Universal health coverage, as World Health Organization Director-General Tedros Adhanom Ghebreyesus has frequently said, is a political choice,1 and yet the role that a country’s political system plays in ensuring essential health services and minimizing financial risk remains poorly understood. During the past three decades many studies have identified links between democracy and lower infant and child mortality rates and between democracy and longer average life expectancy at birth.2 Recent research has found that democracy is associated with longer adult life expectancy at age fifteen and with lower mortality from causes such as cardiovascular disease and transport injuries, which are less targeted by foreign aid and depend on a robust health care infrastructure.3 Although the aforementioned literature has established the link between democracy and population health outcomes, much less is known about how political institutions may influence health services access and financing.
Universal health coverage aims to ensure that all members of a population have access to the high-quality essential health services that they need without suffering financial hardship. The importance of universal health coverage to many countries and international institutions is emphasized by the fact that it is included in the UN Sustainable Development Goals.4 Yet progress on achieving universal health coverage worldwide was already uneven ahead of the 2030 Sustainable Development Goals deadline and may falter further because of the COVID-19 pandemic and its associated economic downturn.
In 2019, before the pandemic began, more than half of the world’s population did not have sufficient access to essential health services. Each year millions of people are pushed into poverty as a result of out-of-pocket health expenses.5 The COVID-19 pandemic is widening those coverage gaps and amplifying their consequences. Deterred access to essential health services during the pandemic and delays in immunization and other health care services have led to resurgences of vaccine-preventable diseases and worsened noncommunicable disease outcomes.6 The pandemic has exposed disparities in access to high-quality health care in many nations, and there have been glaring differences in infection and mortality rates between privileged and marginalized communities.7 Those disparities may increase in the coming months as the economic toll of the pandemic persists in many nations. The International Monetary Fund has estimated that gross domestic product (GDP) growth rate in 2020 will be 5.8 percentage points lower than pre-COVID-19 projected growth in emerging markets and developing countries (excluding China).8 In past recessions, high national debt levels have led some governments to adopt austerity programs, causing reductions in medical services and use.9,10
The theoretical reasoning that democracy should promote progress toward universal health coverage is straightforward: When governed by regular, free, and fair elections, democracies should have a greater incentive than autocracies to provide health-promoting resources and services to the wider population, including the poor. Democracies are also more open to feedback from a broader range of interest groups, are more protective of media freedom, and may be more willing to use that feedback to extend and improve effective essential health services. Autocracies, in contrast, reduce political competition and access to information, which might deter constituent feedback and responsive governance.11–13 For similar reasons, autocracies may be more willing than democracies to curb social spending during economic slowdowns, including investment in universal health coverage. Democracies have a greater incentive to maintain or increase social spending, including on universal health coverage, to shore up electoral support during recessions.14
With this in mind, in this study we assessed the relationships that democratic quality has with universal health coverage and with government health financing, and how economic recessions affect those relationships. We used observational data spanning 170 countries and three decades (1990–2019) to examine the role that democracy plays in the adoption of universal health coverage by governments and in governmental health spending, as well as to assess the specific political characteristics that connect democracy, service provision, and spending. In addition, we compared how economic recessions affect universal health coverage in democracies and autocracies. Although it is too early to determine the effects of COVID-19 on universal health coverage and government health spending, this analysis advocates a plausible scenario of how different types of governments may respond to the economic downturn that is unfolding globally as a consequence of that pandemic.
Study Data And Methods
We assembled panel data spanning 170 countries during the period 1990–2019, with key inputs including an assessment of democratic quality from the Varieties of Democracy project and universal health coverage (UHC) effective coverage and health spending indicators from the Global Burden of Disease research consortium. We completed three investigations. First, we measured the associations that democratic quality had with universal health coverage and government health spending, using panel regression methods. Next, we used synthetic control methods to assess how economic recessions affect the associations that universal health coverage has within democracies and autocracies. Finally, we examined the mechanisms connecting democratic quality and UHC effective coverage, using machine learning methods. We provide an overview of our data and statistical analyses in the following sections; more information is in the online appendix.15
We extracted data from the Varieties of Democracy project, which constructs indices of regime characteristics based on consultation with country experts. The project uses Bayesian item response modeling techniques to aggregate responses across experts on a range of highly specific survey questions. We used version 10 of the Varieties of Democracy data set, capturing information about 201 countries from 1789 to 2019.16 Our independent variable of interest, democratic quality, is an index that ranges from 0 to 1, with 0 representing full autocracy and 1 representing full democracy. This index is constructed on the basis of six components that, in combination, make political leaders responsive to citizens: suffrage, free and fair elections, elected officials, freedom of civil and political organization, and freedom of expression.
We used the UHC effective coverage index produced by the Global Burden of Disease research consortium to measure countries’ progress on Sustainable Development Goals indicator 3.8.1, coverage of essential health services, during the period 1990–2019 for all countries.17,18 This index builds on the Global Burden of Disease 2019 project to capture twenty-three indicators that, when combined, can be used to evaluate progress toward UHC effective coverage. Descriptions of each indicator are in appendix section 1.215 and have been published elsewhere.17
Health spending data were drawn from the Global Health Data Exchange and are available for all countries included in this study from the period 1995–2017.19 These data track government health spending from domestic sources, prepaid private health spending, out-of-pocket health spending, and development assistance for health, as well as total health spending, and are measured in inflation-adjusted purchasing power parity–adjusted dollars.
We included country-level, time-varying covariates that influence UHC effective coverage. We included GDP per capita from the Global Health Data Exchange, measured in 2019 purchasing power parity US dollars adjusted for inflation.20 We included war mortality probabilities and major natural disasters as covariates to control for contemporaneous shocks. The probability for mortality due to war was extracted from the Global Burden of Disease 2019 Covariates database.21 The indicator for natural disasters was based on the International Disaster Database, compiled by the Centre for Research on the Epidemiology of Disasters.22 Finally, we extracted urbanization, defined as the percentage of the population living in an urban setting, from the World Bank’s World Development Indicators.23 To estimate how the current global recession may affect universal health coverage and health spending, we extracted GDP growth projections for 2020 from the International Monetary Fund.8
Analysis 1: Effect Of Democracy On Universal Health Coverage
First, we examined the degree to which country democratic quality was associated with UHC effective coverage and per person government health spending in 170 countries. We used linear and nonlinear regression methods and included country and year fixed effects to control for unobserved time-invariant variation and global time trends and shocks that affected all countries.24 We included the interaction of GDP per capita and democratic quality, which allowed for the association of democratic quality to be distinct for countries with different levels of GDP per capita. We also controlled for the probability of mortality due to war and terrorism, urbanicity, and per person development assistance for health. We used cluster robust standard errors to adjust the standard errors for grouping of residuals by country.
Analysis 2: Responses To Economic Recessions
Second, we used a generalized synthetic control method25 to test the hypothesis that democracies and autocracies respond differently to economic recessions and—more specifically—that in the face of economic recession, autocracies are more apt to cut health services. To do this, we stratified our sample into “democracies” and “autocracies,” as defined by the mode of the Regimes of the World index from Varieties of Democracy.16 According to that index, a country qualifies as democratic if its political leaders are selected via competitive multiparty elections and autocratic if they are unelected or are selected via uncompetitive elections designed so only the opposition could lose. Countries were identified as having undergone a severe economic shock in a particular year if they underwent a percentage decrease in GDP from one year to the next that placed them in the top 10 percent of all such decreases for the period 1990–2016. We excluded countries with an economic recession before 2000 from the sample, and we used only the first shock if a country experienced more than one economic recession.
We identified forty-eight democracies and thirty-six autocracies as having had a major recession between 2000 and 2016, with most recessions occurring around 2008. To construct a synthetic control, or counterfactual, for each of these “treatment” countries, we used a weighted combination of the sixty-four countries that did not experience a recession during the period 1990–2019.
For each country in the treatment group, we calculated a synthetic control. This synthetic control for each treated country was a weighted combination of countries that were unaffected by a recession during the period under study. If the synthetic closely matched the path of the outcome variable during the pretreatment period, then we had greater confidence that it reliably estimated the counterfactual path of that variable during the posttreatment period. The gap between the actual and counterfactual during the posttreatment period provided an estimate of whether the treated countries performed better or worse than would have been the case in the absence of the treatment. The model includes country and year fixed effects, as well as probability of deaths due to war and terrorism, the probability of deaths due to major natural disasters, and development assistance for health as controls for contemporaneous shocks and fluctuations in aid flows. The outcome variable is UHC effective coverage. Standard errors were produced using parametric bootstraps, blocked by country.
Analysis 3: Political Mechanisms Of Universal Health Coverage Improvement
Third, we explored potential pathways by which democratic quality may influence universal health coverage with a random forest analysis. A random forest regression is an ensemble learning method that aggregates the results of many bootstrapped regression trees and is particularly robust against overfitting.26 We aimed to predict variation in UHC effective coverage and per person government health expenditure. First, we regressed out variation in the outcome variables resulting from probability of mortality due to war and natural disasters, urbanicity, per person development assistance for health, GDP per capita, and year and country fixed effects. Then we ran the random forest algorithm using twenty-three subindicators underlying the democratic quality composite index (available in appendix part 1.1)15 to predict variation in the residuals from the first step. Appendix part 2.415 displays a measure of the improvement in the fit of the model with and without each indicator of democratic quality, with a larger number indicating better fit.
This study had four main limitations. First, democratic quality and UHC effective coverage are both estimated variables, which may contain measurement error if, for example, experts use different democracy rating thresholds or country databases report deaths in a systematically biased manner. However, the compilers of the Varieties of Democracy data use a Bayesian item response modeling technique to control for these potential biases. Similarly, the UHC effective coverage estimates benefit from a collaborative network that includes more than 5,500 collaborators in 152 countries.
Second, there is a complicated relationship between democracy level, health financing, and UHC effective coverage. We do not make causal claims because of concerns about the association of democracy with unobserved factors and reverse causality.
Third, the synthetic control analyses used to test the distinct effects that economic recessions have on the health systems of autocracies and democracies compared severe economic recession against all countries without such severe recessions. Because the true effect of economic recession is likely continuous, the estimates presented in this analysis are likely conservative and underestimate the true differences between the two types of countries.
Fourth, these analyses were completed based on data spanning 1990–2019 and therefore reflect trends and relationships observed globally during this period. The current global recession may affect universal health coverage and government health spending differently, especially because the catalyst for the economic downturn—the COVID-19 pandemic—is health related.
Of the 170 countries included in this study, 73 were classified as democratic and the remainder as autocratic.16 Globally, the median democratic quality score increased from 0.04 in 1990 to 0.31 in 2019. In democratic nations, UHC effective coverage increased from 0.51 to 0.67, whereas in autocratic countries, it rose from 0.40 to 0.55. Per person government health spending increased from $326 to $601 in democratic countries, whereas it increased from $77 to $192 in autocratic countries (appendix part 1.2).15 Descriptive statistics on the full sample and World Bank income subgroups are in appendix part 1.2.15
Regression analysis shows that both GDP per capita and democratic quality have statistically significant and positive relationships with UHC effective coverage (both ). Exhibit 1 shows that GDP per capita modifies the association between democratic quality and UHC effective coverage, with the largest association being for low-income countries (defined as GDP ≤$1,035). The average low-income country doubled its democratic quality from 0.04 in 1990 to 0.08 in 2019 (appendix part 1.2);15 we estimate that this increase was associated with a universal health coverage increase of 2.1 percent (95% confidence interval: 1.1, 3.2), holding constant country-specific time-invariant factors, yearly shocks, probability of mortality due to war and terrorism, urbanization, and per person development assistance for health. For wealthier countries, with GDP per capita greater than $9,712 (95% CI: 2,043, 17,688), the association between democratic quality and universal health coverage was not statistically significant.
We used this regression to compare the predicted number of people who would have lost universal health coverage in 2020 under different political scenarios. Our estimates suggest that if the trends and relationships observed during the period 1990–2019 held for the 2020 global recession, an additional 40.1 million (95% CI: 40.0, 40.3) people would have lost universal health coverage because of the 2020 economic downturn if countries completely lacked democratic quality compared with whether they had full democratic quality. This number reflects the predicted difference in the number of people who lost universal health coverage this past year if all countries were democracies versus autocracies, according to our panel regression.
Greater democratic quality was also associated with an increase in government health spending in low-income countries. Exhibit 2 shows that, as in exhibit 1, when a country’s democratic quality doubled (appendix part 1.2), this was associated with an increase in per person government spending of 7.1 percent (95% CI: 3.9, 10.5) for low-income countries and 1.7 percent (95% CI: 0.7, 2.9) for middle-income countries as defined by the World Bank.
Exhibits 3 and 4 show that, on average, economic recessions were associated with reduced UHC effective coverage in autocracies and increased coverage in democracies. Autocracies that experienced an economic recession increased universal health coverage by 1.95 percent (95% CI: 0.18, 1.96; ) less, on average, after five years than we estimate would have occurred in the absence of economic recession. In contrast, democracies that experienced economic recession increased universal health coverage by 2.08 percent (95% CI: 1.68, 3.13; ) more, on average, after five years than we estimate would have occurred in the absence of those shocks (data not shown). Taken together, these findings indicate that democracies outperform autocracies by 4.03 percentage points during severe recessions.
We found that among all of the characteristics of democratic quality, free and fair elections were the most important political predictors of higher universal health coverage, controlling for all covariates (see appendix part 2.4).15 Increased per person government spending was associated with electoral monitoring board capacity and autonomy and reduced electoral violence (see appendix part 2.4).15 The least predictive variables were measures of how government officials are elected and the percentage of the total population with suffrage.
Our research indicates that democratic quality contributes to universal health coverage progress, with significant positive associations for effective coverage and government health spending, especially in low-income countries. The importance of democratic quality for universal health coverage growth and government spending on health appears to increase during economic downturns. Our results show a statistically significant difference between how democracies and autocracies respond to economic recession, with autocracies making less progress toward universal health coverage than democracies.
To put this result into context, these estimates suggest that an additional 40.1 million people would have lost universal health coverage because of the 2020 economic downturn if the world were composed of autocratic governments and if governments responded to recessions in 2020 as they did during the period 1990–2019. Although this is a counterfactual exercise, the Varieties of Democracy project estimates that almost 35 percent of the world’s population resides in an autocratizing country, underscoring the potential negative health effects of future recessions under a different composition of global regimes.27 Furthermore, autocratization has been shown to be associated with slowed progress toward universal health coverage.28
Our finding that political commitment to sustaining universal health coverage in a recession may vary by regime type is consistent with previous research. A previous panel study for seventeen Latin American countries between 1980 and 1993 concluded that relatively little change occurs in social spending for democratic nations as economic growth declines but that authoritarian regimes severely curtail spending, widening the health differences between regime type with deteriorating economic performance.14 It is possible that a pandemic-induced recession may have different effects on health spending than a recession caused by other, nonhealth reasons. Yet preliminary evidence has shown decreased health spending during the COVID-19 pandemic due to reduced expenditure on other health priorities such as cancers.29
Our results indicate that having free, fair, and competitive elections is important for universal health coverage and government health spending, most likely by forcing political leaders to answer to a broad set of citizens, at regular intervals, regarding their expansion. Universal health coverage is generally targeted less by development assistance for health, which focuses on infectious diseases and maternal and child health.19 Without support from foreign aid agencies or pressure from voters, autocracies might have less incentive to maintain investment in universal health coverage during economic slowdowns, such as the global recession caused by the COVID-19 pandemic.14 This result is consistent with previous studies by John Gerring and colleagues, which demonstrated the importance of free and fair elections for population health.3,30 Our random forest analysis linked both electoral board capacity and multiparty elections to universal health coverage progress, suggesting that these mechanisms allow supporters of the opposition to receive universal health coverage in electoral democracies simply because political leaders are more likely to need their electoral support. As such, electoral board capacity and multiparty elections are two areas in which democratic reform could have a positive effect on the global push toward universal health coverage.
There is widespread agreement that progress on universal health coverage will need to be country led and that success will depend on political will. Our results indicate that gaps in universal health coverage are most likely to occur in autocracies during economic recessions when these autocratic nations should be the target of advocacy and monitoring efforts.31 The results of this study suggest that democracy offers one way to improve political will, accountability, and transparency and to sustain country-led progress on universal health coverage. It also suggests that more than just politics may be at stake in the potential shift toward increasing autocratization that has emerged in this pandemic.32
In light of these findings, one option for proceeding is to direct more of the scarce development assistance that exists for universal health coverage promotion and implementation to the nations that have shown a commitment to building accountable institutions and open and transparent democratic processes. The Millennium Challenge Corporation in the US and the Swedish International Development Cooperation Agency have used similar selective approaches to aid to promote good governance. The Millennium Challenge Corporation, for example, uses established, transparent criteria to determine which countries are eligible and then makes large investments toward their nationally determined priorities.33 Countries that are not eligible should be offered targeted assistance to redress those shortcomings.
With high COVID-19 mortality rates in the US, UK, and some European nations, the argument that democracy is good for one’s health has never seemed weaker.34 Many in global health may see authoritarian government as a more effective option for pandemic response. This pessimism about democracy is understandable, but it incorrectly assumes that characteristics that enable governments to contain the COVID-19 pandemic—acting quickly; securing local implementation of testing and other national public health policies; and gaining public compliance with contact tracing, quarantine, and mask mandates—are the same as those required to sustain population health generally. Especially during economic downturns, the continued commitment of policy makers and politicians to funding universal access to affordable, effective health services matters greatly. Democratic quality is an important factor for reaching goals associated with universal health coverage, and democracies are more resilient to the deleterious effects that economic recessions have on health systems. In the current global recession spurred by the COVID-19 pandemic, our results suggest that the importance of democratic quality for preserving progress toward universal health coverage will only increase. Global institutions and development assistance providers may need to adjust their policy recommendations and activities so as to extract the best possible results in countries where political leaders lack sufficient electoral incentives to provide high-quality health care to the population.
Grant funding was provided to Joseph Dieleman and John Mumford by the Bill & Melinda Gates Foundation and to Samantha Kiernan and Thomas Bollyky by Bloomberg Philanthropies. This is an open access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt, and build upon this work, for commercial use, provided the original work is properly cited. See https://creativecommons.org/licenses/by/4.0/.
- 1 . Healthcare for all: every country can do it—an essay by Tedros Adhanom Ghebreyesus. BMJ. 2019;367:l6790. Crossref, Medline, Google Scholar
- 2 . Democracy and population health [Internet]. Cambridge: Cambridge University Press; 2020 Jun 14 [cited
2021 Jun 22]. Available from: https://www.cambridge.org/core/elements/abs/democracy-and-population-health/D8B80F5FCA25361CCEC44CD4A3829B43 Google Scholar
- 3 . The relationships between democratic experience, adult health, and cause-specific mortality in 170 countries between 1980 and 2016: an observational analysis. Lancet. 2019;393(10181):1628–40. Crossref, Medline, Google Scholar
- 4 United Nations. Transforming our world: the 2030 agenda for sustainable development [Internet]. New York (NY): United Nations; 2015 [cited 2021 Jun 22]. (Report No. A/RES/70/1). Available from: https://sustainabledevelopment.un.org/post2015/transformingourworld/publication Google Scholar
- 5 World Health Organization. Primary health care on the road to universal health coverage: 2019 Global Monitoring Report executive summary [Internet]. Geneva: WHO; 2019 [cited
2021 Jun 22]. Available from: https://www.who.int/publications/i/item/WHO-HIS-HGF-19.1 Google Scholar
- 6 . The untold toll—the pandemic’s effects on patients without Covid-19. N Engl J Med. 2020;382(24):2368–71. Crossref, Medline, Google Scholar
- 7 . COVID-19 and racial/ethnic disparities. JAMA. 2020;323(24):2466–7. Crossref, Medline, Google Scholar
- 8 International Monetary Fund. World economic outlook update: policy support and vaccines expected to lift activity [Internet]. Washington (DC): IMF; 2021 Jan [cited
2021 Jun 22]. Available from: https://www.imf.org/en/Publications/WEO/Issues/2021/01/26/2021-world-economic-outlook-update Google Scholar
- 9 Effects of health and social care spending constraints on mortality in England: a time trend analysis. BMJ Open. 2017;7(11):e017722. Crossref, Medline, Google Scholar
- 10 Financial crisis, austerity, and health in Europe. Lancet. 2013;381(9874):1323–31. Crossref, Medline, Google Scholar
- 11 . Health and democracy. Am Econ Rev. 2006;96(2):313–8. Crossref, Medline, Google Scholar
- 12 . Development as freedom. New York (NY): Anchor Books; 2000. Google Scholar
- 13 . The invisible hand of democracy: political control and the provision of public services. Comp Polit Stud. 2001;34(6):587–621. Crossref, Google Scholar
- 14 . Democracy and social spending in Latin America, 1980–92. Am Polit Sci Rev. 1999;93(4):779–90. Crossref, Google Scholar
- 15 To access the appendix, click on the Details tab of the article online.
- 16 V-Dem dataset—version 10 [Internet]. Gothenburg: V-Dem Institute, Varieties of Democracy (V-Dem) Project; 2020 [cited
2021 Jun 22]. Available from: https://www.v-dem.net/en/data/archive/previous-data/v-dem-dataset/ Google Scholar
- 17 Measuring universal health coverage based on an index of effective coverage of health services in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet. 2020;396(10258):1250–84. Crossref, Medline, Google Scholar
- 18 Bill & Melinda Gates Foundation. COVID-19: a global perspective [Internet]. Seattle (WA): Bill & Melinda Gates Foundation; 2020 Sep [cited
2021 Jun 22]. Available from: https://www.gatesfoundation.org/goalkeepers/report/2020-report/ Google Scholar
- 19 Health sector spending and spending on HIV/AIDS, tuberculosis, and malaria, and development assistance for health: progress towards Sustainable Development Goal 3. Lancet. 2020;396(10252):693–724. Crossref, Medline, Google Scholar
- 20 . Developing a comprehensive time series of GDP per capita for 210 countries from 1950 to 2015. Popul Health Metr. 2012;10(1):12. Crossref, Medline, Google Scholar
- 21 Global Burden of Disease Collaborator Network. Global Burden of Disease Study 2019 (GBD 2019) Covariates 1980–2019 [Internet]. Seattle (WA): Institute for Health Metrics and Evaluation; 2020 [cited
2021 Jun 22]. Available from: http://ghdx.healthdata.org/record/global-burden-disease-study-2019-gbd-2019-covariates-1980-2019 Google Scholar
- 22 EM-DAT [home page on the Internet]. Brussels: Université catholique de Louvain, Centre for Research on the Epidemiology of Disasters; 2020 [cited
2021 Jun 22]. Available from: https://public.emdat.be/ Google Scholar
- 23 World Bank. World development indicators [Internet]. Washington (DC): World Bank; [last updated 2021 Jun 30; cited
2021 Jul 1]. Available from: https://datacatalog.worldbank.org/dataset/world-development-indicators Google Scholar
- 24 . How much should we trust estimates from multiplicative interaction models? Simple tools to improve empirical practice. Polit Anal. 2019;27(2):163–92. Crossref, Google Scholar
- 25 . Generalized synthetic control method: causal inference with interactive fixed effects models. Polit Anal. 2017;25(1):57–76. Crossref, Google Scholar
- 26 . Classification and regression by randomForest. R News. 2002;2/3:18–22. Google Scholar
- 27 . State of the world 2019: autocratization surges—resistance grows. Democratization. 2020;27(6):909–27. Crossref, Google Scholar
- 28 . Autocratisation and universal health coverage: synthetic control study. BMJ. 2020;371:m4040. Crossref, Medline, Google Scholar
- 29 The Economist Intelligence Unit. Covid-19: the impact on healthcare expenditure. The Economist [serial on the Internet]. 2020 May 27 [cited
2021 Jun 22]. Available from: https://www.eiu.com/n/campaigns/covid-19-the-impact-on-healthcare-expenditure/ Google Scholar
- 30 . Democracy and human development: issues of conceptualization and measurement. Democratization. 2020;28(2):308–32. Crossref, Google Scholar
- 31 . Democracy’s backsliding in the international environment. Science. 2020;369(6508):1192–6. Crossref, Medline, Google Scholar
- 32 . Autocratization by decree: states of emergency and democratic decline [Internet]. Gothenburg: V-Dem Institute, University of Gothenburg; 2020 Apr [cited
2021 Jun 22]. (Working Paper No. 85). Available from: https://www.v-dem.net/media/filer_public/31/1d/311d5d45-8747-45a4-b46f-37aa7ad8a7e8/wp_85.pdf Google Scholar
- 33 Millennium Challenge Corporation. Report on the criteria and methodology for determining the eligibility of candidate countries for Millennium Challenge Account assistance for fiscal year 2021 [Internet]. Washington (DC): Millennium Challenge Corporation; 2020 Sep 15 [cited
2021 Jun 22]. Available from: https://www.mcc.gov/resources/doc-pdf/report-selection-criteria-methodology-fy21 Google Scholar
- 34 . Preparing democracies for pandemics. BMJ. 2020;371:m4088. Crossref, Medline, Google Scholar