strategies in reaching the population and in increasing LLIN utilization. 2. Determine pre-campaign household LLIN coverage and the median number of LLIN(s) 3. Determine the extent that repeated campaigns targeting children under five achieve universal coverage goals and describe the remaining gaps, comparing regions participating in both 2008 and 2009 campaigns to regions participating only in 2009 campaign. 4. Assess post-campaign LLIN utilization rates in the general population, children under five 5. Determine possession and usage of LLINs distributed during 2008 LLIN campaign, 15 6. Assess success of the campaign in delivering vitamin A, mebendazole, coupons and LLINs

Hypotheses or questions:
1) How does increased availability affect LLIN utilization rates?
2) How well were communications strategies implemented and what was their impact on
3) How well do under-five campaigns reach universal coverage goals and what are the characteristics of the households and individuals that remain to reach this goal?
General approach: The general approach will be observational.
PROCEDURES/METHODS
DESIGN
How study design or surveillance system addresses hypotheses and meets objectives:
The cross-sectional, community based cluster survey design best meets the need to have reliable
population-level data on ITN ownership and usage, as well as campaign participation. The
methods and survey questionnaire will allow comparisons with Demographic and Health Surveys
and Malaria Indicator Surveys done in Senegal.
Audience and stakeholder participation:
The National Malaria Control Program and partners have been involved in discussing the scope of
the survey. The Senegalese population is the primary stakeholders, and will have a chance to
participate or not, and to voice their questions and concerns.

Cost benefit/prevention effectiveness: The cost of this survey is modest in comparison to the
cost of LLINs distributed. This survey does not include a cost / benefit or prevention
effectiveness component. However, the coverage and utilization data derived from this survey
will be used in the overall evaluation of the campaign and compared to the costs per LLIN
delivered and the costs of communication activities.
Study time line: The survey will take place during December 2009, with preliminary results
disseminated within a month after the survey, and the final report available within 3 months.
Accelerated protocol review: N/A
STUDY POPULATION
Description and source of study population and catchment area:
The entire population of Senegal, including both male and female groups of all ages, will be
eligible. Using probability proportional to size sampling at the level of the cluster, and simple
random sampling within the cluster, every household has an equal probability of being selected.
Case definitions: This evaluation will examine ownership and usage of LLINs. Household
ownership of an LLIN is defined as the presence of an LLIN, whereas utilization is defined as
having slept under the LLIN the previous night.

Participant inclusion/exclusion criteria: Participating households will be selected through two
stage probability proportional to size selection, with the final stage being simple random sample.
No selected household or individual there-in will be excluded unless they refuse to participate or
are absent for three visits by the survey team.
Justification of exclusion of any sub-segment of the population:
No sub-segment is excluded.
Estimated number of participants: The sample size will be 3360 households and 30,000
individuals based on nine persons per household.
Sampling, including sample size and statistical power:
Based on an expected frequency of LLIN use in children under 5 of 42% (MIS 2008/9), a design effect of 2, an estimated mean of 1.42 children under five per household, and a 95% participation rate, 243 households are needed per medical region to provide an estimate accurate to +/- 13% for that region. For LLIN utilization in pregnant women, assuming similar parameters except for a mean of 0.16 pregnant women per household, 3213 households will be necessary nationwide to have an estimate accurate to +/- 11%. Selecting 30 households in 8 clusters in each of the 14 medical regions should provide 240 households per region and 3360 households nationwide.
Enrollment: The survey will be described to the head of every selected household, or the
representative, and this person will be asked if they are willing to participate.
Consent Process: The consent form detailing the purpose of the survey, confidentiality of
responses, and non-penalty for non-participation or withdrawal at any time will be built into the
PDA questionnaire and read at the beginning. Consent will be verbal and documentation will be
integrated into the program on the PDA. All participants will be given an opportunity to ask
questions and voice concerns.
VARIABLES/INTERVENTIONS
Variables: Variables to be collected for each household include a list of household residents with
basic demographic information and LLIN use recorded for each person, answers to standard
questions used to develop wealth quintiles according to World Bank methods, information on the
source, type, use and handling of each LLIN, and a questionnaire for each mother or caregiver
covering campaign participation, questions regarding whether each child received the designated
interventions, and visits of community mobilization efforts to promote LLIN use since the
campaign.
To determine the progress towards universal coverage several operational definitions will be used and reported on. One will be based on the WHO operational definition of universal coverage (2 people per LLIN) by calculating the percentage of households where the number of LLIN(s) is greater than or equal to the household size divided by two. Another operational definition that will be considered is setting a threshold based on the median household size divided by two (e.g, 4.5 LLIN per household if the median household size is ~9) and determining the percentage of households reaching that threshold. Utilization will be determined for the general population, for children under five and for pregnant women, by asking about sleeping under an LLIN the night before the survey for each person in the household. Study instruments, including questionnaires, laboratory instruments, and analytic tests:
Data collection will be based on a survey questionnaire programmed into PDAs and no invasive
procedures will be used,
Intervention or treatment:
No interventions or treatment will be provided during this survey.

Outcomes and minimum meaningful differences: The key outcomes of interest are household
ITN ownership, and usage both by the population as a whole and by the vulnerable groups of
children < 5 and pregnant women. We estimate that the campaign resulted in an approximately
20% increase in each of these variables.

Training for all study personnel:
The data collectors will undergo a 5 day training course: 1 day covering the purpose of the
survey, the protocol, principles of research, and interview skills, one day reviewing the
questionnaire in depth and agreeing on phrasing of the questions in common local languages, and
two days to become familiar with the PDAs, using the GPS, and practicing administering the
questionnaire to one another. The fifth day will be a field trial.
DATA HANDLING AND ANALYSIS
Data analysis plan, including statistical methodology and planned tables and figures:
The primary figures will give LLIN ownership at the household level as determined by the
proportion of households with at least one LLIN, the proportion meeting different universal
coverage definitions and the number of LLINs per household, cross tabulated by wealth quintile,
rural/urban locale, socio-economic status, region, participation in the 2008 or 2009 campaigns,
and presence or absence of a child eligible for the campaign. Individual LLIN use in the general
population and among vulnerable groups will be examined by those factors, as well as
demographic factors including the position within the household, LLIN factors such as the
number or type of LLINs in the household, and exposure to various communications strategies
and messages. Additional tables will include campaign participation by logistics such as reach of
campaign communications, distance of household from distribution post, demographic factors and
socio-economic status. Campaign coverage will be calculated for each intervention and compared
to administrative results. Data will be weighted by the probability of selection, adjusted for
cluster.
Data collection, information management and analysis software: The data will be collected on
a PDA programmed in Visual CE (Syware, Cambridge, MA), which will be backed up nightly. At
the end of the survey, all PDAs will be downloaded to an Access database. Data analysis will be
conducted with SAS 9.2 (SAS, Cary, NC).

Data entry, editing and management, including handling of data collection forms, different
versions of data, and data storage and disposition:
Data collection on PDAs allows for validity checking of responses, ensuring following of proper
sequence of questions and completeness of the questionnaires. The lack of paper forms means no
need for data entry. PDAs will be password protected. Because the PDAs are downloaded at the
end of the survey, only one version of the data needs to exist. Final data will be the property of
the President’s Malaria Initiative and the Senegal National Malaria Control Program.

Quality control/assurance: Data checks are built in the PDAs to force answers and reject
impossible answers, and data are entered electronically at the same time as it is generated. Each
questionnaire is backed up as it is completed. Collection of GPS coordinates can document where
interviewers completed questionnaires without the supervisor needing to be physically present at
the selected household.
Handling results in the absence of a reference test: No laboratory testing will be performed.
The survey method is considered the reference method for determining many outcome variables.
Measurement/estimation and adjustment for cross reactivity: N/A
Verifying independence of tests used to confirm results of new test being studied: N/A
Bias in data collection, measurement and analysis:
The primary biases of concern in the research are recall bias, non-response bias and reporting bias.
We are conducting the survey within 6 months of the distribution to diminish recall bias, and
structuring the questionnaire to ensure that as few questions as possible refer to periods of greater
than 6 months. ITN use is defined as use the previous night, and household ownership is based on
observation. Non-response bias will be minimized by requesting that teams visit households at
least three times, making appointments each time, before excluding the household. No
replacements will be made for non-responding households.
Surveys are prone to reporting bias, or social desirability bias. We will minimize this by using
interviewers who are not local to the area being surveyed, and emphasizing the confidential nature
of the results, as well as the lack of incentive for any given response.
Intermediate reviews and analyses: N/A
Limitations of study:
While the sampling strategy achieves the most representative sample possible, it is impossible to
completely eliminate any factor that might result in some bias in the results. Household data are
most reliable for determining ITN ownership and usage, but this survey will not be addressing
treatment seeking behavior, or relating ITN ownership to the outcomes of anemia or parasitemia.
Perhaps the greatest potential weaknesses are recall bias and social desirability. Also, peak malaria
transmission and LLIN use is usually in September – October of each year.

HANDLING OF UNEXPECTED OR ADVERSE EVENTS
Response to new or unexpected findings and to changes in the study environment:
We are attempting to minimize this by conducting the survey in as short a period as possible.
There will be selected alternate clusters in the case that a cluster is inaccessible.

Identifying, managing, and reporting adverse events: We do not expect any adverse events in
the answers to a questionnaire. Our teams will attempt to offer assistance with referral as needed
if seriously ill individuals are encountered.

Anticipated products or inventions resulting from the study and their use: N/A
Disseminating results to public: Findings will be communicated to the National Malaria Control
Program and partners, and a final report will be produced. A manuscript for submission to a peer-
reviewed journal will be prepared.

REFERENCES
1) Ndiaye, S, et al. 2006 2006 Senegal Malaria Indicator Survey. Calverton, Maryland USA:
Centre de recherché pour le développement humain (Sénégal) and Macro International
2) Niaye, S, et al. 2009. 2008/9 Senegal Malaria Indicator Survey. Calverton, Maryland USA:
Centre de recherché pour le développement humain (Sénégal) and Macro International
3) Ndiaye, S, Ayad, M. 2006. 2005 Senegal Demographic and Health Survey (DHS). Calverton,
Maryland USA: Centre de recherche pour le développement humain (Sénégal) and ORC Macro


SCIENTIFIC PEER REVIEW
The protocol will be reviewed by an interagency team and by the Branch Chief, and submitted to
the Human Subjects Committee as required.
CONFLICTS OF INTEREST
We do not anticipate any conflicts of interest surrounding the results of this evaluation.
ANNEX : Standard Malaria Indicator Consent form
INTRODUCTION ET CONSENTEMENT
Bonjour. Mon nom est _______________________ et je travaille pour le Ministère de la Santé, de la Prévention et de l’Hygiène Publique. Nous sommes en train d’effectuer une enquête nationale sur le paludisme et les moustiquaires. Nous souhaiterions que vous participiez à cette enquête. Les informations que vous allez nous fournir aideront le Gouvernement à planifier les services de santé. L’entretien prend généralement entre 10 et 20 minutes. Toute information que vous nous fournirez restera strictement. La participation à cette enquête est volontaire ; s'il y a une question à laquelle vous ne voulez pas répondre, faites-le moi savoir et je passerai à la question suivante. Vous pouvez également arrêter l'entretien à tout moment. Nous espérons cependant que vous participerez à cette enquête car votre opinion est particulièrement importante pour nous. Avez-vous des questions sur l’enquête ? Puis-je commencer l’entretien maintenant ?
INTRODUCTION AND CONSENT

Hello, My name is ____________ and I work for the Ministry of Health, Prevention, and Public Health. We are conducting a national survey on malaria and insecticide-treated bednets. We are asking you to participate in this survey. The information will help the government to plan public health services. The interview will take 10-20 minutes. Any information that you give will be strictly confidential. Participation in this survey is voluntary. If there is any question you don’t want to answer, let me know and I will go to the next question. You can also stop the interview at any time. We hope that you will participate because your opinion is important to us. Do you have any questions about the survey? Can we start the interview?

Source: http://www.allianceformalariaprevention.com/documents/CDC%20protocol%20post-campaign%20survey%20Senegal%2020091111.pdf