| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
RESEARCH ARTICLE |
Office of Population Research, Princeton University, New Jersey.
Address correspondence to Amy Love Collins, PhD, c/o Noreen Goldman, Office of Population Research, Princeton University, Princeton, NJ 08544. E-mail: alc{at}princeton.edu
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Life satisfaction • Well-being • Mobility • Taiwan
Researchers are increasingly attentive to the role that positive mental health plays in physical health outcomes. The term positive mental well-being is often used to reflect frequent experiences of positive affect and the presence of self-esteem, mastery, life satisfaction, and optimism. Positive mental well-being is not the absence of negative well-being; it has distinct physiological and neurobiological correlates (Cacioppo & Berntson, 1999
; Cacioppo, Gardner & Berntson, 1999; Hamer, 1996; Ryff et al., 2006). Thus, the causes and consequences of positive well-being cannot necessarily be inferred from research on negative well-being (Ryff et al.).
Positive well-being is recognized as a protective factor against poor health in young and older populations (Fredrickson, 2001; Fredrickson & Levenson, 1998; Giltay, Geleijnse, Zitman, Hoekstra, & Schouten, 2004; Lyyra Törmäkangas, Read, Rantanen, & Berg, 2006; Maier & Smith, 1999; Ostir, Berges, Markides, & Ottenbacher, 2006; Ostir, Markides, Black, & Goodwin, 2000; Ostir, Markides, Peek, & Goodwin, 2001; Ostir et al., 2002; Pitkala, Laakkonen, Strandberg, & Tilvis, 2004; Pressman & Cohen, 2005; Salovey, Rothman, Detweiler, & Steward, 2000; Scheier & Carver, 1992). In the present study, we examined the health consequences of positive well-being for a middle-aged and elderly population. Specifically, we investigated the relationship between two facets of positive well-being—current life satisfaction and perceptions of future happiness—and functional status over an 8-year follow-up period in Taiwan.
Appraisals of life satisfaction have been shown to be related to functional status in older Western populations (Pressman & Cohen, 2005). Life satisfaction typically refers to quality-of-life judgments that are based on individuals' own criteria for success and happiness (Pavot, Diener, Colvin, & Sandvik, 1991; Shin & Johnson, 1978). Life satisfaction measures vary in their composition, but generally they tap trait levels of affect as well as cognitive assessments of the extent to which a person's life matches his or her expectations (Okun & Stock, 1987). Two cross-sectional studies found that lower life satisfaction was associated with an increased likelihood of physical frailty and residence in a nursing home (Finlayson, 2002; Strawbridge, Shema, Balfour, Higby & Kaplan, 1998). Other studies of younger and middle-aged populations demonstrated a relationship between life satisfaction and fatal injury and work disability (Koivumaa-Honkanen, Honkanen, Koskenvuo, Viinamaki, & Kaprio, 2002; Koivumaa-Honkanen et al., 2000, 2004).
Evidence suggests that affective judgments about the future also contribute to better functional status. In a study of recovery after coronary artery bypass surgery, optimists were faster to accomplish mobility tasks (e.g., sitting in bed, walking around the room), were rated by staff as demonstrating a better recovery 6 to 8 days and 6 months after surgery, and showed better functional status up to 5 years later (Scheier & Carver, 1992; Scheier et al., 1989). A cross-sectional study of community-dwelling older adults with knee pain found that pessimism was significantly related to poorer performance on four daily activities (walking, lifting an object, climbing stairs, and getting into and out of a car) whereas optimism was related to performance only on the walking task. Another cross-sectional study demonstrated that pessimism was significantly related to increased mobility limitations in a sample of older women (Umstattd, McAuley, Motl, & Rosengren, 2007). These studies suggest that perceptions of positive outcomes in the future may benefit physical functioning in both clinical and nonclinical populations.
There are several lines of research demonstrating the pathways through which positive well-being may be beneficial to health. For example, happier people tend to use active, problem-focused coping and are more accepting of stressful life events (Salovey et al., 2000). Similarly, life satisfaction may reveal how well one has adapted or learned to cope with one's health status and environment (Folkman, 1997; Fredrickson, 2001). Higher life satisfaction can also be an indication of how easily an individual disengages from unattainable goals and reengages in new, more attainable goals (Rasmussen, Wrosch, Scheier, & Carver, 2006). This strategy would be vital to coping with health conditions, such as restricted mobility or chronic disease, which may render valued life goals impossible.
Fredrickson (2001) and Salovey and colleagues (2000) have independently suggested that positive emotions and cognitive styles contribute to the building of resources that promote resilience. Positive feelings can broaden individuals' coping abilities by facilitating the gain of social, intellectual, and physical resources (Fredrickson). For example, positive affect increases engagement in social networks and activities, which have been shown to be negatively associated with illness and mortality (Pressman & Cohen, 2005). Positive affect is also associated with confidence in and adherence to health practices (Bardwell, Berry, Ancoli-Israel, & Dimsdale, 1999; Cohen, Doyle, Turner, Alper, & Skoner, 2003; Luoto, Prattala, Uutela, & Puska, 1998; Ryff, Singer, & Dienberg Love, 2004; Salovey et al., 2000; Scheier & Carver, 1992), particularly focusing on and planning for future health outcomes (Salovey et al.).
Another pathway through which frequent positive experiences may benefit health is based on research demonstrating that positive emotions are linked to physiologic reactions. Positive affect is associated with physiological functioning, including autonomic nervous system activation, hypothalamic–pituitary–adrenal (or HPA) axis activation, and changes in cardiovascular response and immune function (Pressman & Cohen, 2005; Steptoe, Wardle, & Marmot, 2005). At the other end of the spectrum, negative emotions and chronic stress may exacerbate illness through neuroimmune and neuroendocrine deficiencies that increase vulnerability to disease (Heim, Ehlert & Hellhammer, 2000; McEwen & Stellar, 1993; Segerstrom & Miller, 2004). Positive emotions can "undo" the immediate effects of negative emotions by reducing psychophysiological reactivity and enhancing coping when individuals are confronted with negative events (Fredrickson & Levenson, 1998; Pressman & Cohen; Shapiro, Jamner, Goldstein, & Delfino, 2001; Tugade, Fredrickson, & Barrett, 2004). In sum, frequent positive experiences may create more favorable physiological profiles.
In the present study, we extend the literature on positive well-being and functional status by examining whether the associations are manifest in a Chinese population. Few of the studies investigating positive well-being and functional status have used national samples, and none, to our knowledge, has been fielded in a non-Western population. Life satisfaction and other global measures of well-being generally have cross-cultural validity between Western and non-Western samples (Diener, Oishi, & Lucas, 2003). In particular, as in Western societies, older adults in China and Taiwan equate successful aging and life satisfaction with economic prosperity, physical health, and social involvement and support (Chen, 2001; Hsu, 2007; Lu & Shih, 1997). Undoubtedly, however, there are some important differences across populations. For example, comparisons between Western and non-Western samples reveal that relationship harmony and interpersonal evaluations are a more central component of life satisfaction in Eastern than in Western cultures (Diener et al.; Lu & Shih).
In this study, we used data from the Survey of Health and Living Status of the Near Elderly and Elderly in Taiwan to test whether life satisfaction and perceptions of future happiness protect older adults from increased mobility limitations over an approximately 8-year period. We used an indicator of mobility to assess health status because older respondents are likely to report measures of functional status and disability more accurately than the presence of chronic diseases, and because these measures are strong predictors of medical care use and survival in older populations. In particular, mobility difficulty is predictive of more severe forms of disability and institutionalization (Fried & Guralnik, 1997; Guralnik et al., 1994).
We included two aspects of positive well-being in our statistical analysis, because it is unclear from existing research to what extent different facets of positive well-being may overlap in their prediction of health. Our data allowed us to examine current life satisfaction and perceptions of future happiness separately. We adjusted for health status and social connectedness at baseline, both of which are related to the development of mobility limitations. We also included baseline depressive symptoms in order to examine the benefits of life satisfaction and perceptions of future happiness independent of negative affect. Few studies of the relationship between positive mental well-being and health include controls for baseline negative affect (Pressman & Cohen, 2005). Given the evidence from previous studies on life satisfaction, affective judgments about future events, and functional status in older adults, we anticipated that both current life satisfaction and perceptions of future happiness would independently contribute to the number of mobility limitations in our sample.
|
METHODS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Measures
Mobility limitations
These limitations included the number of mobility activities for which the respondent had any difficulty (potential range = 0–8). These activities include standing for 15 minutes, squatting, reaching over one's head, grasping with fingers, lifting or carrying 11 to 12 kg, running 20 to 30 m, walking 200 to 300 m, and climbing two to three flights of stairs. We included a variable denoting mobility limitations in 1996 as a baseline health control.
Life satisfaction and perceptions of future happiness
We measured these by using items adapted from the 20-item Life Satisfaction Index (Neugarten, Havighurst, & Tobin, 1961) and modified for use with Taiwanese respondents (0 = low life satisfaction and 8 = high life satisfaction; Cronbach's alpha =.79). Subscales of this index are comparable with other measures of life satisfaction that are typically used in the literature (Liang, 1985; Nezlek, Richardson, Green, & Schatten-Jones, 2002; Stock, Okun, & Benin, 1986). The interviewers made the following statement to each respondent before asking 10 questions (in Mandarin) about his or her life: "I'd like to ask for your current views or feelings about your life. Please tell me whether you agree with the sentences I am about to read."
We used the following eight items from the scale to represent life satisfaction, each of which required a yes or no response: (a) Has your life been better than most people's lives? (b) Are you satisfied with your life? (c) Do you find what you do interesting? (d) Have these few years been the best in your life? (e) If it was possible (to do again or to do over), would you want to start over or live this life again? (f) Do you feel that most of what you do is monotonous and of no interest? (g) Do you feel old and tired? (h) Would you say your life has matched your hopes? All eight items loaded highly on one factor (eigenvalue = 2.66). Similar items were used in a previous study of life satisfaction and aging in Taiwan (Chen, 2001). We measured perceptions of future happiness by using this question: "Do you expect that, in the future, happy things will occur?" (1 = yes; 0 = no).
Sociodemographic controls
These included age, respondent's sex (1 = female; 0 = male), marital status (1 = spouse or companion; 0 = neither), and years of schooling completed (from the 1989 interview for the elderly participants and from the 1996 interview for the near-elderly participants). We also included two variables measuring social connectedness and involvement that are known to be related to health outcomes and positive well-being, specifically (a) the total number of reported ties with children, close relatives, and close friends with which the respondent has monthly contact, and (b) the number of social activities in which the respondent participates (potential range = 0–11).
Baseline health controls
These controls included depressive symptoms (10 items from the Center for Epidemiological Studies–Depression scale, possible range = 0–30; Cronbach's alpha =.82) and cognitive impairment (the number of cognitive tasks that the respondent answered incorrectly out of a possible 11). Cognitive impairment was based on items from the modified Rey Auditory Verbal Learning Test (Lezak, 1983) and a modification of the Digits Backward Test (Wechsler, 1981). We also included one question that asks respondents to rate their overall health in one of five categories (1 = excellent; 5 = poor), which we refer to as self-assessed health ("Regarding your current state of health, do you feel it is excellent, good, average, not so good, or poor?"). Finally, we included smoking status (1 = yes; 0 = no) and the number of chronic conditions the participant reported among the following conditions: high blood pressure, diabetes, heart disease, cancer, respiratory ailments, arthritis, ulcers, liver or gall bladder disease, cataracts, kidney disease, gout, or spinal injury (potential range = 0–12).
Statistical Analysis
The outcome variable was the number of mobility limitations reported during the 2003 interview. The predictors of interest were life satisfaction and perceptions of future happiness, assessed in 1996. We included control variables that are related to health outcomes and positive mental well-being: age, sex, education, marital status, number of social ties, number of social activities, and baseline health status.
We analyzed the data by using a form of Poisson regression because the outcome variable, mobility limitations, is a count variable. With mobility limitations, however, we observed many more zeroes than would be expected if the distribution was Poisson. At follow-up, the mean number of limitations was two but 45% of the respondents had zero limitations, which is well above the 14% expected in a Poisson distribution with that mean. Therefore, we estimated a zero-inflated Poisson regression model that uses two equations estimated simultaneously: one is a logit equation that predicts membership in a latent class that is at risk of experiencing mobility limitations (the "inflate" equation); the other is a Poisson equation that predicts the number of mobility limitations for those participants at risk. Although the inflate equation is typically modeled by predicting those not at risk (Cheung, 2002), we have chosen to predict the complementary outcome to facilitate comparisons with the Poisson equation. In our models, variables that increase the probability of being at risk generally increase the expected number of limitations for those at risk, so the coefficients in the two equations tend to have the same signs. For a fuller treatment of zero-inflated Poisson regression, see Long and Freese (2006), and for a discussion of its application in medical studies, see Cheung.
Several of the previously reviewed prospective studies found that positive emotional well-being influenced health outcomes in respondents that were healthy at baseline (Koivumaa-Honkanen et al., 2000, 2002; Ostir et al., 2000; Ostir, Ottenbacher, & Markides, 2004). Thus, in order to investigate whether the results vary with initial health status, we included an interaction term between life satisfaction and whether or not the respondent had mobility limitations at the baseline interview in 1996. For similar reasons, we included an interaction term between perceptions of future happiness and whether or not the respondent had zero mobility limitations in 1996. To facilitate interpretation of the regression coefficients for life satisfaction and perceptions of future happiness, we estimated two zero-inflated Poisson regression models: Model A includes only the main effects of the covariates and Model B adds the two interaction terms just described to Model A. We analyzed the data by using Stata 9.0 (StataCorp, 2005).
|
RESULTS |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
|
|
|
DISCUSSION |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Our findings are consistent with those from other prospective studies finding associations between positive well-being and physical functioning in healthy individuals. For example, a study by Ostir and colleagues (2000) found that older adults with high positive affect were half as likely as adults with low positive affect to have died or to have acquired limitations in functional status 2 and 7 years later in the presence of controls for baseline functional status, lifestyle indicators (e.g., smoking, drinking), and baseline negative affect scores. The sample for Ostir and colleagues included only respondents who could complete a walking task and had no difficulty with activities of daily living at baseline. In the study by Koivumaa-Honkanen and associates (2004) of life satisfaction and work disability, life dissatisfaction predicted psychiatric sources of work disability in the portion of their sample classified as "ill" at baseline, whereas life dissatisfaction predicted both nonpsychiatric and psychiatric causes of work disability for those classified as "healthy" at baseline. Koivumaa-Honkanen and colleagues (2000) included only those who were free of chronic disease at baseline in their study, demonstrating a relationship between life dissatisfaction and mortality.
Given the multiple pathways through which positive emotions may affect health, our findings raise the question of why life satisfaction and perceptions of future happiness would be associated with health in the longer term only for those of high functional status. One possibility is that positive well-being is beneficial up to a certain threshold of health, at which point other, stronger influences come into play. For example, positive affect states may improve adherence to medical regimens and behaviors for those with no mobility limitations (e.g., better sleep, diet, and exercise), which would reduce the number of mobility limitations they develop, whereas the presence of mobility limitations at baseline may disrupt adherence to health practices. Similarly, if positive well-being improves one's physiological profile over the long term, it would be difficult for these subtle changes to have much of an impact after other negative physiological events have accumulated to the point of disability or disease. Presumably, positive emotions can only prevent or "undo" so much of the cumulative damage that has already taken place. There is also some evidence that high positive affect is dysfunctional in populations with chronic illness, because having higher positive affect may lead individuals to not take their illnesses seriously (Brown, Butow, Culjak, Coates, & Dunn, 2000; Derogatis, Abeloff, & Melisaratos, 1979; Devins et al., 1990). Reporting high life satisfaction and perceptions of future happiness in the face of a high number of mobility limitations may indicate denial or suppression of the seriousness of one's disability.
There are limitations in this study that should be noted. First, we must consider the possibility of reverse causation, in which mobility limitations affect positive well-being. Our analytic strategy deals with this problem by using time lags. Specifically, we examined the effect of life satisfaction and perceptions of future happiness as of 1996 on mobility limitations in 2003, approximately 8 years later, effectively ruling out a direct effect of the outcome on the predictor. It is still possible, of course, that the measures of positive well-being in 1996 were affected by mobility limitations at that time, but our models control for this potential confounder by including mobility limitations in 1996 as a covariate. We also estimated the effect of mobility restrictions in 1996 on life satisfaction and perceptions of future happiness in 2003, controlling for the same measures of positive affect in 1996 (not shown here), and we found no significant effects at the conventional 5% level. We are therefore confident that our estimates reflect genuine effects of positive well-being on mobility limitations and not the other way around. Second, there may be a measurement issue, as researchers have established that affect influences self-reports of health (Diener, Suh, Lucas, & Smith, 1999); for example, levels of mental well-being may drive respondents to underestimate or overestimate their ability to complete specific tasks. But these effects are more likely to be contemporaneous; it is hard to see how mental well-being in 2003 would affect self-reports 8 years earlier, except by means of correlation with previous mental well-being, which is controlled at least in part by the lagged outcome. Finally, it is possible that, despite our careful inclusion of control variables, there remain unobserved confounders that affect both positive well-being in 1996 and mobility limitations in 2003. To the extent that those confounders also determine mobility limitations in 1996, however, their biasing effect would be substantially reduced by our control for baseline limitations.
In conclusion, in a Chinese sample of middle-aged and older adults, we found that current life satisfaction and perceptions of future happiness were determinants of the number of mobility limitations at an 8-year follow-up for those respondents without mobility limitations at baseline (with adjustment for sociodemographic factors, social involvement, baseline health, and depressive symptoms). Higher life satisfaction and perceptions of future happiness may contribute to better coping, to improved health practices, and to better physiological functioning in the longer term. Future research should aim to clarify the multiple pathways through which positive emotions are related to health.
|
Acknowledgments |
|---|
We thank Dana Glei and Rachel Pruchno for helpful comments on the text.
|
Footnotes |
|---|
Received for publication June 26, 2007. Accepted for publication January 30, 2008.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
