key: cord-0770057-ux4gr3tq authors: Sabawoon, W. title: Differences by country-level income in COVID-19 cases, deaths, case-fatality rates, and rates per million population in the first five months of the pandemic date: 2020-07-15 journal: nan DOI: 10.1101/2020.07.13.20153064 sha: b166843f278f55b5946361fe370acd4744d344f5 doc_id: 770057 cord_uid: ux4gr3tq Abstract Objective: To describe differences by country-level income in COVID-19 cases, deaths, case-fatality rates, incidence rates, and death rates per million population. Methods: Publicly available data on COVID-19 cases and deaths from December 31, 2019 to June 3, 2020 were analyzed. Kruskal-Wallis tests were used to examine associations of country-level income with COVID-19 cases, deaths, case-fatality rates, incidence rates, and death rates. Results: A total of 380,803 deaths out of 6,348,204 COVID-19 cases were reported from 210 countries and territories globally in the period under study, and the global case-fatality rate was 6.0%. Of the total globally reported cases and deaths, the percentages of cases and deaths were 59.9% and 75.0% for high-income countries, and 30.9% and 20.7% for upper-middle-income countries. Countries in higher-income categories had higher incidence rates and death rates. Between April and May, the incidence rates in higher-income groups of countries decreased, but in other groups, it increased. Conclusions In the first five months of the COVID-19 pandemic, most cases and deaths were reported from high-income and upper-middle-income countries, and those countries had higher incidence rates and death rates per million population than did lower-middle and low-income countries. Keywords: COVID-19, incidence rate, death rate, case fatality rate, income, and country The disease caused by a coronavirus that recently emerged in Wuhan, China (He et al., 2020) is now known as COVID-19. Most patients with COVID-19 have a respiratory illness with flu-like symptoms including cough, fever, fatigue, and shortness of breath. Some have pneumonia, acute respiratory distress syndrome, respiratory failure, and multiple organ failure, and a few die (Huang et al., The COVID-19 outbreak started a few weeks before the Chinese Spring Festival, which usually occurs in January or February on the Gregorian calendar. In 2020 it was celebrated January 24-30. Each year around the time of that festival, many Chinese people visit their ancestral homes and some people from outside China visit the country for sightseeing and to celebrate the start of the lunar year. In 2020, that seasonal increase in travel exacerbated the spread of SARS-CoV-2 within China and also to countries and territories outside China. The worldwide pandemic was recognized as such on March 11, 2020, after more than 118,000 cases had been reported in 110 countries and there was judged to be a sustained risk of further global spread ). As of mid-March, 146 countries were affected, 154,000 people had been infected, and 5,700 fatalities had been recorded. In absolute terms, the number of cases was very high in China (81,048 cases), followed by Italy trans-national phenomena regarding how economic activity is associated with the spread of the disease and with the resulting burdens on healthcare systems. The findings from such analysis can also be used by policy makers, international organizations, and bilateral agencies to allocate resources for the control of the epidemic according to income-defined groups. We therefore describe the worldwide distribution of the numbers of cases and deaths by country and by country-level income status; describe trends in the numbers of cases and deaths; and examine how country-level income status is associated with COVID-19 case-fatality rate and with the numbers of cases and deaths per million population in the first five months of the pandemic. For that purpose we consider each country's income status, using the World Bank's approach, classifying countries into high-income, upper-middle-income, lower-middle-income, and low-income groups (Fantom and Umar 2016). Two sources of publicly available data were used. First, data on COVID-19 cases and deaths in all affected countries were downloaded from the Our World in Data (OWID) website. The OWID's data on cases and deaths come from the European Centre for Disease Prevention and Control; data on testing were collected from formal reports of countries by the data team of OWID; data on other variables came from the United Nations, the World Bank, the Global Burden of Disease Collaborative (OWID data 2020 a ). Second, data classifying countries by their income status were downloaded from the World Bank website and incorporated into the master dataset. The countries were classified according to the World Bank classification of countries by income, which has been used since 1989. It divides countries into four groups -low income, lower-middle income, upper-middle income, and high income -using gross national income per capita valued annually in US dollars using a three-year average exchange rate. The classification is published at http://data.worldbank.org and revised versions are published each year on July 1 (Fantom and Umar 2016). The World Bank did not provide data on income for seven small areas, all of which reported cases of COVID-19: Jersey, Guernsey, the Falkland Islands, Vatican City State, Montserrat, Bonaire Sint Eus, and Anguilla. Therefore, those seven were categorized as "small areas with unknown income status". The data were exported to STATA 2012 for analysis. Cases on international conveyances (e.g. cruise ships) were excluded from the analysis, because the patients were from many countries, and therefore as groups they could not be linked with any of the World Bank's income categories. Descriptive statistics regarding cases, deaths, case-fatality rates (as percentages), cases per million population, and deaths per million population for each country and for income-defined groups of countries were calculated. Kruskal-Wallis one-way analysis of variance on ranks was used to examine associations of country-level income category with the variables listed above. Data from the small areas with unknown income status were excluded from the Kruskal-Wallis analysis because there was no information on country-level income status, and because the numbers of cases were much smaller than in the other income-groups. One-way analysis of variance (ANOVA) was used to examine associations of country-level income category with the time since the onset of the pandemic. Changes over time in the indices mentioned above were also examined for trends during the period for which data were available. The first cases of COVID-19 were reported to the WHO on December 31, 2019. By June 3 rd , 2020, the disease had spread to 210 countries and territories ( Table 1 ). The mean number of days from the time that the first case was reported to the WHO until June 3, 2020 differed widely among income-defined groups: 100.5 (SD 19.6) days for high-income countries; 94.8 (SD 18.6) for upper-middle-income countries; 89.5 (SD 20.1) for lower-middle-income countries; and 79.1 (SD 17.2) for low-income countries (F= 870.36, df = 3, p < 0.001). In five months, 6,804,286 cases were reported globally. About 90% were reported from high-income or upper-middle-income countries ( Table 2 ). The number of reported cases within each income-defined group varied. The top-10 high-income countries accounted for 48.0% of all COVID-19 cases globally. The top-10 upper-middle-income countries accounted for 26.1% of the cases globally. In contrast, the top-10 lower-middle-income countries accounted for only 6.9%, and the top-10 low-income countries accounted for only 0.6%. The remaining 170 countries and territories reported only 18.4% of all COVID 19 cases globally (Table 1) . Countries in higher income categories reported more cases per day (χ 2 = 514.79, df = 3, p < 0.001) ( Table 2 ). The high-income and upper-middle-income countries also accounted for most of the reported deaths. In five months, 380,803 deaths were reported, and thus the global case-fatality rate was 6.0%. Three quarters of all deaths were reported from high-income countries and one fifth were reported from upper-middle-income countries. Countries in higher income categories reported more deaths per day (χ 2 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 15, 2020. . https://doi.org/10.1101/2020.07.13.20153064 doi: medRxiv preprint = 349.92, df = 3, p < 0.001) ( Table 2) . The overall median daily incidence rate of COVID-19 for all affected countries in the period was 0.9 per million population (interquartile range: 0.0 to 9.3). The incidence rate of the disease varied across countries (Table 1) , and it was higher in countries with higher incomes (χ 2 = 1440.98, df = 3, p < 0.001) ( Table 3 ). The overall median daily death rate per million population was 0.0 (interquartile range: 0.00 to 0.01), which also differed between income-defined groups of countries. The 75 th percentile of the death rates for high-income, upper-middle-income, lower-middle-income, and low-income groups were 0.5, 0.2, 0.0, and 0.0, respectively (χ 2 = 479, df = 3, p < 0.001) ( Table 3) . From the onset of the pandemic until June 3, 2020, the total number of cases increased globally. In the high-income countries, of the total 3,804,286 reported cases, 0.1%, 16.2%, 45.9%, and 34.8% were reported in February, March, April, and May, respectively. A decline in cases was recorded between April and May. In the upper-middle-income group, over the same four months the number of cases increased. Of the total reported cases (1,962,382), 3.6%, 4.2%, 24.4%, and 60.1% were reported in February, March, April, and May, respectively. In the-lower middle-income group, that month-over-month increase in the percentage of cases increased substantially in May. Of the total cases in this income group (532,259), 2.0% were reported in February and 67.1% in May. In low-income countries, the percentage of cases increased between February and May from 0.0% to 71.1%. Similarly, the median (interquartile range) of the daily incidence rates per million population in high-income countries declined between April and May. It was 0.0 (0.0 to 0.1), 5.6 (0.8 to 23.3), 13.0 (0.3 to 45.0), and 2.9 (0.0 to 20.9) per million population in February, March, April, and May, respectively. However, . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 15, 2020. In the small areas with no income data, both cases and deaths declined after April. As of June 3, 2020, a total of 6,804,286 COVID-19 cases including 380,803 (6.0%) deaths had been reported globally. Income-defined groups of countries differed greatly in the numbers of cases and deaths and in their rates, as follows. Among the main findings of this study are the income-related differences in COVID-19 cases and their . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 15, 2020. . https://doi.org/10.1101/2020.07.13.20153064 doi: medRxiv preprint incidence rates per million population. The income-defined category of a country was positively associated with the number of COVID-19 cases and with the daily incidence rate per million population. The large differences in daily incidence rates probably reflect differences in epidemiologic and population characteristics, and differences in clinical and public health practices. First, the most obvious of the epidemiological differences is in the timing of the introduction and early transmission of SARS-CoV-2 in the various countries. Among the income-defined groups, strong connections by air travel probably resulted in the relatively early and widespread transmission of the disease among high-income and upper-middle-income countries. The date on which a country reported its first case(s) to the WHO can be taken as indicator of the date on which the disease began to spread in that country. Given that indicator, it is clear that the temporal order of the introduction and the start of transmission among income-defined categories is exactly the same as the economic ranking of those categories In other words, overall, the disease has been spreading from higher-income countries to lower-income countries. A second possible explanation for the observed differences in the number of cases by income-defined groups of countries involves differences in the availability of and approaches to SARS-CoV-2 testing, including testing patients with illness of various severities. In this context two indices of testing are often used: daily tests per capita and test-positivity rate (the number of positive tests as a percentage of the total number of tests done). High-income and upper-middle-income countries are more likely to test more people with illness of various severities than are lower-middle-income and low-income countries. Kobia and Gitaka (2020) also attributed the low apparent incidence rate of COVID-19 in Africa to insufficient testing. In large parts of Africa and also in most low-income and middle-income countries elsewhere, shortages of test kits, lack of capacity to implement large-scale testing and contact tracing, and lack of capacity to roll out surveillance testing have been reported (Jaffer Shah et al. 2020; Kobia and Gitaka 2020), all of which can cause the true incidence rate of COVID-19 in . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 15, 2020. . https://doi.org/10.1101/2020.07.13.20153064 doi: medRxiv preprint those areas to be underestimated. The test-positivity rate can be used to assess whether enough testing is being done. High test-positivity rates can indicate that only the sickest patients are being tested, which may imply that more people should be tested. The WHO has issued guidance stating that the positivity rate should be below 5% for at least 14 days before social-distancing measures are relaxed. The actual percentage varied both within and among income-defined groups. As shown on the OWID website, testing was limited in lower-middle-income countries and possibly also in low-income countries, which probably resulted in underestimation of incidence rates (OWID data 2020b). Third, the income-related differences might also have been caused in part by differences in the timing, adherence to, and/or implementation of preventive and community and public-health measures among the countries. Such measures include frequent handwashing with soap and water, use of alcohol-based hand rub, wearing masks, social distancing, case detection, isolation, contact tracing, and quarantine of exposed persons, closure of schools, public libraries, cinemas, and clubs, and suspension Also noteworthy are the differences in the COVID-19 case-fatality rate, in the daily reported numbers of COVID-19 cases, and in the numbers of deaths per million population, by income-defined group. Overall, higher-income countries were more severely affected in the time period under study. One contributing factor could be longer life expectancy, i.e. the presence of proportionally more elderly people. The spread of COVID-19 in nursing homes for elderly people has been documented (Bedford et . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 15, 2020. Nonetheless, the most likely explanation for the bulk of the difference between income-defined groups may still be the relatively late establishment of transmission of the virus in lower-income countries. Another main finding of this study is the fact that the COVID-19 incidence rate decreased in high-income countries from April to May 2020, while at the same time it increased in all other income-defined groups of countries. The impact of the pandemic will be large if the incidence rates already recorded in high-income countries eventually also occur in other countries. Some is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 15, 2020. . https://doi.org/10.1101/2020.07.13.20153064 doi: medRxiv preprint upper-middle-income countries. The findings of this study are subject to some limitations. First, the numbers of cases of COVID-19 reported each day are likely to be underestimated, especially in the middle-income and low-income countries. For example, in Afghanistan, COVID-19 cases have been classified as typhoid fever cases and many symptomatic patients did not present themselves to public health services, due to limited testing services, poor quality of medical care, or stigma associated with COVID-19 (Ariana news 2020; CNA news 2020; Sabawoon 2020). Second, the numbers of deaths reported daily may also be underestimated, because of non-follow-up or incomplete follow-up of patients with COVID-19 who died, or because of deaths among people who were infected with SARS-CoV-2 but in whom COVID-19 was not diagnosed. In the first five months of the COVID-19 pandemic, the vast majority of the cases and deaths reported worldwide were from high-income and upper-middle-income countries, with relatively high incidence rates and death rates. Between April and May, the numbers of cases and deaths decreased in high-income countries, but they continued to increase in upper-middle-income, lower-middle-income, and low-income countries. Funding: None. The author declares that he has no conflict of interest. Ethical approval and informed consent: Not required. The author is grateful to Joseph Green (retired from the Graduate School of Medicine at the University of Tokyo) for reading an earlier version of this article and providing . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 15, 2020. . https://doi.org/10.1101/2020.07.13.20153064 doi: medRxiv preprint comments and suggestions. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 15, 2020. . https://doi.org/10.1101/2020.07.13.20153064 doi: medRxiv preprint 8. Bukhari Q, Jameel Y (2020) Will coronavirus pandemic diminish by summer? Accessed April 30. 2020. https://ssrn.com/abstract=3556998 or http://dx.doi.org/10.2139/ssrn.3556998. 9. CNA news: Afghan testing lapse suggests growing and hidden COVID-19 crisis. June 2, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 15, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 15, 2020. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 15, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 15, 2020. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 15, 2020. . https://doi.org/10.1101/2020.07.13.20153064 doi: medRxiv preprint Table1: Numbers of COVID-19 total and daily cases and deaths; their rates per million population, and case-fatality rates from the onset of the pandemic until June 3rd, 2020, by co Days since repo . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted July 15, 2020. 6 3 7 1 1 1 Mexico 88 97,326 676 126 1,960 5 1 15 10,637 67 4 195 8 3 11 1 0 2 China 156 84,159 46 7 212 0 0 0 4,645 2 0 27 2 0 9 0 0 0 Belarus 89 44,255 495 9 913 52 1 97 243 3 0 5 0 0 1 0 0 1 Ecuador 90 40,414 238 53 498 13 3 28 3,438 17 1 38 5 2 11 1 0 2 South Africa 88 35,812 174 51 656 3 1 11 755 3 0 11 2 0 3 0 0 0 Colombia 85 31,833 218 92 635 4 2 12 1,009 10 2 18 3 2 4 0 0 0 Romania 96 19,517 197 93 311 10 5 16 1,279 12 1 24 5 2 8 1 0 1 Argentina 89 18,306 112 35 245 2 1 5 569 6 2 10 3 1 5 0 0 0 Dominican Repub85 17,752 206 90 296 19 8 27 515 6 2 9 3 1 5 1 0 1 Kazakhstan 81 11,796 139 38 218 7 2 12 44 0 0 1 0 0 1 0 0 0 Serbia 86 11,454 89 34 224 13 5 33 245 3 0 5 2 0 3 0 0 1 Armenia 86 10,009 64 30 146 21 10 49 158 1 0 3 1 0 2 0 0 1 Algeria 94 9,626 106 38 165 2 1 4 667 6 2 8 5 3 10 0 0 0 Malaysia 130 7,877 39 2 110 1 0 3 115 0 0 1 0 0 2 0 0 0 Iraq 98 7,387 48 11 83 1 0 2 235 2 0 4 3 1 6 0 0 0 Azerbaijan 89 5,935 49 25 101 5 2 10 71 0 0 1 1 0 2 0 0 0 Guatemala 81 5,586 27 5 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 15, 2020. . https://doi.org/10.1101/2020.07.13.20153064 doi: medRxiv preprint . . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 15, 2020. . https://doi.org/10.1101/2020.07.13.20153064 doi: medRxiv preprint Severe acute respiratory syndrome-related coronavirus: The species and its viruses -a statement of the The proximal origin of SARS-CoV-2 Concerns over the typhoid outbreak in Afghanistan Software for Singapore contact tracing app to be free for global use COVID-19: towards controlling of a pandemic COVID-19: towards controlling of a pandemic The COVID-19 pandemic: effects on low and middle-income countries India 125 207,615 336 1 2,553 0 0 2 5,815 6 0 73 3 2 4 0 0 0 Pakistan 93 80,463 517 104 1,446 2 0 7 1,688 15 1 31 2 1 2 0 0 0 Bangladesh 81 52,445 418 9 1,034 3 0 6 709 5 1 14 1 1 3 0 0 0 Indonesia 87 27,549 325 114 479 1 0 2 1,663 16 8 26 6 3 9 0 0 0 Egypt 106 27,536 126 9 388 1 0 4 1,052 7 0 15 4 2 8 0 0 0 Ukraine 82 24,823 332 68 476 8 2 11 735 8 1 14 3 2 4 0 0 0 Philippines 122 18,997 116 0 258 1 0 2 966 6 0 13 5 2 8 0 0 0 Bolivia 83 10,991 44 9 205 4 1 18 376 3 1 6 3 1 6 0 0 1 Nigeria 86 10,819 75 12 229 0 0 1 314 2 0 6 3 0 4 0 0 0 Moldova 86 8,548 105 35 148 26 9 37 307 3 0 6 3 0 5 1 0 1 Ghana 82 8,297 10 0 152 0 0 5 38 0 0 1 0 0 1 0 0 0 Morocco 86 7,866 86 33 139 2 1 4 206 1