As a household level survey, a malaria indicator survey (MIS) is particularly useful in countries where malaria is a major public health problem. An MIS provides an opportunity to measure the coverage of interventions that primarily target the household level, such as insecticidetreated nets and behaviour change communication, and also helps to understand patterns of antimalarial use among target populations.
The MIS complements other household surveys such as the Demographic and Health Surveys and Multiple Indicator Cluster Surveys in evaluating progress against malaria control targets.
Unlike other surveys, MIS is conducted during the peak malaria transmission season, thus giving a true picture of malaria prevalence among target populations. This is the second MIS to be conducted in Kenya and there are some differences from the fi rst one, conducted in 2007.
For one thing, the 2010 MIS sampled all districts in Kenya, weighting samples by malaria epidemiology, while in 2007, six districts in areas of low or no malaria risk were not included. Second, the 2007 MIS included only children less than fi ve years of age, but this one covered children up through 14 years.
The main objectives of the 2010 KMIS were to measure progress achieved in key malaria indicators since the 2007 KMIS and to provide a baseline for the NMS 2009–2017. The specifi c objectives were:
1. To determine the status of coverage of various key malaria intervention measures (e.g., bed net coverage and use, preventive measures during pregnancy, etc.).
2. To assess the prevalence of malaria among children 3 months to 14 years.
3. To assess the level of anaemia among children 6 months to 14 years.
Kind of Data
Sample survey data [ssd]
Unit of Analysis
Household, Individual, children age 3 months to 14 years, women age 15 to 49
v1.1: Data captured as it was collected from the field.
v1.2: Cleaned data and validated data
Initially the data was in captured using MS-ACCESS, exported to CSpro 4.1 and then exported to SPSS for Analysis
The scope of the 2010 Kenya Malaria Indicator Survey includes:
- Household: age, sex, relationship to the head of the household, characteristics of the household dwelling, such as the water source; type of toilet facilities; materials used for the roof, floors, and walls; and possession and use of mosquito nets.
- Women: background characteristics, reproductive history, use of intermittent preventive treatment (IPT) during pregnancy for recent births, fever prevalence and treatment among children under five, and knowledge and attitudes regarding malaria and child survival.
The survey covered all de jure household members (usual residents) and all women aged between 15-49 years.
Producers and sponsors
Kenya National Bureau of Statistics
Ministry of Planning and National Development
Division of Malaria Control
Ministry of Public Health and Sanitation
World Health Organization
Center for Disease Control and Prevention
United Kingdom Department for Internation International Development
United states President's Malaria Initiative
United Nations Children's Fund
The 2010 KMIS was a representative probability sample designed to produce estimates for the specified domains from household populations in Kenya. The level of malaria endemic in Kenya varies from one area to another and can be classified into five malaria endemic regions. These regions, listed below, served as the domains for the survey.
1. Highland epidemic prone
2. Lake endemic
3. Coast endemic
4. Semi-arid, seasonal risk
5. Low risk
In addition, the five regions are categorized into either urban or rural areas and implicitly provide two domains for analysis, at the national level.
The sampling frame for 2010 KMIS was the National Sample Survey and Evaluation Programme (NASSEP) IV. The frame is a two-stage stratified cluster sample format. The first stage involved selection of primary sampling units (PSUs), which were census enumeration areas (EAs), using the probability proportional to measure of size method, with the districts as the first level of stratification. The second stage involved the selection of households for various surveys. EAs were selected with a basis of one measure of size (MOS) defined as the ultimate cluster with an average of 100 households and constituting one (or more) EAs. The MOS was defined with a lower limit of 50 households and an upper limit of 149 households. Prior to selection, those EAs with fewer than 50 households were merged with the neighboring ones to form the minimum requirements for the MOS.
During listing of selected EAs for the frame, those with more than 149 households were segmented and only one segment randomly picked to constitute a cluster. NASSEP IV has a total of 1,800 clusters with 1,260 being rural areas while the remaining 540 are urban. The frame as undergone regular updates.
Sample Size and Allocation
The sample size of 7,200 households that was used in the 2007 KMIS was maintained for the 2010 KMIS. The precision for key malaria indicators for populations at greater risk of malaria (pregnant women and children aged five years and below) are important for KMIS. The number of pregnant women, at a given time, is smaller than the number of children aged five and below and, therefore, indicators based on pregnant women are the determinants for the sample size.
The allocation of the sample to the domains was done using the power allocation method. This method was appropriate, instead of proportional allocation, to ensure that the domain with the lowest proportion of households was oversampled for valid estimates.
Household and Cluster Sampling
A fi rst-stage selection involved selection of the clusters by KNBS for the specified domains. The clusters were selected from the NASSEP IV frame with equal probability within each frame stratum. The selection of the clusters was expected to retain the probability proportional to measure of size design used in creation of the frame.
A second-stage sampling was conducted at the time of field work using personal digital assistants (PDAs). All households within a cluster were to be listed using PDAs fitted with global positioning units and a simple random sample of 30 households per cluster selected for interviewing.
Every attempt was to be made to conduct interviews in the 30 selected households, and up to three visits were expected be made to ascertain compliance in case of absence of all household members (or any household members in the case of malaria parasite testing) to minimize potential bias. Non-responding households were strictly not to be replaced.
Further details on the sample design are provided in Appendix A of the final report
Ninety-three percent of the targeted households were interviewed. The survey yielded response rates of 93 percent and 94 per cent for eligible children and women, respectively. Response rates for children under five reported by interviewed women were lower (74 percent). Response rates are higher in rural areas than in urban areas.
Sample allocation among the domains was not proportional and therefore, the resulting sample was not self-weighting- Weighting adjustments were done to provide comparable estimates for the domains of study weighting was done using the frame design selection probabilities and adjusted to cater for household and individual non-response and the aggregates weights were normalized and then applied to the data.
Dates of Data Collection
Data Collection Mode
There is one supervisor for each of the 25 data collection teams in the field.
Data Collection Notes
Field staff training for KMIS 2010 was conducted from 7 to 17 July 2010 in Nakuru. A total of 148 participants took part, including 25 team supervisors, 59 research assistants, 56 health workers (28 clinicians and 28 laboratory technologists) and the 8 Provincial Malaria Control Coordinators.
Team supervisors and research assistants were trained on the rationale and methodology of KMIS data collection using PDAs and global positioning system technology. Key concepts in household listing, interviewing skills and filling the questionnaires using PDAs were emphasized.
Health workers were trained on how to conduct informed consent and specimen collection procedures like preparing blood smears and performing rapid diagnostic tests (RDT) for malaria and anaemia testing. Participants also received refresher training on the management of uncomplicated malaria and referral of complicated malaria cases.
As part of the training, the questionnaires were pre-tested in six urban areas in Nakuru that were purposively selected because of their proximity to the training center. The questionnaires were then adapted and finalized for the actual field work.
Among the 148 trained field staff, only 128 were selected for final data collection, while 12 were kept as reserves in case of attrition. The remaining eight were Provincial Malaria Control Coordinators who acted as national coordinators during the fieldwork. Twenty-five teams, each comprising one supervisor, two research assistants and two health workers (a clinician and a laboratory technician) constituted the field staff. Teams were each allocated clusters in the different districts in accordance to their local language competency. Each team was assigned a driver and supplied with logistics for the survey activity. The fieldwork was conducted for approximately 40 days with a one-week break at the beginning of August to allow for the national constitutional referendum activities.
Prior to the fieldwork, the communities residing in the sampled clusters received information about the KMIS through social mobilization and the mass media. This was necessary to alert the communities about the days of the survey and also that children would be tested for malaria. Taking of blood samples is often a sensitive issue requiring adequate information beforehand to avoid misinformation.
The fieldwork commenced on 18 July and after the one-week break (1-7 August) ended on 2 September. Teams spent an average of three days in a cluster with the first day dedicated to mapping the households while the next two days were used to conduct field interviews. Fieldwork was closely supervised by a team of national supervisors from the DOMC including the Provincial Malaria Control Coordinators and KNBS who visited the teams in the field to ensure that the survey was conducted according to the protocol and provide solutions to some of the challenges encountered. The teams were facilitated in the field by KNBS district staff; these included District Statistical Offi cers (DSO) and cluster guides who made sure that the enumeration areas were accurately identifi ed. Village elders were also instrumental in guiding the teams and mobilizing the communities in their respective clusters.
Kenya National Bureau of Statistics
Ministry of Planning and National Development
The questionnaires used in the 2010 KMIS were developed by the Roll Back Malaria Monitoring and Evaluation Reference Group (MERG) in collaboration with ICF Macro. The standard questionnaires were adapted to the Kenyan situation and programmed into PDAs by a team from CDC/Atlanta. The questionnaires were fi rst reviewed by the KMIS Technical Working Group and were translated into Kiswahili. All KMIS interviews were done using the PDAs.
Two types of questionnaires were used: a Household Questionnaire and a Woman’s Questionnaire. The Household Questionnaire captured information on the usual members and visitors, including age, sex and relationship to the head of the household. One purpose of the Household Questionnaire was to identify women aged 15–49 who were eligible for the individual interviews. The questionnaire also collected information on characteristics of the household’s dwelling unit, such as the source of water, type of toilet facilities, materials used for the fl oor, walls and roof of the house, etc. Information on ownership and use of mosquito nets was also collected with the Household Questionnaire. In addition, this questionnaire was used to capture some information on attitudes about malaria and to record the results of the request for doing anaemia and malaria testing on young children.
The Woman’s Questionnaire was administered to consenting women aged 15–49 years to collect data on background characteristics, reproductive history, use of intermittent preventive treatment (IPT) during pregnancy for recent births, fever prevalence and treatment among children under fi ve, and knowledge and attitudes regarding malaria and child survival.
KMIS data were captured using PDAs fitted with GPS. The questionnaires were programmed into the PDAs and tested before the actual field work. They were then periodically transferred to and saved on the supervisor's PDA in each team and at the end of the data collection were downloaded onto a personal computer for merging and analysis.
The data underwent various cleaning processes before analysis. First, the data were corrected for problems in geographic coding by using ArcGIS software to plot the coordinates and identify the misplaced information. The data were further split and merged into various data sets to ease analysis and were converted into SPSS (Statistical Package for Social Science) format. The data sets were then converted into CSPro format and further checks/corrections were made prior to the production of preliminary tabulations.
Estimates of Sampling Error
A series of sampling errors tables are included:
- List of selected variables for sampling errors
- Sampling errors for all Kenya
- Sampling errors for urban areas
- Sampling errors for rural areas
- Sampling errors for highland epidemic zone
- Sampling errors for lake endemic zone
- Sampling errors for coastal endemic zone
- Sampling errors for semi-arid, seasonal risk zone
- Sampling errors for low risk zone
The sampling errors are fully described in Appendix A of "Kenya Malaria Indicator Survey 2010 - Final Report" pp.57-60.
The following data quality tables are produced:
- Household age distribution
- Coverage of testing for malaria and anaemia testing in children
- Prevalence of malaria in children aged 3-59 months
- Anaemia prevalence among children aged 6-59 months
See the tables in Appendix D of the final report.
Data and Data Related Resources
MEASURE DHS believes that widespread access to survey data by responsible researchers has enormous advantages for the countries concerned and the international community in general. Therefore, MEASURE DHS policy is to release survey data to researchers after the main survey report is published, generally within 12 months after the end of fieldwork. with few limitations these data have been made available for wide use.
DISTRIBUTION OF DATASETS
MEASURE DHS is authorized to distribute, at no cost, unrestricted survey data files for legitimate academic research, with the condition that we receive a description of any research project that will be using the data.
Registration is required for access to data.
Datasets are available for download to all registered users, free of charge. To download datasets, you must first register online and request the country(ies) and datasets that you are interested in. When submitting a dataset request, users must include a brief description of how the data will be used.
Datasets are made available with the following conditions:
- Survey data files are distributed by MEASURE DHS for academic research/statistical analysis. Researchers need to provide a description of any research/analysis that will be using the data, before access is granted to the datasets.
- Once downloaded, the datasets must not be passed on to other researchers without the written consent of MEASURE DHS.
- All reports and publications based on the requested data must be sent to the MEASURE DHS Data Archive as a Portable Format Document (pdf) or a hard copy, for us to forward to the country(ies) whose data have been used.
"Kenya National Bureau of Statistics, 2010 Kenya Malaria Indicator Survey (KMIS 2010), Version 1.2, provided by the Kenya National Data Archive. http://statistics.knbs.or.ke/nada/index.php/catalog"
Disclaimer and copyrights
The user of the data acknowledges that the original collector of the data, the authorized distributor of the data, and the relevant funding agency bear no responsibility for use of the data or for interpretations or inferences based upon such uses.
DDI Document ID
Kenya National Bureau of Statistics
Ministry of Planning and National Development
Accelerated Data Program
International Household Survey Network
Review of the metadata
Date of Metadata Production
DDI Document version
Version 02 (October 2013). Edited version based on Version 01 DDI that was done by Kenya National Bureau of Statistics and reviewed by Accelerated Data Program, International Household Survey Network.