Sociopolitical context and COVID-19 fatality rates in OECD countries: a configurational approach

Abstract

Background

The effectiveness of crisis response can be influenced by various structural, cultural, and functional aspects within a social system. This study uses a configurational approach to identify combinations of sociopolitical conditions that lead to a high case fatality rate (CFR) of COVID-19 in OECD countries.

Methods

A Fuzzy set qualitative comparative analysis (QCA) is conducted on a sample of 38 OECD countries. The outcome to be explained is high COVID-19 CFR. The five potentially causal conditions are level of democracy, state capacity, trust in government, health expenditure per capita, and the median age of population. A comprehensive QCA robustness test protocol is applied, which includes sensitivity ranges, fit-oriented robustness, and case-oriented robustness tests.

Results

None of the causal conditions in both the presence and negation form were found to be necessary for high or low levels of COVID-19 CFR. Two different combinations of sociopolitical conditions were usually sufficient for the occurrence of a high CFR of COVID-19 in OECD countries. Low state capacity and low trust in government are part of both recipes. The entire solution formula covers 84 percent of the outcome.  Some countries have been identified as contradictory cases. The explanations for their COVID-19 CFR require more in-depth case studies.

Conclusions

From a governance perspective, the weakness of government in effectively implementing policies, and the citizens’ lack of confidence in their government, combined with other structural conditions, serve as barriers to mounting an effective response to COVID-19. These findings can support the idea that the effects of social determinants of COVID-19 outcomes are interconnected and reinforcing.

Introduction

In response to the COVID-19 pandemic, governments worldwide implemented various prevention and containment measures. However, there is significant variation in the severity of the pandemic between countries [1, 2]. Several studies have attempted to explain why some countries respond to COVID-19 more effectively than others [3,4,5,6,7,8]. Previous research has emphasized the importance of sociopolitical context, as well as economic, demographic, and environmental factors, in determining the success or failure of responses to the pandemic [8,9,10,11,12,13].

During the early stages of the COVID-19 outbreak, when Europe faced a greater disease burden in terms of morbidity and mortality compared to China [2, 14], debates arose regarding which political system performed better in responding to the pandemic [14, 15]. Francis Fukuyama proposed a framework for understanding the differences in the effectiveness of pandemic management from a political perspective. He argued that the effectiveness of crisis response cannot be solely attributed to the type of regime. Both autocracies and democracies can exhibit different levels of performance when dealing with a crisis. Instead, he highlighted the importance of the state’s capacity and the level of trust citizens have in their government as crucial factors that determine the effectiveness of crisis management [14].

A synthesis of evidence from initial evaluations in OECD countries shows that governments have faced significant challenges in responding to the COVID-19 pandemic. These evaluations have emphasized the importance of state capacity and the ability of governments to involve civil society, the private sector, and local actors in addressing the crisis. This inclusiveness is crucial because it enhances transparency in decision-making processes, fosters trust in government actions, and facilitates the effective implementation of crisis management policies [16].

Previous studies have shown that low trust in the government can lead to lower compliance with COVID-19 prevention guidelines, which in turn affects the effectiveness of government measures [8, 10, 17,18,19,20]. Citizens are more likely to adhere to preventive measures, when they perceive the government as well organized [17, 21]. However, the influence of sociopolitical factors on crisis response may vary depending on the context. An analysis of national policy responses to COVID-19 in Europe during the first wave of the pandemic suggests that countries with greater centralization, lower government effectiveness, less freedom, and lower societal trust but with separate ministries of health and health ministers with medical background, took quicker and more decisive actions. It has been argued that governments with high perceived capacity might initially respond passively, falsely believing they can manage the pandemic at a later stage [8].

Despite several multivariate analyses that have been conducted to assess the effects of sociopolitical factors on COVID-19 outcomes, there is limited research exploring how these factors may combine to produce specific outcomes. Our study aims to contribute to identifying these combinations that lead to high COVID-19 mortality by using a set-analytic approach. While multivariate statistical techniques are powerful and rigorous tools for estimating the “net effects” of independent variables on outcomes, set-theoretic analysis offers an alternative approach [22]. From this perspective, explanatory factors do not compete with each other; instead, they combine in various ways to produce an outcome. By using set-analytic techniques, researchers can identify multiple paths leading to the same outcome [22, 23]. It is important to note that interaction effects in quantitative analysis do not provide the same insights as the assessment of combined effects and recipes. While interaction terms test for non-additive effects, set-theoretic analysis directly evaluates the complex combinations of conditions consistently linked to an outcome [23]. Our analysis of the structural factors that contribute to the high CFR of COVID-19 is based on the WHO conceptual framework of social determinants of health (SDH), which highlights the idea that “complexity defines health” [24]. By studying cases as different configurations of causally relevant conditions, we can identify the multiple recipes for the outcome and thereby unravel causal complexity [22].

Our study focused on OECD countries. The COVID-19 pandemic has posed significant challenges for these governments, impacting the health and well-being of their citizens, as well as the economy. It has led to structural and social issues, such as the erosion of public trust in the government [16]. Despite allocating considerable resources to manage and mitigate the crisis, the outcomes of COVID-19 varied substantially across OECD countries [2, 16]. To address these emerging challenges, the OECD compiled data and analysis on various topics related to the health, economic, and societal crisis. Building on this data, we aim to conduct a cross-national comparative analysis using the fuzzy-set QCA. This method allows for a configurational analysis of a small-to-medium N cases or macro-level data by evaluating set-theoretic relations that are inherently asymmetric. The use of set-theoretic methods allows us to utilize the concepts or language of necessity and sufficiency [22].

Given that causal language should be used with caution and QCA identifies configurations that are consistent with ‘usually’ being sufficient to cause outcomes [25, 26], our research question is as follows:

What combinations of sociopolitical conditions are usually sufficient to cause a high COVID-19 fatality rate in OECD countries?

To address this question, we conducted a fuzzy-set QCA on a sample of 38 OECD countries. The outcome we aim to explain is a high COVID-19 case fatality rate (CFR). The five potential causal conditions we considered, based on the social determinants of health conceptual framework and a literature review, are the level of democracy, state capacity, trust in government, health expenditure per capita, and the median age of the population. We have identified two distinct combinations of conditions that are sufficient for the occurrence of a high CFR due to COVID-19 in OECD countries.

The following section provides a summary of the relevant background and literature.

Literature review

In this section, we will first present the SDH conceptual framework that guides our research. Then, we will review previous studies to clarify how our research can contribute to the existing knowledge of the structural determinants of COVID-19 outcomes.

Health outcomes can be influenced by a wide range of interconnected factors at multiple levels, including biological, ecological, social, and environmental factors. The WHO commission on social determinants of health (CSDH) has developed a conceptual framework for action that illustrates how these factors interact with and influence each other in complex ways, shaping health outcomes and ultimately impacting health equity [24].

The first element of CSDH conceptual framework is the socio-economic and political context. The context broadly includes structural, cultural, and functional aspects of a social system. It highlights how socio-economic and political contexts create social stratification in societies and assign individuals to different social positions. This, at the individual level, leads to differential exposure to health-damaging conditions and varying vulnerability in terms of health conditions and access to material resources [24, 27]. Another component of the CSDH framework is social capital. There are three types of social capital: bonding, bridging, and linking social capital. Bonding social capital refers to the trust and cooperative relationships among members of a network who share a similar social identity. On the other hand, bridging social capital refers to respectful relationships and mutual understanding between individuals and groups who acknowledge their differences in socio-demographic terms. Finally, linking social capital refers to “norms of respect and networks of trusting relationships between people who are interacting across explicit, formal or institutionalised power or authority gradients in society” [24, 28]. Linking social capital is essential for understanding the complex interplay between social determinants and health outcomes. The notion of linking social capital highlights the role of the state in promoting health equity, so that a key task for health policies is to foster collaborative relationships between citizens and institutions [28]. In the context of COVID-19, it has been argued that linking social capital can affect the government’s ability to effectively control the pandemic. Trust in the government can influence the dissemination of public health and healthcare messages about risk reduction, which can potentially impact adherence to non-pharmacological interventions (NPIs) [17, 29].

The CSDH framework conceptualizes the health system itself as a social determinant of health. The role of the health system is particularly important when it comes to the issue of access, which demonstrates how the capacity of the health sector can impact differences in exposures, vulnerability, and the consequences of illness for people’s health and their social and economic circumstances [24].

Overall, the CSDH conceptual framework emphasizes the configurational nature of health, demonstrating how the socio-political and economic context of a society interact with and influence each other in complex ways, shaping individuals’ behavior and impact their health status [24, 30, 31].

Considering the complexity of the COVID-19 pandemic and the sociopolitical factors that may influence its outcomes [9, 16, 32], we have formulated our research question based on the CSDH conceptual framework and the existing evidence on the structural determinants of COVID-19 outcomes, which is briefly reviewed below.

At the beginning of the COVID-19 pandemic, debates emerged about which political system is more effective in handling the crisis [6, 17, 33,34,35]. A number of researchers have investigated the relationship between democracy and government responses to COVID-19. After controlling for political, institutional, economic, and demographic factors, these studies revealed that non-free countries were more likely to have lower levels of deaths compared to free countries. Some of them suggest that a conditional effect should be considered when interpreting these findings [13, 33, 36,37,38]. On the other hand, an assessment of governments policy responses across more than 130 countries indicated that although autocratic regimes have implemented more stringent policy measures compared to democratic governments, however, they have been less effective in implementing these measures [15]. Several studies have also found that democracy is associated with better COVID-19 outcomes [12, 15, 39]. In this regard, Francis Fukuyama stated, “When the pandemic subsides, I suspect that we will have to discard simple dichotomies. The major dividing line in effective crisis response will not place autocracies on one side and democracies on the other.  Rather, there will be some high-performing autocracies, and some with disastrous outcomes. There will be a similar, though likely smaller, variance in outcomes among democracies. The crucial determinant in performance will not be the type of regime, but the state’s capacity and, above all, trust in government.” [14]. Numerous researchers have also highlighted the role of state capacity and public trust in response to COVID-19, while taking into account country characteristics such as the level of democracy, age, population density, health system resource capacity, and income level [6, 17, 33,34,35]. A comparative case study analysis of different responses to COVID-19 in China and South Korea proposed four types of state capacity, including information capacity, decision-making and implementation capacity, coercive capacity, and mobilization and cooperation capacity. The study suggests that the combinations of state capacity vary between democracies and authoritarian regimes. The research has implications for managing the COVID-19 pandemic under different political systems and concludes that both democracies and authoritarian regimes can employ different state capacities to effectively respond to crises [40].

Overall, what we know about the effectiveness of a country’s response to the pandemic is that various factors beyond the political system can influence it, including state capacity (e.g., government effectiveness), public trust, and health systems governance and financing [6, 8, 32, 36, 41]. A study of 185 countries revealed that those with better governance had more successful control over the pandemic. These governments can foster public trust, which in turn result in higher levels of public cooperation and compliance behavior [35]. In the initial wave of the pandemic, a study examining why COVID-19 has resulted in more deaths in certain European regions compared to others revealed that the extent to which expert recommendations to control the spread of the virus have been endorsed and followed varies among governments and societies. It has been argued that low trust and political polarization play a significant role in these regional differences, along with factors such as age, density, accessibility, and the preparation of the health system. These factors have had an impact on contagion and, specifically, on COVID-19-related deaths [17]. There is evidence that trust-building has been a strategy and core component of governance and policy responses to the COVID-19 crisis within the OECD [16]. Studies suggest that political trust plays a crucial role in explaining effective pandemic response [10, 16]. Conversely, trust can affect people’s perception of risk and their behavior. When individuals trust others, they are less likely to view them as a potential health threat. Furthermore, if people trust the government’s competency in handling the pandemic, their perception of pandemic-related risks may decrease. This, in turn, may lead to individuals perceiving personal protective measures as less important, increasing the likelihood of more widespread infections [42, 43].

Several comparative studies using a configurational approach have been conducted to understand how different social, political, and economic factors in various contexts may interact and associate with COVID-19 outcomes [3, 7, 44,45,46].

An fuzzy-set QCA analysis of 80 countries yielded eight combinations of five factors, including a delayed public-health response, past epidemic experience, proportion of elderly in population, population density, and national income per capita linked to either high or low rates of early years of life lost (YLL) [3]. A fuzzy-set qualitative comparative analysis of COVID-19 community resilience in 16 countries indicated different combinations of robustness, physical and social preparedness, resources and social capital, and government response linked to a good recovery from COVID-19 [45]. Another QCA study investigated the contextual and government response factors to the first-wave of the COVID-19 pandemic in 25 OECD countries. The potential causal conditions were obesity rates, proportions of elderly people, inequality, country travel openness and COVID-19 testing regimes. Results showed three combinations of conditions linked to low COVID-19 mortality [7].

Overall, in our review of the literature, we have identified several attempts to evaluate the effects of socioeconomic and political factors on COVID-19 outcomes. However, only a few researchers have employed a configurational approach, which involves examining the cases as configurations of conditions and focusing on the combined effects of the structural conditions. As mentioned in the previous section, it seems that the dominant approach in cross-national studies in the context of COVID-19 is the “variable-oriented” approach, in which researchers seek to assess the “net effects” of independent variables on the outcomes, not their multiple combined effects [22]. However, it may be useful to conduct a study that is more capable of addressing effects that are theoretically likely to be configurational, as demonstrated by the CSDH conceptual framework [24, 32]. Therefore, the aim of this study is to explain high CFR due to COVID-19 among OECD countries by utilizing a configurational approach, which will be described in the following section.

Methods

Study design

We conducted a fuzzy-set qualitative comparative analysis (QCA) on a sample of 38 OECD countries. The fuzzy-set QCA builds on a combination of the fuzzy-set theory as a mathematical system for addressing the degree of membership in sets and Boolean and fuzzy algebra to determine which combinations of conditions are linked to a specific outcome [22]. The Fuzzy-set QCA method uses specific terminology. Instead of “independent variable,” the term “condition” is used. The term “outcome” is used in place of “dependent variable” to refer to the phenomenon being explained. The results of a QCA are referred to as “solution formulas” or “solution terms” rather than “equations” [47].

When applying QCA, the key issue is not which variable is the best predictor of the outcome (i.e., having the biggest net effect), but how different conditions combine and generate the same outcome. This idea is implemented in a truth table, which examines all logically possible combinations of causal conditions [22].

Two measures are used to evaluate the set-theoretic relations and interpret the results of the fuzzy-set QCA: consistency and coverage. Set-theoretic consistency measures the extent to which cases that share a particular combination of conditions (e.g., socioeconomic and political context) agree in displaying the outcome in question (e.g., high fatality rate). On the other hand, set-theoretic coverage assesses the degree to which a cause or causal combination accounts for instances of an outcome. When there are multiple paths leading to the same outcome, the coverage of a specific causal combination may be small. Therefore, coverage measures the empirical relevance or importance of a solution term [22].

Data collection

Data were collected from different international data sources as follows.

Outcome: high case fatality rate (CFR) of COVID-19

To measure the effectiveness of government responses to the pandemic, the case fatality rate (CFR) in each country that is the proportion of individuals diagnosed with a disease who die from that disease was used. According to the World Health Organization (WHO), it is an indicator of severity among detected cases [1].

Data on CFR were retrieved from Our World in Data (OWID) on June 30, 2022 that provides daily updates of the data on coronavirus pandemic worldwide [2].

Conditions

Democracy

Data on democracy were extracted from the democracy indices dataset via Gapminder for the year 2020 [48]. The index is a self-described measure of the state of democracy in 165 countries and territories that prepared by the Economist Intelligence Unit (EIU). The countries have been categorised into one of four regime types: full democracies, flawed democracies, hybrid regimes and authoritarian regimes [49].

State capacity

The government effectiveness index with estimates ranging between − 2.5 (weak) and + 2.5 (strong), is commonly used as a measure of state capacity [50]. It is constructed from 43 different indicators, which assess the perception of the quality of public services, the quality of the civil service and its independence from political pressures, quality of policy formulation and implementation, and the credibility of the government’s commitment to such policies [51, 52].

For each country, the average score for government effectiveness estimates was calculated for the years 2020 to 2021.

Trust in government

The data on trust in government was obtained from the OECD dataset. This indicator measeures the percentage of people who report having confidence in their national government [53]. To calculate the average score for trust in government, data from the years 2020 and 2021 was used for each country.

Population median age

Data on the median age of the population are available from the Our World in Data (OWID) based on World Population Prospects (UN Population Division) for 2020 [2].

Health expenditure per capita

Data for health expenditure per capita were retrieved from the OECD dataset on Health spending (USD per capita) for 2020.¹ Health spending measures the final consumption of health care goods and services (i.e. current health expenditure) including personal health care and collective services, but excluding spending on investments [54].

Calibration

The calibration of the degree of membership in fuzzy sets is central to fuzzy-set QCA as a set-analytic method. The first step in fuzzy-set QCA is to conceptualize and label the sets. Then, fuzzy sets need to be calibrated, which means the raw scores should be transformed into fuzzy membership scores [22]. The sets calibration process is discussed in more detail below.

To construct the fuzzy-set measures of countries with high levels of democracy, this analysis uses the EIU democracy index threshold adjusted for four types of regimes based on their democracy score (i.e., full democracies, flawed democracies, hybrid regimes, and authoritarian regimes). A six-value fuzzy set was used (see Appendix S1 and Table 1) [22].

Table 1 Degree of membership in the set of fully democratic countries

To construct the other fuzzy sets, the literature does not provide clear categories to identify three qualitative thresholds. Thus, the calibration of fuzzy sets was grounded in the structure of the data and case knowledge. The full membership threshold was set near the 80th percentile, the full non-membership threshold was set near the 20th percentile, and the cross-over point was set at about the 50th percentile on a value with no observation [22]. The selected calibration thresholds are shown in Table 2.

Table 2 Fuzzy set calibration thresholds

Data analysis

After calibrating the fuzzy sets, four analysis steps were followed; (1) testing for necessary conditions, (2) constructing the truth table and addressing the contradictions, (3) analyzing sufficiency, and (4) repeating the analysis for the negation of the outcome [22, 47].

The analysis was carried out using fsQCA 3.0 software [55] and the R package SetMethods [56].

Results

This section examines the necessary and sufficient relations between the five sociopolitical conditions and the high COVID-19 case fatality rate. Then, the robustness test of the results is conducted.

To increase transparency and replicability of the analysis [47] the raw data and fuzzy data matrix are available in Appendices S2 and S3.

Table 3 shows a brief description of the raw data.

Table 3 Description of raw data for outcome and conditions

Analysis of necessity

The data analysis started with the necessary conditions. The analysis of the necessity for the presence of the outcome and for its negation showed that none of the causal conditions in both the presence and negation forms exceeded the consistency threshold of 0.9 [57, 58]. Thus, no individual conditions were considered necessary for high or low levels of COVID-19 fatality rate.

Analysis of sufficiency

Using fuzzy-set QCA, we can identify configurations that are consistent with being usually sufficient to cause outcomes through truth table analysis. Each row of the truth table represents a specific combination of conditions (or configuration in QCA terminology). The first step in the construction and analysis of a truth table is to specify the frequency threshold. Given that the total number of cases in this analysis is relatively small, the combination with at least one country with a membership greater than 0.5 was retained for sufficiency analysis. To evaluate which of these combinations are subsets of - in other words, sufficient for - the outcome, the next step is selecting a consistency threshold. Following Ragin’s recommendations, the consistency threshold value was set at 0.75 [22, 58].

Table 4 shows the distribution of cases across the 13 configurations with a frequency threshold of at least one case.

Table 4 Truth table for high COVID-19 CFR

Truth table analysis provides three solutions for the outcome: the complex solution, the parsimonious solution, and the intermediate solution (i.e., a subset of the most parsimonious solution and a superset of the most complex solution). As recommended, we focus on the intermediate solution, which strikes a balance between complexity and parsimony [22]. The formulas for all three solutions for the outcome and negation of the outcome are presented in Appendix S4.

Table 5 shows the intermediate solution for the presence of high CFR of COVID-19, after excluding contradictory cases. This solution was obtained by including two assumptions in line with literature and expert opinion, namely that state capacity and trust in government are key factors in effective pandemic response [14, 22].

Table 5 Intermediate Solution for the presence of high CFR of COVID-19

Table 5 shows that the final solution consists of two different configurations that are sufficient for high CFR, with a consistency score of 0.87, which covers 84% of the outcome set. It can be expressed as a linear formula (the “+” sign represents logical OR):

(Low State capacity * Low Trust * Low Health expenditure) + (Democracy * Low State capacity * Low Trust * High Age)

This solution can cover, or explain, 84% of OECD countries with a high CFR of COVID-19. The following discussion explores these findings in detail by applying the truth table solutions to actual cases or groups of similar cases.

Sensitivity and robustness tests

The fuzzy-set QCA is sensitive to researcher decisions on calibration, consistency cutoff, and frequency of cases threshold, thus several different robustness tests against such choice should be examined [47, 58].

To calculate the sensitivity ranges within which causal recipes for the initial solution (IS) stays the same, some functions in the R package SetMethods were used: rob.inclrange(), rob.ncutrange(),rob.calibrange, and rob.fit() function [59].

Table 6 shows the summary of sensitivity analysis to evaluate the robustness of the results. According to the sensitivity ranges for (1) thresholds of calibration, (2) raw consistency and (3) frequency cutoff, the initial solution might be sensitive to changes in calibration anchors especially for state capacity and trust in government. It is also sensitive to changes in the raw consistency threshold. The IS stays the same only if the raw consistency threshold is specified within the 0.72–0.79 range. Therefore, a plausible alternative solution could be obtained if we modify the raw consistency threshold to 0.8. When it comes to the frequency cutoff, Table 6 shows that the initial solution is sensitive to changes. However, with 38 cases, it is not justified to place the frequency threshold at, for example, 2. Following the protocol for testing robustness, parameters should be tested as different as possible, but within substantively plausible ranges [59]. In other words, even if the frequency threshold is sensitive to a change to 2, it is not substantively plausible to place the frequency cutoff there.

Overall, for evaluating the robustness of the results, three plausible alternative solutions were created that change the consistency threshold and calibration anchors for state capacity and trust in government out of the determined sensitivity ranges.

Table 6 Robustness protocol report

Using the rob.fit() function, we obtained the four robustness parameters. The RF_cov, RF_cons, RF_SC_minTS, and RF_SC_maxTS are all close to 1. Therefore, the IS demonstrates a high level of robustness to the changes that were tested [59]. Figure 1 shows that a considerable number of cases are located on the diagonal, thus provides a graphical support for the findings of the robustness fit parameters.

The RCR_typ shows that all potential typical cases (countries that are both in line with the statement of sufficiency and good empirical instances of the outcome and conjunction of conditions) are members of both IS and minTS, thus robust.

The RCR_dev indicates that all deviant consistency cases are shaky. Shaky deviant cases are those cases that, when making plausible alternative analytic decision, are no longer deviant, but rather become individually irrelevant. Therefore, while these cases are not robust, they still confidence in the initial solution as if alternative decisions are being made, the solution actually becomes less prone to deviant cases [59].

Finally, the value of 3 for Rob_Case_Rank indicating that there are not cases that simultaneously fall into both categories ‘shaky’ and ‘possible’ as can be seen in Fig. 1.

Fig. 1

Robustness XY plot

The results of the sensitivity and robustness tests show that our findings are robust against plausible analytic changes regarding the selection of calibration anchors and consistency cut-off.

Discussion

The primary goal of this study was to identify configurations of socio-political conditions that are jointly sufficient for the occurrence of a high case fatality rate due to COVID-19 in OECD countries. The outcome to be explained is high CFR of COVID-19. We used five socio-political characteristics as conditions.

In this section, countries are discussed individually or in groups according to the truth table configurations and WHO regions. The OECD countries were classified according to WHO regions (see Appendix S2). The analysis of the truth table revealed that out of the six OECD countries in the region of the Americas, all four countries in Latin America (Mexico, Chile, Colombia, and Costa Rica) display combinations of conditions that are sufficient to result in a high CFR of COVID-19 (Table 4). These countries, which vary in terms of their levels of democracy, all share common challenges such as low state capacity, a low level of trust in government, and low health expenditure. They all suffer from a high COVID-19 fatality rate. Studies of the COVID-19 response in Latin America have also indicated that, due to contextual and socioeconomic factors such as mistrust of the state and the state’s crisis of legitimacy, high levels of inequality, and a high rate of informal work, the region has faced challenges in confronting the outbreak [60, 61]. The other two countries in the Americas region, Canada and the United States (US), were identified as deviant cases. They have a high COVID-19 CFR, but they do not fit with the solution terms linked to the high CFR (Table 4). In other words, the high CFR of COVID-19 in the US and Canada was not explained by our studied conditions. The explanation of COVID-19 outcomes in Canada and US requires more in-depth case studies. A comparison of health policy responses to the pandemic in Canada, the United States, Ireland, and the United Kingdom (UK) during the first wave highlights differences in the countries’ health system organization and financing, political leadership, and governance structures. For instance, in the US, the lack of universal health coverage has led to obstacles in accessing healthcare. Additionally, resistance against scientific leadership has likely undermined the crisis response. Lockdown measures in Canada, the US, and the UK were either implemented too late or not strict enough. Noncompliance with containment measures was also a significant issue, influenced by pandemic fatigue and concerns about the negative impact of measures such as lockdowns and social distancing on the economy. Canada, the US, and the UK also faced additional challenges in implementing effective pandemic responses, including rapidly scaling up testing capacity, implementing effective test, trace, and isolate systems, ensuring an adequate supply of personal protective equipment and other essential supplies, and creating surge capacity [62]. The high CFRs of Covid-19 in the US, Canada, and UK may be partially attributed to these factors.

European countries experienced different levels of COVID-19 mortality. In the WHO European region, 13 out of 28 OECD countries have made up configurations that are consistent with being sufficient to cause high CFR (Table 4). All, except Belgium, are located in Eastern, Southeastern, and southern Europe (e.g., Poland, Hungary, Czechia, Lithuania, Latvia, Greece, Spain, Italy, Slovakia). Slovenia, Israel, and Turkey were identified as deviant cases. They appear in solution terms linked to the outcome of a high CFR but do not indicate a high CFR. In other words, in certain contexts, combinations of the conditions we studied may contribute to high CFR of COVID-19, while in others, they may not. Further research would be useful to explore how the governments in these countries approached the pandemic. On the other hand, Sweden fits with a solution term connected to the non-occurrence of the outcome, but it meets the outcome of high CFR. Sweden’s COVID-19 strategy has received international attention and criticism due to its higher mortality rates compared to neighboring countries such as Finland, Norway, and Denmark. In this analysis, Sweden and Norway display the same combination of causal conditions. Previous studies indicate that the similarities between Sweden and Norway, including COVID-19 risk factors, socioeconomics, demographics, life expectancy, comorbidity, governmental and administrative systems, healthcare services, and education, make them an interesting case study. In contrast to Norway, Sweden chose not to close its borders, implemented limited virus tracking, and relied on voluntary school closures and remote work. The Swedish COVID-19 Commission prioritized personal freedom by favoring voluntary measures over lockdowns, but it also acknowledged that earlier and stricter actions were necessary during the initial outbreak. This has raised questions about how the different mitigation measures implemented by each country have influenced their respective mortality rates [63, 64].

Our findings are consistent with previous research, indicating that the COVID-19 crisis in Eastern Europe has differed from that in the West. The high CFRs of COVID-19 in these countries, can be partially explained by a combination of weak state capacity, low trust in government, and low health expenditure, according to this analysis. Governments in East Central Europe, including Hungary, Poland, and the Czech Republic, quickly implemented strict lockdown measures at the start of the pandemic. However, they encountered more challenges in handling subsequent waves of the virus. Four major vulnerabilities have been highlighted in East Central Europe during the COVID-19 crisis: the crisis of care, lack of resources for social solidarity, democratic erosion, and dependent capitalism. The region also faced challenges in providing economic support to its populations, especially vulnerable individuals and businesses. Despite their initial achievements, governments in Eastern Europe were less capable of responding effectively to the crisis and mitigating its social consequences. As a result, infection and death rates were higher compared to Western European countries [65, 66].

Among the countries in the Northwestern region, Belgium is the only one that exhibits a configuration associated with the outcome of high CFR. The government of Belgium claims that its higher death rate from COVID-19 is due to its rigorous reporting of deaths. However, there seem to be shortcomings in the availability of personal protective equipment (PPE) and the implementation of testing strategies, as well as pressure on the medical system. It has been argued that Belgium’s initial response to COVID-19 was inadequate, particularly in terms of testing, despite the relative advantages provided by its social welfare systems [7, 66].

In the Western Pacific region, none of the four studied cases (i.e., Australia, Japan, Korea, and New Zealand) was associated with high levels of COVID-19 CFR. These countries with well-resourced health systems, high public trust, and effective government management have been relatively successful in combating COVID-19. The geographical factor has also contributed to their success, especially for Australia, New Zealand, and Japan. Their approach has centered on aggressive suppression and addressing the economic impact. To minimize community transmission, they have implemented strict measures like border closures and lockdowns. Despite the economic challenges, they have provided support for wages and businesses. [67,68,69]. Evidence shows that trust in government increased during the pandemic in Australia and New Zealand, underscoring the importance of government effectiveness. It has been argued that trust in government can be both an outcome and an antecedent of government effectiveness [70]. In South Korea, strong social mobilization and cooperation capacity were also achieved through state-society synergy. South Korea effectively prevented and controlled infections by cooperating with the private sector through a strategy known as the 3T (test, trace, and treat) strategy. The successful practices in South Korea highlight the significance of efficient case identification and management, comprehensive contact tracing, and isolation as potent strategies [40, 69, 70].

While it has been highlighted in the literature that low trust or weak state capacity, for example, can increase the odds of COVID-19 infection or mortality, from a policy perspective it is far more useful to know under what circumstances these factors contribute to COVID-19 outcomes. Configurational analysis allows for answering such questions. Overall, applying fuzzy-set QCA to OECD data yields two causal recipes that are robustly sufficient for the high CFR of COVID-19 (Table 5). Of these two configurations associated with high CFR, the first solution is more comprehensive, explaining more OECD countries. It shows that regardless of the level of democracy, when governments exhibit low effectiveness in policy implementation, face low levels of public trust, and have limited health spending, the overall response to COVID-19 can be compromised, leading to higher CFR. The second causal recipe shows that democratic countries with an older population, lower levels of state capacity, and low trust in government experienced higher COVID-19 CRF. These results are consistent with those of other studies conducted across OECD and European regions, which indicated that effectiveness of national governance, health system capacity, older populations, and levels of trust in government are associated with COVID-19 mortality [5, 71,72,73]. It is interesting to note that both causal recipes share a low level of state capacity combined with low trust in government, resulting in a high COVID-19 CFR. This supports Fukuyama’s (2020) view that the divergence in country performance regarding the pandemic needs to be explained by various factors that interact in complex ways. In this context, state capacity and trust in government are crucial determinants from a governance perspective [14]. The findings of our configurational analysis can also provide evidence in support of the conceptual framework of SDH, which emphasizes that social determinants of health are interconnected and reinforcing [24].

Limitations

This analysis was conducted in the context of OECD countries. The generalizability of any study is reduced if its findings are context-specific. There is also the problem of operationalization. Although relatively well-known indicators of democracy, state capacity, trust in government, and the outcome of COVID-19 were used, there are alternative indicators that their use may influence the results to some extent.

Conclusions

Addressing the COVID-19 pandemic as a complex social phenomenon requires an analytical approach that recognizes its configurational nature. This study utilizes the conceptual framework of social determinants of health and the fuzzy-set QCA method to explain the high COVID-19 case fatality rates in OECD countries. This technique provides researchers with tools to examine cases as configuration of conditions and to investigate the relationships between combinations of causally relevant conditions and outcomes [22].

We have identified two different combinations of sociopolitical conditions that are linked to high COVID-19 CFR. These findings suggest that the potentially causes of COVID-19 high CFR are interconnected, thus effective management of the pandemic requires intersectoral coordination and action. The two causal recipes for high COVID-19 CFR are similar in that they both include a low state capacity and low trust in the government. It has been argued that a low level of government effectiveness in policy implementation may decrease public trust across different political systems, resulting in lower levels of compliance behavior [10, 35]. On the other hand, public trust can significantly contribute to the efficiency of governance systems. Trust is essential for fostering cohesion and reaching a shared understanding of policy challenges between citizens and the government [74]. Governments are more capable of achieving their policy goals when citizens trust them, as trust in government can reduce the cost of implementing a policy by enhancing voluntary compliance and reducing compliance costs [66, 74]. However, to comprehend the mechanisms through which factors such as public trust and state capacity affect COVID-19 outcomes, further research in different political systems and consideration of other structural characteristics are needed.

Implications for policy, practice, and future research

Based on this analysis, the causal recipe with the highest empirical importance (e.g., highest coverage) for a high CFR of COVID-19 in OECD countries consists of a combination of low state capacity, low trust in government, and low health expenditure. It means that governments in OECD countries without a well-funded health system to handle health crises should prioritize improving their efficiency and responsiveness. They should also promote social dialogue with citizens to build trust and create a social and political environment that encourages individuals to comply with government policies and change their health-related behaviors (see Appendix 4). The debate on behavior change has a long history. Promoting behavior change on a population level, particularly in low resource settings, requires a complex interplay of policy and strategy, creation of supportive environments, active community engagement, and the reorientation of health services. To achieve this, governments must commit to transparent, accountable, and inclusive public policy processes that actively involve the community. This commitment is crucial for ensuring the fair and effective implementation of response measures, and the efficient dissemination of public health information. it can also facilitate interaction among various societal groups or organizations, fostering collaboration in responding to health crises and supporting vulnerable populations [28].

There are several suggestions for future research. The set-theoretic approach allows researchers to ask context-specific questions that go beyond the scope of conventional quantitative analyses. For instance, “Under what conditions is there a connection between a specific causal condition and an outcome?“ [23]. Accordingly, researchers can explore the contexts in which a connection exists between lockdown measures and COVID-19 outcomes. They can investigate whether this relationship differs between countries that provide higher levels of social protection for enterprises and employment, compared to those that offer lower levels. According to the literature, sociopolitical factors such as public trust and the rule of law can be examined as contributing conditions. Furthermore, previous experiences with outbreaks can be considered in the analysis. The findings of these action-oriented inquiries could inform policymakers in developing targeted interventions for specific contexts, thereby enhancing the effectiveness of their responses.

¹Note 1. In the preliminary analysis, Russia was included, and the solution terms remained unchanged.