Identification
Survey ID Number
KEN-KNBS-KAIS-2007-v01
Title
Kenya Aids Indicator Survey 2007
Country
| Name | Country code |
|---|---|
| KENYA | KEN |
Study type
Integrated Survey (non-LSMS) [hh/is]
Series Information
The Second household survey to be conducted by KNBS
Abstract
Obtaining nationally representative estimates on behavioural, clinical, and biologic indicators for HIV/AIDS is critical for evaluating a country's response to the HIV epidemic. National population-based surveys with HIV testing provide national-level prevalence estimates and the opportunity to link HIV status with behavioural, social, demographic and other biological information.
The 2007 Kenya AIDS Indicator Survey (KAIS) is Kenya's first survey of its type and provides comprehensive information on HIV and other sexually transmitted infections (STIs). These data provide the information needed for advocacy and for planning appropriate interventions for HIV prevention, treatment and care. The 2007 KAIS builds upon previous national-level HIV estimates from the first population-based survey with HIV testing, the 2003 Kenya Demographic and Health Survey (KDHS); this allows us to compare prevalence estimates and important behavioural indicators between 2003 and 2007.
Findings from the 2007 KAIS are summarized below and described in detail in this report. The general background characteristics of respondents are provided in Appendix B.1. Estimates presented in the report and their corresponding sample sizes and 95% confidence intervals are presented in Appendices B.2-B.15. Estimates have been weighted appropriately for the two-stage sample design, with a noted exception in Chapter 15, where we present uptake of test results. The report presents the results of univariate and bivariate analyses; analyses are not adjusted for confounding factors. Multivariate analysis of KAIS data, adjusted for possible confounders, will be presented in other dissemination materials, such as peer-reviewed scientific publications. Throughout the report, the term significant indicates a p-value1 less than 0.05. Marginally significant indicates a p-value between 0.05 and 0.10, inclusive; and not significant indicates a p-value greater than 0.10.
The 2007 Kenya AIDS Indicator Survey (KAIS) is Kenya's first survey of its type and provides comprehensive information on HIV and other sexually transmitted infections (STIs). These data provide the information needed for advocacy and for planning appropriate interventions for HIV prevention, treatment and care. The 2007 KAIS builds upon previous national-level HIV estimates from the first population-based survey with HIV testing, the 2003 Kenya Demographic and Health Survey (KDHS); this allows us to compare prevalence estimates and important behavioural indicators between 2003 and 2007.
Findings from the 2007 KAIS are summarized below and described in detail in this report. The general background characteristics of respondents are provided in Appendix B.1. Estimates presented in the report and their corresponding sample sizes and 95% confidence intervals are presented in Appendices B.2-B.15. Estimates have been weighted appropriately for the two-stage sample design, with a noted exception in Chapter 15, where we present uptake of test results. The report presents the results of univariate and bivariate analyses; analyses are not adjusted for confounding factors. Multivariate analysis of KAIS data, adjusted for possible confounders, will be presented in other dissemination materials, such as peer-reviewed scientific publications. Throughout the report, the term significant indicates a p-value1 less than 0.05. Marginally significant indicates a p-value between 0.05 and 0.10, inclusive; and not significant indicates a p-value greater than 0.10.
Kind of Data
Sample survey data [ssd]
Unit of Analysis
Households Indviduals within Households Community
Version
Version Date
2012-10-15
Coverage
Geographic Coverage
The survey covered all the districts in Kenya. The data representativeness are at the following levels -National -Urban/Rural -Provincial -District
The 2007 KAIS was conducted among a representative sample of households selected from all eight provinces in the country, covering both rural and urban areas. A household was defined as a person or group of people related or unrelated to each other who live together in the same dwelling unit or compound (a group of dwelling units), share similar cooking arrangements, and identify the same person as the head of household. The household questionnaire was administered to consenting heads of sampled, occupied households. All women and men aged 15-64 years in selected households who were either usual residents or visitors present the night before the survey were eligible to participate in the individual interview and blood draw, provided they gave informed consent. For minors aged 15-17 years, parental consent and minor assent were both required for participation. Participants could consent to the interview and blood draw or to the interview alone. The inclusion criteria may have captured non-Kenyans living as usual residents or visitors in a sampled household. Military personnel and the institutionalized population (e.g. imprisoned) are typically not captured in household-based surveys, but may have been included in the 2007 KAIS if at home during the survey.
The 2007 KAIS was conducted among a representative sample of households selected from all eight provinces in the country, covering both rural and urban areas. A household was defined as a person or group of people related or unrelated to each other who live together in the same dwelling unit or compound (a group of dwelling units), share similar cooking arrangements, and identify the same person as the head of household. The household questionnaire was administered to consenting heads of sampled, occupied households. All women and men aged 15-64 years in selected households who were either usual residents or visitors present the night before the survey were eligible to participate in the individual interview and blood draw, provided they gave informed consent. For minors aged 15-17 years, parental consent and minor assent were both required for participation. Participants could consent to the interview and blood draw or to the interview alone. The inclusion criteria may have captured non-Kenyans living as usual residents or visitors in a sampled household. Military personnel and the institutionalized population (e.g. imprisoned) are typically not captured in household-based surveys, but may have been included in the 2007 KAIS if at home during the survey.
Universe
The survey covered all household members (usual residents), all women aged 15-49 years resident in the household, and all children aged 0-4 years (under age 5) resident in the household.
Producers and sponsors
Primary investigators
| Name | Affiliation |
|---|---|
| Kenya National Bureau of Statistics | Ministry of Planning and National development |
Producers
| Name |
|---|
| Ministry of Health |
Funding Agency/Sponsor
| Name | Abbreviation |
|---|---|
| KEMRI/CDC | |
| University of California San Francisco | UCSF |
| Liverpool VCT |
Sampling
Sampling Procedure
Administratively, Kenya is divided into eight provinces. Each province is divided into districts, each district into divisions, each division into locations, each location into sub-locations, and each sublocation into villages. For the 1999 Population and Household Census, the Kenya National Bureau of Statistics (KNBS) delineated sub-locations into small units called Enumeration Areas (EAs) that constituted a village, a part of a village, or a combination of villages. The primary sampling unit for Kenya's master sampling frame, and for the 2007 KAIS, is a cluster, which is constituted as one or more EAs, with an average of 100 households per cluster. The master sampling frame for the 2007 KAIS was the National Sample Survey and Evaluation Programme IV (NASSEP IV) created and maintained by KNBS. The NASSEP IV frame was developed in 2002 based on the 1999 Census. The frame has 1800 clusters, comprised of 1,260 rural and 540 urban clusters. Of these, 294 (23%) rural and 121 (22%) urban clusters were selected for KAIS.
The overall design for the 2007 KAIS was a stratified, two-stage cluster sample for comparability to the 2003 KDHS. The first stage involved selecting 415 clusters from NASSEP IV and the second stage involved the selection of households per cluster with equal probability of selection in the rural-urban strata within each district. The target of the 2007 KAIS sample was to obtain approximately 9,000 completed household interviews. Based on the level of household nonresponse reported in the 2003 KDHS (13.2% of selected households), 10,375 households in 415 clusters were selected for potential participation in the 2007 KAIS. Table 1.4 shows the provincial distribution of households and clusters originally sampled for the 2007 KAIS.
The overall design for the 2007 KAIS was a stratified, two-stage cluster sample for comparability to the 2003 KDHS. The first stage involved selecting 415 clusters from NASSEP IV and the second stage involved the selection of households per cluster with equal probability of selection in the rural-urban strata within each district. The target of the 2007 KAIS sample was to obtain approximately 9,000 completed household interviews. Based on the level of household nonresponse reported in the 2003 KDHS (13.2% of selected households), 10,375 households in 415 clusters were selected for potential participation in the 2007 KAIS. Table 1.4 shows the provincial distribution of households and clusters originally sampled for the 2007 KAIS.
Response Rate
Overall, participation rates in the 2007 KAIS were high. We calculated household response rate as the number of households consenting to the household interview divided by the total number of sampled households that were located and occupied. The individual interview response rate was calculated as the number of individuals who completed interviews divided by the number of individuals eligible for the individual interview based on the household census. Only those participating in the individual interview were eligible to participate in the blood draw. We calculated blood draw coverage as the number of blood draws divided by the number of all individuals eligible for the individual interview; the blood draw response rate reflects the number of successful blood draws divided by the number of individuals who completed individual interviews.
Data Collection
Dates of Data Collection
| Start |
|---|
| 2007 |
Data Collection Mode
Face-to-face [f2f]
Data Collectors
| Name | Abbreviation | Affiliation |
|---|---|---|
| Kenya National Bureau of Statistics | KNBS | Ministry of Planning and National Development |
Questionnaires
Questionnaires
Household Questionnaire
· Household census
· Parental survivorship
· Household characteristics
· Mosquito net use
· Support to households for sick and recently deceased adults, and OVCs
Individual Questionnaire
· Socio-demographic characteristics
· HIV/STI knowledge and attitudes
· Marriage and sexual partnerships
· Fertility and family planning
· Uptake of HIV prevention, care and treatment services
· Household census
· Parental survivorship
· Household characteristics
· Mosquito net use
· Support to households for sick and recently deceased adults, and OVCs
Individual Questionnaire
· Socio-demographic characteristics
· HIV/STI knowledge and attitudes
· Marriage and sexual partnerships
· Fertility and family planning
· Uptake of HIV prevention, care and treatment services
Data Processing
Other Processing
Data processing included a number of steps to prepare data collected in the field for analysis. The initial steps included editing questionnaires, both in the field and at KNBS, and double-data entry of all questionnaire responses to minimise errors. Data were entered using Census and Survey Processing System (CSPro) version 3.3.3 Once all survey responses were transferred to electronic format, the next step was to ensure full concordance between the two data entry databases, using paper questionnaires to resolve any discrepancies in transcription. A series of internal consistency and range checks helped to identify any illogical responses and to verify that responses adhered to skip patterns in the questionnaire. Data validation programs for data cleaning were written in Stata version 8.04 and corrections were entered directly in CSPro at KNBS.
A concurrent process of cleaning the raw laboratory data was conducted at the NHRL. The final, cleaned questionnaire database at KNBS was merged with the laboratory results database at the NHRL using unique survey identification numbers to ensure accurate matches (>99.9% of identification numbers were matched). After successfully merging the questionnaire and laboratory results databases, cluster and household identification numbers were serialized from 1-402 and from 1-25, respectively. Original cluster and household numbers, barcodes, and individual survey identification numbers were stripped from the database prior to weighting and analysis to ensure anonymity of survey participants.
All results presented in the report are based on weighted data to account for the survey sampling design and participation rates. The weights are used to correct for unequal probability of selection, to produce results that are representative of the larger population from which the sample was drawn and to adjust for survey non-response. The final weights were derived from the design weights of the NASSEP IV sampling frame and subsequently adjusted for non-response. Three weights were calculated for analyses: a household weight, an individual interview weight and a blood draw weight.
This report presents the results of univariate and bivariate analyses; analyses are not adjusted for confounding factors. Multivariate analyses of KAIS data will be presented in other dissemination materials, such as peer-reviewed scientific publications. Data analysis was conducted using Statistical Analysis System (SAS) version 9.135, which has procedures to account for multi-stage stratified sampling designs and can produce reliable standard errors and confidence intervals. With the exception of Chapter 3 (Comparison of HIV Prevalence in the 2003 KDHS and 2007 KAIS), statistical significance was assessed based on chi-square p-values. In Chapter 3, we assumed the estimates from each time period (the 2003 KDHS and the 2007 KAIS) were independent and used the z-test to compare two weighted estimates and to determine if differences between 2003 and 2007 were statistically significant. Throughout the report, the term significant indicates a p-value6 less than 0.05. Marginally significant indicates a p-value between 0.05 and 0.10, inclusive; and not significant indicates a p-value greater than 0.10.
A concurrent process of cleaning the raw laboratory data was conducted at the NHRL. The final, cleaned questionnaire database at KNBS was merged with the laboratory results database at the NHRL using unique survey identification numbers to ensure accurate matches (>99.9% of identification numbers were matched). After successfully merging the questionnaire and laboratory results databases, cluster and household identification numbers were serialized from 1-402 and from 1-25, respectively. Original cluster and household numbers, barcodes, and individual survey identification numbers were stripped from the database prior to weighting and analysis to ensure anonymity of survey participants.
All results presented in the report are based on weighted data to account for the survey sampling design and participation rates. The weights are used to correct for unequal probability of selection, to produce results that are representative of the larger population from which the sample was drawn and to adjust for survey non-response. The final weights were derived from the design weights of the NASSEP IV sampling frame and subsequently adjusted for non-response. Three weights were calculated for analyses: a household weight, an individual interview weight and a blood draw weight.
This report presents the results of univariate and bivariate analyses; analyses are not adjusted for confounding factors. Multivariate analyses of KAIS data will be presented in other dissemination materials, such as peer-reviewed scientific publications. Data analysis was conducted using Statistical Analysis System (SAS) version 9.135, which has procedures to account for multi-stage stratified sampling designs and can produce reliable standard errors and confidence intervals. With the exception of Chapter 3 (Comparison of HIV Prevalence in the 2003 KDHS and 2007 KAIS), statistical significance was assessed based on chi-square p-values. In Chapter 3, we assumed the estimates from each time period (the 2003 KDHS and the 2007 KAIS) were independent and used the z-test to compare two weighted estimates and to determine if differences between 2003 and 2007 were statistically significant. Throughout the report, the term significant indicates a p-value6 less than 0.05. Marginally significant indicates a p-value between 0.05 and 0.10, inclusive; and not significant indicates a p-value greater than 0.10.
Access policy
Access authority
| Name | Affiliation | URL | |
|---|---|---|---|
| Kenya National Bureau of Statistics | Ministry of Planning, National Development and Vision 2030 | directorgeneral@knbs.or.ke | www.knbs.or.ke |
Contacts
| Name | Affiliation | URL | |
|---|---|---|---|
| Director General | KNBS | directorgeneral@knbs.or.ke | www.knbs.or.ke |
Citation requirements
"Kenya Aids Indicator Survey 2007, Version 1.0 of the KNBS dataset "
Disclaimer and copyrights
Disclaimer
The data users shall acknowledge that any available intellectual property rights, including copyright in the data are owned by the KNBS. The data user or client is encouraged to provide KNBS with a copy of such report, paper or article.
Copyright
(c) 2012, Kenya National Bureau of Statistics
Metadata production
DDI Document ID
DDI-KEN-KNBS-KIHBS-2005-2006-v02
Producers
| Name | Abbreviation | Affiliation | Role |
|---|---|---|---|
| Kenya National Bureau of Statistics | KNBS | Ministry of Planning, National Development and Vision 2030 | Documentation of the Study |
Date of Metadata Production
2007-12-03
DDI Document version
KIHBS 2005-2006 V02: This is the IHSN reviewed version which contains some updates compared to version v01