This blog post is part of Luminare, our blog series exploring innovative solutions to data collection, quality assurance, biomarker measurement, data use, and further analysis.
Have you ever wondered how to write a Stata program for vaccination coverage or struggled to construct mortality rates using DHS data? Well, DHS Program staff are busy writing SPSS and Stata code for all indicators listed in the Guide to DHS Statistics, and you can use this code to jump-start your exploration of the data. And as they are completed, the code will be posted on GitHub for open access to the public.
The DHS Program GitHub site contains two repositories: DHS-Indicators-Stata and DHS-Indicators-SPSS. Users can download the code from these repositories or clone the repository to their own Github site. Users can also suggest changes to the code that will be reviewed by DHS Program staff before acceptance.
Don’t see what you need? The programming for all indicators listed in the Guide to DHS Statistics will be available by September 2020. The Guide corresponds to the topics/chapters that are typically found in a DHS survey final report in addition to the modules for malaria and HIV prevalence. As of July 2019, about half of the indicators have been coded and shared in Stata including indicators covering child health, family planning, and reproductive health. SPSS code will follow later in 2019 and 2020, along with the remainder of the indicators. Review the Readme text file for more details.
This blog post is part of Luminare, our new blog series exploring innovative solutions to data collection, quality assurance, biomarker measurement, data use, and further analysis.
Two needs are often expressed by both DHS host countries and donors: 1) for data to be made available more frequently, and 2) for the continued strengthening of implementing agencies’ capacity to conduct surveys. Among various innovations that The DHS Program has pursued to respond to these needs is the Continuous Survey (CS) model.
What is a Continuous Survey?
In a CS, data are collected and reported annually by a permanently maintained office and field staff. A smaller sample size is designed to provide estimates at the national level and for urban/rural residence every year. For regional-level estimates, data are pooled over multiple consecutive phases. Through both the smaller sample and continuously maintained staff, the model can lower costs and institutionalize the implementing agency’s ability to conduct a DHS survey. In 2004, Peru became the first country to conduct a CS, and the effort is still ongoing.
How did Senegal implement a Continuous Survey?
Inspired by the Peru experience, USAID and The DHS Program piloted the CS model in Africa. Senegal was chosen for its long survey history and the capacity of the local implementing agency, Agence Nationale de la Statistique et la Démographie (ANSD). The Senegal Continuous Survey (SCS) expanded on the original model to include an annual facility-based Continuous Service Provision Assessment (C-SPA), in addition to the household-based Continuous Demographic and Health Survey (C-DHS). The SCS was conducted in five phases, spanning the period from 2012 to 2018.
Covers from final reports from each of the five phases of the SCS.
What were the successes and challenges of the SCS?
ANSD partnered with Le Soleil newspaper to create an 8-page spread highlighting results from the 2016 SCS.
The SCS demonstrated many successes. Senegal is the only country in Africa to annually collect nationally representative demographic and health data, allowing Senegal to monitor progress towards the SDGs every year. This was also the first time a country releases both facility and household data at the same time. This model of releasing C-SPA data annually and in conjunction with the C-DHS resulted in flourishing data use for both surveys.
The SCS greatly strengthened capacity in Senegal. ANSD is now capable of conducting DHS and SPA surveys with only limited technical assistance. ANSD has the initiative to move beyond the pilot to implement the 2018 SCS with limited technical assistance and is already continuing the annual surveys.
Most surveys encounter challenges, and, in the Senegal experience, CS-specific design challenges emerged. Some stakeholders were concerned about the approach of pooling two consecutive years of CS data to generate a large enough sample size for regional-level estimates. Additionally, a census and an updated health facility master list in Senegal during the SCS pilot period resulted in new sampling frames for both the C-DHS and the C-SPA, and subsequent challenges in data interpretation. Finally, survey dissemination activities overlapped with the next phase’s design and implementation activities, increasing the burden on ANSD.
The CS model demands an overlap of activities. While one phase moves toward dissemination, planning is already occurring for the next phase of data collection, as evidenced in the SCS pilot experience.
Lessons learned from the SCS experience will inform The DHS Program’s continued efforts to innovate in the areas of data collection and use.
This blog post is part of Luminare, our new blog series exploring innovative solutions to data collection, quality assurance, biomarker measurement, data use, and further analysis. This is the second post in the series that focuses on innovations using DHS.rates data.
Ever struggled to calculate fertility or child mortality indicators from survey data? Want to customize the reference period? DHS.rates can do it for you!
What is DHS.rates?
The DHS.rates is a user-friendly R package to calculate fertility
and childhood mortality rates based on DHS datasets. First released in March
2018, the current version of DHS.rates calculates the Total Fertility Rate,
General Fertility Rate, Age-Specific Fertility Rates, Neonatal Mortality Rate,
Post-Neonatal Mortality Rate, Infant Mortality Rate, Child Mortality Rate, Under-5
Mortality Rate and mortality probabilities. For each indicator, the package
calculates standard error, design effect, relative standard error, and
confidence intervals. Data users can customize rates:
periods other than DHS standard reference periods
Based on calendar
years so the end of the reference period is not the date of the survey
sub-populations or domains other than those produced by The DHS Program
Based on other surveys
other than DHS if the required variables are available
Not an R user? Try
theweb-application, DHS.rates Shiny
This web-application provides all the DHS.rates functions without needing to download or use R. The DHS.rates Shiny web application includes two main tabs, fert and chmort. After uploading the relevant survey dataset, the application calculates fertility or childhood mortality rates according to the DHS methodology.
Just as with the R package, Shiny web application users can customize the reference period as well as the end date of the reference period. By adding a variable to “Class of the rate”, users can do the calculations for different subpopulations other than the ones produced by The DHS Program. Users also can change any of the fields on the screen allowing them to use the application with other surveys other than the DHS.
This blog post is part of Luminare, our new blog series exploring innovative solutions to data collection, quality assurance, biomarker measurement, data use, and further analysis. This is the first post in the series that focuses on innovations to improve the quality of anthropometry data.
Anthropometry, the measurement of the human body, gives a snapshot of the malnutrition situation in a country. Yet, the collection of accurate height and weight measurements, especially for young children, is difficult during data collection. To address this challenge, The DHS Program has tested new innovations to enhance the quality of the anthropometry data in the 2018 Nigeria Demographic and Health Survey (NDHS).
The 2018 NDHS is the sixth DHS survey conducted in Nigeria. The National Population Commission (NPC), in collaboration with the National Malaria Elimination Programme (NMEP) of the Federal Ministry of Health (FMOH), implemented the survey. Thirty-seven field teams closely monitored by coordinators and quality controllers collected data from August 14, 2018, to December 29, 2018. Each field team included a supervisor, field editor, two male interviewers, three female interviewers, and a biomarker team consisting of a lab scientist and nurse.
Introducing a Biomarker Checklist to strengthen supervision
The DHS Program has developed a Biomarker Checklist to assess the performance of and provide feedback to field staff. The checklist includes a core set of tasks required to collect biomarker data. Each task is a crucial action that, if missed, can result in poor quality data. The Biomarker Checklist is administered by supervisors and coordinators during collection of biomarkers in households.
The Biomarker Checklist was tested during the 2018 NDHS pre-test and biomarker main training using a mixed-method design which included administering anonymous questionnaires and conducting key informant interviews with the supervisory staff who used the checklist to assess its’ viability and usefulness. Feedback on the Biomarker Checklist was positive, so the Biomarker Checklist was used during the 2018 NDHS data collection. A Biomarker Checklist will be publicly available at a later date.
“it [Biomarker Checklist] has helped a lot because you are able to monitor what the biomarkers are doing so [you] can actually give corrective action”
Reducing errors while still in the field through re-measurement of children
The shift from paper questionnaires to a CAPI-based data collection approach provides an opportunity to easily identify children who may have been mismeasured and return to the household to measure these children again. To identify children with incorrect measurements requires performing a complicated calculation – a child’s body measurements are compared against a healthy population by transforming their measurements into anthropometry Z-scores. Extreme measurement results are then detected by calculating anthropometry Z-scores and flagging cases with higher or lower Z-scores than expected. The DHS Program has developed a program to automatically calculate anthropometry Z-scores and flag extreme cases in the CAPI system. A user-friendly interface on the tablet produces a report with the children who need to be re-measured.
The DHS Program relies on field check tables that are run periodically during data collection. While an important data quality tool, a major limitation of the field check tables is that enough data need to accumulate before problems can be identified. At that point, the information can only be used to improve collection of data moving forward; these corrections do not fix data previously collected. The real-time ability to re-measure children while still in the field is a major step forward and can easily be applied to other CAPI surveys.
Re-measuring anthropometry in a random subsample of children
A random re-measurement of height and weight in a subsample of children was also piloted in Nigeria. The DHS Program has developed a CAPI program that randomly selects one household in each cluster for the biomarker team to revisit. The program then compares differences between measurements and reports precision, or how close the first measurement is to the second measurement, as acceptable or unacceptable.
The aim is measurements will be of better quality as a result of instituting random re-measurement. The biomarker team may be more careful and not rush measurements if they know poor measurements will be exposed. The data produced on precision can be used as a motivational tool for biomarker teams and provide an opportunity to identify and re-train in cases where there is a high degree of discrepancy between measurements.Precision estimates will also help better assess data quality post-data collection.
A review of the anthropometry data from the 2018 NDHS indicates it meets data quality targets. Results from the 2018 NDHS Key Indicators Report (KIR) show that 37% of children under 5 are stunted. Stunting generally increases with age, peaking at 47% for children age 24-35 months. Overall, 7% of children under 5 are wasted, while 23% of children are underweight.
Lessons learned in the implementation of the quality assurance activities in Nigeria are being used to conduct similar activities in DHS surveys in other countries.
About the survey
The 2018 NDHS was implemented by the National Population Commission (NPC) in collaboration with the National Malaria Elimination Programme (NMEP) of the Federal Ministry of Health, Nigeria. Funding for the 2018 NDHS was provided by the United States Agency for International Development (USAID), Global Fund, Bill and Melinda Gates Foundation (BMGF), the United Nations Population Fund (UNFPA), and World Health Organization (WHO). ICF provided technical assistance through The DHS Program, a USAID-funded project that provides support and technical assistance in the implementation of population and health surveys in countries worldwide.
Additional information about the 2018 NDHS may be obtained from the headquarters of the National Population Commission (NPC), Plot 2031, Olusegun Obasanjo Way, Zone 7, Wuse, P.M.B. 0281, Abuja, Nigeria (telephone: 234-09-523-9173; fax: 243-09-523-1024; email: firstname.lastname@example.org; internet: www.population.gov.ng).
This blog post is part of Luminare: The DHS Program Blog Series on Innovation. You can find additional posts in the Luminare series here.
While The DHS Program is known for comparability and standard methods, it would not be relevant today without innovation. We’ve made big leaps – like moving from paper to tablet-based interviewing and from basic print publications to web and mobile data dissemination. We’ve integrated complex biomarker testing and developed weeks-long curricula in data analysis and use. We’ve also made hundreds of smaller, less flashy improvements, such as use of WhatsApp to communicate with field teams and the use of checklists to improve biomarker collection. Innovation – large and small – is part of life at The DHS Program, as we are constantly seeking new ways to solve problems, increase efficiency, and improve data quality while meeting the needs of an increasingly diverse audience.
Over the course of the next several years, we will be undertaking a systematic review of new ideas, from new biomarker assays to non-traditional partnerships. This new blog series is just one of the ways that we will be exploring and sharing innovations. We will also be holding topical consultations with experts, reviewing the academic literature, attending key conferences, and interviewing key informants such as external survey experts, staff, consultants, and implementing agency representatives.
In DHS-8, we have an even stronger mandate to innovate, and for the first time ever, we have a formal innovation strategy which is led by our new Senior Science Advisor, Joanna Lowell. Her job, in part, is to lead the “deliberate exploration and strategic incorporation of state-of-the-art advances in survey tools methodologies, and partnerships.”
We welcome your ideas. You can reach out to Joanna directly at Innovations@dhsprogram.com and look out for announcements regarding other ways to contribute.