“…Covariates at the baseline included age (four quintile), sex (binomial), educational attainment (college graduate or higher), marital status (married, unmarried, or widowed/separated), 22 having children (binomial 14 ), smoking habits (never, past, current), drinking habits (never, past, current), and having more than one underlying diseases (binomial) from hypertension, diabetes, asthma, bronchitis/pneumonia, atopic dermatitis, periodontitis, angina pectoris, cardiac infarction, cerebral vascular disease, chronic obstructive pulmonary disease, malignancy, chronic back pain, depression, and any other psychiatric disease 41 . Equivalent household income (four quintile 42 ), employment status (employer and self-employed, regular employee, others including nonregular employee), fear of Coronavirus-2019 Scale (FCV-19; four quintile), and mean effective reproduction number of COVID-19 in the month before the survey were considered for each prefecture as time-varying covariates 43,44 . The Japanese version of the FCV-19 was validated in previous studies 45 .…”