Global-scale modeling of early factors and country-specific trajectories of COVID-19 incidence: a cross-sectional study of the first 6 months of the pandemic
Abstract Background Studies examining factors responsible for COVID-19 incidence have been mostly focused at the national or sub-national level. A global-level characterization of contributing factors and temporal trajectories of disease incidence is currently lacking. Here we conducted a global-sca...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2022-10-01
|
| Series: | BMC Public Health |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12889-022-14336-w |
| _version_ | 1828343241781542912 |
|---|---|
| author | Sujoy Ghosh Saikat Sinha Roy |
| author_facet | Sujoy Ghosh Saikat Sinha Roy |
| author_sort | Sujoy Ghosh |
| collection | DOAJ |
| description | Abstract Background Studies examining factors responsible for COVID-19 incidence have been mostly focused at the national or sub-national level. A global-level characterization of contributing factors and temporal trajectories of disease incidence is currently lacking. Here we conducted a global-scale analysis of COVID-19 infections to identify key factors associated with early disease incidence. Additionally, we compared longitudinal trends of COVID-19 incidence at a per-country level, and classified countries based on COVID-19 incidence trajectories and effects of lockdown responses. Methods This is an observational cross-sectional study covering COVID-19 incidence over the first 6 months of the pandemic (Jan 1, 2020 to June 30, 2020). A retrospective analysis was performed using publicly available data for total confirmed COVID-19 cases by country, and using recent data on demographic, meteorological, economic and health-related indicators per country. Data was analyzed in a regression modeling framework. Longitudinal trends were assessed via linear and non-linear model fitting. Competing models of disease trajectories were ranked by the Akaike’s Information Criterion (AIC). A novel approach involving hierarchical clustering was developed to classify countries based on the effects of lockdown measures on new COVID-19 caseloads surrounding the lockdown period. Results Univariate analysis identified 11 variables (employments in the agriculture, service and industrial sectors, percent population residing in urban areas, population age, number of visitors, and temperatures in the months of Jan-Apr) as independently associated with COVID-19 infections at a global level (variable p < 1E-05). Multivariable analysis identified a 5-variable model (percent urban population, percent employed in agriculture, population density, percent population aged 15–64 years, and temperature in March) as optimal for explaining global variations in COVID-19 (model adjusted R-squared = 0.68, model p < 2.20E-16). COVID-19 case trajectories for most countries were best captured by a log-logistic model, as determined by AIC estimates. Six predominant country clusters were identified when characterizing the effects of lockdown intervals on variations in COVID-19 new cases per country. Conclusions Globally, economic and meteorological factors are important determinants of early COVID-19 incidence. Analysis of longitudinal trends and lockdown effects on COVID-19 highlights important nuances in country-specific responses to infections. These results provide valuable insights into disease incidence at a per-country level, possibly allowing for more informed decision making by individual governments in future disease outbreaks. |
| first_indexed | 2024-04-13T23:40:57Z |
| format | Article |
| id | doaj.art-92b5cbc641824ac1b8a864d8ac45d7b7 |
| institution | Directory Open Access Journal |
| issn | 1471-2458 |
| language | English |
| last_indexed | 2024-04-13T23:40:57Z |
| publishDate | 2022-10-01 |
| publisher | BMC |
| record_format | Article |
| series | BMC Public Health |
| spelling | doaj.art-92b5cbc641824ac1b8a864d8ac45d7b72022-12-22T02:24:32ZengBMCBMC Public Health1471-24582022-10-0122111310.1186/s12889-022-14336-wGlobal-scale modeling of early factors and country-specific trajectories of COVID-19 incidence: a cross-sectional study of the first 6 months of the pandemicSujoy Ghosh0Saikat Sinha Roy1Centre for Computational Biology & Program in Cardiovascular and Metabolic Disorders, Duke-NUS Medical SchoolDepartment of Economics, Jadavpur UniversityAbstract Background Studies examining factors responsible for COVID-19 incidence have been mostly focused at the national or sub-national level. A global-level characterization of contributing factors and temporal trajectories of disease incidence is currently lacking. Here we conducted a global-scale analysis of COVID-19 infections to identify key factors associated with early disease incidence. Additionally, we compared longitudinal trends of COVID-19 incidence at a per-country level, and classified countries based on COVID-19 incidence trajectories and effects of lockdown responses. Methods This is an observational cross-sectional study covering COVID-19 incidence over the first 6 months of the pandemic (Jan 1, 2020 to June 30, 2020). A retrospective analysis was performed using publicly available data for total confirmed COVID-19 cases by country, and using recent data on demographic, meteorological, economic and health-related indicators per country. Data was analyzed in a regression modeling framework. Longitudinal trends were assessed via linear and non-linear model fitting. Competing models of disease trajectories were ranked by the Akaike’s Information Criterion (AIC). A novel approach involving hierarchical clustering was developed to classify countries based on the effects of lockdown measures on new COVID-19 caseloads surrounding the lockdown period. Results Univariate analysis identified 11 variables (employments in the agriculture, service and industrial sectors, percent population residing in urban areas, population age, number of visitors, and temperatures in the months of Jan-Apr) as independently associated with COVID-19 infections at a global level (variable p < 1E-05). Multivariable analysis identified a 5-variable model (percent urban population, percent employed in agriculture, population density, percent population aged 15–64 years, and temperature in March) as optimal for explaining global variations in COVID-19 (model adjusted R-squared = 0.68, model p < 2.20E-16). COVID-19 case trajectories for most countries were best captured by a log-logistic model, as determined by AIC estimates. Six predominant country clusters were identified when characterizing the effects of lockdown intervals on variations in COVID-19 new cases per country. Conclusions Globally, economic and meteorological factors are important determinants of early COVID-19 incidence. Analysis of longitudinal trends and lockdown effects on COVID-19 highlights important nuances in country-specific responses to infections. These results provide valuable insights into disease incidence at a per-country level, possibly allowing for more informed decision making by individual governments in future disease outbreaks.https://doi.org/10.1186/s12889-022-14336-wCOVID-19GlobalRegressionFactorsLockdownModeling |
| spellingShingle | Sujoy Ghosh Saikat Sinha Roy Global-scale modeling of early factors and country-specific trajectories of COVID-19 incidence: a cross-sectional study of the first 6 months of the pandemic BMC Public Health COVID-19 Global Regression Factors Lockdown Modeling |
| title | Global-scale modeling of early factors and country-specific trajectories of COVID-19 incidence: a cross-sectional study of the first 6 months of the pandemic |
| title_full | Global-scale modeling of early factors and country-specific trajectories of COVID-19 incidence: a cross-sectional study of the first 6 months of the pandemic |
| title_fullStr | Global-scale modeling of early factors and country-specific trajectories of COVID-19 incidence: a cross-sectional study of the first 6 months of the pandemic |
| title_full_unstemmed | Global-scale modeling of early factors and country-specific trajectories of COVID-19 incidence: a cross-sectional study of the first 6 months of the pandemic |
| title_short | Global-scale modeling of early factors and country-specific trajectories of COVID-19 incidence: a cross-sectional study of the first 6 months of the pandemic |
| title_sort | global scale modeling of early factors and country specific trajectories of covid 19 incidence a cross sectional study of the first 6 months of the pandemic |
| topic | COVID-19 Global Regression Factors Lockdown Modeling |
| url | https://doi.org/10.1186/s12889-022-14336-w |
| work_keys_str_mv | AT sujoyghosh globalscalemodelingofearlyfactorsandcountryspecifictrajectoriesofcovid19incidenceacrosssectionalstudyofthefirst6monthsofthepandemic AT saikatsinharoy globalscalemodelingofearlyfactorsandcountryspecifictrajectoriesofcovid19incidenceacrosssectionalstudyofthefirst6monthsofthepandemic |