2020 Data Sources

Mills, I. D. “The 1918-1919 Influenza Pandemic—The Indian Experience.” Indian Economic & Social History Review 23, no. 1 (March 1986): 1–40. https://doi.org/10.1177/001946468602300102.

For more information about vaccine coverage estimates by the Institute of Health Metrics and Evaluation (IHME), see the “Explore the Data” section in this report.

United States Census Bureau. “Week 11 Household Pulse Survey: July 9 - July 14.” U.S. Census Bureau (website), July 22, 2020. https://www.census.gov/data/tables/2020/demo/hhp/hhp11.html.

COVID and measles simulation data is from The Institute for Disease Modeling (IDM), Global Health Division, Bill & Melinda Gates Foundation.

Frey, Kurt, and Brittany Hagedorn. The Effect of Regional Under-Age-5 Campaigns on the SARS-CoV2 Outbreak in LMIC. Institute for Disease Modeling (website), June 30, 2020. https://covid.idmod.org/data/Effect_of_campaigns_COVID_transmission_in_LMIC.pdf.

The specific application of the measles models presented here has not yet been published, but the modeling methodologies employed have been described in previous publications. For more information, see:

Thakkar, Niket, Syed Saqlain Ahmad Gilani, Quamrul Hasan, and Kevin A. McCarthy. “Decreasing Measles Burden by Optimizing Campaign Timing.” Proceedings of the National Academy of Sciences 116, no. 22 (May 2019): 11069–11073. https://doi.org/10.1073/pnas.1818433116.

Data used in the model comes from publicly available sources such as WHO for case data and SIA calendar, the World Bank for birth rates and population, and WHO/UNICEF Estimates of National Immunization Coverage (WUENIC) for vaccination coverage estimates.

For more information about the other models referenced, see:

LSHTM or DynaMICE model:
Verguet, Stéphane, Mira Johri, Shaun K. Morris, Cindy L. Gauvreau, Prabhat Jha, and Mark Jit. “Controlling Measles Using Supplemental Immunization Activities: A Mathematical Model to Inform Optimal Policy,” Vaccine 33, no. 10 (March 3, 2015): 1291–1296. https://doi.org/10.1016/j.vaccine.2014.11.050.

Penn State model:
Chen, Shi, John Fricks, and Matthew J. Ferrari. “Tracking Measles Infection Through Non‐Linear State Space Models.” Journal of the Royal Statistical Society: Series C (Applied Statistics) 61, no. 1 (January 2012): 117–134. https://doi.org/10.1111/j.1467-9876.2011.01001.x.

Gentilini, Ugo, Mohamed Almenfi, Pamela Dale, Ana Veronica Lopez, Ingrid Veronica Mujica, Rodrigo Quintana, and Usama Zafar. Social Protection and Jobs Responses to COVID-19: A Real-Time Review of Country Measures (June 12, 2020). COVID-19 Living Paper. World Bank (website). documents1.worldbank.org/curated/en/590531592231143435/pdf/Social-Protection-and-Jobs-Responses-to-COVID-19-A-Real-Time-Review-of-Country-Measures-June-12-2020.pdf.

Gopinath, Gita. “Reopening from the Great Lockdown: Uneven and Uncertain Recovery.” IMFBlog (blog). International Monetary Fund (website), June 24, 2020. https://blogs.imf.org/2020/06/24/reopening-from-the-great-lockdown-uneven-and-uncertain-recovery.

International Labour Organization. ILO Monitor: COVID-19 and the World of Work. 3rd ed. ILO (website), April 29, 2020. https://www.ilo.org/wcmsp5/groups/public/@dgreports/@dcomm/documents/briefingnote/wcms_743146.pdf.

International Monetary Fund. “World Economic Outlook Database.” IMF (website). Accessed August 2020. https://www.imf.org/external/pubs/ft/weo/2020/01/weodata/index.aspx.

Kazzaz, Zachary, “Emergency Disbursements During COVID-19: Regulatory Tools for Rapid Account Opening and Oversight. Preprint, submitted July 20, 2020. https://dx.doi.org/10.2139/ssrn.3656651.

Pollack, Todd, Guy Thwaites, Maia Rabaa, Marc Choisy, Rogier van Doorn, Duong Huy Luong, Dang Quang Tan, et al. Emerging COVID-19 Success Story: Vietnam’s Commitment to Containment. Exemplars in Global Health (website). Accessed August 2020. https://www.exemplars.health/emerging-topics/epidemic-preparedness-and-response/covid-19/vietnam.

Impact of Global Recessions on GDP chart: Dashed line represents the baseline scenario. The shaded area is bounded by alternative scenarios as defined by the World Bank and the International Monetary Fund (IMF). Sources:

World Bank Group. Global Economic Prospects, June 2020. World Bank Group (website), June 2020. https://openknowledge.worldbank.org/bitstream/handle/10986/33748/9781464815539.pdf.

International Monetary Fund. World Economic Outlook Update. IMF (website), June 2020. https://www.imf.org/~/media/Files/Publications/WEO/2020/Update/June/English/WEOENG202006.ashx?la=en.

Size of Economic Stimulus in Response to COVID-19 chart: Overseas Development Institute. Country Policy Responses to COVID-19, as of 12 August 2020. ODI (website), August 12, 2020. https://set.odi.org/wp-content/uploads/2020/08/Country-fiscal-and-monetary-policy-responses-to-coronavirus_12-Aug-2020-.pdf.

GDP ($) and population data are 2018 values from the World Development Indicators database. In the chart, South Africa is grouped with other G20 countries, but is included in the calculation of average stimulus funding for both G20 countries and sub-Saharan African countries, respectively.

“World Development Indicators.” World Bank (website). Accessed August 2020. https://datacatalog.worldbank.org/dataset/world-development-indicators.

Forecast Global and Regional Poverty Trends chart: See the “Explore the Data” section for notes on poverty. Rates of change are based on IHME’s poverty estimates comparing number of people living at or below the extreme poverty line (US$1.90 a day 2011 purchasing power parity: PPP) and the lower-middle-income line (US$3.20 a day 2011 PPP) annually between 2017 and 2020.

World Health Organization. “The Access to COVID-19 Tools (ACT) Accelerator.” WHO (website). Accessed July 2020. https://www.who.int/initiatives/act-accelerator.

Pronker, Esther S., Tamar C. Weenen, Harry Commandeur, Eric H. J. H. M. Claassen, and Albertus D. M. E. Osterhaus. “Risk in Vaccine Research and Development Quantified.” PloS ONE8 , no. 3 (March 20, 2013): e57755. https://doi.org/10.1371/journal.pone.0057755.

How Many Lives Could Equitable Vaccination Save? chart: Chinazzi, Matteo, Jessica T. Davis, Natalie E. Dean, Kunpeng Mu, Ana Pastore y Piontti, Xinyue Xiong, M. Elizabeth Halloran, Ira M. Longini Jr., Alessandro Vespignani. Estimating the Effect of Cooperative Versus Uncooperative Strategies of COVID-19 Vaccine Allocation: A Modeling Study . Laboratory for the Modeling of Biological and Socio-technical Systems (MOBS LAB), Northeastern University (website), September 2020. https://www.mobs-lab.org/uploads/6/7/8/7/6787877/global_vax.pdf.

To estimate the effects of the COVID-19 pandemic on the SDG indicators included in the Goalkeepers Report and their accompanying projections to 2030, IHME implemented an approach as shown in Figure 1 for capturing both the short-term effects of the COVID-19 pandemic, as well as the long-term effects that reflect its impact on economic production and development and the consequences of these on the SDG indicators.

These projections are based on a hybrid Susceptible-Exposed-Infected-Recovered or SEIR model, in which the model analyses data on cases, hospitalizations, and deaths to estimate the death rate in the past and next eight days. The SEIR model then fits a statistical model of transmission using a range of drivers, including cell phone mobility (predicted by mandates and the underlying trend), mask use, pneumonia death rate seasonal pattern, testing per capita, altitude, particulate matter (PM2.5), and population density.

For the purposes of this report, projections were extended to December 31, 2021. To project the most likely or reference scenario for this time period, the projections incorporate re-imposition of social distancing mandates on a location-specific basis for six weeks when the daily death rate for that location reaches eight deaths per million population. This threshold corresponds to the 90th percentile of the threshold for mandates in the first round, and the median for the second round of mandates. This reflects the increased reluctance of decision makers to impose mandates due to the economic consequences. Re-imposition of mandates reduces projected mobility based on the historical, location-specific reduction in mobility, which in turn reduces transmission in the SEIR model and subsequently infections and deaths from COVID-19.

These methods are described in:

IHME COVID-19 Forecasting Team, and Simon I. Hay. “COVID-19 Scenarios for the United States.” Preprint, submitted July 14, 2020. https://www.medrxiv.org/content/10.1101/2020.07.12.20151191v1.

Out-of-sample predictive validity tests show that the model has the lowest error rate among seven publicly released models, as described in:

Friedman, Joseph, Patrick Liu, Emmanuela Gakidou, and IHME COVID-19 Model Comparison Team. “Predictive Performance of International COVID-19 Mortality Forecasting Models.” Preprint, submitted July 14, 2020. https://www.medrxiv.org/content/10.1101/2020.07.13.20151233v3.full.pdf.

To capture the impact of COVID-19 and the subsequent global economic recession on key indicators, IHME modeled the current and projected effects of the pandemic on GDP per capita to 2030. To estimate the impact of COVID-19 and COVID-19 policy responses on economic production (i.e., GDP per capita) in the short term (2020 and 2021), IHME developed a two-step process.

First, IHME assessed the relationship between estimated COVID-19 deaths and changes in population mobility (as described in the previous subsection) against key GDP correlates, each measured monthly through June 2020. Numerous GDP correlates were considered. For the final analysis, IHME relied on five key GDP correlates: tourist arrivals, consumer price index, electricity production, employment rates, and short-term interest rates. Sources include the CEIC Global Database, International Monetary Fund, Organisation for Economic Co-operation and Development, World Bank, and World Bank World Development Indicators.

Using the estimated relationship between these correlates and COVID-19 deaths and mobility, and projected COVID-19 deaths and mobility patterns from IHME, IHME estimated expected relative changes for these five correlates.

For step two, IHME assessed the historical relationship between changes in these same five correlates and GDP growth rates, using annual data from 2005 through 2019. Using this relationship and projected 2020 and 2021 growth rates for the correlates, IHME estimated COVID-19-sensitive GDP growth rates for 2020 and 2021. For estimating annual economic growth beyond 2021 and developing alternative scenarios, IHME used its existing IHME GDP forecasting ensemble framework. See:

Global Burden of Disease Health Financing Collaborator Network. “Health Sector Spending and Spending on HIV/AIDS, Tuberculosis, and Malaria, and Development Assistance for Health: Progress Towards Sustainable Development Goal 3.” Lancet (April 23, 2020). https://www.thelancet.com/action/showPdf?pii=S0140-6736%2820%2930608-5.

The drop in growth in 2020 and 2021 leads to slower growth rates from 2021 onward and indirectly includes long-term effects of the recession on economic development.

IHME reflected revised GDP projections in its projections of the Socio-demographic Index (SDI) which incorporates income, fertility, and education. Using the existing framework, IHME also estimated the downstream consequences of reduced GDP on health spending and development assistance of health. Key drivers were identified for each indicator and were in turn used for the projections to 2030. This approach allows for the forecasted economic COVID-19 impacts on GDP to be reflected in the projections to 2030. To do so, IHME employed its MR-BRT meta-regression tool in a two-stage model. In the first stage, IHME fits an indicator spline on the covariate, typically SDI. In the second stage, the residuals of the first stage model are fit with a linear model for each country on calendar year (country-specific slope).

The two model stages are then combined to create predictions. The “better” and “worse” scenarios are constructed from SDI scenarios (15th and 85th percentile) combined with the 15th and 85th percentile of the country-specific slopes in the second stage. In some instances, SDI, or an alternative covariate, is used in both stages of the MR-BRT model.

Health services have also faced disruption due to the implementation of social distancing mandates or lockdowns. IHME relied on new primary data collection to estimate these effects of the pandemic on health indicators using smartphone-based and computer-assisted surveys for the following target populations:

• General population, totaling 70,808 respondents from 82 countries. Main indicators include health care utilization and medication use.
• Women of reproductive age, totaling 13,893 respondents from 76 countries. The main indicator is use of contraceptives.
• Caregivers of children under the age of 2 years, totaling 7,230 respondents from 76 countries. The main indicator is vaccine coverage.
• Pregnant women or women who have given birth in the last six months, totaling 2,129 respondents from 73 countries. Main indicators include antenatal care, in-facility delivery, and presence of a skilled birth attendant.
• General population in malaria endemic countries, totaling 14,615 respondents from 20 countries. Main indicators include insecticide-treated bed net (ITN) usage and artemisinin-based combination therapy (ACT) usage.

Data was sample-weighted according to the age, sex, and educational attainment profiles by country. Survey data was used to estimate the change in the key indicators or underlying drivers between the pre-pandemic period (December 2019–February 2020) and since the start of the pandemic (March–June 2020).

In addition to survey data, IHME identified available administrative data that reported on health service disruption. In the time available, IHME was only able to access administrative data for 2020 from the following sources: Nigeria National Health Analytical Tool; South Africa National Department of Health; Morbidity and Mortality Weekly Report series of the U.S. Centers for Disease Control and Prevention; and data from the United Kingdom published in the scientific journal Eurosurveillance.

IHME related the level of health service interruption (e.g., relative ratio of vaccine doses delivered) to measures of cell phone mobility developed as part of the COVID-19 projections. These mobility estimates are based on data from Facebook, Google, Descartes Labs, and SafeGraph; this data has been shown to be highly related to the imposition of social distancing mandates by governments. To estimate the relationship between health service interruption and mobility, IHME used a random-spline meta-regression. This captures the nonlinear relationship between the level of interruption and mobility changes, and it allows for variation in this relationship by country. This function is then combined with the projections of mobility to December 31, 2021, to estimate changes in indicators or drivers for 2020 and 2021. These are then used to adjust the long-term projections as described earlier in this section.

Given reporting lags, there was limited data, particularly from administrative data sources for both economic and health indicators. While IHME collected important survey data, it was limited by the mode of collection (smartphone, CATI) and by small sample sizes for several indicators. In several cases, IHME relied on proxy indicators of service disruption (e.g., overall medication use for ART), but where possible, IHME has assessed whether there were statistically significant differences between the proxy indicators and the indicator of interest. There is also substantial uncertainty regarding whether and how indicators will rebound (e.g., unemployment, catch-up vaccination). Data was largely limited to the beginning of the pandemic (before May/June 2020), when COVID-19 cases and deaths were increasing and mobility was decreasing. IHME was not able to incorporate debt explicitly into its economic analysis. However, the long-term effects of debt due to COVID-19 are captured indirectly through slower economic growth rates.

There is also considerable uncertainty in the COVID-19 projections, particularly for Africa. While IHME’s COVID-19 projections are the best performing among public models, median absolute percent error is still relatively large at 13 percent at six weeks. This is a reflection of the fact that current drivers in the COVID-19 model do not fully capture epidemic variation and the overall effectiveness of countries’ pandemic response. Furthermore, reference scenario projections are a function of policy response; it has been difficult to predict when governments implement social distancing mandates in response to subsequent waves. Finally, IHME has projected COVID-19 only until the end of 2021.

Extreme poverty rates measure the fraction of a country’s population that is estimated to live on less than US$1.90 per day, measured in 2011 purchasing power parity (PPP) adjusted dollars. To estimate a complete time series of extreme poverty for all countries, all available data was first extracted from the World Bank and supplemented with data extracted from the United Nations’ World Institute for Development Economics Research and country-specific surveys. Second, IHME modeled this extracted data using an approach that builds from available data and uses information from across time, geography, and predictive covariates (GDP per capita, female education, kilocalorie consumption, natural resource exports, and government expenditure).

IHME models the mean consumption rate for each country and year, and the consumption distribution (the Lorenz curve) for each country, in order to estimate the value of consumption for each percentile of the population of each country and year through 2021. While no survey data was available beyond 2019, IHME uses this model to estimate poverty rates for 2020 and 2021 because it is more sensitive to economic shocks, like those currently being experienced in most countries. IHME forecasted extreme poverty rates (US$1.90) and lower-middle-income poverty estimates (US$3.20) for 2022 to 2030 by estimating the year-over-year change in the poverty rate using an ensemble model. This model is based on GDP per capita, fertility, government expenditure, and education forecasts; it only indirectly captures the other impacts of the global economic recession.

IHME measures stunting prevalence as height-for-age more than two standard deviations below the reference median on the height-age growth curve based on WHO 2006 growth standards for children 0–59 months. Projections to 2030 used SDI as a key driver, which incorporates projections of income per capita and the effects of the COVID-19 pandemic.

Roberton, Timothy, Emily D. Carter, Victoria B. Chou, Angela R. Stegmuller, Bianca D. Jackson, Yvonne Tam, Talata Sawadogo-Lewis, and Neff Walker. “Early Estimates of the Indirect Effects of the COVID-19 Pandemic on Maternal and Child Mortality in Low-Income and Middle-Income Countries: A Modelling Study.” Lancet Global Health 8, no. 7 (May 12, 2020): e901–08. https://doi.org/10.1016/S2214-109X(20)30229-1.

The maternal mortality ratio (MMR) is defined as the number of maternal deaths among women ages 15–49 years during a given time period per 100,000 live births during the same time period. It depicts the risk of maternal death relative to the number of live births and essentially captures the risk of death in a single pregnancy or a single live birth. Short-term effects (2020–2021) incorporated the effect of reductions in in-facility delivery. IHME estimated the level of interruption in in-facility delivery, using survey data on the level of interruption in all health provider visits as a proxy. IHME did not, however, find statistically significant differences between the relative level of interruptions in in-facility delivery and any health provider visits in the pooled sample. Projections to 2030 used SDI as a key driver, which incorporates projections of income per capita and the effects of the COVID-19 pandemic.

IHME defines under-5 mortality rate as the probability of death between birth and age 5. It is expressed as number of deaths per 1,000 live births. Projections were based on a combination of key drivers, including Global Burden of Disease (GBD) risk factors, selected interventions (e.g., vaccines, ITNs, ACTs), and SDI.

IHME defines neonatal mortality rate as the probability of death in the first 28 completed days of life. It is expressed as the number of deaths per 1,000 live births. Short-term effects (2020–2021) incorporated the effect of reductions in in-facility delivery. IHME estimated the level of interruption in in-facility delivery using survey data on the level of interruptions in all health provider visits as a proxy for in-facility delivery. IHME did not, however, find statistically significant differences between the relative level of interruptions in in-facility delivery and any health provider visits in the pooled sample. Projections were based on a combination of key drivers, including GBD risk factors, selected interventions (e.g., vaccines), and SDI.

IHME estimates the HIV rate as new HIV infections per 1,000 population. Forecasts of HIV incidence were based on forecasted antiretroviral therapy (ART), prevention of maternal-to-child transmission (PMTCT) coverage, and incidence as inputs into a modified version of Avenir Health’s Spectrum software. Adult ART is forecasted using the expected spending on HIV curative care—which in turn was forecasted based on income per capita, including the effects of the COVID-19 pandemic—and ART prices. Additionally, the short-term effects (2020–2021) of the COVID-19 pandemic on ART coverage was estimated using survey data. Due to sample size limitations, any medication interruption was used as a proxy for interruptions in ART medication by country. No statistically significant differences were found between ART medication and any medication interruption in the pooled sample.

The observed result of increasing incidence beyond the COVID-19-disrupted years is due to the observed rates of change in no-treatment incidence based on pre-2019 data, and the expected future ART. While the COVID-specific effects affect ART coverage in 2020–21, the overall projected change to 2030 is based on the observed past trends, the length of time a country has been experiencing an epidemic, and how ART overall is expected to change, which is in turn dependent on future HIV spending. The scenarios are derived using the 15th, 50th, and 85th percentiles, respectively, of a modeled past rate of change in no-treatment incidence.

IHME estimates new and relapse tuberculosis cases diagnosed within a given calendar year (incidence) using data from prevalence surveys, case notifications, and cause-specific mortality estimates as inputs to a statistical model that enforces internal consistency among the estimates.

Methodological improvements to the GBD 2019 study have led to lower historical estimates of TB incidence compared with last year’s estimates. Improvements include incorporating excess mortality rates as a function of health care access and quality (HAQ) and improved methods using the mortality-to-incidence (MI) ratio. These approaches have resulted in improved estimates in countries including Rwanda, Uganda, Nigeria, Botswana, and South Africa.

IHME has also revised the modeling approach to forecasting TB incidence. In addition to historical trends, projections to 2030 used SDI as a key driver, which incorporates projections of income per capita and the effects of the COVID-19 pandemic. IHME also incorporated the short-term effects (2020–2021) of COVID-19 on TB incidence by using the level of interruption in any medication use as a proxy for interruption in TB treatment by country from the survey data and applied that effect using the historical relationship between HAQ and TB incidence. Unfortunately, there was insufficient sample size in the survey data to estimate a direct effect on TB medication. It was also not possible under the timeframe, nor was there data available to incorporate, the direct effect of reduced testing on TB incidence.

IHME estimates the malaria rate as the number of new cases per 1,000 population. Short-term effects (2020–2021) were measured via survey data on artemisinin-based combination therapy (ACT) coverage interruption and relative changes in the number of insecticide-treated bed nets (ITNs) received or purchased since the pandemic compared to before. Projections to 2030 were derived using a two-stage model. First, coverage of ACT and ITNs is forecast as a function of malaria development assistance for heath (DAH), which is predicted in turn by projections of income per capita. After fitting a spline on intervention coverage in the first stage, IHME then uses the residuals from the first stage to fit a country-specific linear model on calendar year. For countries outside of sub-Saharan Africa, where there is no available data on intervention coverage, SDI is used in the first stage, and calendar year in the second stage.

IHME measures the sum of the prevalence of 15 NTDs per 100,000 that are currently measured in the annual Global Burden of Disease study: human African trypanosomiasis, Chagas disease, cystic echinococcosis, cysticercosis, dengue, food-borne trematodiases, Guinea worm, soil-transmitted helminths (hookworm, trichuriasis, and ascariasis), leishmaniasis, leprosy, lymphatic filariasis, onchocerciasis, rabies, schistosomiasis, and trachoma. Short-term effects (2020–2021) varied by NTD. For the preventive chemotherapy NTDs, IHME assumed modest increases in prevalence as a result of missing a single round of mass drug administration (MDA) (lymphatic filariasis, onchocerciasis, soil-transmitted helminths, and schistosomiasis). For NTDs that rely on active case detection as a primary strategy for control, IHME assumed a discontinuation of active and passive case detection, which results in increased prevalence (leishmaniasis, human African trypanosomiasis, and Chagas). IHME further assumed that 15 percent of individuals infected with rabies would not receive post-exposure prophylaxis (PEP) and minimal adjustments for dengue due increased geographic spread. IHME assumed increased leprosy prevalence due to moderate shifts in severity in grades 1 and 2 due to lack of treatment coverage. IHME assumed no change in prevalence for NTDs transmitted through contaminated food with a long latency period (food-borne trematodiases, cystic echinococcosis, and cysticercosis). No adjustments were made for prevalence of blindness or low vision due to trachoma or to Guinea worm disease. Projections to 2030 used SDI as a key driver, which incorporates projections of income per capita and the effects of the COVID-19 pandemic.

NTD Modelling Consortium. “The Potential Impact of Programmes Interruptions Due to COVID-19 on 7 Neglected Tropical Diseases: A Modelling-based Analysis.” Preprint, submitted July 24, 2020. https://doi.org/10.21955/gatesopenres.1116665.1.

IHME estimates the proportion of women of reproductive age (15–49 years) who have their need for family planning satisfied with modern contraceptive methods. Modern contraceptive methods include the current use of male or female sterilization, male or female condoms, diaphragms, cervical caps, sponges, spermicidal agents, oral hormonal pills, patches, rings, implants, injections, intrauterine devices (IUDs), and emergency contraceptives. Short-term effects (2020–2021) were measured via survey data. Due to sample size limitations, IHME used any medication interruption as a proxy by country. There were no statistically significant differences found between contraception and any medication interruption in the pooled sample. IHME incorporated questions on method mix and changes in demand into the survey but ultimately was not able to incorporate these into the analysis due to small sample sizes. Projections to 2030 used SDI as a key driver, which incorporates projections of income per capita and the effect of the COVID-19 pandemic.

The universal health coverage (UHC) effective coverage index is a new metric composed of 23 effective coverage indicators that cover population-age groups across the entire life course (maternal and newborn age groups, children under age 5, youths ages 5–19 years, adults ages 20–64, and adults ages 65 years old or older). These indicators fall within several health service domains: promotion, prevention, treatment, rehabilitation, and palliation.

Health system promotion indicators include met need for family planning with modern contraception.

Health system prevention indicators include the proportion of children receiving the third dose of the diphtheria-tetanus-pertussis vaccine and children receiving the first dose of measles-containing vaccine. Antenatal care for mothers and antenatal care for newborns are also considered indicators of health system prevention and treatment of diseases affecting maternal and child health.

Indicators of treatment of diseases affecting maternal and child health and communicable diseases are the mortality-to-incidence (MI) ratios for lower respiratory infections, diarrhea, and tuberculosis, as well as coverage of antiretroviral therapy among those with HIV/AIDS. For non-communicable diseases they are the MI ratios for acute lymphoid leukemia, asthma, epilepsy, appendicitis, paralytic ileus and intestinal obstruction, diabetes, stroke, chronic kidney disease, chronic obstructive pulmonary disease, cervical cancer, breast cancer, uterine cancer, colorectal cancer, and the risk-standardized death rate due to ischemic heart disease.

A novel weighting scheme was developed for the analysis: Each individual indicator was weighted by its theoretical potential impact on reducing disability-adjusted life years (DALYs) within each location and year to create the new UHC effective coverage index. The UHC effective coverage index differs from the UHC index produced for the 2019 Goalkeepers Report, which has led to different estimates in the 2020 Goalkeepers Report when compared to the 2019 Goalkeepers Report. To produce forecasts of the UHC index from 2020 to 2030, a meta-stochastic frontier model for UHC was fit, using total health spending per capita projections as the independent variable. Country- and year-specific inefficiencies were then extracted from the model and forecasted to 2030 using a linear regression with exponential weights across time for each country level. These forecasted inefficiencies, along with forecasted total health spending per capita estimates, were substituted into the previously fit frontier to obtain forecasted UHC for all countries for 2020–2030. Short-term effects (2020–2021) were included by adjusting our estimates for 2020 and 2021 downward based on adjustment factors from the survey data, using any missed medication as a proxy for reductions in UHC.

IHME measures the age-standardized prevalence of the current use of smoked tobacco among ages 15 and older. IHME collates information from all available surveys that include questions about frequency of tobacco use (e.g., daily, occasional), either currently or within the last 30 days, and information on the type of tobacco product smoked (including cigarettes, cigars, pipes, hookahs, as well as local products). IHME converts all data to its standard definition so that meaningful comparisons can be made across locations and over time. Projections to 2030 used SDI as a key driver, which incorporates projections of income per capita and the effects of the COVID-19 pandemic.

IHME’s measurement of immunization coverage reports on the coverage of the following vaccines separately: three-dose diphtheria-tetanus-pertussis (DTP3), measles second dose (MCV2), and three-dose pneumococcal conjugate vaccine (PCV3). IHME measured the short-term (2020–2021) effects via survey data based on missed visits for vaccination and administrative data on vaccine doses. In collaboration with WHO, IHME was able to synthesize data on the number of vaccine doses delivered by month in early 2019 and 2020 from 85 countries. To estimate the change in vaccine coverage since the onset of the pandemic for each country, the number of doses delivered in each month between March and May 2020 was compared to those delivered in the same month in 2019, adjusting for pre-pandemic year-to-year changes observed in January and February. Survey and administrative data were triangulated with qualitative information on the level of vaccine system disruption compiled by WHO, including two recent WHO pulse polls, WHO Essential Health Services polls, and reports from WHO regional offices. Data sources that were implausible based on the reported level of disruption within a country were excluded. Projections to 2030 used SDI as a key driver, which incorporates projections of income per capita and the effects of the COVID-19 pandemic.

IHME measured households with piped sanitation (with a sewer connection or septic tank); households with improved sanitation but without a sewer connection (pit latrine, ventilated improved latrine, pit latrine with slab, composting toilet); and households without improved sanitation (flush toilet that is not piped to sewer or septic tank, pit latrine without a slab or open pit, bucket, hanging toilet or hanging latrine, no facilities), as defined by the Joint Monitoring Programme for Water Supply and Sanitation. Projections to 2030 used SDI as a key driver, which incorporates projections of income per capita and the effects of the COVID-19 pandemic.

RuLIS - Rural Livelihoods Information System “Data by Indicator.” RuLIS (website), Food and Agriculture Organization of the United Nations (FAO). Accessed August 2020. www.fao.org/in-action/rural-livelihoods-dataset-rulis/data/by-indicator/en/. Most recent year available was used for select countries, ranging from 2005–2017.

For methodology, see:

Food and Agriculture Organization of the United Nations. Rural Livelihoods Information System (RuLIS): Technical Notes on Concepts and Definitions Used for the Indicators Derived from Household Surveys. Report. Rome: FAO, 2018. www.fao.org/3/ca2813en/CA2813EN.pdf.

UNESCO Institute for Statistics. “Sustainable Development Goal 4.” UIS (website). Data accessed August 2020. http://data.uis.unesco.org/Index.aspx?DataSetCode=SDG4_DS.

The UNESCO Institute for Statistics (UIS) updated its Protocol for Reporting Indicator 4.1.1 in February 2020. This change in protocol addressed the selection of data sources when there are more than one available for a given country and indicator, thus avoiding having multiple data sources in the time series. It also altered the criteria for use of the results from National Learning Assessments, which are now restricted to assessments using Item Response Theory (IRT). This has reduced the number of data points compared to the data published in the SDG 4 Data Book: Global Education Indicators 2019.

UIS, and Global Alliance for Monitor Learning. Protocol for Reporting Indicator 4.1.1. UIS and Global Alliance for Monitor Learning, February 2020. gaml.uis.unesco.org/wp-content/uploads/sites/2/2019/05/GAML6-WD-2-Protocol-for-reporting-4.1.1_v1.pdf.

UIS. SDG 4 Data Book: Global Education Indicators 2019. UIS, 2019. uis.unesco.org/sites/default/files/documents/sdg4-databook-global-ed-indicators-2019-en.pdf.

Concerning the UIS analysis on post-COVID learning loss, several estimates seek to model the impact of COVID-19 on students’ achievement of minimum proficiency in reading at the end of primary school education. UIS analysis adapts the effects of three studies to show the projected percentage of students achieving minimum proficiency, in a scenario where 68 percent of countries implement remediation activities (the percentage of countries that are planning remedial activities according to a UNESCO, UNICEF, and World Bank survey). For the estimates from the World Bank’s “Simulating…” study and Gustafsson and Nuga’s (noted below), the effect of the pandemic on minimum proficiency is assumed in 2020, with a linear trajectory back to the pre-COVID projections; this effect would disappear as students exposed to the COVID-19 shock exit the education system. The Kaffenberger analysis emphasizes the cumulative effect of lost learning over students’ school careers (i.e., for students who were affected by the pandemic in their early grades and eventually reach the end of primary); this effect is assumed to continue until around 2025. The estimates are calculated as a weighted average between remedial and nonremedial scenarios laid out in the papers. These estimates illustrate the potential scale of learning losses and also reflect the variety of potential scenarios that could take place.

Gustafsson, Martin, and Carol Nuga. How Is the COVID-19 Pandemic Affecting Educational Quality in South Africa? Evidence to Date and Future Risks. Policy brief, July 15, 2020. https://cramsurvey.org/wp-content/uploads/2020/08/Nuga.-Gustafsson_policy-brief.pdf.

Kaffenberger, Michelle. “Modeling the Long-Run Learning Impact of the COVID-19 Learning Shock: Actions to (More Than) Mitigate Loss.” RISE Insight Series, Research on Improving Systems of Education, June 4, 2020. https://doi.org/10.35489/BSG-RISE-RI_2020/017.

Azevedo, João Pedro, Amer Hasan, Diana Goldemberg, Syedah Aroob Iqbal, and Koen Geven. Simulating the Potential Impacts of Covid-19 School Closures on Schooling and Learning Outcomes: A Set of Global Estimates. Washington, D.C.: World Bank Group, June, 2020. https://doi.org/10.1596/1813-9450-9284.

The chart is adapted from:

UN Women. Progress of the World’s Women 2019-2020: Families in a Changing World. Report. New York: UN Women, 2019. https://www.unwomen.org/-/media/headquarters/attachments/sections/library/publications/2019/progress-of-the-worlds-women-2019-2020-en.pdf?la=en&vs=3512.

The data is the most recent available for 88 countries and territories (2001–2017). The age group is 15 or older where available (18 or older in Ghana). In a number of cases, data is for those ages 10 or older or 12 or older. In the case of Thailand (2015) they are for those ages 6 or older, and in the United Republic of Tanzania (2014) for those ages 5 or older. Data for Bulgaria, Denmark, Latvia, the Netherlands, Slovenia, and Spain corresponds to time spent on unpaid care among those ages 20 to 74 only. In the case of Qatar, only urban areas are covered in the analysis. Differences across countries should be interpreted with caution, given heterogeneity across surveys and countries in definitions, methodology, and sample coverage. For further information on the country-level data, see:

UN Statistics Division. “SDG Indicators: United Nations Global SDG Database.” UNSD. Updated August 2019. https://unstats.un.org/sdgs/indicators/database/.

UN Women, and Gender and COVID-29 Working Group. “Will the Pandemic Derail Hard-Won Progress on Gender Equality?” New York: UN Women, 2020. https://www.unwomen.org/-/media/headquarters/attachments/sections/library/publications/2020/spotlight-on-gender-covid-19-and-the-sdgs-en.pdf?la=en&vs=5013.

Demirgüç, Asli, Leora Klapper, Dorothe Singer, Saniya Ansar, and Jake Hess. The Global Findex Database 2017: Measuring Financial Inclusion and the Fintech Revolution. Washington, DC: World Bank, 2018. https://globalfindex.worldbank.org/.

The “Richest/Poorest” comparison refers to what the World Bank calculates as account ownership of the richest 60% of households and poorest 40% of households, respectively.