Abstract

Obtaining reliable size estimates for key populations is crucial for the Rwanda Biomedical Center/Institute of HIV/AIDS, Disease Prevention and Control (RBC/IHDPC) and their partners to design an effective HIV response in line with the national HIV strategy. Estimating the size of key populations at higher risk for HIV not only allows for an understanding of the magnitude of the response that is needed, but also helps in more accurately projecting the future of the epidemic in Rwanda. To be effective, it is important to produce consistent and comparable estimates over time. The following study utilized a single household survey to estimate the size of several key populations, including sex workers, men who have sex with men (MSM), injecting drug users (IDU), and clients of sex workers. These populations include several groups outlined in the National Strategic Plan for HIV and AIDS as most at risk for HIV infection, specifically sex workers and MSM.

Geographic coverage

National

Analysis unit

• Household
• Individuals

Kind of data

Sample survey data [ssd]

Sampling procedure

The ESPHS used a two-stage sample design, implemented in a representative sample of 2,125 households selected nationwide in which all women and men age 15 years and above where eligible for an individual interview. Each of these households was visited to obtain information using the Household Questionnaire. All women and all men age 15 years and above were eligible to be individually interviewed, if they were either usual residents of the household or visitors present in the household on the night before the survey. A total of 4,669 women and men were successfully interviewed.

The sampling frame used was the preparatory frame for the Rwanda Population and Housing Census (RPHC) 2012, provided by the National Institute of Statistics of Rwanda (NISR). The sampling frame is a complete list of natural villages covering the whole country (14,837 villages). Two strata were defined: the city of Kigali and the rest of the country. One hundred and thirty Primary Sampling Units (PSU) were selected from the sampling frame (35 in Kigali and 95 in the other stratum). To reduce clustering effect, only 20 households were selected per cluster in Kigali and 15 in the other clusters. As a result, 33 percent of the households in the sample were located in Kigali.

The list of households in each cluster was updated upon arrival of the survey team in the cluster. Once the listing had been updated, a number was assigned to each existing household in the cluster. The supervisor then identified the households to be interviewed in the survey by using a table in which the households were randomly pre-selected. This table also provided the list of households pre-selected for each of the two different definitions of what it means “to know” someone.

Mode of data collection

Face-to-face [f2f]

Research instrument

The survey used two types of questionnaires: a household questionnaire and an individual questionnaire. The same individual questionnaire was used to interview both women and men. In addition, two versions of the individual questionnaire were developed, using two different definitions of what it means “to know” someone. Each version of the individual questionnaire was used in half of the selected households.

Household questionnaire: The Household Questionnaire was a short version of the 2011 Rwanda DHS questionnaire. It was primarily used to list all the usual members and visitors in the selected households and to collect some basic information on the characteristics of each person listed, including age, sex, status of residence, and marital status. The main purpose of the Household Questionnaire was to identify women and men who were eligible for the individual interview. The Household Questionnaire also collected information on characteristics of the household’s dwelling unit, such as the source of water, type of toilet facilities, and ownership of various durable goods. This information was used to create an index representing the wealth of the households. The wealth index is a proxy for long-term standard of living of the households and is used in the following analysis as a background characteristic of the respondents who are members of these households.

Individual questionnaire: The individual questionnaire was organized accordingly and included six sections: - Respondent’s background; - Known population; - Summation; - Target population; - Proxy respondent; and - Stigma.

Cleaning operations

The processing of the ESPHS data began shortly after the fieldwork commenced. Completed questionnaires were returned periodically from the field to the SPH office in Kigali, where they were entered and checked for consistency by data processing personnel who were specially trained for this task. Data were entered using CSPro, a programme specially developed for use in DHS surveys. All data were entered twice (100 percent verification). The concurrent processing of the data was a distinct advantage for data quality, because the School of Public Health had the opportunity to advise field teams of problems detected during data entry. The data entry and editing phase of the survey was completed in late August 2011.

Response rate

The number of occupied households successfully interviewed was 2,102, yielding a household response rate of 99%. From the households interviewed, 2,629 women were found to be eligible and 2,567 were interviewed, giving a response rate of 98%. Interviews with men covered 2,102 of the eligible 2,149 men, yielding a response rate of 98%. The response rates do not significantly vary by type of questionnaire or residence.

Abstract

The 2019-20 Rwanda Demographic and Health Survey (2019-20 RDHS) follows those implemented in 1992, 2000, 2005, 2010, and 2014-15. A nationally representative sample of 500 clusters and 13,000 households were selected. All women age 15-49 who were usual residents of the selected households or who slept in the households the night before the survey were eligible for the survey.

The primary objective of the 2019-20 RDHS is to provide up-to-date estimates of basic demographic and health indicators. Specifically, the 2019-20 RDHS: • collected data on fertility levels and preferences; contraceptive use; maternal and child health; infant, child, and neonatal mortality levels; maternal mortality; gender; nutrition; awareness about HIV/AIDS; self-reported sexually transmitted infections (STIs); and other health issues relevant to the achievement of the Sustainable Development Goals (SDGs) • obtained information on the availability of, access to, and use of mosquito nets as part of the National Malaria Control Program • gathered information on other health issues such as injections, tobacco use, and health insurance • collected data on women’s empowerment and domestic violence • tested household salt for iodine levels • obtained data on child feeding practices, including breastfeeding, and conducted anthropometric measurements to assess the nutritional status of children under age 5 and women age 15-49 • conducted anemia testing of women age 15-49 and children age 6-59 months • conducted malaria testing of women age 15-49 and children age 6-59 months • conducted HIV testing of women age 15-49 and men age 15-59 • conducted micronutrient testing of women age 15-49 and children age 6-59 months

The information collected through the 2019-20 RDHS is intended to assist policymakers and program managers in evaluating and designing programs and strategies for improving the health of the country’s population.

Geographic coverage

National coverage

Analysis unit

• Household
• Individual
• Children age 0-5
• Woman age 15-49
• Man age 15 to 59

Universe

The survey covered all de jure household members (usual residents), all women aged 15-49, all men age 15-59, and all children aged 0-5 resident in the household.

Kind of data

Sample survey data [ssd]

Sampling procedure

The sampling frame used for the 2019-20 RDHS is the fourth Rwanda Population and Housing Census (RPHC), which was conducted in 2012 by the National Institute of Statistics of Rwanda (NISR). The sampling frame is a complete list of enumeration areas (EAs) covering the whole country provided by the National Institute of Statistics, the implementing agency for the RDHS. An EA is a natural village or part of a village created for the 2012 RPHC; these areas served as the counting units for the census.

The 2019-20 RDHS followed a two-stage sample design and was intended to allow estimates of key indicators at the national level as well as for urban and rural areas, five provinces, and each of Rwanda’s 30 districts for some limited indicators. The first stage involved selecting sample points (clusters) consisting of EAs delineated for the 2012 RPHC. A total of 500 clusters were selected, 112 in urban areas and 388 in rural areas.

The second stage involved systematic sampling of households. A household listing operation was undertaken in all selected EAs from June to August 2019, and households to be included in the survey were randomly selected from these lists. Twenty-six households were selected from each sample point, for a total sample size of 13,000 households. Because of the approximately equal sample sizes in each district, the sample is not self-weighting at the national level, and weighting factors have been added to the data file so that the results will be proportional at the national level.

For further details on sample selection, see Appendix A of the final report.

Mode of data collection

Computer Assisted Personal Interview [capi]

Research instrument

Five questionnaires were used for the 2019-20 RDHS: the Household Questionnaire, the Woman’s Questionnaire, the Man’s Questionnaire, the Biomarker Questionnaires, and the Fieldworker Questionnaire. These questionnaires, based on The DHS Program’s standard Demographic and Health Survey (DHS-7) questionnaires, were adapted to reflect the population and health issues relevant to Rwanda.

Cleaning operations

The processing of the 2019-20 RDHS data began almost as soon as the fieldwork started. As data collection was completed in each cluster, all electronic data files were transferred via the Internet File Streaming System (IFSS) to the NISR central office in City of Kigali. These data files were registered and checked for inconsistencies, incompleteness, and outliers. The field teams were alerted to any inconsistencies and errors. Secondary editing, carried out in the central office, involved resolving inconsistencies and coding the open-ended questions. The NISR data processor coordinated the exercise at the central office. The biomarker paper questionnaires were compared with electronic data files to check for any inconsistencies in data entry. Data entry and editing were carried out using the CSPro software package. The concurrent processing of the data offered a distinct advantage because it maximized the likelihood of the data being error-free and accurate. Timely generation of field check tables allowed for effective monitoring. The secondary editing of the data was completed in the second week of September 2020.

Response rate

A total of 13,005 households were selected for the sample, of which 12,951 were occupied. All but two occupied households (12,949) were successfully interviewed, yielding a response rate of 100.0%. In the interviewed households, 14,675 women age 15-49 were identified for individual interviews; interviews were completed with 14,634 women, yielding a response rate of 99.7%. In the subsample selected for the male survey, 6,503 households were selected, of which 6,472 were occupied. All but one occupied household (6,471) were successfully interviewed, yielding a response rate of 100.0%. In this subsample, 6,544 men age 15-59 were identified and 6,513 were successfully interviewed, yielding a response rate of 99.5%. In the subsample selected for the micronutrient survey, 3,501 households were selected, of which 3,492 were occupied. All but one of the occupied households (3,491) were successfully interviewed, yielding a response rate of 100.0%.

Sampling error estimates

The estimates from a sample survey are affected by two types of errors: nonsampling errors and sampling errors. Nonsampling errors are the results of mistakes made in implementing data collection and data processing, such as failure to locate and interview the correct household, misunderstanding of the questions on the part of either the interviewer or the respondent, and data entry errors. Although numerous efforts were made during the implementation of the 2019-20 Rwanda Demographic and Health Survey (2019-20 RDHS) to minimize this type of error, nonsampling errors are impossible to avoid and difficult to evaluate statistically.

Sampling errors, on the other hand, can be evaluated statistically. The sample of respondents selected in the 2019-20 RDHS is only one of many samples that could have been selected from the same population, using the same design and sample size. Each of these samples would yield results that differ somewhat from the results of the actual sample selected. Sampling errors are a measure of the variability among all possible samples. Although the degree of variability is not known exactly, it can be estimated from the survey results.

Sampling error is usually measured in terms of the standard error for a particular statistic (mean, percentage, etc.), which is the square root of the variance. The standard error can be used to calculate confidence intervals within which the true value for the population can reasonably be assumed to fall. For example, for any given statistic calculated from a sample survey, the value of that statistic will fall within a range of plus or minus two times the standard error of that statistic in 95% of all possible samples of identical size and design.

If the sample of respondents had been selected by simple random sampling, it would have been possible to use straightforward formulas for calculating sampling errors. However, the 2019-20 RDHS sample was the result of a multi-stage stratified design, and, consequently, it was necessary to use more complex formulas. Sampling errors are computed using SAS programs developed by ICF. These programs use the Taylor linearization method to estimate variances for survey estimates that are means, proportions, or ratios. The Jackknife repeated replication method is used for variance estimation of more complex statistics such as fertility and mortality rates.

Note: A more detailed description of estimates of sampling errors are presented in APPENDIX B of the survey report.

Data appraisal

Data Quality Tables

• Household age distribution
• Age distribution of eligible and interviewed women
• Age distribution of eligible and interviewed men
• Completeness of reporting
• Births by calendar years
• Reporting of age at death in days
• Reporting of age at death in months
• Standardization exercise results from anthropometry training
• Height and weight data completeness and quality for children
• Height measurements from random sub-sample of measured children
• Number of enumeration areas

Abstract

The Estimating the Size of Populations through a Household Survey (EPSHS), sought to assess the feasibility of the network scale-up and proxy respondent methods for estimating the sizes of key populations at higher risk of HIV infection and to compare the results to other estimates of the population sizes. The study was undertaken based on the assumption that if these methods proved to be feasible with a reasonable amount of data collection for making adjustments, countries would be able to add this module to their standard household survey to produce size estimates for their key populations at higher risk of HIV infection. This would facilitate better programmatic responses for prevention and caring for people living with HIV and would improve the understanding of how HIV is being transmitted in the country.

The specific objectives of the ESPHS were: 1. To assess the feasibility of the network scale-up method for estimating the sizes of key populations at higher risk of HIV infection in a Sub-Saharan African context; 2. To assess the feasibility of the proxy respondent method for estimating the sizes of key populations at higher risk of HIV infection in a Sub-Saharan African context; 3. To estimate the population size of MSM, FSW, IDU, and clients of sex workers in Rwanda at a national level; 4. To compare the estimates of the sizes of key populations at higher risk for HIV produced by the network scale-up and proxy respondent methods with estimates produced using other methods; and 5. To collect data to be used in scientific publications comparing the use of the network scale-up method in different national and cultural environments.

Geographic coverage

National

Analysis unit

• Household
• Individual

Sampling procedure

The Estimating the Size of Populations through a Household Survey (ESPHS) used a two-stage sample design, implemented in a representative sample of 2,125 households selected nationwide in which all women and men age 15 years and above where eligible for an individual interview. The sampling frame used was the preparatory frame for the Rwanda Population and Housing Census (RPHC), which was conducted in 2012; it was provided by the National Institute of Statistics of Rwanda (NISR).

The sampling frame was a complete list of natural villages covering the whole country (14,837 villages). Two strata were defined: the city of Kigali and the rest of the country. One hundred and thirty Primary Sampling Units (PSU) were selected from the sampling frame (35 in Kigali and 95 in the other stratum). To reduce clustering effect, only 20 households were selected per cluster in Kigali and 15 in the other clusters. As a result, 33 percent of the households in the sample were located in Kigali.

The list of households in each cluster was updated upon arrival of the survey team in the cluster. Once the listing had been updated, a number was assigned to each existing household in the cluster. The supervisor then identified the households to be interviewed in the survey by using a table in which the households were randomly pre-selected. This table also provided the list of households pre-selected for each of the two different definitions of what it means "to know" someone.

For further details on sample design and implementation, see Appendix A of the final report.

Mode of data collection

Face-to-face [f2f]

Research instrument

The Estimating the Size of Populations through a Household Survey (ESPHS) used two types of questionnaires: a household questionnaire and an individual questionnaire. The same individual questionnaire was used to interview both women and men. In addition, two versions of the individual questionnaire were developed, using two different definitions of what it means “to know” someone. Each version of the individual questionnaire was used in half of the selected households.

Cleaning operations

The processing of the ESPHS data began shortly after the fieldwork commenced. Completed questionnaires were returned periodically from the field to the SPH office in Kigali, where they were entered and checked for consistency by data processing personnel who were specially trained for this task. Data were entered using CSPro, a programme specially developed for use in DHS surveys. All data were entered twice (100 percent verification). The concurrent processing of the data was a distinct advantage for data quality, because the School of Public Health had the opportunity to advise field teams of problems detected during data entry. The data entry and editing phase of the survey was completed in late August 2011.

Response rate

A total of 2,125 households were selected in the sample, of which 2,120 were actually occupied at the time of the interview. The number of occupied households successfully interviewed was 2,102, yielding a household response rate of 99 percent.

From the households interviewed, 2,629 women were found to be eligible and 2,567 were interviewed, giving a response rate of 98 percent. Interviews with men covered 2,102 of the eligible 2,149 men, yielding a response rate of 98 percent. The response rates do not significantly vary by type of questionnaire or residence.

Sampling error estimates

The estimates from a sample survey are affected by two types of errors: (1) non-sampling errors, and (2) sampling errors. Non-sampling errors are the results of mistakes made in implementing data collection and data processing, such as failure to locate and interview the correct household, misunderstanding of the questions on the part of either the interviewer or the respondent, and data entry errors. Although numerous efforts were made to minimize this type of error during the implementation of the Rwanda ESPHS 2011, non-sampling errors are impossible to avoid and difficult to evaluate statistically.

Sampling errors, on the other hand, can be evaluated statistically. The sample of respondents selected in the ESPHS 2011 is only one of many samples that could have been selected from the same population, using the same design and identical size. Each of these samples would yield results that differ somewhat from the results of the actual sample selected. Sampling errors are a measure of the variability between all possible samples. Although the degree of variability is not known exactly, it can be estimated from the survey results.

A sampling error is usually measured in terms of the standard error for a particular statistic (mean, percentage, etc.), which is the square root of the variance. The standard error can be used to calculate confidence intervals within which the true value for the population can reasonably be assumed to fall. For example, for any given statistic calculated from a sample survey, the value of that statistic will fall within a range of plus or minus two times the standard error of that statistic in 95 percent of all possible samples of identical size and design.

If the sample of respondents had been selected as a simple random sample, it would have been possible to use straightforward formulas for calculating sampling errors. However, the ESPHS 2011 sample is the result of a multi-stage stratified design, and, consequently, it was necessary to use more complex formulae. The computer software used to calculate sampling errors for the ESPHS 2011 is a SAS program. This program uses the Taylor linearization method for variance estimation for survey estimates that are means or proportions.

A more detailed description of estimates of sampling errors are presented in Appendix B of the survey report.