key: cord-0835328-66hr1g7j
authors: Benítez, María Alejandra; Velasco, Carolina; Sequeira, Ana Rita; Henríquez, Josefa; Menezes, Flavio M.; Paolucci, Francesco
title: Responses to COVID-19 in five Latin American countries
date: 2020-08-27
journal: Health Policy Technol
DOI: 10.1016/j.hlpt.2020.08.014
sha: 7b6951e92bf12fe85e00d75d5fa0adc18e2f3ca4
doc_id: 835328
cord_uid: 66hr1g7j

BACKGROUND: COVID-19 reached Latin-American countries slightly later than European countries, around February/March, allowing some emergency preparedness response in countries characterized by low health system capacities and socioeconomic disparities. OBJECTIVE: This paper focuses on the first months of the pandemic in five Latin American countries: Brazil, Chile, Colombia, Ecuador and Peru. It analyses how the pre-pandemic context, and the government's responses in terms of containment and mitigation and economic measures have affected the COVID-19 health outcomes. METHODS: Extensive qualitative document analysis was conducted focused on publicly-available epidemiological data and federal and state/regional policy documents since the beginning of the pandemic. RESULTS: The countries were quick to implement stringent COVID-19 measures and incrementally scaled up their health systems capacity, although tracing and tracking have been poor. All five countries have experienced a large number of cases and deaths due to COVID-19. The analysis on the excess deaths also shows that the impact in deaths is far higher than the official numbers reported to date for some countries. CONCLUSION: Despite the introduction of stringent measures of containment and mitigation, and the scale up of health system capacities, pre-pandemic conditions that characterize these countries (high informal employment, and social inequalities) have undermined the effectiveness of the countries’ responses to the pandemic. The economic support measures put in place were found to be too timid for some countries and introduced too late in most of them. Additionally, the lack of a comprehensive strategy for testing and tracking has also contributed to the failure to contain the spread of the virus.

In this paper we analyse how the pre-pandemic context, the mitigation and containment measures, together with the health interventions, technologies and the economic response, have affected the COVID-19 outcomes. We also examine the five countries' profiles regarding epidemiologic and demographic characteristics, health system capacity and socio-economic development, with a view to understanding how these variables have impacted the effectiveness of the responses to COVID-19. Our focus is on the first months of the pandemic.

The five selected countries capture different realities across Latin America in terms of population size, area, density, demographic and socio-economic characteristics, health system financing and coverage, and other development indicators. For the five countries studied, we conducted extensive documentary analysis focused on federal and state/regional policies and interventions implemented in these countries since January 2020. We also analyze publicly available epidemiological data (released by the governments).

system characteristics of the five selected countries. It also presents country data on some of the health conditions that have been associated with poor clinical outcomes from COVID-19 Table 1 summarizes key socio-economic characteristics of the countries. In terms of population, Brazil has by far the largest population, followed by Colombia that has around a quarter of its population. Peru, Chile and Ecuador, in that order, are the ones with lower populations. The large population size may make it more difficult to ramp up the capacity of the health sector. In terms of density, Colombia and Ecuador are clearly outliers, which may impact the ability to mitigate the spread of COVID-19 and maintain social distancing in public areas (such as public transport). With respect to wealth, income inequality and access to basic social infrastructure, Chile is clearly an exception, with the highest GDP per capita and considerably better development indicators, but with a relevant level of inequality, as all the other countries analyzed. We note that a high level of informality is present in all five countries, which implies that for a significant fraction of the population staying at home is not an option unless there is an appropriate level of income support by the government. Also, poor access to sanitation infrastructure and clean water, in all countries except Chile, makes it difficult to take the preventive hygiene required to reduce contagion in certain areas. Table 2 presents indicators of health system typology, population covered by health insurance, health expenditure and healthcare resources for each country. All five countries are characterized by the existence of a public funded system of healthcare. Brazil is the only country with a national health system with comprehensive free access to healthcare services, while Chile, Ecuador and Peru have mixed systems in terms of health insurance (public and private). Colombia has a Social Health Insurance system, where individuals contribute a fixed amount of their incomes and have access to a defined health plan.

While Brazil and Chile spend a larger fraction of the GDP on health than the OECD average, health expenditures per capita are substantially lower than in developed countries. We also note that the fractions of total public expenditure over the total expenditure on health for the five countries are also far under OECD and the out of pocket (OOP) health expenditure is higher, except for Colombia.

The health system capacity indicators also raise concerns about the ability of the health sector to achieve a surge in capacity. 10, 11, [13] [14] [15] [16] [17] [18] [Insert Table 2 here] COVID-19 outcomes are related to the health status of the population. An increase in severity and the likelihood of poor clinical outcomes have been linked to patients' age, comorbidities and overweight. In particular, the older population and those with cardiovascular disease, diabetes and obesity present higher risks (19, 20) . Table 3 presents population health risk and life expectancy indicators for the five countries and the OECD. The five Latin American countries have a younger population than the OECD, but

Chile, Brazil and Colombia have a higher prevalence of obesity and diabetes. Brazil and Chile have also the highest incidence of cardiovascular diseases. In terms of risk behaviors (smoking and alcohol consumption), the five countries perform better than OECD, except Chile with a very high prevalence of smoking. 10, 11, 21) [Insert Table 3 here]

The high levels of income inequality and informality, alongside health sectors that are under-staffed and under-resourced, suggest that the challenges faced by these five Latin

American countries are different from those faced by developed countries and represent a high risk scenario when facing a pandemic like COVID-19. The information presented raises concern about the efforts that these countries must dedicate in the response to COVID-19, which will be analyzed in the following.

This section outlines how the five countries responded to the pandemic. The public health response and interventions were partly drawn from the OECD containment and mitigation strategies (22) . Table 4 provides a summary of the measures adopted in three domains: mitigation and containment, economic and health. Details about each policy can be found in Appendix 1.

Regarding mitigation and containment measures, Colombia, Ecuador, and Peru aimed at containing the spread, while Chile's objective was to mitigate it (reducing the rate of contagion) (23) . Brazil's objective seems unclear, with Brazil's president dismissing COVID-19 as a 'measly cold' at the end of March (24) and later arguing publicly with the Health Minister (who was later fired) over the need for social distancing (25) . Chile, Colombia, Ecuador, and Peru adopted measures that were decided centrally. Brazil, probably influenced by the USA, left the heavy lifting to states and cities, with no known attempts to achieve a nationally consistent approach. 1

This meant that, excluding Brazil, all four countries closed schools, prohibited massive gatherings and, except for Colombia, implemented curfews. Only Colombia, Ecuador, and Peru declared mandatory nation-wide lockdown. Chile used selective (i.e., by location) and dynamic lockdowns based on incidence rates, confirmed cases per km 2 and health risk of the population. Chile's quarantine strategy, however, did not consider access to basic services and sociodemographic vulnerability (26) . In Brazil, there has not been a quarantine mandated by the federal government, only the recommendation for social distancing. [Insert Table 4 here]

In the economic area, all countries have adopted policies to achieve mainly two objectives. First, to allow for compliance of the measures imposed, which is highly relevant for this paper, and second, to minimize the negative effects of the containment and mitigation measures to the economy (at the macro and micro level). In practice, this has translated into measures to increase liquidity for enterprises (micro and Small and Medium-sized Enterprises (SME)) and families, to protect labor and the economic activity, and to support the vulnerable population (e.g. income support and food baskets for the informal sector and low socio-economic population). All countries under analysis implemented the range of measures described, except for Ecuador which did not reduce the interest rate, impose tax reductions or support for informal workers.

A third area of response corresponds to the health-related measures, which complement the mitigation and containment strategy. In the studied countries the focus has been on capacities for treating patients. Therefore, hospitals have been reinforced by increasing ICU beds and ventilators, personnel and inputs such as personal protective equipment and other supplies. Nevertheless, extensive testing and tracking strategies have been almost nule or implemented only in one region (in Colombia was achieved in Medellín), except for Chile that adopted it recently. Those measures are necessary to timely isolate the infected population and reduce the spread of the virus.

In the following subsections, we analyze the response to the pandemic in three areas: (1) the timing and stringency of the mitigation and containment as well as the economic measures applied;

(2) compliance of the measures, assessing mobility and some possible determinants of it, such as "pandemic management" and the socio-economic context of the countries; and finally, (3) the health system response. These three areas can be identified as determinants leading to the poor outcomes the countries are experiencing when tackling the pandemic (see Section 4). The analysis of the measures is conceptually framed on the general recommendation published by the OECD (22, 27) , and the International Development Bank (IDB) for Latin-American countries (28) . Figure 1 presents the main mitigation and containment measures in a timeline. In general, countries applied containment and mitigation measures early on (closely after the first confirmed case), being Peru the first country of the group to implement a national lockdown (7 days after the first confirmed case). In contrast, Brazil was the latest of the group to declare state of emergency and closing borders (both towards the end of March), which, together with an ineffective health screening at the international airports, and the carnival celebrations (29) could have contributed to the rapid spread of the virus, especially in international hubs such as Rio de Janeiro and São Paulo. The level of stringency of the measures can be another determinant of the health outcomes. To measure it, the Oxford index is used and presented in Figure 2 (30)), as cases and deaths were increasing steadily as well as health services demand.

Ecuador and Peru evidenced a decrease in the index in the last days of the period studied, as both countries started to lift some measures in May, even though none of them observed a sustained decrease in cases. Ecuador ended its lockdown on May 4 and started a new stage of "social distancing", where each canton decides about containing measures using the traffic light system for restrictions. This flexibility has been applied also in regions with large numbers of cases and ICU patients, as Pichincha, which might explain the poor health outcomes (see Section 4). On the other hand, in Peru, some industries and services (mining, construction, tourism and retail) resumed their activity in May.

Regarding the rest of the countries, Chile announced the "Paso a paso" (step by step) plan for reopening (based on health indicators) in the middle of July, Brazil opened borders to international air travel tourists at the end of the same month, and Colombia allowed economic activity and internal travel resume in early August, though the activities exempted from the quarantine have been increasing over time. [Insert Figure 2 here] Although all the countries took most of the recommended mitigation and containment measures at early stages of the pandemic and stringency was relatively high, there is not a clear link between the time and stringency of the measures and the reported health outcomes (daily new cases and positivity rate), as both indicators continued to increase even after policies were implemented. We only observe an increase in the number of cases in Peru after restrictions were lifted from May onwards.

As mentioned, the economic measures are important to ensure that individuals can afford to comply with the imposed measures, especially those who live on daily income, and to counteract the spillover effects of the pandemic and the mitigation and containment measures (e.g. to prevent the collapse of financial and payment systems, promote the rapid reactivation after the crisis, protect employment and activity, and protect the vulnerable (37). As it was described in section 2, Latin American countries´ context (i.e. low income, high levels of poverty, inequality and informality) conditions the pandemic response, relevating the importance of supporting the vulnerable. IDB specifically recommended for Latin America emphasizing on the fiscal situation (i.e. temporary resources, reassign and develop policies that make more efficient expenditure when possible), as well as the protection of labor, enterprises and vulnerable populations due to the economic crisis generated by the pandemic (28) . This section will analyze the measures in the light of the first objective, focusing on the economic support to those that face more difficulties to comply. We examine the timing and the strictness (in this case referring to the amount and scope of the responses) the measures were implemented. Figure 3 shows the economic measures in a timeline. In general, a very responsive support channeled to the formal sector was observed throughout the countries. All countries reduced interest rates at least one time (except for Ecuador), created special conditions for credit during the pandemic (for micro, SME and families), and established measures to protect labor at early stages. However, support for informal workers and the most vulnerable (in cash and in kind) was slower, despite its importance for the household livelihoods. Considering the days after the first case was reported and the average date the measures were taken in each country, Colombia proved to be the fastest with an average of 16 days, almost half of the average of the rest of the countries.

Moreover, there was a delay in the support for households even when comparing with the date when first mandatory lockdowns were taken. It took 6 days to Chile from the date the first quarantine was implemented in a low-income county until the payment of the first income supports for informal workers and 50 days until the delivery of the first food baskets. In the case of Colombia, it took 15 and 9 days, respectively, and for Brazil, 11 and 64 days (in Brazil, considering the date when strict restrictions of movement were imposed in Rio de Janeiro and São Paulo). On the other hand, Ecuador took 12 days to deliver the food baskets since lockdown was imposed and Peru 54 days from lockdown until informal sector support (no support for the informal sector was applied in Ecuador and no food baskets in Peru The sizes and content of the economic packages differ between countries and relate to its pre pandemic reality (i.e. Chile facing social unrest and Ecuador indebtment problems with IMF) and to the evolution of the virus. Indeed, in the majority of the countries have been increasing since March until now (i.e. Brazil announced in early July an important credit program to support SME and Chile in late July passed an act with measures to support middle class people).

The Oxford tracker of COVID-19 government responses gives also information about the amount of economic stimulus spent by each country. We calculate the amount spent per capita for each country, finding that Brazil is the one among the five countries studied with the biggest effort, spending around USD$ 880 per inhabitant (around 10% of the country GDP), followed by Chile, with almost USD$ 800 (5.2%). Colombia and Peru spent almost USD$ 500 (7.6%) and USD$ 380 (5.6%) per population, while

Ecuador is by far the country with the lowest stimulus, both per capita and as proportion of GDP, with USD$ 24.8 spent for each habitant, that is 0.4% of its GDP (31).

Using the Oxford Stringency index for the economic measures, Figure 4 plots the evolution of the magnitude and extensiveness of the support given to households (income and credit) against cases and positivity rates. The figure evidences a low economic support for the population in Brazil (score of 50%), and higher scores for Colombia, Ecuador and Peru (around 75%) that have not changed since April. Chile started with low scores (less than 40%) and increased to 75% in the middle of June. As economic support would help people stay at home, especially those without a formal or no job, and living on daily/weekly income, a negative correlation would be expected. However, for the majority of the countries there is no clear relation between the increase of the stringency of the economic support to households and the number of new cases or positivity rate in the following period. On the contrary, cases increased or maintained after the latest increase in the index. That can suggest that the income support was not as effective as intended, not only due to the delay in its implementation as mentioned, but also because of the insufficient magnitude of the support. Only Chile shows a relevant decrease in the positivity rate and in the magnitude of new cases that coincides with a sharp increase in the economic index, and with the period when mitigation and containment measures and the tracing and tracking strategy were strengthened. This might suggest that applying those strategies in a strong magnitude together lead to better outcomes. 

As commented before, the reduction in the spread of the virus is crucial, even more in countries with a low level of capacity (ICU beds, ventilators, health workforce, among others). Thus, high levels of compliance are needed. Data on mobility reported for the countries indicate that this has not been the case. In Ecuador, based on data from a Brazil is an example of mismanagement. Public health response was not coordinated, and there was no federal policy enforcing physical distancing and isolation, or even guidelines to the states, since the central government could not agree on the strategy. It resulted in two changes of health ministers in one month (15-05 and 16-04).

Population received conflicting and mixed messages (41) , impacting the public health response, compliance levels and country's capacity to contain the spread of the virus.

Moreover, there was a testing time lag, lack of transparency, authoritarianism and censorship (42) about the truth burden the country is facing has led to an outrage across the political spectrum, particularly from medical professional associations and research institutes.

To some extent a similar situation occurred in Chile between the government and the COVID-19 advisory council in the first period of the pandemic, as well as between the government and majors and other authorities. At the beginning of the pandemic, the sanitary authorities emphasized the good performance of the country compared with other countries. The government announced plans to return to work in late April and talked about "new normality" (43) . Indeed, civil servants began their return in late April.

A few days later, the number of cases started to grow rapidly, and more stringent measures were taken. This episode eroded trust and compounded with other data reporting transparency issues can partly explain the resignation of the minister of health (June 14).

Trust in authorities is also important for citizens to comply. Fetzer et al. (44) , who surveyed 188 countries, found that 43% of the population reported that the government has not been truthful about COVID-19, while more than 60% of the population of Brazil, and 70% in Chile and Colombia has that perception about their respective country.

Despite countries' efforts to improve their communications, there have been continuous changes in the information provided to the population, the methodologies used to calculate cases, deaths and other relevant indicators. For example, in Brazil the government cancelled the publication of epidemiology reports, a task that was restored after the intervention from the supreme court. In Chile, the ministry was forced to correct the deaths after the publication of studies reporting important data gaps between Civil

Registry data and COVID-19 official reports (45) . These events affecting transparency and communication, undermined the public trust of the ordinary citizen experiencing financial and health hardships, especially during the first stages of the pandemic, but also research and academic institutions that are trying to understand the situation.

The analysis of compliance cannot be dissociated from socioeconomic factors. We have mentioned that a pre-pandemic commonality across the countries (to different extents) is having a high informality rate, poverty, consequently overcrowding, low sanitation systems, among others. We argue that strict measures are less effective in areas with low performing socioeconomic indicators, which can be evidenced when comparing different regions within the countries which have been exposed to similar measures, but obtain different results.

In Chile's Metropolitan region, which gather around 40% of the population of the country and concentrated around 70% of COVID-19 cases to date, those counties most affected in terms of cases and deaths are the ones with lower incomes (average income near USD 1,000) (46) and living in overcrowding conditions (11%) (47) . As the lockdowns in the country were implemented following a dynamic strategy, which means that the measure was applied at similar levels of incidence of the virus between the counties, the Chilean strategy allows for the comparison of the effects of the lockdown between counties with different income levels. Figure 5 shows two sets of counties, those with higher income (Vitacura, Lo Barnechea Providencia, Las Condes and Ñuñoa) and those with low income levels (Independencia, San Ramón, La Granja and Recoleta) (46) . It is observed that a decrease in the number of new daily cases (considering date of first symptoms) after one week of quarantine for the high-income counties, while those with lower levels of income show an increase in the number of daily new cases. That fact is consistent with the delay in the income support measures for vulnerable groups previously mentioned, for whom the lack of timely economic responses prevents them from staying at home, undermining the impact of lockdowns. In Colombia, the most affected region in terms of cases and deaths per population is Amazonas (32.9 cases and 1.3 deaths per 1,000 inhabitants), located in a province where informality reaches 90% of the labor sector and overcrowding and poverty reaches levels of 16% and 35%, respectively (48) . The department of Atlántico has been very affected too, together with Cartagena (in the department of Bolívar), although much lower than Amazonas. Both departments have high levels of overcrowding (4.0 in Atlántico and 4.7 in Bolívar ), workers in the informal sector (around 55%) and poverty (24 and 36%) (49). In opposition, Antioquia (where Medellín is located) and Valle del Cauca, which present lower rates of death per population, have lower levels of overcrowding (2.7 and 1.4%) and poverty (21.2 and 20.4% respectively) (50).

In Peru, Lima has the largest rate of cases per population (25.9 per 1,000) and is one of the departments with the largest rate deaths per population (1.1 per 1,000 inhabitants).

Lima Metropolitan area has greater population density and a significant informal sector dependent on daily cash payouts. Some of the remote regions do not have the health infrastructure nor health workforce to surge capacity. As an example, in Iquitos (capital of Loreto region, with no connection by road), there were reports of the collapse of the health services, with ICU beds at capacity, and 17 doctors have died of COVID-19 (51) in March. In this region, a significant part of the population is indigenous associated with poorer health outcomes compared to the general population (52) .

In Ecuador, the province of Guayas presents the larger total number of cases and deaths (around 18,000 and 1,700) and Guayas the rate of deaths per population ( 

In addition to the mitigation, containment and economic measures, as commented before, the five countries also applied measures related to health. Mainly, they made a great effort to increase their health systems capacities, which were far lower than those of developed countries, as presented in Section 2. A second focus of the health measures was regarding the testing strategy, while all countries fall back in tracking, despite its importance to contain the spread of the virus (54) . Both are analyzed in more detail in the following, and some technological developments, while minor, are also highlighted.

As recommended by OECD (22) and IDB (28) Nonetheless, the amounts spent differ among countries. Chile announced early (March 19) an increase in the health budget in an amount that corresponds to a 2% of total public budget (USD USD100 per inhabitant approximately). Peru also distributed resources to different institutions of the health system to prepare early in March. Ecuador, later than the others, also increased the health budget (USD 11 per person) (Appendix 1). Brazil allocated less than 1 dollar per capita in the middle of March for actions related to stopping the spread of COVID-19, although later on the country increased substantially the health package (Appendix 1). The number efforts for the first period of the pandemic regarding ICU beds, ventilators and laboratories can be seen in Table 5 . (59) . Brazil has the highest rate of ICU beds per population, but the regional distribution is uneven, thus, some states have already reached 100% occupancy in the public system (60) . In Peru, the national ICU occupancy level reached around 93% (August 11) and some reports across the country outline that some services were at capacity (61). Chile increased its occupancy levels to 88% in late May, and have maintained that level since then.

Metropolitan,Tarapacá, Antofagasta regions were at high levels of occupation in June, reaching levels around 95% of occupancy of ICU beds, despite a steady increase in the ICU bed count (62).

Even though the analyzed countries have made great efforts to increase the capacity of their health systems, many countries around the world introduced ban export and restriction measures in relation to medical supply products, hand sanitizers and disinfectants, at the same time they introduced a mix of import restriction and liberalization measures too. For low and middle-income countries that were affected slightly later on by the pandemic, these restrictions meant reduced global market access to COVID 19 related products and unfair competition with high-income countries. The global market context meant some medical supplies and protective equipment arrived later, and that governments had to rely on the local industries and innovation to produce these life-saving supplies (63).

Testing and tracing strategies are an important complement to containment and mitigation measures and as well as to the strengthening of the health systems in fighting COVID-19 (54) . OECD includes these strategies as one of the five objectives for the pandemic health response, emphasizing the use of technologies for detection, prevention, response and recovery (23, 28) . None of the studied countries has had an extensive and systematic testing and tracking approach at a national level for all the period of the pandemic.

In regard to testing, low levels of tests performed and time lag to receive the results, mean that the real number of cases and deaths can be highly underestimated in these countries. As well, it undermines the efficacy of containment and mitigation strategies, as people who have not received their test results and feel fairly well, together with asymptomatic cases not tested, could be moving around and spreading the virus. Related to the time lag, the evidence for Colombia indicates that the time between the symptoms and the diagnosis takes an average of 11.1 days, with 70% of the cases receiving a confirmation over 7 days after the first symptoms (64). In Chile, the time until the confirmation was estimated to be around 4 to 9 days (65). Unfortunately, there are no official reports of the time that takes the processing of the results in any of the five countries studied, neither is it evidence for the rest of the countries. However, some innovative testing strategies allow high levels of testing in some groups of the population using less resources. In Chile, pool testing has been implemented for analyzing the incidence of COVID-19 in elderly centers. That strategy combines laboratory samples of a group and analyzes them as one sample, saving resources, while allowing to find if there is one positive case and consequently knowing if other measures have to be implemented.

Regarding tracking of patients, none of the countries had reported to be doing it constantly and thoroughly at early stages, although there are some local initiatives that had resulted in important outcomes as presented later for the case of Medellín in Colombia. Only in early June, Chile has announced a more aggressive tracking and surveillance of COVID-19 cases, which also coincides with the time the cases and positivity rates started to decrease, as noted previously. The strategy includes primary health care workforce and an increase in personnel in charge of the tracing (reaching almost 4,000 in late July), as well as in the number of places in health residences (for people that cannot safely quarantine in their homes and for those forced to quarantine), with a capacity of 11,000 people in August, according to the official information (33). The tracing involves call centers dependent on the primary care institutions and on the regional health secretaries (Ministry of Health). They call the confirmed or probable cases (reaching 80% of them in the Metropolitan Region in early August) as well as their contacts (62.4% of the cases were reported to come from a close contact in early August), educating, testing and isolating when necessary.

Resource scarcity, barriers and the fast-moving spread of the virus and its severity, also led to a surge of innovations to solve problems and to be scalable (66). Some interesting cases are local initiatives to produce supplies. In Peru, the government resorted to local procurement strategies for the manufacture and supply of masks and PPEs for the whole country (67). The production of mechanical ventilators combined efforts from local universities, and some units repaired by the Armed Forces. In Chile, some machines for anesthesia were converted into ventilators and some ventilators were used to supply oxygen to more than one patient. The Production Support Corporation under the Ministry of Economics opened tenders to stimulate the production of sanitary resources, which resulted in late July in the first locally made ventilators (68). On July 30th, the Nature journal published a paper on the development of low cost tests for COVID 19 (69). The research developed by Chilean researchers achieves high levels of accuracy and the test costs around US$1. Another interesting and low cost way of detecting COVID-19 outbreaks is the study of wastewater, which has also been being done in Chile (33).

Despite a slow implementation and some resistance (pre-pandemic), telehealth has now been boosted and strongly encouraged in the countries studied. Building on telehealth commission work (70) The city of Medellin, Colombia, implemented an intensive health technology plan in January, based on a platform (´Medellín me cuida´), where citizens can register and add information related to comorbidities, location, family, contacts, workplace, among others. More than 2 million people had registered (72) . The platform can connect with the travel card and, for example, infected people can be banned from using public transportation. It is also connected to the police system, allowing them to easily know if a person is authorized to circulate. Since the first case, the municipality has been applying COVID-19 tests at home, following the cases with daily calls, and tracking close contacts, in coordination with the civil police. The platform also allows text messaging to people located close to a positive case, the use the information to predict future contagious and risk zones and detect people that must be tested because of their association to a positive case. Medellín has the largest testing rate per infected cases in

Colombia (72).

Similarly, Ecuador has developed a digital application that identifies the zones with high levels of cases and the level of movement of COVID-19 (using GPS). The government had open access to the information about the number of cases (movement data is not open) to all the citizens to be aware of the magnitude of active cases in different areas (73) . As well, Brazil has recently added to its app (Coronavirus-SUS) the functionality of alerting people that has been exposed in the previous 14 days to a confirmed case (74).

This section addresses how COVID-19 has impacted the five countries. First, regarding its health impact, we analyze the so-called "direct effects" based on officially reported data by the health authorities, related to the number of positive cases and deaths (and the age and gender de-aggregations when possible) and the number of patients in ICU.

We aim at showing where each country stands in comparison with the others. We also analyze "Indirect effects", where we study excess mortality, as they may capture spillover effects such as non-COVID related deaths, for which we also add some information about non-COVID spillover effects, but also it may show if direct effects are underreported when data availability and transparency cast doubt on official numbers (still, direct). Second, we present some indicators related to the economic impact of the pandemic on the five countries, considering their characteristics and context (i.e. reliance on certain economic sectors and dependency on commodities).

While this study relied on official data published by the Ministry of Health of each country, there are several caveats. First, there are many cases under investigation in Brazil and Ecuador. Besides, data inconsistency between the data reported by the Ministry of Health and the analysis of researchers and academics using public data have been evidenced in Brazil, Chile and Ecuador.

In respect to the data breakdowns, Chile, Peru, Ecuador and Brazil publish aggregated data and also include some level of disaggregation (i.e. age, gender, region) for the daily (or periodic) cases and/or deaths. In contrast, Colombia is the only country with detailed micro data for each case (e.g. including demographic variables and information about the health status (hospitalized, recovered, death)). Appendix 2 presents a summary of how and what COVID-19 data has been collected in this study.

Brazil and Ecuador do not publish information on the number of ICU patients, while Chile and Peru publish the total number of ICU patients daily, but there is no information on how many patients have been hospitalized daily due to COVID-19, or the length of stay in the hospitals. Since June, Chile has begun to publish data of total discharges for each week, but the level of aggregation does not allow to analyze daily admissions or length of hospitalization either.

This subsection uses epidemiological data to describe some of the health outcomes from the beginning of the pandemic until August 9. General country-level patterns are presented, as well as the breakdown per gender and age.

Brazil is the country with the highest number of COVID-19 confirmed cases, accounting for around 3 million. The total number of positive cases in Brazil is over five times

Peru's cumulative cases, around eight times Chile's and Colombia's and more than 30 times Ecuador's total cases. Currently, the only country that seems to be bending the curve is Chile, which, in addition, has been decreasing levels of positivity rates (under 10% in early August). In contrast, Colombia, which remained with a low and stable number of cases for about three to four months, began to increase the number of cases at a fast pace at the end of June, and in early August, is the country that presents the steepest slope. Peru, after relenting the increase in cases is now presenting a higher speed of increase. Although all countries have increased their health capacity, large numbers of cases could have impacted significantly the occupancy of ICU beds. Figure 8 shows the initial and current ICU beds, as well as patients in those units for the countries where the data was available (Peru, Chile and Colombia). Chile quickly surpassed initial capacity by May, while its efforts in increasing health system capacity have managed to meet demand, and since June, when the country reached its peak occupancy rate, has been observing, for more than a month, a decreasing number of patients in ICU units. In contrast, Colombia and Peru are still increasing the number of patients that require that level of attention, especially the former, that is doing so rapidly. Nevertheless, Colombia's high capacity has implied that the country is far from reaching its limit. That is not the case for Peru, on the border of collapsing, with only around 100 ICU beds available. In regard to deaths, Brazil is the country with the highest number of fatal cases, with over 100,000 deaths, while Chile, Colombia and Peru exceeded the 10,000 death mark. While we had observed an even distribution of the cases among genders, in Figure 9 Panel B, we observe that in Colombia and Peru, the deceases are concentrated among men, accounting nearly 60% and 70% of deaths, respectively. In Panel C of the same figure, where deaths are shown for 60 years old and over, we observe the same as in other countries (20) , that the most affected population is gathered in this group. Nevertheless, the percentages vary for the countries, as in Brazil the population over 60 years old represents around 70% of deaths, while in Chile this number is significantly higher, at 80%. Over time, mortality for 60 and over has decreased substantially for the two countries, as this indicator was 90% and 85% respectively. While all five countries have been adversely affected so far, as all of them are currently presenting a large number of daily cases and deaths, the cumulative impact of at least the first wave of the pandemic is going to be greater than the impact presented here. Colombia, despite presenting a far lower number of cases and deaths in the first months since the first case, is currently experiencing a rapid increase in those numbers, probably related to the relaxation of measures and lack of an extensive testing and tracing strategy, similar to Peru´s situation. On the contrary, Chile has managed to sustain in the last month a decreasing rate of new cases and deaths, as well as ICU patients. As presented in section 3, that can be explained since during June Chile increased the extent and amount of economic measures, covering the formal and informal sector and vulnerable population, and started an aggressive national strategy of testing, tracing and tracking of confirmed cases and its contacts, reaching a high level of coverage of them along the country. Similarly, during July Chile´ containment and mitigation measures stringency increased.

The deaths due to COVID-19 might be higher than the official number reported by each country due to two reasons. First, some deaths for which the real cause was COVID-19 may be underreported because there was not a positive test that confirmed the diagnosis. In all the countries analyzed, the deaths reported by governments as COVID-19

deaths are only those with a positive test. Second, the pandemic can have an additional impact on deaths, that is, an indirect effect, increasing mortality for other diagnostics, caused by lower access to health. This can respond to a reduction in the resources for routine and non-emergency healthcare (which are being redirected to COVID-19 patients) and fear and avoidance to seek medical care, as well as to the consequences of containment and mitigation measures on household income that may be leading to situations of food insecurity (78), impacting health outcomes for children and general population. In contrast, the measures can also have some positive effects, decreasing the number of deaths for other causes, for example, reducing the number of preventable accidents (i.e. car accidents) and lower contagion of other viruses and infections.

To assess the overall mortality impact of COVID-19, Figure We contrast the excess deaths data with the official COVID-19 report, to estimate the magnitude of the additional impact of COVID-19 in the selected countries. The results

show that the impact in deaths in some countries is far higher than the reported officially to date. Peru shows the highest difference between excess deaths and the official COVID-19 deaths, with almost 25,000 deaths. Indeed, the sanitary authorities have announced that they are reviewing the total number with the National System of Deceases of Peru and there are still 15,000 unrevised until August 8 (79) . In terms of the difference magnitud, Peru is followed by Ecuador (21,990) and the Brazilian cities (8,180).

Those numbers represent 130%, 386% and 30% of total COVID-19 reported deaths for the period. On the other hand, Chile presents the lowest difference for the period (127 deaths), however, the upper bound for Chile (with a confidence interval of 95%) is around 2,500 (close to 30% of total COVID-19 reported deaths for the period). As it is not possible to determine whether those deaths correspond only to uncounted COVID-19 deaths or to the indirect effect by priority allocation to COVID-19 and its restrictions, we provide some information to evidence on the effects of COVID-19 on other health areas. The data support a shift in health services utilization and provision of immunization routine programs for Chile, Colombia, Ecuador and Peru. This information is presented in Appendix 4. In Chile, the data shows a reduction in emergency admissions due to respiratory and cardiovascular diseases. Regarding the notification of diseases, in Bogotá (Colombia) there is some evidence of a decrease in the notifications of respiratory diseases, whilst Ecuador also reports a reduction in the notification of vaccine-preventable diseases (chickenpox, mumps). Ecuador evidenced a decrease in the level of the vigilance of diseases of mandatory notification compared with previous years. In Peru, there is also a reduction in the coverage of immunization programs for four different vaccines below 10%. BCG vaccine has a higher coverage as it is administered to newborns immediately after the birth.

This subsection briefly examines the economic implications of COVID-19. Some of these implications followed directly from changes in consumer behavior as a response to the pandemic, that occur even in the absence of lockdowns and social distancing rules, but also there is an economic impact of the restrictions implemented. Other implications are more indirect, and results from changes in the world economy.

Strict mitigating and containing measures as lockdowns affect all sectors, but specially the service, commerce and tourism (37) . Figure 11 shows the unemployment rate in the countries. Although countries have adopted measures to protect employment, data shows that unemployment rates increased in all five countries, with spikes in Colombia The panorama is not auspicious, and the pandemic is still on course. Inevitably employment and salaries will be affected and, thus, informality and poverty, which, in turn, makes it more difficult to contain the virus. The Organization for Economic Cooperation and Development talks about an "unprecedented global economic crisis" for the region, which was already in a difficult situation, estimating an increase in the average poverty rate of Latin-American countries from 30.3% to 34.7% (88) .

Unemployment data reflects on some economic indicators such as the economic activity and GDP. Figure 12 Panel A, shows the index of economic activity for the countries, while Panel B shows the projections of GDP growth. We observe that the decrease in activity has been steeper in Chile, Colombia and Peru, countries that had stable indexes before 2020. As a consequence, GDP projections also dropped heavily for the five countries, turning to negative since March 2020. Ecuador and Peru present the largest decreases up to August, between 8 and 9 points. Countries that were expected to have a higher increase in the annual GDP before the pandemic, such as Colombia and Chile, are now expecting to have a decrease in GDP, but less negative in Ecuador and Peru, as well as Brazil.

As the countries also heavily rely on commodity prices, their reduction as well as decrease in exports has affected the economic indicators presented (89) . Oil price, partly influenced by a decrease in world demand, shows the largest decrease, affecting Brazil's economy but specially Colombia for whom oil represents more than 50% of exports.

On the other hand, the evidence shows that exports also decrease due to pandemic. Exports to China are expected to drop 24.4% (88) , affecting heavily the revenues of Brazil, Chile and Peru, whose economies rely heavily on China's demand (88) . [Insert Figure 12 here]

As analyzed before, governments responded with economic packages which sizes differ, according to their economic reality. In all cases, those efforts implied an increase of fiscal deficit and indebtments. The fiscal deficit is expected to be 9.5% of GDP on average but the proportion debt to GDP can reach more than 90% for Brazil and over 60%

for Colombia and Ecuador (96) .

Thus, the scenario is negative, as captured by rating agencies, which have been constantly modifying the outlooks of these countries in the last months. For example, S&P modified the outlook of Colombia and Chile from stable to negative (March 26 and April 27, respectively) and for Brazil from positive to stable (April 6) (97).

This paper described and analyzed COVID-19 economic and health impacts in Brazil, Chile, Colombia, Ecuador, and Peru. We presented pre-pandemic data on socioeconomic development, epidemiologic and demographic characteristics, and health system resources and performance in the five countries. In addition, we conducted extensive documentary analysis focusing on federal and state public health and economic responses covering since the arrival of the virus to the continent in late February until mid-August.

All five countries adopted strict measures early on to contain the first wave of COVID-19, including lockdowns (national or focalized) and curfews. However, the effectiveness of the measures was undermined by the existing fragility of the health systems, which are characterized by insufficient investment in health resources, regional disparities, modest information systems and poor communication and coordination. Indeed, the health systems have been overwhelmed in the first 100 days of the pandemic, with ICU beds reaching nearly 100% occupancy in some regions.

The existence of a large informal sector affected the ability of individuals to comply with the containment and mitigation measures further undermining their effectiveness.

While the five countries introduced income support measures, these measures were by and large too timid or too late to achieve high levels of compliance. Moreover, there was a lack across the five countries of a comprehensive strategy for early detection, isolation, surveillance, and tracking of patients and close contacts.

Our analysis provides an explanation for why, despite their early response, the five countries are facing high incidence rates and deaths per population, especially in regions with lower socioeconomic conditions (dense and overcrowded neighbourhoods and remote areas) and greater concentrations of informal workers. Moreover, the evidence presented on excess mortality and health spillover effects shows that the human cost of the pandemic is far higher than what is currently accounted for in official data. This is due not only to unreported cases and ineffective information systems, but also to the hidden impact on healthcare by the diversion of resources to the COVID-19 response.

To make matters worse, the five countries are also experiencing economic hardship. In addition to the impact on sectors such as retail and tourism, which have been prevalent worldwide, these countries were particularly impacted by the fall in exports of commodities to China. This negative economic outlook will likely persist for a number of years. Of concern is the impact on the vulnerable members of the population, with limited access to social welfare and to well-resourced health services.

Overall, this study highlighted the importance of early emergency preparedness and the need to improve the capacity of health systems to mitigate the spread of the virus.

Health system reform that aligns with the WHO health systems building blocks would create resilient health systems that could respond better for disease outbreaks, as well as natural and human-made disasters (98). In addition, the insights gained from this study reinforce the importance of trustworthy and robust government institutions to lead a The scope of this study is time-bounded and constrained by the information on the public domain. This implies, for example, that the role of technology is underexploited.

The epidemiological data quality and availability is a significant limitation that highlights the disease monitoring shortcomings in the selected countries. This is an important caveat for a well-planned recovery strategy. To ensure decision-making and policies are driven by evidence and focused on the most vulnerable population, further research is required preferably using mixed methods and building on an interdisciplinary approach to assess the mid-term and long-term effects of COVID-19 across the different societal sectors, and the actions needed to recover from this pandemic. The urgency of such pursuing such research cannot be underestimated as the five countries relax restrictions and face a resurgence of infections and deaths' in lieu of talking about a second wave.

(2) Worldometers. Countries in the world by population (2020) Note: Only containment measures (school closing, workplace closing, cancel public events, restrictions on gatherings, close public transport, stay at home requirements, restrictions on internal movement and international travel controls) are plotted in the graph. The graph also includes data of daily new cases and positivity rate (five-day average). Positivity rate calculated as daily confirmed cases per number of tests processed. Limits of new cases axis in Ecuador were fixed from 0 to 3,000, because corrections from sanitary authority reporting a reduction or abnormal increase on the total cumulative cases, that result in negative cases on 6/05, 8/05 and 11/05 and cases above 10,000 on 24/04. Days with missing data for Ecuador corrected using data of cumulative cases in the previous and following dates. Note: Only economic measures (income support and debt/contract relief) are plotted in the graph. The graph also includes data of daily new cases and positivity rate (five-day average). Limits of new cases axis in Ecuador were fixed from 0 to 3,000, because corrections from sanitary authority reporting a reduction or abnormal increase on the total cumulative cases, that result in negative cases on 6/05, 8/05 and 11/05 and cases above 10,000 on 24/04. Days with missing data for Ecuador corrected using data of cumulative cases in the previous and following dates. Note: For each county growth is defined as the variation (%) in the average of daily new cases on the week number 2 after quarantine declaration (week 0) and the average of new cases in the 2 weeks before quarantine was implemented. The average of new cases is corrected by the number of tests taken each week. Pink bars correspond to the counties that increased the daily average of new cases. Blue bars, those that decreased it. Note: For Brazil the initial number of ICU beds has been calculated using the data of deaths per 100,000 population. NA=not available data. Note: There is no daily data disaggregated by gender and age for Brazil, nor by age for confirmed cases for Peru. Upper bound Confirmed cases 2020

Boletín del aseguramiento en salud

Latin America's health systems brace for a battering

Comorbidity and its impact on 1590 patients with COVID-19 in China: a nationwide analysis

Obesity and impaired metabolic health in patients with COVID-19

Age-standardized death rate from cardiovascular diseases, per 100,000 individuals

Flattening the COVID-19 peak: Containment and mitigation policies

(24) NYT. Bolsonaro isolated and defiant, dismisses Coronavirus threat

The Washington Post. Brazil's Bolsonaro fires Health Minister Mandetta after differences over coronavirus response. Press release 2020

Vulnerabilidad urbana y accesibilidad en las comunas con cuarentena parcial. Press reléase 2020

Containment: Health systems responses to COVID-19 in the OECD

La política pública frente al Covid-19. Recomendaciones para América Latina y el Caribe

Pandemia e fronteiras brasileiras: análise da evolução da Covid-19 e proposições

Ministro Blumel encabeza reunión de coordinación con alcaldes de las comunas que entran en cuarentena. Press release 2020

Oxford COVID-19 Government Response Tracker

Blavatnik School of Government

Ministry of Health Brazil. Coronavirus panel

and Ministry's of Information Technology and Communication website

Ministry of Health Peru. Sala Situacional COVID-19 Perú

Coronavirus: antecedentes sanitarios y económicos para la discusión. Punto de Referencia N° 532, Centro de Estudios Públicos

Presidente revela que enfermos de covid-19 rompen cuarentena y se mueven en Guayaquil y Quito. Press release 2020

Impact of the first days of massive quarantine in Metropolitan Region

NYT. Highways of Peru Swell With Families Fleeing Virus. Press release 2020

Overlapping values: religious and scientific conflicts during the COVID-19 crisis in Brazil

Judge orders Bolsonaro to resume publishing Brazil COVID-19 data. Press release 2020

Press release 2020

Global Behaviors and Perceptions at the Onset of the COVID-19 Pandemic

Informe sobre la evolución de la epidemia de

Seremi de Desarrollo Social y Familia Metropolitana

Ministry of Housing and Urbanism. Estadísticas Habitacionales

National Administrative Department of Statistics Colombia. Pobreza Multidimensional Región Amazonía-Orinoquía. Boletín técnico [Internet]

(49) National Administrative Department of Statistics Colombia. Statistics by topic [Internet]

Incidencia de la pobreza monetaria

Se necesitarían al menos 1,200 respiradores al 26 de abril, pero solo hay 500 en el país. Press release 2020

Por una nueva convivencia. La sociedad peruana en tiempos del COVID-19: Escenarios, Propuestas de Política y Acción Pública

Active case finding with case management: the key to tackling the COVID-19 pandemic. The Lancet

Se necesitarían al menos 1,200 respiradores al 26 de abril, pero solo hay 500 en el país. Press release 2020

Gobierno de Brasil entrega 16 ventiladores más a Amazonas

Nuevo record de casos COVID-19 en Colombia: 1,888 para un total de 48,746. Press release 2020

Capacidad de UCIs en Colombia

Las hospitalizaciones en Quito están en niveles críticos y las UCI prácticamente copadas. Press release 20200

Camas comienzan a escasear en América Latina. Press reléase 2020

¿Por qué la curva de Medellín es menor que la de otras capitales

El presidente Lenin Moreno presentó al país la app SaludEc. Press release 2020

Coronavirus-SUS application will alert close contacts of Covid-19 patients

Ministry of Health Brazil. Boletines epidemiológicos

Capacidad de UCIs en Colombia

Chilean Society of Intensive Medicine. Daily survey intensive national reality

Joint Statement on COVID-19 Impacts on Food Security and Nutrition

COVID-19: este miércoles se volverá a actualizar cifra de fallecidos, adelanta Minsá

Ministry of Health Peru. National Informatic System of Deaths

Excess deaths tracker

Administrative National Administrative Department of Statistics of Colombia

Brazilian Institute of Geographic and Statistics. Panel de indicadores

OECD. COVID-19 in Latin America and the Caribbean: Regional socio-economic implications and policy priorities. OECD Policies Responses to Coronavirus

License: Creative Commons Attribution CC

Portal Brasileño de datos abiertos

Bloomberg consensus estimates as of

Latin American Mid-Year Outlook

(98) WHO. Monitoring the building blocks of health systems: a handbook of indicators and their measurement strategies

State of State Reform in Latin America

Life Satisfaction and Confidence in National Institutions: Evidence from South America

COVID-19): contributing to a global effort

Note: HDI: Human Development Index

Note: NA: not available. Hb.= inhabitants

Table 3: Health Risk factors for Peru, Chile, Brazil and Colombia (3, 10, 11, 21)

Source:* (10) and

Note: Hb. = inhabitants. Note: Blue boxes if the country adopted the measure at national level, and white boxes if not adopted at that level. NA means no available information

Daily. ICU patients and beds occupancy. Tax relief X X X X No disconnection from basic services X X X X Income support X X X X X Labor protection X X X X X Food baskets X X X X X Informal workers support X X X X Support for the vulnerable X X X X X