Enhancing Subjective Well-Being through Physical Activity for the Elderly in Korea: A Meta-Analysis Approach

The purpose of the current study was to meta-analytically investigate the psychological impacts of physical activity for the elderly population in Korea. The findings from 21 studies, using the comprehensive meta-analysis (CMA) program, indicated that participation in physical activity had a low, but meaningful, impact on the subjective well-being of elderly individuals. Of three exercise dose components, the duration of physical activity was the most influential, followed by the frequency and intensity of the physical activity. Of six subjective well-being measures, self-efficacy was most strongly associated with physical activity, followed by life satisfaction, leisure satisfaction, exercise satisfaction, successful aging, and happiness. Results of moderator analyses indicated that the influence of physical activity became more pronounced as the proportion of males increased. Meanwhile, of the three exercise dose components, only the relationship between the duration and subjective well-being was influenced as the percentage of participants without a spouse or partner increased. Overall, regular participation in physical activity is an effective way of promoting subjective well-being among older adults in Korea. Relevant guidelines regarding physical activity prescription and behavioral management strategies are discussed.


Introduction
According to the United Nations' forecast, Korea is facing the biggest increase in the proportion of the elderly, with Korean society projected to become hyperaged by 2026 [1][2][3]. The rate of elderly population increase is faster than others, with the 65+ age group expected to make up more than 38% of the population in Korea by 2050 [2,3]. Trends in aging of the population are not uncommon, especially for Asian countries. To a lesser or similar extent, Asian countries as Singapore, Taiwan, China, and Thailand are also expected to become hyperaged societies in the near future [3]. Such demographic transitions might come with higher social and economic costs and issues in quality of life if not well-prepared. Thus, successful aging is a critical social and national issue for fast-aging countries such as Korea, given that aging is associated with decreased body functions and reduced income [4,5].
From the quality of life perspective, individual life quality is determined by several key factors, including physical well-being, relationships with other people, social activities, and recreation [6]. Among others, physical well-being is considered to be the most critical contributing factor, with positive evaluate the overall effect size of the elderly's participation in physical activity on various dependent variables, including life satisfaction, successful aging, leisure and exercise satisfaction, self-efficacy, and happiness. Second, the influence of the three individual components of physical activities was investigated, namely the frequency, intensity, and duration of the physical activity, on a set of exercise benefits. Lastly, the moderating roles of gender and marital status on the relationship between physical activity and exercise benefits were explored. The research model of the study is presented in Figure 1. investigated, namely the frequency, intensity, and duration of the physical activity, on a set of exercise benefits. Lastly, the moderating roles of gender and marital status on the relationship between physical activity and exercise benefits were explored. The research model of the study is presented in Figure 1.

Data Sources and Search
Both published articles and unpublished theses and dissertations (covering 1995-2016) were included in the meta-analysis (see Figure 2). Duplicated unpublished studies were excluded when a certain thesis or dissertation was published in a journal. To find relevant studies, a search was conducted using three major national electronic databases, including PubMed, the National Assembly Library (NAL), Korea Education and Research Information Service (KERIS), and Korean-Studies Information Service System (KISS), using the keywords "physical activity for the elderly (in/for older adults)", "exercise for older adults", "sport participation among older adults", and "leisure activity participation of older adults". In addition, a reference list describing previous reviews of the related studies was inspected.

Data Sources and Search
Both published articles and unpublished theses and dissertations (covering 1995-2016) were included in the meta-analysis (see Figure 2). Duplicated unpublished studies were excluded when a certain thesis or dissertation was published in a journal. To find relevant studies, a search was conducted using three major national electronic databases, including PubMed, the National Assembly Library (NAL), Korea Education and Research Information Service (KERIS), and Korean-Studies Information Service System (KISS), using the keywords "physical activity for the elderly (in/for older adults)", "exercise for older adults", "sport participation among older adults", and "leisure activity participation of older adults". In addition, a reference list describing previous reviews of the related studies was inspected. investigated, namely the frequency, intensity, and duration of the physical activity, on a set of exercise benefits. Lastly, the moderating roles of gender and marital status on the relationship between physical activity and exercise benefits were explored. The research model of the study is presented in Figure 1.

Data Sources and Search
Both published articles and unpublished theses and dissertations (covering 1995-2016) were included in the meta-analysis (see Figure 2). Duplicated unpublished studies were excluded when a certain thesis or dissertation was published in a journal. To find relevant studies, a search was conducted using three major national electronic databases, including PubMed, the National Assembly Library (NAL), Korea Education and Research Information Service (KERIS), and Korean-Studies Information Service System (KISS), using the keywords "physical activity for the elderly (in/for older adults)", "exercise for older adults", "sport participation among older adults", and "leisure activity participation of older adults". In addition, a reference list describing previous reviews of the related studies was inspected.  Two investigators independently reviewed all the titles and abstracts to identify potentially relevant articles and dissertations for further review. Disagreements between two investigators were resolved by a consensus. Studies were considered eligible if they measured either the frequency, intensity, and/or duration of study participants' physical activities, included any kind of quality of life or subjective well-being measures, such as life satisfaction, successful aging, leisure satisfaction, exercise satisfaction, self-efficacy, or happiness, and included study participants aged 65 years and older.
As reported in Figure 2, an initial screening and subsequent screening yielded a total of 42 studies, including 22 journal articles and 20 dissertations. From the remaining studies, 21 studies were excluded if there was no relevant statistical information, such as zero-order correlation coefficients and data allowing computation of effect sizes [19,20]. The final sample for the current meta-analysis consisted of 21 studies (N = 7614), 12 journal articles and nine unpublished dissertations. Table 1 reports a list of the individual studies that were included in the meta-analysis.

Coding
To ensure the accuracy of coding, two of the authors coded all sample characteristics and relevant statistical data (e.g., estimates of the effects of physical activity and exercise) from each study. Differences, other than errors, were resolved by discussion with a third investigator when necessary. Coded items included the level of physical activity participation and outcome variables as well as the sample size, the gender proportion of the study, the percentage of the sample with a partner (married or cohabiting), publication type, and publication year.
The independent variable, physical activity dose, was coded in terms of the frequency, intensity, and duration. Also, the elderly's subjective well-being (type) was coded into six categories, namely, exercise satisfaction, leisure satisfaction, life satisfaction, self-efficacy, successful aging, and happiness. The proportions of gender and marital status from each study were also coded. The type of marital status was coded as single (or divorced) or married (or with a domestic partner). In addition, correlations between physical activity and subjective well-being (i.e., physical activity-related outcomes) were coded for each study.

Statistical Procedures
A random-effects meta-analysis was conducted using Comprehensive Meta-Analysis (CMA) software [37]. A correlation between variables was considered significant if the 95% confidence interval (CI) excluded zero. The stability of the effects was tested using the Fail-Safe sample size (N) [38]. A result of the meta-analysis was considered robust to publication bias if the effect size is greater than "5K + 10", where k is the number of studies [38]. Hunter and Schmidt's Q test was used to test the significance of the estimated variance of the population correlation [39]. Moderator effects were tested using z-tests to examine the differences in the weighted correlations [39].

Aggregated Effect of Physcial Activity on Subjective Well-Being
The mean correlation between physical activity and individual-level psychological outcomes (aggregated over outcomes) was r = 0.21, 95% CI = [0.17, 0.26], p < 0.001, which showed a small but significant positive effect of physical activity across the subjective well-being range among the elderly Korean population (see Table 2). Effect sizes ranged from 0.096 to 0.338 throughout the 21 studies. Heterogeneity was considerable, with a high I 2 value of 91.77%, indicating the potential presence of moderators for the physical activity effect on the elderly's subjective well-being [40]. There was no conclusive evidence of sample bias (Egger's Test of the Intercept B 0 = −0.028, p > 0.05), and the funnel plot of standard error and log rate ratio was fairly symmetrical (see Figure 3). The Fail-safe N was 4070, suggesting that there would need to be an additional 4,070 studies added to the analysis in order for the cumulative effect to become significantly nonsignificant.  As shown in Table 2, the mean correlations between physical activity and six psychological outcome measures were also calculated: life satisfaction (k = 9), leisure satisfaction (k = 5), self-efficacy (k = 5), happiness (k = 4), successful aging (k = 3), and exercise satisfaction (k = 1). The strongest correlation was found with self-efficacy, r = 0. 34  As shown in Table 2, the mean correlations between physical activity and six psychological outcome measures were also calculated: life satisfaction (k = 9), leisure satisfaction (k = 5), self-efficacy (k = 5), happiness (k = 4), successful aging (k = 3), and exercise satisfaction (k = 1). The strongest correlation was found with self-efficacy, r = 0. 34 Only the effect of physical activity for self-efficacy was found to be moderate [41].

The Duration, Frequency, and Intensity of Physcial Activity on Subjective Well-Being
The effects of physical activity's three components on subjective well-being were examined ( Table 3). The correlation was stronger with the duration of physical activity (r = 0.27), followed by the frequency (r = 0.20) and intensity (r = 0.17). The 95% confidence interval for all three components did not include zero, indicating that all three correlations were statistically significant. The effects of the three PA components on each subjective well-being measure were also examined. As expected, the results were almost identical to the aggregated effect of physical activity. The effects of the duration, frequency, and intensity were stronger on self-efficacy (r = 0.37, r = 0.31, and r = 33, respectively) than other subjective well-being measures. Only the intensity did not exhibit a significant effect on leisure satisfaction and happiness.

Moderator Analyses
Moderator analyses using meta-regression were performed based on the ratio of male subjects (Table 4 and Figure 4) and the ratio of the absence of a spouse/partner (Table 5 and Figure 5) per each study, given the suggested heterogeneity in this study. For gender differences, the ratio of male subjects (percentages of male participants) per study was entered as a predictor of the effect sizes, in order to examine whether the relationship between physical activity and the aggregated subjective well-being was influenced by the proportion of gender in each study. The results indicated that the aggregated influence of physical activity was more pronounced as the proportion of males increased (z = 10.08, p < 0.001). Of the three components of physical activity, the influence of the duration (z = 11.69, p < 0.001) and intensity (z = 7.13, p < 0.001) on the aggregated subjective well-being increased as the proportion of male subjects increased. However, the gender proportion did not influence the relationship between the frequency and the aggregated subjective well-being.
Secondly, the percentage of those living without a spouse or partner (i.e., the absence of a spouse) was entered as a predictor of the effect sizes. Counter to what might be expected, the influence of the duration increased as the percentage of participants without a spouse or partner increased (z = 2.23, p = 0.025). Regarding the other components, the absence of a spouse did not moderate the relationship between the frequency/intensity and the aggregated subjective well-being.

Moderator Analyses
Moderator analyses using meta-regression were performed based on the ratio of male subjects (Table 4 and Figure 4) and the ratio of the absence of a spouse/partner (Table 5 and Figure 5) per each study, given the suggested heterogeneity in this study. For gender differences, the ratio of male subjects (percentages of male participants) per study was entered as a predictor of the effect sizes, in order to examine whether the relationship between physical activity and the aggregated subjective well-being was influenced by the proportion of gender in each study. The results indicated that the aggregated influence of physical activity was more pronounced as the proportion of males increased (z = 10.08, p < 0.001). Of the three components of physical activity, the influence of the duration (z = 11.69, p < 0.001) and intensity (z = 7.13, p < 0.001) on the aggregated subjective well-being increased as the proportion of male subjects increased. However, the gender proportion did not influence the relationship between the frequency and the aggregated subjective well-being.

Discussion
The current meta-analysis showed small to moderate, but significant, effects of physical activity on the subjective well-being of the Korean elderly population. In particular, physical activity had a stronger influence on the self-efficacy of the elderly, among others. The results suggested a positive association between physical activity and subjective well-being, which was consistent with the previous studies [42]. Many elderly people in Korea tend to rely on dietary and health medicines and supplements to promote and maintain their health, preferring non-physical activities to promote socio-psychological well-being, as opposed to physical leisure and recreational activities. However, the results of the current study demonstrated the efficacy and utility of physical activity in the promotion of elderly Koreans' subjective well-being. Overgaard et al. argued that physically active individuals who consumed energy-and nutrient-balanced diets did not gain additional benefits on health from the consumption of dietary supplements [43]. Therefore, the concept of physical activity prescription should be embraced by health and human services agencies in Korea.
Of the various subjective well-being indicators, the self-efficacy of elderly people could be enhanced by physical activity to a greater extent, followed by life satisfaction, leisure satisfaction, (perceived) successful aging, and happiness. Thus, the results indicated that all subjective well-being

Discussion
The current meta-analysis showed small to moderate, but significant, effects of physical activity on the subjective well-being of the Korean elderly population. In particular, physical activity had a stronger influence on the self-efficacy of the elderly, among others. The results suggested a positive association between physical activity and subjective well-being, which was consistent with the previous studies [42]. Many elderly people in Korea tend to rely on dietary and health medicines and supplements to promote and maintain their health, preferring non-physical activities to promote socio-psychological well-being, as opposed to physical leisure and recreational activities. However, the results of the current study demonstrated the efficacy and utility of physical activity in the promotion of elderly Koreans' subjective well-being. Overgaard et al. argued that physically active individuals who consumed energy-and nutrient-balanced diets did not gain additional benefits on health from the consumption of dietary supplements [43]. Therefore, the concept of physical activity prescription should be embraced by health and human services agencies in Korea.
Of the various subjective well-being indicators, the self-efficacy of elderly people could be enhanced by physical activity to a greater extent, followed by life satisfaction, leisure satisfaction, (perceived) successful aging, and happiness. Thus, the results indicated that all subjective well-being measures were positively affected by participation in physical activities. Consequently, health policymakers and practitioners should promote physical activity among the elderly, for example, by providing proper exercise prescription. According to Mazzeo and Tanaka, there are five exercise prescription elements, i.e., warm-up, exercise intensity, exercise duration, frequency, and exercise type/mode [44]. In most cases, elderly Koreans do not receive proper exercise prescription due to the lack of guidelines for medical or sport medicine professionals. The results of the current study suggested that, of the three exercise dose indicators, the duration of physical activity was the most important aspect, followed by the frequency and intensity of physical activity, in the promotion of the subjective well-being of the elderly person. Per exercise prescription guidelines [44], future studies should also examine the efficacy and influence of physical activity type/mode on the elderly's subjective well-being and other health benefits.
According to the American College of Sports Medicine (ACSM)'s guidelines concerning physical activity programs for older adult populations [45], there are seven factors that influence behavioral management strategies in physical activity interventions, including social support, self-efficacy, active choices, health contracts, perceived safety, regular performance feedback, and positive reinforcement. Given the results of the current study, a behavioral management strategy to expedite initial participation and sustain continuous participation in physical activity should be developed to address the "duration" aspect of the physical activity.
Of the six subjective well-being measures investigated in this study, especially, self-efficacy is a factor that increases self-esteem and reduces anxiety levels, and it can be enhanced by participation in physical activity. On the other hand, self-efficacy is one of the factors that positively influences effective behavioral management strategies in physical activity interventions [45]. Thus, it can be argued that elderly participation in physical activity triggers a virtuous cycle in the development of physical activity-based health promotion.
The results of our moderator analyses indicated that elderly males and those without a spouse or partner materialized the psychological benefits of physical activity to a greater extent. These results were somewhat related to the male-gatekeeping, gendered culture in sports and physical activity in Korea, where elderly women experience a greater level of participation constraints and difficulty in accessing physical activity. Therefore, policymakers should develop strategies to support facilitators and reduce participation barriers to promote the physical activity-based healthy lifestyle of older adults [46].
There are some noteworthy limitations in this study. As previously stated, the current study purposefully examined the previous studies conducted in Korean populations. Therefore, the results of the current study may not be generalized to other contexts, or there might be some different dose-response relationships between physical activity and subjective well-being measures. In addition, the current study only included studies utilizing a cross-sectional research design, as opposed to experimental studies. Thus, future studies should consider conducting a meta-study on experimental studies concerning the dose-response relationship. Furthermore, the current study conducted moderator analyses using the proportion of gender and marital status per study. Consequently, future studies should consider other socio-economic variables as moderators in the relationship between physical activity and subjective well-being.

Conclusions
In Korea, many older adults believe that dietary supplementation is a better way to maintain their health; elderly people have limited confidence in the health benefits of physical activity. However, the results of the current meta-study clearly indicate that regular physical activity brings psychological health benefits to older adults. In particular, physical activity in elderly people brings a greater level of subjective well-being in terms of self-efficacy, which is an important element of sustainable participation in physical activity. Of the three physical activity doses, the duration of physical activity is the most critical element when designing a physical activity prescription for older adults. In the Korean context, male older adults and elderly people without a spouse or partner experience higher levels of physical activity-related benefits. Consequently, health and human services agencies should develop policies and invest resources to maximize the benefits of physical activity for elderly women and couples.