Data Release Information Sheet
Data summary
Project name: Malaria Control Policy Assessment (MCPA) project in Uganda
Relevant publications:
Roberts DA, Ng M, Ikilezi G, Gasasira A, Dwyer-Lindgren L, Fullman N, Nalugwa T, Kamya M, Gakidou E. Benchmarking health system performance across regions in Uganda: a systematic analysis of levels and trends in key maternal and child health interventions and outcomes, 1990-2011. BMC Medicine 2015, x(x). doi: xxxxxxxx[NF1]
Institute for Health Metrics and Evaluation (IHME). Assessing Impact, Improving Health: Progress in Child Health Across Regions in Uganda. Seattle, WA: IHME, 2014.
Citation: Institute for Health Metrics and Evaluation (IHME), Infectious Diseases Research Collaboration (IDRC). Uganda Malaria Control Policy Assessment Results 1990‒2011. Seattle, WA: IHME, 2015.
Date of release: December 3, 2015
Summary: This dataset contains estimates of maternal and child health (MCH) indicators in Uganda at the regional and national levels. These estimates were produced by the Institute for Health Metrics and Evaluation (IHME) and the Infectious Diseases Research Collaboration (IDRC) by using multiple data sources and applying complex modeling approaches. Trend estimates in this dataset include under-5 mortality, indicators of childhood nutrition (prevalence of underweight and stunting among children under 5), and a range of MCH interventions including malaria control, childhood immunizations, and other key MCH interventions such as antenatal care, skilled birth attendance, and exclusive breastfeeding. Regional estimates for a number of socioeconomic indicators, including women’s educational attainment and household characteristics, are also available.
File inventory
File name / Description / Date finalizedIHME_UGA_MCPA_1990_2011_FULL_RESULTS_Y2015M12D03.CSV / Contains estimates for all indicators, at the regional and national levels, from 1990 to 2011 / December 1, 2015
IHME_UGA_MCPA_1990_2011_{REGION}_Y2015M12D03.CSV / Contains estimates for all indicators, for a given region or at the national levels, from 1990 to 2011 / December 1, 2015
IHME_UGA_MCPA_1990_2011_{INDICATOR}_Y2015M12D03.CSV / Contains all estimates for a given indicator, at the regional and national levels, from 1990 to 2011 / December 1, 2015
IHME_UGA_MCPA_1990_2011_CB_Y2015M12D03.CSV / Contains the codebook for the full dataset / December 1, 2015
IHME_UGA_MCPA_1990_2011_INDICATOR_DEFINITIONS_Y2015M12D03.CSV / Contains abbreviations, full names, and definitions for each indicator / December 1, 2015
Data structure
The data are structured so that each row represents estimates by location-year-indicator (e.g., Kampala, 2000, under-5 mortality).
Acknowledgments
This research was carried out as part of the MCPA project in Uganda, a collaboration between IDRC at Makerere University and IHME at the University of Washington. Funding for this project came from the Bill & Melinda Gates Foundation. This work has benefited greatly from key inputs and support from the Ministry of Health and the National Malaria Control Programme in Uganda. We also thank the AIDS Control Program and the Uganda Bureau of Statistics for a range of survey data; the World Health Organization office in Uganda for providing access to immunization data; Uganda’s Medical Stores Limited and Joint Medical Store for granting access to drug distribution data; and Abt Associates for providing IRS data. We are most grateful to these organizations, especially for their willingness to facilitate data access and provide crucial content knowledge. We thank the MCPA Advisory Group, which consisted of international and local stakeholders who contributed toward refining the project’s research concept and framework. We thank all members of the MCPA team, including Mary Lakiyo at IDRC and Kelsey Pierce, Annie Haakenstad, Caterina Guinovart, and Ellie Colson at IHME, who contributed to the development and management of the project, as well as analyses.
Methodological statement
Data sources
SURVEYS
Demographic and Health Survey (DHS) 1995, 2000-2001, 2006, 2011
Malaria Indicator Survey (MIS)2009-2010
AIDS Indicator Survey (AIS)2011
Integrated Household Survey (IHS)1992-1993
Uganda National Household Survey (UNHS)1999-2000, 2002-2003, 2005-2006, 2009-2010
Uganda National Panel Survey (UNPS)2009-2010, 2010-2011
Uganda National Service Delivery Survey (UNSDS)2004, 2008
Netmark Survey reports 2000, 2006
CENSUSES
National population census1991, 2002
Overview of the analytical approach
In order to comprehensively assess regional and national levels and trends for key MCH outcomes and interventions, analyses took place in four main steps:
(1) Mapping: Due to frequent redistricting and data availability, we conducted our analysis at the regional level, using the 10 sub-regions delineated in the 2011 Uganda Demographic and Health Surveys.
(2) Collating data: We identified 27 health outcomes, interventions, and socioeconomic indicators based on their relevance to Uganda’s health priorities and data availability. We then conducted a comprehensive search of all available regional survey data for Uganda, with 17 surveys and two population censuses meeting inclusion criteria. Surveys were excluded if they did not measure any of the study’s indicators or we could not link units of observation to a given region.
(3) Generating source-specific estimates: We produced regional estimates for MCH and socioeconomic indicators using each data source. We calculated regional intervention coverage estimates for each survey-year of available data. Aside from information extracted from Netmark survey reports, all estimates were produced using microdata and accounting for each survey’s complex multistage design. We applied synthetic life table methods to pooled complete birth history data from sources where complete birth histories were collected (i.e., DHS) to generate direct estimates of under-5 mortality. For sources where summary birth histories only were collected (i.e., AIDS Indicator Survey [AIS] and both censuses), we applied the combined version of the maternal age cohort and maternal age period methods to generate indirect estimates of under-5 mortality.
(4) Estimating regional trends from 1990 to 2011: We used a two-step modeling approach to generate regional trends from 1990 to 2011 for each indicator. In the first stage, we fit the following linear mixed-effects model with random intercepts and slopes for each region.
Observations are indexed to region and year . For modeling coverage estimates, which were bounded between 0 and 1, logit transformation was applied. On the other hand, for variables such as years of maternal educations log transformation was used. We used a one-knot natural cubic spline with two basis functions (h1 and h2) to act as a smoother. The random effects (γ0i, γ1i, and γ2i) allow the levels and trends to vary between regions.
In the second step, the predicted trend from this linear model acts as the mean prior for Gaussian process regression (GPR), which is implemented with a Matern covariance function. GPR takes into account the model variance as well as the relative sampling uncertainty of the observed data to estimate a posterior mean function. We generated trends with uncertainty for each indicator by drawing 1,000 times from the posterior distribution and back-transforming to the original scale. The point estimate was based on the median of the draws, and 95% confidence intervals (CIs) were obtained by taking the 2.5th and 97.5th percentiles of the samples.
Under-5 mortality. To estimate trends in under-5 mortality for each region in Uganda, we applied the following model:
where is under-5 mortality in region , year , as measured by source . The terms are fixed effects; and are the global intercept and slope, respectively, and is an adjustment coefficient included for non-DHS sources to account for an observed discrepancy between under-5 mortality estimates derived from DHS surveys and those derived from other surveys. All other terms are random effects. Specifically, and are a region-level random intercept and slope, respectively, and are both assigned conditional autoregressive priors; these terms allow for each region to deviate from the global level and linear trend in under-5 mortality. is a year-level random intercept assigned a first-order random walk prior; this term allows for non-linearity in the global time trend. Similarly, is a region-year level random intercept with the prior given by the interaction between a conditional autoregressive prior for spatial trends and a first order random walk prior for temporal trends. This random effect allows for non-linearity in the region-specific time trends. Finally, is a source-year level random effect assigned an independent and identically distributed normal prior and is included to account for autocorrelation in estimates of under-5 mortality derived from the same source in the same region. Weakly informative normal priors were assigned to all fixed effects and weakly informative gamma priors were applied to the log precision of all random effects. To generate predictions from this model, we approximate the posterior distribution of by setting and to 0. The median, 2.5th, and 97.5th percentiles of this distribution are inverse-logit transformed to generate the point estimates and confidence intervals for in each region and year.
Overall intervention coverage. We constructed an overall intervention coverage metric to examine levels and trends across multiple key MCH indicators that reflect the priorities of Uganda’s health system. This metric included 11 interventions: three malaria interventions, four childhood vaccines, and four other MCH indicators. Each indicator was equally weighted for the overall intervention coverage metric, which was based on the average of the 11 indicators.
Codebooks
Variable names and labels can be found in the file “IHME_UGA_MCPA_1990_2011_CB_Y2015M12D03.CSV,” and definitions for each indicator can be found in the file “IHME_UGA_MCPA_1990_2011_INDICATOR_DEFINITIONS_Y2015M12D03.CSV.”
Contact information
To request further information about these results or other work conducted at IHME, please contact:
Institute for Health Metrics and Evaluation
2301 Fifth Ave., Suite 600
Seattle, WA 98121
USA
Telephone: +1-206-897-2800
Fax: +1-206-897-2899
Email:
[NF1]Will fill in on 12/3