Page 213 - GHES-3-2

Global Health Economics and
Sustainability
Health behaviors during COVID-19 pandemic

between medical/fitness expenditure on behaviors of PPE wearing, awareness of COVID-19 risk mitigation, number of
purchasing during a 3-day window following the start of chronic diseases and reported poor household subsidy. For
the outbreak and ease of maintaining expenses during and all models, a final cohort of 9,878 participants was included
following the COVID-19 pandemic among residents aged in our final analyses. A flowchart of the data procedure is
45+ of mainland China. In this study, we use the recently presented in Figure 1.
released China Health and Retirement Longitudinal Study
(CHARLS) wave 5 dataset, which includes many health- 2.2. Variables
related behaviors during the COVID-19 outbreaks in 2.2.1. Measurement of PPE purchasing behavior
China in 2020.
In the Module-V COVID section of the 2020 CHARLS
In the next section, a detailed description of the secondary dataset, we identified the outcome variable for PPE
data we used and our statistical methods for our analyses will purchase. This question asked study participants whether
be provided. In the Results section, outputs of our models they had purchased an increased amount of face masks,
of choice will be provided along with corresponding tables. hand sanitizer, and disinfectant to stockpile during the
Finally, our findings will be interpreted and elucidated 3 days following the lockdown of Wuhan, China (Zhao
in the Discussion and Conclusion sections, along with a et al., 2023). Responses were coded as “Yes” or “No” based
discussion of the limitations in this study. on the participants’ answers.
2. Data and methods 2.2.2. Measurement of ease in covering daily expenses

2.1. Study design In the Module G1 Household Income section, the
dependent variable assessing the ease in covering daily
This investigation included CHARLS survey data collected expenses with household income was measured by the
in 2020 following the onset of the COVID-19 outbreak. question, “From the breakout of COVID-19 to now, can
Datasets within the CHARLS database ranged from the income of the respondent’s household cover daily
income to health conditions, encompassing survey answers expenses?” Responses were categorized as either “difficult”
from participants aged 18+ (Zhao et al., 2023). However, or “easy” based on the respondent’s assessment.
its primary aim was to collect data from participants aged
45+ (Zhao et al., 2014). For clarity, its first survey (i.e., 2.2.3. Measurement of medical expenditure
wave 1) was disseminated in 2011 – 2012. Subsequently, the The first exposure, household medical expenditure (both
following waves 2, 3, 4, and 5 were collected in 2013, 2015, direct and indirect), was measured using questionnaires
2018, and 2020, respectively. The survey spanned across from the Module G1 Household Income section.
150 counties/districts and 450 villages/urban communities, Indirect medical expenses included costs associated
which included a total of 17,708 individuals within 10,257 with transportation, nutrition, and other family-related
households, encompassing both middle-aged and older expenditures incurred due to medical treatment,
adults. Specifically, wave 5 includes updated information excluding amounts covered by Medicare (Zhao et al.,
following the COVID-19 outbreak. Detailed information 2023). Respondents who reported no such expenses were
regarding the purpose, design, sample, and questionnaires classified as “No expenditure,” while those who incurred
of this repository is accessible in other articles (Zhao et al., expenses were classified as “Had expenditure.”
2014). The CHARLS research team has received ethical
approval from the institutional review board at Peking 2.2.4. Measurement of fitness expenditure
University Health Science Center (approval number: The second exposure, household fitness expenditure,
IRB00001052 – 11015). was assessed through Module G1 (Household Income
For our statistical analyses, among the 19,395 participants section) of the questionnaires, covering expenditures
in 2020, we excluded a total of 7,983 subordinate participants categorized to fitness activities, fitness equipment, and
of each household, a total of 128 participants younger than health supplements (Zhao et al., 2023). Respondents
age 45, 9 participants not sampled within the COVID-19 who reported no such expenses were classified as “No
questionnaire, 29 participants not sampled within the expenditure,” while those who incurred expenses were
household income questionnaire, a total of 442 participants classified as “Had expenditure.”
not sampled in the household expenditure questionnaire,
and finally, a total of 926 participants who did not provide 2.2.5. Confounding and covariates
information on the following covariates: Age, gender, The selection of covariates and/or confounding variables
government COVID-19 subsidy, reported building structure, was primarily based on the present literature that has also
reported social activities within the past month, mask used the CHARLS dataset for other analyses (Gong et al.,

Volume 3 Issue 2 (2025) 205 https://doi.org/10.36922/ghes.6619