Pre-pandemic physical activity as a predictor of infection and mortality associated with COVID-19: Evidence from the National Health Insurance Service

Introduction During the coronavirus disease 2019 (COVID-19) pandemic, many populations have experienced reduced physical activity (PA) levels, weight gain, and increased anxiety and depression. However, according to a previous study, engaging in PA has a positive effect on damages caused by COVID-19. Therefore, this study aimed to investigate the association between PA and COVID-19 using the National Health Insurance Sharing Service Database in South Korea. Methods Logistic regression analysis was used to analyze the association of PA with COVID-19 and mortality. The analysis was adjusted for body mass index, sex, age, insurance type, comorbidity, and region of residence at baseline. Disability and lifestyle (weight, smoking, and drinking status) were adjusted consecutively. Results The results indicated that engaging in insufficient PA as per the WHO guidelines predicts a higher risk of COVID-19 when controlling for personal characteristics, comorbidity, lifestyle, disability, and mortality. Discussion This study revealed the need to engage in PA and manage weight to reduce the risk of infection and mortality associated with COVID-19. Because engaging in PA is an important component of weight management and can help restore physical and mental health after the COVID-19 pandemic, it should be emphasized as a pillar of recovery after COVID-19.

. Introduction Park et al. .
/fpubh. . previously unaffected individuals; for example, the anxiety and depression of people whose family, colleagues, classmates, or neighbors were affected by quarantine were increased (8). In addition, the risks of anxiety, depression, stress, and sleep disorders in COVID-19 patients were increased (9). Post-traumatic stress symptoms were occasionally experienced after infection, but results concerning physical health were limited (10). However, it is unclear how factors relating to lifestyles are linked to the prognosis of COVID-19. Engaging in physical activity (PA) has played an important role in improving psychological and physical health. It was found that engaging in regular moderate-to-vigorous PA (MVPA) is associated with reducing anxiety and negative self-perceptions, as well as improving physical health (11). In addition, engaging in MVPA is associated with losing and managing body weight that may predispose individuals to several types of chronic diseases, such as obesity and high blood pressure (12). However, a recent study shows that people spend more time engaging in sedentary behavior and less time engaging in PA than before the pandemic (13). Individuals engaging in PA is found to be more associated with lower risk of COVID-19 and mortality than those who do not meet the recommended PA level (150 min of MVPA at least once a week) (14,15), although these results did not consider other factors of health such as personal characteristics, comorbidity, and disability level.
Thus, we aimed to examine the associations of MVPA with COVID-19 and mortality. To produce more robust study results, we tried to include controlling variables beyond MVPA, using nationally representative data.

. . Database
The National Health Insurance Service (NHIS) of South Korea is a social insurance system for the entire nation, and registration is compulsory; approximately 97% of the Korean population is currently registered (16). The NHIS assists people with scheduling medical checkups every 2 years and records their results automatically. These data include various information such as demographic information, payment specification, consultation statement, diagnosis statements, and prescriptions. With these records they developed the National Health Insurance Sharing Service for researcher to support various studies providing sample cohort DB, customized cohort DB, health screening cohort, etc.
This study used NHIS-COVID19 DB that included 4,363 adult COVID-19 patients in South Korea between January 1, 2020, and July 14, 2020, who had medical records between 2015 and 2018, the most recent data before the pandemic. We selected 67,125 adults for the control group in NHIS DB who also had medical checkup data. We used the most recent records in our study; the final data included COVID-19 status, demographic information, comorbidity, disability status, and lifestyle, including PA and body mass index (BMI). This study conformed to the Guidelines on De-identification of Personal Data of Korea and was approved by the Kyung Hee University's Institutional Review Board (IRB No. KHSIRB20-301[EA]) as a review exemption study. Thus, the requirement for informed consent was waived.
. . Variables . . . PA PA was measured using a self-report questionnaire from NHIS. Moderate PA (MPA) was measured with the following question: "During the last week, how many times a week and for how many hours a day did you engage in physical activity at a moderate level for more than 10 min (e.g., fast walking, doubles tennis, riding bicycle, cleaning)?" Vigorous PA (VPA) was assessed with the following question: "During the last week, how many days a week and for how many hours a day did you engage in physical activity at a vigorous level for more than 10 min (e.g., running, aerobic, fast riding a bicycle)?" PA was categorized into two groups according to PA guidelines: 150 min of MVPA at least once a week (1 min of VPA = 2 min of MPA). The items have been widely used in the literature (17).

. . . Covariates
We adjusted for sex, age, region of residence, economic status, the number of comorbidities, disability, smoking status, drinking status, and weight, which are reportedly associated with COVID- 19 (18, 19). Residence was categorized into the five regions (Seoul, Daegu, Gyeonggi, Gyeong-buk, and other) of South Korea (from January 1, 2020 to August 14, 2020) with the most confirmed COVID-19 cases.
Economic status was measured using health insurance premiums. Health insurance premiums are categorized into five quintiles. In South Korea, every person must pay part of their income as an insurance premium. Thus, a higher quintile indicates a higher economic status. Basic livelihood security recipients were included in the medical aid.
Comorbidity refers to an underlying condition (e.g., diabetes, hypertension) that may cause and affect other diseases. In this study, the number of comorbidities was investigated with the question: "Among the following diseases, which diseases have you been diagnosed with, or have you been treated for?" with the examples of comorbidity, stroke, heart disease (e.g., myocardial infarction, angina pectoris), hypertension, diabetes, dyslipidemia, pulmonary tuberculosis, and other diseases including cancer.
Disability status included the presence, severity, and type of disability, including non-disabled, physical disability, encephalopathy, visual impairment, hearing impairment, and others. Disabled persons were registered with the Ministry of Health and Welfare of South Korea; hence, the NHIS data included their disability status. Severity and type of disability were measured according to the Act on Welfare of Persons with Disability.
Smoking was measured with the question: "In your life, have you ever smoked over five packs of cigarettes (100 pieces)?" Drinking was measured using the question, "How many times do you drink a week?"

. . Statistical analysis
The effects of PA and BMI on infection and mortality associated with COVID-19 were analyzed using logistic regression analysis, which was adjusted for sex, age, insurance type, comorbidities, and region of residence at baseline. Disability and lifestyle (e.g., need for weight management, smoking, and drinking status) were adjusted conjointly. Cases with missing data were excluded from the analysis. The 95% confidence interval (CI) was estimated using the SAS PROC PHREG. Statistical significance was set at p < 0.05. Statistical analyses were conducted using the SAS software (SAS Institute Inc., Cary, NC, USA).

. . Associations between PA and COVID--associated mortality
According to logistic regression Model 1 of COVID-19associated mortality (Table 3), which was adjusted for characteristics and comorbidity, not engaging in sufficient PA (OR: 1.548, 95% CI: 1.051-2.279, p < 0.05) predicted a higher risk of COVID-19associated mortality than engaging in sufficient PA. According to logistic regression Model 2 of COVID-19-associated mortality, which was additionally adjusted for the need for weight management, smoking, and drinking status, not engaging in sufficient MVPA (OR: 1.623, 95% CI: 1.078-2.445, p < 0.05) still predicted a higher risk of COVID-19-associated mortality than engaging in sufficient MVPA. According to logistic regression Model 3 of COVID-19-associated mortality, which was additionally adjusted for disability status, engaging in sufficient MVPA (95% CI: 0.96-1.882, p = 0.085) did not predict COVID-19-associated mortality. However, the presence of disability and the levels of severity of disability predicted COVID-19associated mortality.

. Discussion
This study aimed to determine the effect of MVPA on COVID-19 and the association between BMI and COVID-19, considering disability status. Investigating NHIS data on COVID-19 revealed several associations. Engaging in insufficient MVPA was associated with higher risk of infection and mortality associated with COVID-19, depending on confounding variables.
The importance of engaging in an active lifestyle was found to be influential on the risk of COVID-19 over the pandemic period. The result of this study aligns with those of a recent study, which reported the decrement of PA engagement over the period of COVID-19 (20) and the lower likelihood of developing COVID-19 (14). However, the current study has taken a step forward by utilizing the most recent data available before the pandemic. It is widely known that engaging in adequate MVPA is associated with positive health outcomes. However, the current study adds another piece of information on the role of engaging in MVPA. For example, a special focus should be paid to individuals with disabilities regardless of the type of disabilities. During the pandemic, disadvantaged populations may experience issues with accessing health information and PA programs in local communities. For example, statistics show a 4% decrement in PA participation by individuals with disabilities in 2021, compared to the participation rate in 2020 (21). Thus, it might be important to dedicate effort toward improving their participation. A recent review emphasized the role of supportive environments to stimulate one's autonomy, competence, and relatedness (i.e., social supports from close people) in PA settings (22). Given that older individuals are more susceptible to infection, these efforts must be focused on them. For example, local communities may provide newly developed programs to educate PA leaders, who may motivate older adults in their communities.
Engaging in insufficient MVPA could play an important role in reducing mortality associated with COVID-19. Previously, insufficient MVPA was reported as a negative predictor of all-cause mortality (23). However, the relation has not been confirmed in the patients with COVID-19. Interestingly, the estimates from the mortality results are higher than from the infection results, which suggests the importance of regular MVPA participation. Since the infection is an on-going phenomenon, engaging in MVPA should be recommended in South Korea. Particular attention should be paid to people, who are aged or physically disabled. It is noted that individuals with disability show much higher risks of mortality than those without disability. Thus, barriers to PA should be eliminated,       particularly perceived barriers, such as feeling uncomfortable, a lack of time, and other priorities (24). A technique to overcome these barriers could be providing clear intervention based on the stage of intention of individuals with disability. It is important for researchers and regional practitioners to consider levels of motivation and volition (i.e., psychological willingness to participate in PA) (25) to provide proper PA guidelines. This study has some limitations. First, this study's sampling period (8 months) was relatively short to determine the tendency of COVID-19, as the situation continues to evolve. At the beginning of this study, in August 2020, there were 4,222 confirmed cases and only 141 people had died of COVID-19 in South Korea. However, there are now more confirmed cases and deaths in South Korea. Second, it is difficult to generalize the results of South Korea and apply them to the rest of the world. Furthermore, the fatality rate in South Korea was deferred from the rate of the world. Thus, the results of this study may not be generalized. Finally, COVID-19 variants continue to evolve and affect individuals differently (26). As a result, the results from this study could not be applied to different variants. Thus, further studies must include various cases and variants to strengthen the associative findings between PA and COVID-19. Lastly, the current study did not examine the roles of light PA and sedentary behavior due to the limited data from the secondary source. Future research may include those variables to clarify the results of the current study.
In conclusion, this study examined the role of sufficient MVPA on both COVID-19 and associated mortality and found that engaging in sufficient MVPA may play a role in reducing the risk of COVID-19 and associated mortality. The main implication of the results is that an active lifestyle should be promoted in the community. In addition, urgent implementation is needed for people who are older or physically disabled.

Data availability statement
Data can be accessed with the permission of NHIS. Datasets of NHIS can be found here: https://nhiss.nhis.or.kr/bd/ay/bdaya001iv. do.

Author contributions
Conceptualization and analysis: I-HO and S-YP. Writing: SP and HK. All authors contributed to the article and approved the submitted version.

Funding
This work was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2022S1A5A8050449). This study was supported by the Translational R&D Program on Smart Rehabilitation Exercises (TRSRE-PS04), National Rehabilitation Center, Ministry of Health and Welfare, Korea.