Trajectories of Leisure Activity and Disability in Older Adults Over 11 Years in Taiwan

Abstract

We aimed to identify leisure activity (LA) trajectories and examined the association among baseline characteristics, LA trajectories, and the later disability among older Taiwanese adults. Data were from the Taiwan Longitudinal Study on Aging Survey for the years 1996-2007 (N = 3,186). LA trajectories were identified by using latent class growth curve modeling. Regression analyses were applied to predict the relationships among baseline characteristics, LA trajectories, and disability. Four LA trajectories—consistent high, consistent low, increasing, and decreasing—were identified. Lower depressive symptom was related to consistently active in LAs. Younger age and fewer comorbidities were related to develop an increasing LA trajectory. Participants in the consistent-high or increasing LA trajectories were more likely to be functionally independent, but those in the decreasing LA subgroup were more at risk of developing disability. The findings suggested that long-term changes in LA over time have benefits on physical health in older population.

Keywords

disability leisure activities Taiwan trajectories

Introduction

Leisure activities (LAs), including both physical and recreational subtypes, have been shown to benefit health (Janke, Payne, & Van Puymbroeck, 2008; World Health Organization [WHO], 2010). The literature supports a positive relationship between physical LAs and better functional performance, lower prevalence of cardiovascular disease, and decreasing mortality rate (Boyle, Buchman, Wilson, Bienias, & Bennett, 2007; Palmer, Espino, Dergance, Becho, & Markides, 2012; Petersen et al., 2012; Tak, Kuiper, Chorus, & Hopman-Rock, 2013). There is also a growing body of research that indicates that recreational LAs increase social connectedness, thus improving the individual’s health and quality of life (Agahi, Lennartsson, Kareholt, & Shaw, 2013; Janke et al., 2008; O’Neil & Dogra, 2016).

LA is commonly understood to be different from work and used to pursue an individual’s relaxation, social achievement, or personal development (Shaun, 2010). Patterns of LA vary by age group. The patterns of LA participation in late life appear to be diverse, ranging from physical activities, lifelong learning, and caring for grandchildren, to any activities which fulfill higher physical, cognitive, and psychosocial needs (Rojek, Shaw, & Veal, 2006). It may be considered that a broad category of LAs which includes both physical and recreational LAs could be closer to the real-life experience of older adults. The WHO (2010) also recommends that older adults, an age group expected to grow worldwide, should be encouraged to participate in both physical and recreational activities in any context, whether at the individual, family, or community level, to optimize opportunities for health. Therefore, the effects of LAs on older adults’ health need to be fully examined.

Past studies have noted the heterogeneous patterns of LA behavioral changes (Jackson & Dunn, 1988; McGuire, Yeh, O’Leary, & Dottavio, 1989; Searle, Mactavish, & Brayley, 1993). The findings originally provided by Jackson and Dunn (1988) and replicated by other researchers suggest that although the majority of people tend to continue the same level of LAs throughout their life span, even in old age, some people change their LA behaviors and may discontinue, replace, or add new LAs (Jackson & Dunn, 1988; McGuire et al., 1989; Searle et al., 1993). Zimmer, Hickey, and Searle (1995, 1997) extended the work by Jackson and Dunn and found that physical impairments were associated with different patterns of LA changes among older adults with arthritis. They suggested that other factors, such as age, may also be associated with patterns of change in LAs and that older adults with physical impairment should be encouraged to continue LAs to pursue better well-being (Zimmer et al., 1995, 1997). These studies have highlighted the need to investigate how LA patterns may change when studying the factors that interplay with leisure behaviors and health outcomes.

Disability is a commonly used indicator of the health of older adults (Nagi, 1965; Verbrugge & Jette, 1994). According to the disablement process framework proposed by Verbrugge and Jette (1994), multiple factors act in moderating the speed of progression toward disability. Previous studies have found that demographic characteristics and preexisting health status play the main roles in the development of disability in older adults. For example, being older, being female, having less educational attainment, and having worse physiological and psychological well-being have been found to lead to a higher probability of disability (Yu, Chen, Chiang, Tu, & Chen, 2015). These preexisting factors are unlikely to change in old age. However, certain modifiable factors could be targeted to encourage older adults to change their lifestyles to maximize their functional capacities and prevent deterioration (Yu et al., 2015). One of these modifiable factors is LAs. LAs have been consistently found to have an effect on maintaining an individual’s functional ability. It is possible that LAs benefit older adults’ physiological well-being, in turn helping to reduce the risk of developing disabilities (Boyle et al., 2007; Tak et al., 2013; Yu et al., 2015). Another reason is that LAs improve older adults’ psychosocial function (O’Neil & Dogra, 2016), and such improvements may help to decrease the risk of disability (Agahi et al., 2013).

Although many studies have found positive relationships between LAs and older adults’ health, to our knowledge, the effects of changes in LA patterns over time on disability are not fully understood (Janke et al., 2008; Strain, Grabusic, Searle, & Dunn, 2002). Past studies have reported the heterogeneity of LA patterns over time among older adults, but the measures used in those studies were limited only to the physical subtype (Barnett, Gauvin, Craig, & Katzmarzyk, 2007, 2008; Pan, Hsu, Chang, & Luh, 2014). Other research has found that recreational LAs also affect the disablement process among older adults (Janke et al., 2008). Taken together, the previous studies suggest that heterogeneity exists in both physical and recreational subtypes of LAs, but there is insufficient evidence to test both domains in one concept. To fill this gap, we hypothesized that the combination of physical and recreational subtypes of LAs and their changes over time, or trajectories, might lead to different disability outcomes in older adults.

In Taiwan, adults aged 65 and above currently represent 11% of the overall population, with that proportion projected to exceed 20% by 2025 (National Development Council in Taiwan, 2012). The Survey for the Elderly in Taiwan showed that around 30% of older Taiwanese did not participate in any type of LA in their daily life, and less than 50% of older adults had exercised regularly in the previous 2 weeks (Ministry of the Interior in Taiwan, 2009; National Health Research Institutes in Taiwan, 2009). The government of Taiwan has recently launched a long-term care policy that encourages older adults to actively participate in LAs, whether physical, recreational, or social, to achieve the goals of active aging and delaying disability among the rapidly growing older population (The Executive Yuan in Taiwan, 2015). Understanding determinants that affect older adults’ LA behaviors seems important to efficient policy.

As the previous studies suggest researchers should note the heterogeneous patterns of LA behavioral changes and the WHO recommends that older adults engage in both physical and recreational activities as much as possible (Jackson & Dunn, 1988; McGuire et al., 1989; Searle et al., 1993; WHO, 2010), our study extends previous knowledge by combining physical and recreational LAs and examining the effects of both of the LA trajectories on older adults’ health. The aims of this study were (a) to identify distinct subgroups of LA trajectories among an older population in Taiwan drawn from a national longitudinal survey, (b) to examine the related factors which might predict the development of different LA trajectories in older adults, and (c) to assess the impact of the LA trajectories on the later development of disabilities in activities of daily living (ADLs) and instrumental activities of daily living (IADLs) in Taiwanese older adults.

Method

Study Population

Data were taken from the Taiwan Longitudinal Study on Aging (TLSA), a national representative survey initiated in 1989 and followed up in 1993, 1996, 1999, 2003, and 2007. A total of six waves of surveys were completed by trained interviewers, and cases of nonresponse were linked to national death registrations for the purpose of recording deaths or loss to follow-up during the surveys. The TLSA collected three aged-in cohorts to represent the distribution of both community- and institution-dwelling older adults in Taiwan. The first cohort was a sample of 4,049 adults aged 60 and older, who were interviewed in the first wave in 1989. Two supplemental cohorts aged 50 to 66 (2,462 older adults in 1996 and 1,599 older adults in 2003) were added in the third and fifth waves. The TLSA had a high response rate of around 90%.

In the current study, we analyzed the first two cohorts of adults aged 50 and older in the first wave of 1996 and traced their LA changes until 2007 to meet the statistical requirements for estimating a trajectory. The third cohort was not included because they completed only two waves of surveys, which was insufficient for the estimation of trajectories. In addition, to ensure the robustness of our results, we included only respondents who completed at least three of the four surveys (1996, 1999, 2003, and 2007). In the present study, a total of 3,186 respondents were analyzed, including 1,256 respondents from the original sample of 4,090 in 1989 and 1,930 respondents from the supplemental sample of 2,462 in 1996.

Measures

LA trajectories

The main variables for estimating LA trajectories were measured by asking respondents whether they regularly engaged in any kinds of LAs in their daily life, including listening to the radio, reading newspapers or books, playing chess or poker, chatting with friends, gardening, taking a walk, engaging in outdoor activities, and participating in group activities such as dancing or Tai-Chi. Each of the items for LAs was dichotomized into either no (0) or regular (1) engagement in the activity in daily life. A sum score ranging from 0 to 8 was calculated consistently in every wave (1996, 1999, 2003, and 2007). The rates of nonresponse for LAs in the surveys were 3.45% (1996), 5.81% (1999), 2.82% (2003), and 4.46% (2007). Multiple imputations were applied for missing data, as recommended for dealing with missing survey data (Kmetic, Joseph, Berger, & Tenenhouse, 2002; Rubin, 1987; Schafer, 1997). The multiple imputation analyses were generated using SAS software, version 9.2, of the SAS System for PROC MI (SAS Institute Inc., Cary, NC, USA).

Disabilities

The outcome variables, difficulties in ADLs and IADLs, were assessed in the last wave of 2007 to identify the impact of LA trajectory on disabilities. A total of 12 items, six on ADLs and six on IADLs, were assessed. They were as follows: bathing, eating, dressing, standing up from a chair and bed, indoor walking, toileting, managing money, shopping, taking public transportation, doing light housework, doing heavy housework, and telephoning. Each of the items of ADL and IADL disability was coded as no difficulty (0) or any difficulty (1) in performing the task. We combined ADL and IADL disabilities together into a simple sum to enhance the sensitivity of scaling, as recommended by a previous study (Spector & Fleishman, 1998). A sum score of ADL and IADL disabilities ranging from 0 to 12 was calculated for the wave of 2007.

Baseline variables

Other measures related to older adults’ disabilities, such as demographic variables, physiological and psychological well-being, and social relations, were extracted from the baseline wave of 1996. The demographic variables included age, gender, and years of education. Physiological conditions were assessed by the number of comorbidities out of 14 disorders: hypertension, diabetes mellitus, heart disease, stroke, cancer, pulmonary disease, arthritis, gastric ulcer, liver disease, hip fracture, cataract, renal disease, gout, and spinal spurs. Psychological condition was assessed with the 10-item version of the Center for Epidemiological Studies Depression Scale (CESD; Radloff, 1977), which represents levels of depressive symptoms along a score range of 0 to 30. Social relations were assessed by the individual’s social network and social support. The social network score was calculated according to the weekly frequency of contact with relatives and friends. Social support was assessed with four items measuring the level of satisfaction with emotional support. The social support scores ranged from 4 to 20, with higher scores representing greater satisfaction with support.

Analysis

Identifying the optimal number of different LA trajectories

Growth mixture modeling was used to identify the LA trajectories of older adults in Taiwan from 1996 to 2007. We used linear latent class growth curve modeling to estimate four waves of LA changes over time (1996, 1999, 2003, and 2007). Several goodness of fit indices have been suggested to test the optimal number of trajectories in a model, three of them being the Bayesian information criteria (BIC), the Vuong–Lo–Mendell–Rubin likelihood ratio test (VLMR-LRT), and the Lo–Mendell–Rubin adjusted likelihood ratio test (LMR-adj-LRT; Nylund, Asparouhov, & Muthén, 2007). A lower value of BIC represents a better model fit. The VLMR-LRT and LMR-adj-LRT indices were used to test whether class k was better fitted to the data than class k−1 with a significant p value <.05. The software package used for analysis was Mplus version 7 (Muthén & Muthén, 1998-2012).

Examining the factors associated with LA trajectories and the impact of LA trajectories on later ADL and IADL disabilities

To assess the relationships of the baseline factors which might predict older adults developing different LA trajectory patterns, we conducted a multinomial logistic regression analysis and estimated the odds ratio (OR) and the 95% confidential interval (CI) for each factor in the equation. To test the main hypothesis, that distinct LA trajectories affect later ADL and IADL disabilities, we used a multivariate linear regression model and estimated the coefficients and p value. A significant p value was set at <.05. The sample weight was adjusted in the equation based on the recommendation from the TLSA. The analyses were generated using SAS software, version 9.2, of the SAS System for PROC LOGISTIC and PROC REG (SAS Institute Inc.).

Results

Identification of Four LA Trajectories Among a Sample of Older Adults in Taiwan

We specified a linear latent class growth curve model for estimating four waves of LA changes over time. As shown in Table 1, we examined two to five LA trajectories in a sample of older individuals who survived over 11 years in Taiwan. The optimal number we considered was four subgroups of LAs, for two reasons. First, although the BIC value declined as the number of trajectories increased, we found a relatively lower value with four trajectories, and the values became diminished afterward. Second, the p values of VLMR-LRT and LMR-adj-LRT were significant for both three and four trajectories. As the indices recommended, three trajectories might fit better than two, according to the significant p value. However, four trajectories might also fit better than three, as indicated by the significant p value. In summary, the lowest BIC value and the significant p value were found for four LA trajectories, so we decided that four trajectories would be the best choice for determining LA trajectories among a sample of older individuals surviving over 11 years in Taiwan. Figure 1 shows the membership probabilities and the patterns of LA trajectories among a sample of 3,186 older adults in Taiwan. The LA trajectories included four distinct subtypes: consistent-high LA, increasing LA, decreasing LA, and consistent-low LA, with 722 (22.66%), 84 (2.64%), 201 (6.31%), and 2,179 (68.39%) adults in each subgroup, respectively.

Table 1.

Model Testing for the LA Trajectories Across Two to Five Classes in a Sample of Older Individuals Over 11 Years in Taiwan.

No. of trajectories	BIC	VLMR-LRT	LMR-adj-LRT	Membership probabilities in the models
2	44,915.4	NA	NA	2,284 (71.69)	902 (28.31)
3	44,885.6	<.001***	<.001***	2,289 (71.85)	745 (23.38)	152 (4.77)
4	44,867.5	<.001***	<.001***	2,179 (68.39)	722 (22.66)	201 (6.31)	84 (2.64)
5	44,863.2	.058	.062	1,893 (59.42)	827 (25.96)	234 (7.35)	120 (3.77)	112 (3.52)

Note. Statistics of membership probabilities in the models are shown as frequency (%). LA = leisure activity; BIC = Bayesian information criteria; VLMR-LRT = Vuong–Lo–Mendell–Rubin likelihood ratio test; LMR-adj-LRT = Lo–Mendell–Rubin adjusted likelihood ratio test; NA = not available.

p < .05. **p < .01. ***p < .001.

Figure 1.

Four LA trajectories among a sample of older individuals who survived an 11-year period in Taiwan.

The consistent-high LA trajectory had an average number of four to five LAs and remained high throughout the study period. The increasing LA trajectory had a low number of LAs at the beginning but steadily increased afterward, even exceeding the consistent-high LA subgroup in 2007. In contrast, the older adults in the decreasing LA subgroup started with a relatively high level of LA, similar to those in the consistent-high LA subgroup, but their LA levels rapidly declined at the end of the observation period. Finally, the consistent-low LA subgroup had a consistently low number, with an average of two LAs throughout the study period.

The Relationship of Baseline Factors to Distinct LA Trajectories

A total of 3,186 older adults were analyzed to identify the effects of the baseline factors on four distinct LA trajectories in Taiwan. The sample characteristics for each LA trajectory at baseline are presented in Table 2, and the results of baseline factors on predicting different trajectories in older adults are shown in Table 3. In Table 2, all examined factors differed significantly among the four LA trajectories. Older adults in the consistent-high LA trajectory were more likely to be male (70.78%) and to have higher educational attainment (M = 8.73, SD = 4.40), reported fewer depressive syndrome (M = 3.21, SD = 3.89), and had higher scores for social networking (M = 21.49, SD = 18.58) and satisfaction with their social support (M = 17.03, SD = 2.60). Those in the increasing LA trajectory were younger (M = 58.73, SD = 6.15) and had fewer comorbidities (M = 0.80, SD = 1.12) at baseline. Those in the decreasing LA trajectory had a higher probability of being older (M = 66.58, SD = 7.58) and more comorbidities (M = 1.56, SD = 1.53). Finally, older adults in the consistent-low LA subgroup were more likely to be female (58.19%) and to have less educational attainment (M = 3.52, SD = 3.91), worse depressive syndrome (M = 5.76, SD = 5.75), and the lowest scores for social networking (M = 18.57, SD = 16.82) and satisfaction with their social support (M = 15.99, SD = 2.95).

Table 2.

Sample Characteristics at Baseline in Each LA Trajectory (N = 3,186).

		Trajectories
Sample characteristics	Total N = 3,186	Consistent-high LA (n = 722)	Increasing LA (n = 84)	Decreasing LA (n = 201)	Consistent-low LA (n = 2,179)
Age***	63.89 (8.17)	62.36 (7.68)	58.73 (6.15)	66.58 (7.58)	64.35 (8.29)
Female***	1,593 (50.00)	211 (29.22)	46 (54.76)	68 (33.83)	1,268 (58.19)
Educational years (0-17)***	4.98 (4.62)	8.73 (4.40)	5.90 (4.50)	7.01 (4.09)	3.52 (3.91)
Comorbidities (0-14)***	1.40 (1.52)	1.17 (1.36)	0.80 (1.12)	1.56 (1.53)	1.48 (1.57)
CESD (0-30)***	5.07 (5.40)	3.21 (3.89)	4.88 (5.00)	4.35 (4.52)	5.76 (5.75)
Social network***	19.33 (16.99)	21.49 (18.58)	18.89 (12.59)	20.01 (13.64)	18.57 (16.82)
Social support (4-20)***	16.28 (2.89)	17.03 (2.60)	16.14 (2.81)	16.77 (2.73)	15.99 (2.95)

Note. Statistics are shown as mean (SD) for continuous variables and as frequency (%) for categorical variables. The significances among the four LA trajectories were examined by one-way ANOVA or chi-square test. LA = leisure activity; CESD = Center for Epidemiological Studies Depression Scale.

p < .05. **p < .01. ***p < .001.

Table 3.

Effects of Baseline Factors on LA Trajectories Among a Sample of Older Adults Over 11 Years in Taiwan (N = 3,186).

Sample characteristics^a	Trajectories^b
Sample characteristics^a	Consistent-high LA (n = 722)	Increasing LA (n = 84)	Decreasing LA (n = 201)
Age	0.996 [0.983, 1.009]	0.938 [0.908, 0.969]	1.052 [1.031, 1.073]
Female	0.711 [0.584, 0.866]	1.325 [0.875, 2.008]	0.644 [0.459, 0.904]
Educational years (0-17)	1.290 [1.260, 1.322]	1.136 [1.081, 1.194]	1.204 [1.160, 1.249]
Comorbidity (0-14)	0.988 [0.922, 1.060]	0.747 [0.620, 0.961]	1.060 [0.955, 1.175]
CESD (0-30)	0.940 [0.918, 0.961]	1.010 [0.970, 1.050]	0.983 [0.951, 1.107]
Social network	1.002 [0.997, 1.007]	0.994 [0.980, 1.007]	1.002 [0.994, 1.010]
Social support (4-20)	1.076 [1.038, 1.116]	0.998 [0.928, 1.073]	1.065 [1.005, 1.130]

Note. Statistics are shown as OR (95% confidence interval). LA = leisure activity; CESD = Center for Epidemiological Studies Depression Scale; OR = odds ratio.

The statistics of OR (95% confidence interval) were waved to the nearest thousandth (0.001) due to the potentially overestimated wave off to the second decimal place.

The consistent-low LA trajectory was set as the reference group.

Table 3 presents the results of multinomial logistic regression. We used the trajectory of consistent-low LA as the reference and analyzed the relationships of the baseline factors that predicted that older Taiwanese would develop different LA trajectories. As expected, older adults in the consistent-high LA subgroup were more likely to be male (OR = 0.711, CI = [0.584, 0.866]) and to have higher educational attainment (OR = 1.290, CI = [1.260, 1.322]), less depression (OR = 0.940, CI = [0.918, 0.961]), and greater satisfaction with social support (OR = 1.076, CI = [1.038, 1.116]) than their counterparts in the consistent-low LA subgroup. Those with a younger age (OR = 0.938, CI = [0.908, 0.969]), higher educational attainment (OR = 1.136, CI = [1.081, 1.194]), and fewer comorbidities (OR = 0.747, CI = [0.620, 0.961]) had a higher probability developing the trajectory of increasing LA. Finally, as compared with the referenced consistent-low LA subgroup, older adults in the decreasing LA trajectory tended to be older (OR = 1.052, CI = [1.031, 1.073]), male (OR = 0.644, CI = [0.459, 0.904]), more highly educated (OR = 1.204, CI = [1.160, 1.249]), and more satisfied with their social support (OR = 1.065, CI = [1.005, 1.130]).

LA Trajectories and Later ADL and IADL Disabilities

The results of whether different LA trajectories predicted the development of later ADL and IADL disabilities in a sample of Taiwanese older adults who survived for 11 years are shown in Table 4. On comparison with the referenced consistent-low LA subgroup, the older adults in the consistent-high and increasing LA trajectory groups had lower probabilities of becoming disabled; the estimates were as follows: β = −0.99 (SE = 0.13) and β = −0.88 (SE = 0.27), respectively. The older adults in the decreasing LA trajectory group were more likely to become disabled (β = 1.27, SE = 0.21).

Table 4.

The Impact of Distinct LA Trajectories on Later ADL and IADL Disabilities Among a Sample of Older Adults Over 11 Years in Taiwan (N = 3,186).

	Trajectories
	Consistent-high LA (n = 722)	Increasing LA (n = 84)	Decreasing LA (n = 201)	Consistent-low LA^a (n = 2,179)
Development of disabilities^b	−0.99 (0.13)***	−0.88 (0.27)**	1.27 (0.21)***	Ref.

Note. Statistics are shown as estimates (SE). LA = leisure activity; ADL = activities of daily living; IADL = instrumental activities of daily living.

The consistent-low LA trajectory was set as the reference group.

The model was adjusted for baseline age, gender, educational years, number of comorbidities, Center for Epidemiological Studies Depression Scale levels, social support, and baseline number of disabilities.

p < .05. **p < .01. ***p < .001.

Discussion

It is well recognized that LAs benefit older adults’ health. The major contribution of this study is that it links the baseline characteristics, distinct LA trajectories, and the later development of ADL and IADL disabilities among older adults in Taiwan based on an 11-year follow-up survey. To our knowledge, this is the first study to examine four distinct LA trajectories among Taiwanese older adults: consistent high, increasing, decreasing, and consistent low. Our study showed that participants with different baseline characteristics developed different LA trajectories. Among the older participants, those with consistent-high and increasing LA trajectories were more likely to retain their functional independence later in life. We extend previous knowledge that connected long-term changes in LAs over time, or LA trajectory, to older adults’ physical function, and the results supported the benefits of LAs on the physical health of older adults in Taiwan.

This study identified four distinct LA trajectories. Most of the participants had trajectories of consistent-low LA (68.39%), followed by those with the trajectories of consistent-high LA (22.66%), decreasing LA (6.31%), and increasing LA (2.64%). It should be noted that most Taiwanese older adults have consistent-low lifestyles. Previous studies reported that low LA might lead to negative health outcomes in older adults (Agahi et al., 2013; Tak et al., 2013; Thorp, Owen, Neuhaus, & Dunstan, 2011). Our longitudinal analysis showed that almost 70% of the older adults had consistent-low trajectories. This high percentage might raise concerns for the future health care and long-term care systems in Taiwan. A recent national cross-sectional survey indicated that up to 30% of older Taiwanese did not participate in any type of LA in their daily life (Ministry of the Interior in Taiwan, 2009). Considering the low percentages of older adults participating in daily LAs found in longitudinal and cross-sectional surveys, the health division of the Taiwan government should actively promote the engagement in LAs to older adults to optimize their health.

The present study also found heterogeneous patterns of LA trajectories over time. Our results were fairly consistent with those of previous studies, even though the measures used in the present study incorporated both physical and recreational LAs instead of only the physical subtype or exercise, as done in previous studies (Barnett et al., 2008; Pan et al., 2014). The number and the patterns of LA trajectories in our study were similar to past findings of consistently low, consistently high, increasing, and decreasing sub-patterns (Barnett et al., 2008; Pan et al., 2014). However, we included recreational LAs in an attempt to test the broader concept of activities for improving older adults’ health recommended by the WHO (2010). Researchers have also suggested that not only physical LAs but also recreational subtypes confer benefits on older adults’ health (Janke et al., 2008; O’Neil & Dogra, 2016; Sampaio & Ito, 2013). As per the recommendations of the WHO, older adults are encouraged to engage in physical and recreational activities of any kind to optimize their health. We believe that the incorporation of these two types of LAs more accurately reflects the lifestyles of older adults.

Furthermore, previous studies have found that the majority of people tend to continue the same level of LA throughout their life span, even into old age (Jackson & Dunn, 1988; McGuire et al., 1989; Searle et al., 1993). Our study validated this finding with empirical evidence that most of the older Taiwanese in this study continued a stable pattern of LAs from middle to old age, in consistent-high or consistent-low LA trajectories. This study adds to the literature new evidence that the two different consistent LA trajectories had opposite consequences toward later functional disability. Older participants who had a consistent-high LA trajectory were found to have better physical health in later life.

For older Taiwanese adults to have ceased or decreased the number of LAs they participated in as they aged (decreasing LA trajectory), one possible reason can be explained by the concept of selective optimization with compensation (SOC; Baltes, 1997; Baltes & Baltes, 1990). SOC suggests that older adults may adapt to the aging process by selecting fewer activities that are more meaningful, maximizing their participation in the selected activities, and compensating for lost abilities. However, our study found the consequence of a decreasing number of LAs (decreasing LA trajectory) among older adults to be negative impact to physical health. While SOC (Baltes, 1997; Baltes & Baltes, 1990) suggests that decreases in the number of LAs may be a practical choice, our study warns that this choice could result in disadvantages for physical function in later life. In contrast, older Taiwanese who increased their LAs while aging (increasing LA trajectory) had a higher probability of delaying disability. Based on our findings, not only the pattern of changes in LA behaviors is important but also the number of types of LA is an important factor in older adults’ disability development. Future studies should consider investigating the amount to which individuals participate in their LAs. Policy makers will need to understand the possible reasons for greater and lesser engagement in LAs to promote LAs as a strategy for better health in late life.

Our study findings suggest that LA trajectories affected the later development of disability among older Taiwanese, even when baseline disability is included. Past studies have mentioned that baseline functional disability might not necessarily be the determinant of LA changes over time (Strain et al., 2002; Thomas, 2011). Thomas (2011) found that baseline physical limitations had no relationship with the development of different LA trajectories, especially for social LAs. Strain et al. (2002) reported that baseline disability had little effect on older adults continuing LAs. Our study showed that initial disability may have a limited influence on the relationship between LA classes and later disability. In contrast, whether older people had initial disability or not in four LA trajectory groups, later development of different LA trajectories played an important role in determining older adults’ chances of later disability. Some disabled older adults were still able to maintain or increase their LA engagement, and those who did still benefited from better physical function later in life. It merits attention due to the potential benefit of encouraging older adults to participate more in LAs to increase their physical health, even if they already have disability. Further investigation into the dynamic relationships between changes in LA and changes in disability is strongly recommended.

Older adults with a history of either consistent-high LAs or an increasing number of LAs throughout the study period had less likelihood of being functionally dependent in later life. In other words, it is not only that high LA participation among older Taiwanese adults was associated with less concurrent disability but also that older Taiwanese who had lower LA participation at earlier time periods but then steadily increased their participation eventually attained similar benefits in terms of functional health as those older Taiwanese who maintained higher LA throughout the study period. In contrast, the group of people who began with higher LA but rapidly decreased their participation ended up with worse functional health, and the impact of decreasing LA participation on disability was similar to that of ongoing low LA participation. We might conclude that the benefit of LA to functional independence in older adults not only takes place in current time but also has a positive effect on older adults’ physical health in later life. It is important to encourage older adults, no matter when, to increase their LA engagement to benefit from less functional disability.

The disablement process proposes that multiple factors act to moderate the speed of progression toward disability (Verbrugge & Jette, 1994). LA is one of the factors that has been considered beneficial to older adults’ physical health (Boyle et al., 2007; Tak et al., 2013; Yu et al., 2015). Our study findings further suggest that patterns of LA over time may also play an important role in the disablement process. Past studies have suggested that it is the changes in patterns of LA, and not the static status of LA engagement, that should be measured (Jackson & Dunn, 1988; McGuire et al., 1989; Searle et al., 1993). Our study not only validates LA as an important disablement factor but also validates the importance of changes in LA patterns. Older Taiwanese adults who had consistently high LA participation were less likely to develop disability. Older adults who started out engaging in lower amounts of LA but then steadily increased their level of activity as they aged enjoyed the same positive effect of postponing physical disability. Our study replicated the framework of the disablement process (Verbrugge & Jette, 1994) and indicated that change in LA plays an important role in slowing the progression toward disability.

We suggest that multiple factors affected the development of LA trajectories in Taiwanese older adults and that these trajectories led to different capacities of physical function at the 11-year follow-up. First, older adults who began with less depression were more likely to be actively engaged in LAs throughout the study period and in the end to have more potential to be functionally independent. A previous study reported that better psychosocial well-being was associated with frequent LA engagement later in life (Kaplan, Newsom, McFarland, & Lu, 2001). Our study further verified that mental health plays an important role in keeping an older population active in LAs over time, which would finally lead to better physical function. The policy makers of public health should increase efforts to screen for psychological well-being in the older population to encourage better physical health.

Second, the management of chronic disease should be also addressed due to its protective effect on keeping older adults on an increasing LA trajectory, which will result in better physical function. As expected, people with chronic conditions were less likely to engage in LAs (Ashe, Miller, Eng, & Noreau, 2009), so they had less potential for entering a trajectory of increasing LA. Although the causal relationship of whether LAs act on chronic disease or vice versa requires further study, we might conclude, based on our findings, that the management of chronic disease is essential for older adults to maintain an active lifestyle and thus better function.

Third, the issues of age, LA trajectory, and later physical disability merit discussion. In the present study, participants in the increasing LA subgroup were younger than those in the decreasing and consistent-low LA subgroups; that is, young-old people had more potential than their old-old counterparts to increase their participation in LAs later in life. In addition, participants in the increasing LA trajectory subgroup tended to have better physical health at the 11-year longitudinal follow-up. Although age inevitably increases during the aging process, we noted that the effect of age on LAs reached a plateau across the different age groups. In one study of adults, age was positively related to increases in LA participation (Barnett et al., 2007, 2008), but it was found to have a negative relationship to LA participation in a sample of older adults in the study of Strain et al. (2002) and in ours. This discrepancy may indicate that people constantly increase their levels of LA during adulthood, even continuing their more active lifestyles into young-old age, as was the case in the increasing LA subgroup identified in our study, but the advantage of age on increasing the number of LAs diminished afterward. This can be explained by the nature of biological decline among older adults and social disengagement in older-old adults (Carstensen, Isaacowitz, & Charles, 1999) because old-old people tend to constrain their activities, whether physical or recreational. Nevertheless, in spite of the nature of age affecting the development of LA trajectories and later physical function in older adults, enrichment of LA participation appears to be a good strategy for promoting the health of older adults. The younger-old generation in Taiwan should have more awareness of the advantage of LAs on health and keep maintaining an active lifestyle in late life.

Fourth, higher educational attainment was found to be positively related to engagement in LAs in the active LA subgroup as compared with the consistent-low LA subgroup. The cumulative advantages of education on health are well supported in the literature. For example, people with greater educational attainment tend to lead healthier lifestyles and have access to richer resources, thus increasing their likelihood of maintaining an active LA trajectory and finally achieving better health (Kim & Durden, 2007; Ross & Wu, 1996). The benefits of education to older adults in keeping active rather than developing an inactive lifestyle in the long run are further supported by our findings.

Besides the importance of the above related baseline characteristics, we also noted that gender and satisfaction with social support played roles in determining the long-term effects on LA trajectory and disabilities. Participants who were male and satisfied with their social support in the beginning of the survey period were more likely to develop trajectories of consistent-high and decreasing LAs, but the later development of physical disabilities was opposite in the consistent-high and decreasing LA subgroups. This diversity should be considered carefully. For the issue of gender, findings on the impact of gender on LA have been mixed in past studies (Barnett et al., 2007, 2008; Tsunoda et al., 2013), but we found that males tended to be more active than females at baseline in the present study. Nevertheless, females, rather than males, were notably inactive in LA participation in our study and in most surveys (Dumith, Hallal, Reis, & Kohl, 2011). Females should be encouraged to increase their involvement in LAs of any kind to achieve better functional outcomes in later life. Social support helped older adults maintain long-term active LA participation and finally resulted in better functional health. It is expected that people who feel satisfied with their social relationships will have better social connectedness and therefore stronger involvement in LAs over the long term, followed by better health later in life (O’Neil & Dogra, 2016). Surprisingly, participants in the decreasing LA trajectory were also noted to have greater satisfaction with social support at baseline, and we knew that people in the decreasing LA trajectory were at greater risk of functional disabilities in the end. We conducted a supplementary analysis in an attempt to explain the potential factors that would offset the effects of social support on long-term LA in the decreasing LA trajectory subgroup. Our supplementary analysis indicated that the effect of depressive symptoms on determining older adults’ development of a later LA trajectory was greater than that of social support, explaining an additional 3.10% of the variance. Thus, even if older people started with similar levels of satisfaction with social support, those with poor mental health would be at a higher risk of decline in LA involvement, which in turn would lead to functional disability, as compared with their counterparts who had better mental health. Again, public health practice should increase efforts to screen for mental health to help older people maintain a consistently active lifestyle.

In brief, we tried to connect the baseline characteristics, LA trajectories, and final physical function in a sample of Taiwanese older adults. However, whether LA acts as a mediator between baseline characteristics and disabilities could not be confirmed in this study. It is strongly recommended that further studies focus on the mechanism among these related factors in this targeted population.

Limitations

Several limitations of this study should be noted. First, the main variable, LA, was a combination of both physical and recreational subtypes in this study. Previous studies have noted discrepancies in measuring different subtypes of LAs. To examine the potential discrepancies, we conducted an analysis and found that the trends for only physical or only recreational LA engagement among older adults were similar to trends for both types of LA combined, as shown in Supplemental Table 1. Second, the sample included only those participants who survived through 2007 and completed the survey at least 3 times. It is likely that the participants in our study tended to be healthier and engaged in more active lifestyles than those who were lost to follow-up due to death or major disability. Thus, the estimation of LA trajectories may have been overestimated. Third, other potential factors that might affect the LA trajectories of older adults, such as the environmental component of access to health care facilities, were not included due to the limitations of secondary data analysis. Further research should address the contributions of these factors to LA trajectories among older adults in Taiwan. Fourth, we employed separate analytical methods to link the baseline characteristics, LA trajectories, and later physical disability. Further studies should test the relationships of these factors in the same model simultaneously to prevent potential methodological bias. In addition, we note that the regression models in our study showed a reasonable statistical significance. Although an exciting outcome was found for older adults who began with low LA but increased the number of LAs gradually (increasing LA trajectory), the small sample size in this subgroup may have resulted in overestimation. Future studies with larger sample sizes are recommended to obtain more reliable results. Finally, although we included baseline disabilities in the model for the purpose of controlling for physical health at the initial stage, we recommend that future researchers remain cautious about causal interpretations of these relationships.

Conclusion

In Taiwan, the National Long-Term Care Needs Survey estimated that an average of 14.95% of Taiwanese older adults aged 65 and older in 2011 experienced disability (Lee et al., 2013). As it is projected that the older population in Taiwan will grow rapidly, the number of disabled older adults could double in the next two decades (Lee et al., 2013; National Development Council in Taiwan, 2012). To face this issue, one priority in the long-term care policy in Taiwan should be to determine factors that might postpone disability in older adults.

A strategic understanding of specific factors than can postpone disability in older Taiwanese adults may make the delivery of health care services more productive and efficient. In the present study, we determined that LAs played an important role in determining the development of disability in older adults. Keeping consistently engaged in active LAs and increasing the levels of LA engagement late in life increase the probability that older adults will retain their functional independence. Therefore, older adults should be encouraged to participate in LAs, whether physical or recreational, depending on the context of the individual, for the benefit of their functional capacity, family, and community.

Supplemental Material

Supplemental_Table – Supplemental material for Trajectories of Leisure Activity and Disability in Older Adults Over 11 Years in Taiwan

Supplemental material, Supplemental_Table for Trajectories of Leisure Activity and Disability in Older Adults Over 11 Years in Taiwan by Hsiao-Wei Yu, Tung-liang Chiang, Duan-Rung Chen, Yu-Kang Tu, and Ya-Mei Chen in Journal of Applied Gerontology

Footnotes

Acknowledgements

The authors thank the anonymous reviewers and the editor for their constructive comments.

Authors’ Note

The current study has been approved by the Research Ethics Committee of National Taiwan University (2013HS064).

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors wish to express their gratitude to the Ministry of Science and Technology in Taiwan for its generous financial support (MOST101-2410-H-002-212-MY2, MOST103-2410-H-002-146) and the Health and Welfare Data Science Center at Ministry of Health and Welfare for its gracious help with data access (H102054).

Author Biographies

Hsiao-Wei Yu is a PhD candidate of health policy and management at National Taiwan University. Her research interests include healthy aging and measurement of physical function among older adults.

Tung-liang Chiang is a professor of health policy and management at National Taiwan University. His research field includes health care policy and medical sociology.

Duan-Rung Chen is a professor of health behaviors and community sciences at National Taiwan University with interests in medical sociology, health behavior, and community intervention.

Yu-Kang Tu is an associate professor of epidemiology at National Taiwan University. His research interests include life course epidemiology and meta-analysis.

Ya-Mei Chen is an assistant professor of health policy and management at National Taiwan University with key interests in healthy aging, long-term care policy, and health promotion among older adults.

References

Agahi

Lennartsson

Kareholt

Shaw

B. A.

(2013). Trajectories of social activities from middle age to old age and late-life disability: A 36-year follow-up. Age Ageing, 42, 790-793. doi:10.1093/ageing/aft140

Ashe

M. C.

Miller

W. C.

Eng

J. J.

Noreau

(2009). Older adults, chronic disease and leisure-time physical activity. Gerontology, 55, 64-72. doi:10.1159/000141518

Baltes

P. B.

(1997). On the incomplete architecture of human ontogeny: Selection, optimization, and compensation as foundation of developmental theory. American Psychologist, 52, 366-380.

Baltes

P. B.

Baltes

M. M.

(1990). Successful aging: Perspectives from the behavioral sciences. New York, NY: Cambridge University Press.

Barnett

T. A.

Gauvin

Craig

C. L.

Katzmarzyk

P. T.

(2007). Modifying effects of sex, age, and education on 22-year trajectory of leisure-time physical activity in a Canadian cohort. Journal of Physical Activity & Health, 4, 153-166.

Barnett

T. A.

Gauvin

Craig

C. L.

Katzmarzyk

P. T.

(2008). Distinct trajectories of leisure time physical activity and predictors of trajectory class membership: A 22 year cohort study. International Journal of Behavioral Nutrition and Physical Activity, 5, Article 57. doi:10.1186/1479-5868-5-57

Boyle

P. A.

Buchman

A. S.

Wilson

R. S.

Bienias

J. L.

Bennett

D. A.

(2007). Physical activity is associated with incident disability in community-based older persons. Journal of the American Geriatrics Society, 55, 195-201. doi:10.1111/j.1532-5415.2007.01038.x

Carstensen

L. L.

Isaacowitz

D. M.

Charles

S. T.

(1999). Taking time seriously: A theory of socioemotional selectivity. American Psychologist, 54, 165-181.

Dumith

S. C.

Hallal

P. C.

Reis

R. S.

Kohl

H. W.

III . (2011). Worldwide prevalence of physical inactivity and its association with human development index in 76 countries. Preventive Medicine, 53, 24-28. doi:10.1016/j.ypmed.2011.02.017

10.

The Executive Yuan in Taiwan. (2015). White paper on aging society. Retrieved from http://www.ey.gov.tw/News_Content4.aspx?n=0AD1AB287792C301&sms=4ACFA38B877F185F&s=AD43B6E1D7406D7C

11.

Jackson

E. L.

Dunn

(1988). Integrating ceasing participation with other aspects of leisure behavior. Journal of Leisure Research, 20(1), 31-45.

12.

Janke

M. C.

Payne

L. L.

Van Puymbroeck

(2008). The role of informal and formal leisure activities in the disablement process. International Journal of Aging & Human Development, 67, 231-257.

13.

Kaplan

M. S.

Newsom

J. T.

McFarland

B. H.

(2001). Demographic and psychosocial correlates of physical activity in late life. American Journal of Preventive Medicine, 21, 306-312.

14.

Kim

Durden

(2007). Socioeconomic status and age trajectories of health. Social Science & Medicine, 65, 2489-2502. doi:10.1016/j.socscimed.2007.07.022

15.

Kmetic

Joseph

Berger

Tenenhouse

(2002). Multiple imputation to account for missing data in a survey: Estimating the prevalence of osteoporosis. Epidemiology, 13, 437-444.

16.

Lee

Y. C.

Lin

L. C.

S. C.

Cheng

W. H.

L. Y.

, & National Long-Term Care Insurance Group. (2013). Plan and vision of national long-term care insurance in Taiwan. Community Development Journal, 141, 26-44.

17.

McGuire

F. A.

Yeh

C.-K.

O’Leary

J. T.

Dottavio

F. D.

(1989). Integrating ceasing participation with other aspects of leisure behavior: A replication and extension. Journal of Leisure Research, 21(4), 316-326.

18.

Ministry of the Interior in Taiwan. (2009). Report of the Senior Citizen Condition Survey. Taipei, Taiwan: Ministry of the Interior in Taiwan.

19.

Muthén

L. K.

Muthén

B. O.

(1998-2012). Mplus user’s guide (7th ed.). Los Angeles, CA: Author.

20.

Nagi

S. Z.

(1965). Some conceptual issues in disability and rehabilitation. Washington, DC: American Sociological Association.

21.

National Development Council in Taiwan. (2012). Population Statistic Projection: 2012-2060. Taipei, Taiwan: National Development Council in Taiwan.

22.

National Health Research Institutes in Taiwan. (2009). National Health Interview Survey. Miaoli, Taiwan: National Health Research Institutes in Taiwan.

23.

Nylund

K. L.

Asparouhov

Muthén

B. O.

(2007). Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling: A Multidisciplinary Journal, 14, 535-569. doi:10.1080/10705510701575396

24.

O’Neil

Dogra

(2016). Different types of sedentary activities and their association with perceived health and wellness among middle-aged and older adults: A cross-sectional analysis. American Journal of Health Promotion, 30, 314-322.

25.

Palmer

R. F.

Espino

D. V.

Dergance

J. M.

Becho

Markides

(2012). The role of physical activity and diabetes status as a moderator: Functional disability among older Mexican Americans. Age Ageing, 41, 752-758. doi:10.1093/ageing/afs106

26.

Pan

L. Y.

Hsu

H. C.

Chang

W. C.

Luh

D. L.

(2014). Trajectories of physical activity and risk factors among Taiwanese older adults. International Journal of Behavioral Medicine, 22, 62-69.

27.

Petersen

C. B.

Gronbaek

Helge

J. W.

Thygesen

L. C.

Schnohr

Tolstrup

J. S.

(2012). Changes in physical activity in leisure time and the risk of myocardial infarction, ischemic heart disease, and all-cause mortality. European Journal of Epidemiology, 27, 91-99. doi:10.1007/s10654-012-9656-z

28.

Radloff

L. S.

(1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385-401. doi:10.1177/014662167700100306

29.

Rojek

Shaw

S. M.

Veal

A. J.

(2006). A handbook of leisure studies. New York, NY: Palgrave Macmillan.

30.

Ross

C. E.

C. L.

(1996). Education, age, and the cumulative advantage in health. Journal of Health and Social Behavior, 37, 104-120.

31.

Rubin

D. B.

(1987). Multiple imputation for nonresponse in surveys. New York, NY: John Wiley.

32.

Sampaio

P. Y.

Ito

(2013). Activities with higher influence on quality of life in older adults in Japan. Occupational Therapy International, 20, 1-10. doi:10.1002/oti.1333

33.

Schafer

J. L.

(1997). Analysis of incomplete multivariate data. London, England: Chapman & Hall.

34.

Searle

M. S.

Mactavish

J. B.

Brayley

R. E.

(1993). Integrating ceasing participation with other aspects of leisure behavior: A replication and extension. Journal of Leisure Research, 25, 389-404.

35.

Shaun

(2010). Leisure studies: Themes and perspectives. Thousand Oaks, CA: SAGE.

36.

Spector

W. D.

Fleishman

J. A.

(1998). Combining activities of daily living with instrumental activities of daily living to measure functional disability. Journals of Gerontology, Series B: Psychological Sciences & Social Sciences, 53, S46-S57.

37.

Strain

L. A.

Grabusic

C. C.

Searle

M. S.

Dunn

N. J.

(2002). Continuing and ceasing leisure activities in later life: A longitudinal study. The Gerontologist, 42, 217-223.

38.

Tak

Kuiper

Chorus

Hopman-Rock

(2013). Prevention of onset and progression of basic ADL disability by physical activity in community dwelling older adults: A meta-analysis. Ageing Research Reviews, 12, 329-338. doi:10.1016/j.arr.2012.10.001

39.

Thomas

P. A.

(2011). Trajectories of social engagement and limitations in late life. Journal of Health and Social Behavior, 52, 430-443. doi:10.1177/0022146511411922

40.

Thorp

A. A.

Owen

Neuhaus

Dunstan

D. W.

(2011). Sedentary behaviors and subsequent health outcomes in adults: A systematic review of longitudinal studies, 1996-2011. American Journal of Preventive Medicine, 41, 207-215. doi:10.1016/j.amepre.2011.05.004

41.

Tsunoda

Soma

Kitano

Tsuji

Mitsuishi

Yoon

J. Y.

Okura

(2013). Age and gender differences in correlations of leisure-time, household, and work-related physical activity with physical performance in older Japanese adults. Geriatrics & Gerontology International, 13, 919-927. doi:10.1111/ggi.12032

42.

Verbrugge

L. M.

Jette

A. M.

(1994). The disablement process. Social Science & Medicine, 38, 1-14. doi:10.1016/0277-9536(94)90294-1

43.

World Health Organization. (2010). Global recommendations on physical activity for health. Geneva, Switzerland: Author.

44.

H. W.

Chen

D. R.

Chiang

T. L.

Y. K.

Chen

Y. M.

(2015). Disability trajectories and associated disablement process factors among older adults in Taiwan. Archives of Gerontology and Geriatrics, 60, 272-280. doi:10.1016/j.archger.2014.12.005

45.

Zimmer

Hickey

Searle

M. S.

(1995). Activity participation and well-being among older people with arthritis. The Gerontologist, 35, 463-471.

46.

Zimmer

Hickey

Searle

M. S.

(1997). The pattern of change in leisure activity behavior among older adults with arthritis. The Gerontologist, 37, 384-392.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.03 MB