A Multiplicity of Environmental, Economic and Social Factor Analyses to Understand COVID-19 Diffusion

doi:10.21203/rs.3.rs-717943/v1

Download PDF

Research

A Multiplicity of Environmental, Economic and Social Factor Analyses to Understand COVID-19 Diffusion

https://doi.org/10.21203/rs.3.rs-717943/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 30 Nov, 2021

Read the published version in One Health →

Version 1

posted

You are reading this latest preprint version

Background: Research on the impact of the environment on COVID-19 lacks a full-comprehensive perspective, and neglecting the multiplicity of the human-environment system can lead to misleading conclusions. We attempted to reveal all environmental-to-human and human-to-human determinants that influence the transmission of COVID-19.

Methods: We estimated the daily case incidence ratios (CIR) of COVID-19 for prefectures across mainland China, and used a mixed-effects mixed-distribution model to study the association between the CIR and 114 considered factors related to climate, atmospheric environmental quality, terrain, population, economic, human mobility as well as public health control measures (PHCMs).

Results: CO, O₃, PM₁₀ and PM_2.5 were positively linked with CIR, but the effect of NO₂ was negative. The temperature had no significant association with CIR, and the daily minimum humidity was a significant negatively predictor. National emergency response level was negatively associated with CIR until with a lag of 15 days. Higher accumulated destination migration scale flow from the epicenter and lower distance to the epicenter (DisWH) were associated with a higher CIR, however, the interaction between DisWH and the time was positive. In China, the more economically developed and more densely populated cities have a higher probability of CIR occurrence, but they may not have a higher CIR intensity.

Conclusion: The COVID-19 pandemic’s diffusion patterns are caused by a multiplicity of environmental, economic, social factors as well as public health control measures PHCMs. First, multiple pollutants carried simultaneously on particulate matter act in a synergistic way to affect COVID-19 transmission. Second, the temperature has a limited impact on the spread of the epidemic. Third, PHCMs must last for at least 15 days or longer before the effect has been apparent. Fourth, preventing the introduction of the epidemic between cities seemed to be more complex and difficult than controlling the growth of the epidemic within the city. China's relatively cautious response to the epidemic could provide lessons for countries that are still experiencing a health emergency and help to prevent future pandemics similar to COVID -19.

Environmental Engineering

Environmental Policy

COVID-19

Human-environment system

Determinant

Public health control measures

Mixed-effects mixed-distribution model

Table 1 Data type, variables and data source.

Data type	Variable		Abbreviation	Units	Scale	Data source
Disease	COVID-19 data	Case incidence ratios (cases per 10 million persons)	CIR	0.00001%	Daily	National Health Commission of the People’s Republic of China (http://www.nhc.gov.cn/xcs/yqtb/list_gzbd.shtml) and provincial or municipal health commissions
Natural environment	Climate data	Minimum temperature	MinT	0.1℃	Daily	http://data.cma.cn
		Maximum temperature	MaxT	0.1℃	Daily
		Mean temperature	MeanT	0.1℃	Daily
		Relative humidity	Rh	%	Daily
		Minimum relative humidity	MinRh	%	Daily
	Atmospheric environmental quality data	Carbon monoxide, CO	CO, CO_1, …, CO_9	mg/m³	Daily	https://datacenter.mee.gov.cn/
		Nitrogen dioxide, NO₂	NO2, NO2_1, …, NO2_9	μg/m³	Daily
		Ozone, O₃	O3, O3_1, …, O3_9	μg/m³	Daily
		Fine particles, PM_2.5	PM25, PM25_1, …, PM25_9	μg/m³	Daily
		Inhalable coarse particles, PM₁₀	PM10, PM10_1, …, PM10_9	μg/m³	Daily
		Sulfur dioxide, SO₂	SO2, SO2_1, …, SO2_9	μg/m³	Daily
		Air Quality Index	AQI, AQI_1, …, AQI_9		Daily
	Terrain data	Mean DEM	MeanDEM	m	Daily	http://www.gscloud.cn
Human environment	Population data	Household population	Pop	10000 Person	2019, Fixed value	http://tjj.shandong.gov.cn/tjnj/nj2020/zk/indexch.htm, etc.
		Population density	PD	Person/km²	2019, Fixed value	http://www.mohurd.gov.cn/xytj/index.html
	Economic data	Gross domestic product per capita	GDP	100 million RMB	2019, Fixed value	http://tjj.shandong.gov.cn/tjnj/nj2020/zk/indexch.htm, etc.
	Human mobility	Destination migration scale flow from Wuhan = destination proportion in population flow from Wuhan * migration scale	Popmob，Popmob1, … , Popmob9, Popmobsum		Daily	http://qianxi.baidu.com/
		The distance of each city from Wuhan	DisWH	km	Fixed value	Distance measurement based on GIS
	Public health control measures	National emergency response	Reslevel, Reslevel1, …, Reslevel20		Daily	National Health Commission of the People’s Republic of China (http://www.nhc.gov.cn/xcs/yqtb/list_gzbd.shtml) and provincial or municipal health commissions

Table 2 Lasso variable selection results. Estimate lasso with lambda=1427.596 (lopt). Lasso is the variable coefficient estimated by Lasso, and Post-est is the variable coefficient of OLS regression on the selected variables.

Selected	Lasso	Post-est	Selected	Lasso	Post-est
Pop	-0.7565	-0.9184	SO2_7	-0.0283	-0.0421
PD	0.4994	0.5822	SO2_8	-0.0276	-0.0385
DisWH	-0.4102	-0.4808	SO2_9	-0.0551	-0.0557
GDP	0.9788	1.0951	NO2	-0.0762	-0.1032
MeanDEM	-0.0050	-0.0084	NO2_1	-0.0086	-0.0083
MinT	-0.0123	-0.0145	NO2_2	-0.0143	-0.0299
MinRh	-0.0120	-0.0330	NO2_3	-0.0046	-0.0082
Popmob0	-0.4786	-0.5358	NO2_4	-0.0027	-0.0061
Popmob1	-0.3684	-0.3436	NO2_5	-0.0087	-0.0125
Popmob2	-0.0497	-0.1001	NO2_6	-0.0231	-0.0287
Popmob3	-0.1051	-0.1209	NO2_8	-0.0149	-0.0223
Popmob4	-0.2876	-0.2543	NO2_9	-0.0059	0.0152
Popmob5	-0.0412	-0.2639	O3	-0.0145	-0.0204
Popmob7	0.0166	0.2176	O3_1	-0.0003	-0.0003
Popmob9	1.5534	1.5899	O3_2	-0.0043	-0.0042
Popmobsum	0.0152	0.0153	O3_6	-0.0049	-0.0057
Reslevel2	0.1354	0.0305	O3_7	-0.0006	-0.0006
Reslevel3	0.5769	0.4942	O3_9	-0.0028	-0.0044
Reslevel5	0.0473	-0.1455	PM10	0.0014	0.0022
Reslevel6	0.2925	0.3833	PM10_1	0.0006	0.0011
Reslevel8	0.5027	0.5793	PM10_2	0.0012	0.0014
Reslevel9	0.9803	1.5329	PM10_3	0.0007	0.0012
Reslevel15	-0.2266	-0.7235	PM10_4	0.0016	0.0020
Reslevel17	-0.1993	-0.2165	PM10_5	0.0006	0.0010
Reslevel20	-2.5200	-2.4571	PM10_6	0.0003	0.0007
CO	1.7358	2.9689	PM10_7	0.0011	0.0015
CO_2	0.0384	0.9589	PM10_9	0.0020	0.0038
CO_9	-0.1670	-2.0334	PM25_7	0.0009	0.0005
			PM25_8	0.0115	0.0149

Table 3 Parameter estimates, standard errors and fit statistics for mixed-effects mixed-distribution model to CIR occurrence and intensity.

Parameter	Occurrence			Intensity
	Estimate	SE	p	Estimate	SE	p
Intercept	-2.9447	0.2349	<.0001	3.1101	0.1528	<.0001
Day	0.1765	0.0068	<.0001	-0.0115	0.0037	0.0018
Day2	-0.0030	0.0001	<.0001	-0.0003	0.0000	<.0001
Pop	0.0618	0.0237	0.0096	-0.1784	0.0158	<.0001
PD	0.0813	0.0244	0.0010	-0.0266	0.0167	0.1116
GDP	0.0460	0.0181	0.0114	0.0688	0.0113	<.0001
Day_GDP	0.0023	0.0002	<.0001	-0.0002	0.0001	0.0099
MeanDEM	-0.0052	0.0009	<.0001	-0.0011	0.0007	0.0862
DisWH	-0.1865	0.0136	<.0001	-0.0852	0.0096	<.0001
Day_DisWH	0.0028	0.0002	<.0001	0.0027	0.0002	<.0001
Popmobsum	0.0019	0.0003	<.0001	0.0026	0.0002	<.0001
Reslevel15	-0.5810	0.0192	<.0001	-0.1162	0.0095	<.0001
MinT	-0.0008	0.0005	0.1106	-0.0004	0.0003	0.1979
MinRh	0.0026	0.0012	0.0350	-0.0031	0.0007	<.0001
CO_9	0.3844	0.0820	<.0001	0.0171	0.0516	0.7411
NO2_9	-0.0258	0.0029	<.0001	-0.0101	0.0015	<.0001
O3_9	0.0075	0.0010	<.0001	0.0035	0.0006	<.0001
PM10_9	0.0016	0.0004	0.0002	0.0015	0.0005	0.0021
PM25_8	0.0038	0.0007	<.0001	0.0005	0.0004	0.1564
SO2_9	0.0013	0.0039	0.7423	-0.0002	0.0026	0.9424
Residual				0.5681	0.0123	<.0001
Random effect	0.9592	0.1021	<.0001	0.4287	0.0416	<.0001
Covariance	0.2819	0.0529	<.0001
Fit statistics	Uncorrelated	Correlated	Difference in -2 log likelihood			p
AIC	44686.5	44634.17
-2 Log Likelihood	44600.5	44546.17		54.33		<.0001