COVID-19: weather, demographics and infection timeline.
The dynamics of COVID-19: weather, demographics and infection timeline.
Spencer EA, Jefferson T, Heneghan C.
Published on June 16, 2020
Transmission Dynamics of COVID-19
||Pedrosa RHL. The dynamics of Covid-19: weather, demographics and infection timeline. medRxiv 2020.04.21.20074450. 2020
||Central database case records.
||One author is part of project "Science and technology indicators in the state of São Paulo", Research Grant Fapesp N. 2019/20093.
||Temperature, absolute humidity, population density, timeline
In a multivariable model only population density and the timeline remained statistically significant.
- Population density and timeline emerged as the dominant effects on the infection rate.
- Weather variables were significant in some of the single-variable models but had opposite effects on the infection rate for the two groups of regions, US states vs countries
- Once the timeline and/or demographic variables were introduced as covariables, weather variables lost statistical significance.
What did they do?
Confirmed Covid-19 cases were obtained from databases of reported cases from Johns Hopkins University’s in the USA and from the European Centre for Disease Prevention and Control. The analysis included the 110 countries and states that had at least 10 days of data and had reached 100 accumulated cases, the latest end-date being April 10th.
The authors developed multivariable linear models in which the dependent variable was the early daily growth rate of Covid-19, estimated by the coefficient of the best exponential fit to the evolution curve of cases in the period of 10 days starting on the date when a geographical region reached the 100th case. This coefficient is denoted by k.
Exposures investigated were: average temperature and absolute humidity values during the 25 days starting 15 days before the region reached the 100th case; the date when the 100th case occurred; and the (log of) population density for the counties with higher participation in the spread of Covid-19 (for U.S. states only).
The evidence is based on early limited data. Many countries did not have accurate test data on the early phase of the outbreak which limits the findings.
|Clearly defined setting
||Demographic characteristics described
||Follow-up length was sufficient
||Transmission outcomes assessed
||Main biases are taken into consideration
What else should I consider?
Evidence on the effect of temperature and humidity is inconsistent across the literature. Many of these early studies need replication with more accurate test data, more accurate meteorological variables and longer follow-up. Countries starting their outbreaks later showed higher temperatures and more humid weather. Pollution was not assessed.
About the authors
Carl is Professor of EBM & Director of CEBM at the University of Oxford. He is also a GP and tweets @carlheneghan. He has an active interest in discovering the truth behind health research findings
Dr Elizabeth Spencer; MMedSci, PhD. Epidemiologist, Nuffield Department for Primary Care Health Sciences, University of Oxford.
Tom Jefferson, epidemiologist.