The Relative Roles of Ambient Temperature and Mobility Patterns in Shaping the Transmission Heterogeneity of SARS-CoV-2 in Japan

Abstract

We assess the effects of ambient temperature and mobility patterns on the transmissibility of COVID-19 during the epidemiological years of the pandemic in Japan. The prefecture-specific daily time-series of confirmed coronavirus disease 2019 (COVID-19) cases, meteorological variables, levels of retail and recreation mobility (e.g., activities, going to restaurants, cafes, and shopping centers), and the number of vaccinations were collected for six prefectures in Japan from 1 May 2020 to 31 March 2022. We combined standard time-series generalized additive models (GAMs) with a distributed lag non-linear model (DLNM) to determine the exposure–lag–response association between the time-varying effective reproductive number (Rt), ambient temperature, and retail and recreation mobility, while controlling for a wide range of potential confounders. Utilizing a statistical model, the first distribution of the mean ambient temperature (i.e., −4.9 °C) was associated with an 11.6% (95% confidence interval [CI]: 5.9–17.7%) increase in Rt compared to the optimum ambient temperature (i.e., 18.5 °C). A retail and recreation mobility of 10.0% (99th percentile) was associated with a 19.6% (95% CI: 12.6–27.1%) increase in Rt over the optimal level (i.e., −16.0%). Our findings provide a better understanding of how ambient temperature and mobility patterns shape severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) transmission. These findings provide valuable epidemiological insights for public health policies in controlling disease transmission.

Nonlinear and Multidelayed Effects of Meteorological Drivers on Human Respiratory Syncytial Virus Infection in Japan

“…To establish a robust and reliable time-series statistical model, multiple stages were incorporated into it [ 11 , 53 , 54 ]. Prior to constructing the model, we checked the probability distribution of the dependent variables and number of weekly newly confirmed HRSV infection cases (the normality of probability distribution was assessed by the Shapiro–Wilk test) ( Figure S2 ), followed by an assessment of the relationships (e.g., linearity) between number of weekly newly confirmed HRSV infection cases and each independent variable.…”

“…The general algebraic definition of the time-series statistical models is formulated as follows: where is the outcome time-series; is the expected time-series of the number of weekly newly confirmed HRSV infection cases in prefecture i on week t ; the term corresponds to the overall intercept; denotes the cross-basis function of DLNMs with exposure and multilagged effects modelled by a natural cubic spline function and a linear function of multiple meteorological drivers (i.e., mean ambient temperature, relative humidity, precipitation, and wind speed) in prefecture i in week t , respectively. We also modeled baseline risk along with shared long-term seasonal variations and cycles and short-term trends by incorporating natural cubic splines of time (7 degrees of freedom (df) per year) as term , year as term , and number of public holidays per week as term , with the fixed effects variables as possible confounders [ 11 , 53 , 54 , 59 , 60 , 61 ]. denotes prefectural characteristics or regional variable indicators in prefecture i .…”

“… denotes prefectural characteristics or regional variable indicators in prefecture i . Furthermore, the autocorrelation of residuals in the case of infectious disease was pathogen-specific and needed to be accounted for; therefore, autoregressive terms at order one (as term ) was incorporated into the statistical models, accounting for potential serial correlation [ 53 , 54 , 59 , 60 , 61 ]. The teem denotes the logarithm of the yearly population (per 100,000) by prefecture as the offset term [ 62 ].…”

“…The term indicate errors. To quantify the total contribution, independent effects, and relative importance of multiple meteorological drivers, we included all variables in the same model [ 53 , 63 ]. By including all variables of interest in the same regression equation, we were able to strengthen the interpretation of the effects as independent or additive, based on accumulated empirical knowledge.…”

Clinical Features of COVID-19 in Pediatric Rheumatic Diseases: 2020–2022 Survey of the Pediatric Rheumatology Association of Japan

“…The coronavirus disease 2019 (COVID-19), caused by severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), has shown a remarkably high rate of spread since December 2019 [ 1 ]. According to the data reported as of August 2022 by the World Health Organization, more than 598.0 million and 6.4 million people worldwide have been infected and died, respectively [ 2 ].…”