Health Education Research Advance Access originally published online on August 8, 2005
Health Education Research 2006 21(1):146-156; doi:10.1093/her/cyh050
The relationship between active living and health-related quality of life: income as a moderator
Departments of 1 Community and Behavioral Health and 2 Health and Sport Studies, College of Public Health, University of Iowa, 1215 Westlawn, Iowa City, IA 52242, USA
3 Correspondence to: J. J. Peterson; E-mail: jana-peterson{at}uiowa.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMethodResultsDiscussionConflict of interest statementReferences |
|---|
This study used a moderator model to examine the relationship between active living and the physical components of health-related quality of life [i.e. overall physical component of quality of life (PQOL), physical functioning and ability to fulfill physical role] among a randomly selected sample of rural residents (n = 407) from the Midwestern US. Results showed that active living was associated with greater increases in health-related quality of life for those reporting lower income. The effect size of the relationship between active living and the PQOL for the low-income group was over 2 times the effect size for the high-income group. For physical functioning, the effect size of active living for the low-income group was greater than 3 times the effect size for the high-income group. Although active living behaviors have been demonstrated to be less prevalent among those of low socioeconomic status, this group may have the most to gain from these activities. Findings highlight the need for increased and specifically targeted promotion of active living interventions.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConflict of interest statementReferences |
|---|
The benefits of regular physical activity on health are well documented, and include decreased prevalence of coronary heart disease (Powell et al., 1987
) and cancer (Albanes et al., 1989), as well as decreased all-cause mortality (Leon and Connett, 1991; Paffenbarger et al., 1986). Although health-related quality of life (HRQOL) has joined morbidity and mortality as a health outcome of interest in recent years, much of the research on the HRQOL benefits of physical activity has focused on populations with specific disease states, e.g. coronary heart disease (Hage et al., 2003), osteoarthritis (Dias et al., 2003), kidney disease (Brodin et al., 2001) and liver disease (Hickman et al., 2004). The description of the relationship between physical activity and HRQOL in the general population is less adequate (Brown et al., 2003). The few studies that have been performed indicate a positive relationship between physical activity and HRQOL (Rejeski et al., 1996; Ellingson and Conn, 2000; Rejeski and Mihalko, 2001; Brown et al., 2003).
Similarly, while the link between socioeconomic status (SES) and health is well established (Rose and Marmot, 1981; Marmot et al., 1984; Marmot et al., 1991), few studies investigating SES have focused on HRQOL as an outcome. Existing evidence indicates that HRQOL correlates with SES (Hemingway et al., 1997). Previous investigations have shown that health behaviors explain a moderate amount of SES disparities in morbidity and mortality, but emphasize that health behaviors could not explain all of the difference due to SES (Rose and Marmot, 1981; Lantz et al., 1998; Lantz et al., 2001). Low physical activity, in particular, has been shown to significantly predict deleterious health change (Lantz et al., 2001).
The objective of this study was to utilize a moderator model (Figure 1) to examine whether there is a differential relationship between active living and the physical aspects of HRQOL (PQOL) at various income levels. Often, investigations infer that a third variable moderates a demonstrated relationship, but do not explicitly test a moderator model via a linear regression interaction term. Several authors have used moderator models to demonstrate the moderating effects of SES on other health relationships (Lachman and Weaver, 1998; Kempen et al., 1999; Brummett et al., 2003; Peterson and Hughey, 2002, 2004). This investigation fills an important gap in research, since health benefits of physical activity for disparate income groups within the general population have not been empirically demonstrated.
|
|
Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConflict of interest statementReferences |
|---|
Participants
This study was part of a larger evaluation of rural community health in two Midwest communities, with year 2003 populations of approximately 2500 and 2200. A database of adult members of the two communities was compiled from the telephone listings and county auditor's housing records. Those with addresses that fell within a 3-mile radius of the center of each town were included as prospective participants. Participants, randomly selected from this database, were contacted first by letter and then by telephone call. Criteria for ineligibility included residence in a care facility, current hospitalization, residence outside of the defined community boundaries and lack of fluency in spoken English. All data was collected during in-person interviews in 2003. The response rate was 25%.
The age range of the participants (n = 407) was 25–88 years, with a mean age of 56.5 years. The sample was 57.0% female. The majority of the participants, 98.8%, were white/non-Hispanic, while the remaining 1.2% was white/Hispanic. Educationally, 93.9% of participants had obtained a high school diploma and 26.0% had at least a 4-year college degree. For household income, 2.3% reported less than US$10 000, 21.4% reported US$10 000–24 999, 36.4% reported US$25 000–49 999, 28.1% reported US$50 000–74 999 and 11.8% reported $75 000 or more.
Measures
Physical activity
Physical activity was assessed using the Baecke physical activity questionnaire (Baecke et al., 1982). The questionnaire consists of three indices each measuring a different type of physical activity, labeled occupation, sport and leisure/transportation. This is a widely used questionnaire considered valid and reliable in adult populations (Kriska and Caspersen, 1997). Test–re-test reliability scores for occupation, sport and leisure/transportation indices were 0.88, 0.81 and 0.74, respectively (Baecke et al., 1982). Validity studies have revealed correlations with 3-day activity diaries of 0.66 for men and 0.44 for women (Pols et al., 1995) and correlations with VO2max of 0.54 in a population of both men and women (Jacobs et al., 1993). In addition to occupation, sport and leisure/transportation, a fourth type of physical activity was also collected using the caretaker/household index of a modified Baecke physical activity questionnaire, known as the Kaiser physical activity survey (Ainsworth et al., 2000).
The caretaker/household index is comprised of 11 items measuring time spent in elder and child care tasks, preparing meals, light cleaning, and heavy cleaning activities (Ainsworth et al., 2000). The occupation index consists of seven questions that query time spent sitting, standing, walking, lifting heavy loads, feeling tired and sweating while at work, as well as rating relative physical activity at work compared to others of the same age (Baecke et al., 1982). The sport index consists of three items about sports and exercise that ask about frequency of sports and exercise, sweating during sports and exercise, and amount of recreational physical activity performed compared to others the same age. The leisure/transportation index is comprised of four items that query frequency of walking and riding a bicycle, time spent walking or bicycling for transportation, and time spent watching television (the last of which is negatively scored).
We calculated the mean of the sport and leisure/transportation indices, and refer to this variable as active living. We focused on active living to align our study with previous research (Lavizzo-Mourey and McGinnis, 2003; Sallis et al., 2005). The definition of active living used in this study, encompassing a combination of physical activity performed at leisure and for transportation, is consistent with the definition currently utilized by others in the field (Lavizzo-Mourey and McGinnis, 2003; Sallis et al., 2005). In addition, we calculated the mean of the scores on the occupation and caretaker/household indices, and refer to this variable as home/work. The home/work score was utilized as a control variable in all analyses.
HRQOL
HRQOL was measured via the Medical Outcomes Study (MOS) Short Form Health Survey (SF-36). Nationally and internationally, the SF-36 is the most widely used HRQOL survey (Ware et al., 2002). It has been validated extensively on general and at-risk populations, and demonstrates high reliability and construct validity (Ware and Sherbourne, 1992). The SF-36 contains 36 items scored in eight subscales, each measuring a separate health concept: physical functioning (10 items), role limitations due to physical health problems (four items), bodily pain (two items), general health (five items), vitality (four items), social functioning (two items), role limitations due to emotional problems (three items) and mental health (five items). In addition, the measure contains one unscaled item on health change over the past year. For this study, the outcome variables were the physical aspects of HRQOL (PQOL). PQOL variables include the four physical subscales of the SF-36, e.g. physical functioning, physical role limitations, bodily pain and general health, and the corresponding physical component score.
All SF-36 variables were scored according to the norm-based scoring system for the US 1998 general population norms, facilitating meaningful data interpretation (Ware et al., 2002). To calculate the norm-based score for each subscale, the MOS score of 0–100 is first determined for each. The MOS score is standardized to a z-score by subtracting the MOS US general population mean and dividing the result by the population SD. Each z-score is transformed to a norm-based score with a mean of 50 and SD of 10 in the general population. To calculate the summary physical component score from the eight subscales, each scale z-score is multiplied by a predetermined physical factor coefficient and the eight products were summed. This aggregate score was standardized to a general population mean of 50 and standard deviation of 10. MOS guidelines for missing data were followed (Ware and Kosinski, 2001).
Other variables
The proxy for SES utilized in this study is household income. It has been suggested that recent income is at least as highly correlated with mortality outcomes as other traditional indicators of SES (Daly et al., 2002). Other study variables include age, sex and body mass index (BMI). Household income, which was collected categorically as one of eight self-reported income ranges, and age were determined by self-report. BMI was calculated from anthropometric measurements of height and weight taken during the interview.
Procedures
Ethical approval was granted from the university Institutional Review Board. Extra care was taken to maintain the confidentiality of the participants, due to decreased anonymity within a rural community, and informed consent was obtained from each participant.
Active living and HRQOL were collected by means of participant self-administered written surveys. Household income, sex and age were collected as part of a face-to-face interview, and BMI was measured by a research assistant. Interviews were held in the communities at central locations. The complete interview consisted of a number of measures and required the participant to spend approximately 90 minutes to complete the full battery of instruments. Participants were compensated with US$25.
Analysis
As consistent with recommendations in the literature for the testing of moderator effects, we regressed the dependent variable on the independent variable, moderator and their product term (Baron and Kenny, 1986). Specifically, a hierarchical regression analysis with the SF-36 physical component score as the dependent variable was performed. Because age, sex and BMI are factors highly correlated with HRQOL (Jenkinson et al., 1993; Ford et al., 2001; Gallant and Dorn, 2001; Groessl et al., 2004), these factors were entered first. Home/work physical activity was entered next to control for physical activity that is performed in addition to active living behaviors. Income, active living, and the interaction of income and active living were then entered hierarchically.
|
Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConflict of interest statementReferences |
|---|
The difference in PQOL across income groups in our study was notable. After adjusting for all covariates in the study, the average physical component score for the low-income group was 44.98, compared to 50.67 and 51.20 for the medium- and high-income groups. After controlling for all covariates, the average physical functioning scores for the low-, medium- and high-income groups were 44.92, 50.12 and 51.73, respectively. The average role-physical scores for the low-, medium- and high-income groups were 45.59, 51.58 and 51.81, respectively, after controls. As the MOS norm-based scoring system has a general population mean of 50 with a SD of 10, the difference between the low-income group is notably dissimilar to both the general population mean and the current study's medium/high-income group scores on all three measures.
As shown in Table I, results of the hierarchical regression analysis support the moderator model for the interaction of income and active living on physical component score. Age was entered into the model first and was a significant predictor of the physical component score, accounting for 19% of the variance in the score. Sex, the second factor added to the model, did not contribute any unique explanation. BMI was significant, with a unique contribution of 7% of the variance in physical component score, while home/work did not contribute any unique explanation. Income and active living were significant predictors, contributing 6 and 4% of the variance, respectively. Finally, the interaction between income and active living was also significant after controlling for the above factors.
|
In order to further examine the moderator effect on PQOL, four hierarchical regression analyses were then performed, with each of the four SF-36 physical subscales serving as the dependent variable. The income by active living interaction was significant for the physical functioning and role limitations due to physical role (role-physical) scales, but not for the bodily pain or general health scales. Results of the physical functioning and role-physical models can be seen in Table II. For the hierarchical regression on the physical functioning scale, age, sex, BMI and home/work were all significant predictors of physical functioning, explaining 24, 1, 7 and 2% of the variance in the scale score, respectively. After controlling for the above factors, income, active living and the income by active living interaction were added to the model. All three were significant predictors. Income explained 8% of the variance in physical functioning and leisure-time physical activity explained 7% of the variance. The interaction term was significant after controlling for all other variables. For the regression on role-physical, age was a significant contributor, accounting for 18% of the variance. Sex was insignificant, explaining none of the model variance. BMI, a significant factor, contributed 2% of the variance. Home/work contributed no unique variance to the model. After controlling for the above factors, income described 5% of the variance in role-physical. Active living was not significant when added next, and did not contribute any unique effects to the model. However, when the income by active living interaction was added to the model, the unique effects of active living became significant. In addition, the interaction was a significant predictor of role-physical.
|
To increase the interpretation of this moderator effect, analysis of covariance (ANCOVA) was performed to examine the nature of the relationship between active living and the three dependent variables (i.e. physical component score, physical functioning, and role-physical) for people of lower-, middle- and higher-income groups. The use of post-hoc ANCOVA is consistent with prior research that has investigated moderator effects of SES on health (Lachman and Weaver, 1998
). For the post-hoc analyses only, active living was recoded into three equal-sized categories. These categories were labeled low, medium and high active living. Household income was also recoded into three similarly-sized categories, specifically: less than US$25 000 (n = 92), US$25 000–49 999 (n = 144) and US$50 000 and above (n = 155). As in the hierarchical regression analysis, age, sex, BMI and home/work were included in the analyses as covariates. The relationship between active living and the PQOL components are demonstrated graphically. Physical component score (Figure 2) for individuals reporting low levels of active living differed significantly from those reporting high levels of active living for low- (P = 0.01), medium- (P < 0.05) and high-income (P < 0.01) groups. The effect size for increase in physical component score for individuals with low income 
was approximately double the magnitude of the effect size for those in the medium- 
and high-income 
groups.
|
Physical functioning (Figure 3) scores for individuals of all income levels differed significantly between low-, medium- and high-active individuals. For the participants within the low-income category, physical functioning for those reporting low activity was significantly lower than physical functioning for those reporting medium activity (p = 0.001), and those reporting medium activity had significantly lower physical functioning than those reporting high activity (P < 0.001). The effect size of active living and physical functioning for those with low income was
For those in the medium-income category, low activity corresponded with significantly lower physical functioning than medium activity (P < 0.05), while medium activity corresponded with significantly lower physical functioning than high activity (P < 0.05), 
For the high-income group, physical functioning scores for low-active living were significantly different from medium-active living (P < 0.05) and physical functioning scores for medium active living were significantly different from high-active living (P = 0.001). The effect size of the relationship between active living and physical functioning for those with high income was 
The effect size of active living level on physical functioning for the low-income group was a magnitude of 3–4 times greater than the effect sizes for the medium- and high-income groups.
|
The interaction between active living and income for role-physical is shown in Figure 4. Notably, the ANCOVA results were not significant in the post-hoc analysis because of the reduced variability in active living and income that occurred from our recoding into categories. However, the pattern of the interaction is similar to the patterns for both the physical component score and physical functioning. As can be observed in Figure 4, the increase in role-physical from low-active living to medium and high-active living was greatest for the low-income group.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConflict of interest statementReferences |
|---|
Our study suggests a differential relationship between active living and HRQOL at various levels of income. Specifically, income was found to be a significant moderator of the relationship between active living and PQOL, demonstrated in the regression on the summary physical component score, the physical functioning subscale and the role-physical subscale. This was evidenced in the post-hoc analyses, where it was shown that the effect size for active living on physical component score for the low-income group was more than twice the magnitude of the active living effect size for the physical component score of the medium- and high-income groups. The effect size for active living on physical functioning of the low-income group was between 3–4 times the effect size associated with the medium- and high-income group physical functioning.
The interaction effect was seen for physical functioning and role-physical subscales, but not for bodily pain and general health. This suggests that the differential association between active living and PQOL for individuals of lower income compared to those of higher income is specific to certain physical health aspects, i.e. greater functioning and decreased disability. Physical functioning and role-physical, the two SF-36 factors that have the greatest factor loading with physical composite score, are essentially the ‘most physical’ of the SF-36 subscales. While bodily pain has only slightly lower factor loading with physical component score, general health has moderate correlations with the rotated components of both physical component score and its companion mental component score, and can therefore be considered both a physical and a mental SF-36 subscale (Ware and Kosinski, 2001).
In recent years, SES has been shown to be an important determinant of health. The Whitehall Study of British civil servants investigated the social gradient of SES and health. Even though the Whitehall study population did not include the extremes of SES, the effects of SES on health were impressive. The Whitehall investigators found that employment grade positively affects health, both in overall mortality and for individual disease states, such as heart disease, stroke and cancer (Marmot et al., 1984). Those in the lowest civil service employment grade had 4 times the mortality rate of those in the highest administrative employment grade. In addition, this occurred as a gradient, with the second highest employment grade experiencing a higher mortality rate than the highest employment grade. This pattern was repeated across all grades (Marmot and Shipley, 1996). Hemingway et al. (Hemingway et al., 1997) showed a direct relationship between civil service grade of employment and scale scores on the SF-36. This effect was largest for physical functioning. In addition, the decrease in physical functioning with decreasing employment grade occurred in both individuals with and without disease, suggesting that this effect occurs independent of disease. Others have shown a similar relationship between physical functioning and SES (Pinsky et al., 1987; Berkman et al., 1993; Seeman et al., 1994).
The differences in PQOL across income groups in our study were notable. As presented in the results, after adjusting for covariates, the average physical component score for the low-income group was 44.98, compared to 50.67 and 51.20 for the medium- and high-income groups. While we are unaware of disease status of participants in different income groups, it is helpful to compare the association between income and physical component score in this study to published effects of pathology on the physical component score. The difference of more than five points between those in the low-income group and those in the two higher-income groups is greater than general population estimates of the unique effects of many diseases, including back pain (3.75), angina (3.67), diabetes (3.44), chronic lung disease (3.12) and arthritis (2.77) (Ware and Kosinski, 2001). The individuals within our sample reporting less than US$25 000 annual household income experienced a greater disparity in PQOL than the deficits experienced by individuals with many chronic diseases.
One possible explanation for the effect of SES on health is a psychobiological mechanism. Long-term social stress has been linked to chronically elevated cortisol levels in animal studies (Sapolsky and Mott, 1987; Sapolsky et al., 1997) and human studies (Hellhammer et al., 1997; Kristenson et al., 1998; Pruessner et al., 1999). Low SES is associated with multiple psychosocial stressors that affect health. In the Whitehall II study, decreased employment gradient position was linked to increasingly low control of work activities, lack of work variety, low job satisfaction, increased hostility, low social contact, distressing events, financial difficulties and low control over health outcomes (Marmot et al., 1991). Further, abnormal cortisol levels have been linked to the stress of low SES (Kristenson et al., 1998). Chronic elevations of cortisol, as well as epinephrine and norepinephrine (catecholamines), have been linked to decreased health status (Kristenson et al., 2004), offering an explanation for the relationship between SES and health.
Just as cortisol and catecholamines are part of the adaptive physiological response to stress, their release is an important part of the adaptive physiological response to exercise. A chronic adaptation to exercise includes lower levels of circulating catecholamines. In addition, a decline in state anxiety after exercise is well documented in the literature (Astrand et al., 2003). It is possible that the effects of physical activity counteract the maladaptive cortisol and catecholamine response to prolonged stress, which would account for the greater correlation between active living and PQOL for individuals of low income that were observed in the present study.
There are several limitations to this study. The study is cross-sectional in design, so only weak causal inferences can be drawn. This allows the possibility that PQOL serves the causal role in the observed relationships, with active living prevalence differing as an effect of differing PQOL. Also, both PQOL and physical activity were measured via self-report and we had no objective measure of health status, level of functioning or activity. In addition, there was no measure of disease or medical diagnosis in the study, so we were unable to factor out effects of disease on PQOL. Finally, the three SF-36 variables included in the study had skewed distributions (e.g. skewness values of –1.31 to –1.59). Previous research using the SF-36 to study socioeconomic influences on health-related quality of life justified use of parametric methods, given that, although distributions for the SF-36 variables were non-normal, non-parametric methods revealed no differences in trends (Hemingway et al., 1997).
There are additional study limitations regarding characteristics of the sample. The sample size in this study was small for a population-based study. Although the sample size for the present study was limited to only 407 individuals, it represents over 10% of the entire study population because of the low population density of the rural communities in the study. Future research, however, particularly those studies drawing samples from large, urban populations, should strive for larger samples to adequately represent these populations. Additional characteristics of the sample that may limit generalizability of the study results include limited diversity in education, race, ethnicity and the rural nature of the population. It is possible that these factors may moderate the relationship between active living and quality of life. While these are limitations, high levels of education and little racial diversity are typical of rural Midwestern communities, and it is an important to study the health of rural communities as the 2000 US Census reported that 21% of the US population lives in rural areas.
Despite these limitations, our findings are consistent with previous research that has shown that physical activity and other health behaviors explained some, but not all, of the health differences among SES groups (Rose and Marmot, 1981; Lantz et al., 1998; Lantz et al., 2001). This previous work did not explicitly test a moderator model that included activity, SES and health outcomes. Our results suggest that income acts as a moderator in its relationship with active living and PQOL.
We do not disagree with the sentiment that public health should continue to focus on economic policy, to work toward changing social and economic causes of poor health. Socioeconomic disparity creates deleterious health outcomes and this is a major injustice. In spite of this, efforts should be made to reduce negative health consequences in light of economic disadvantage that does exist. The results of this study indicate that active living can have meaningful effects on the physical circumstance of those with lower income. Health promotion should continue to focus on active living in low-income communities via health promotion programming and via resources for an infrastructure that supports active living. Among those of low SES, active living behaviors are less prevalent (Lynch et al., 1997), yet this group may have the most to gain from these activities. This provides a clear rationale for increased and specifically targeted promotion of active living. Results of the current investigation suggest that community interventions that increase active living prevalence in economically disadvantaged communities may have a positive impact on HRQOL. Research should continue to focus specifically on elucidating methods to successfully promote active living to groups of lower SES in order to reduce health disparities along socioeconomic lines.
|
Conflict of interest statement |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConflict of interest statementReferences |
|---|
None declared.
|
Acknowledgments |
|---|
The authors would like to thank the community members who participated in this project and Missy Peterson for her editorial assistance. This study was supported by a grant from the Centers for Disease Control and Prevention to the University of Iowa's Prevention Research Center (grant U48/CCU720075).
|
References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionConflict of interest statementReferences |
|---|
Ainsworth, B.E., Sternfeld, B., Richardson, M.T. and Jackson, K. (2000) Evaluation of the Kaiser physical activity survey in women. Medicine and Science in Sports and Exercise, 32, 1327–1338.
Albanes, D., Blair, A. and Taylor, P.R. (1989) Physical activity and risk of cancer in the NHANES I population. American Journal of Public Health, 79, 744–750.
Astrand, P., Rohahl, K., Dahl, H.A. and Stromme, S.B. (2003) Textbook of Work Physiology: Physiological Bases of Exercise, 4th edn. Human Kinetics, Champaign, IL, pp. 358–361.
Baecke, J.A., Burema, J. and Frijters, J.E. (1982) A short questionnaire for the measurement of habitual physical activity in epidemiological studies. American Journal of Clinical Nutrition, 36, 936–942.
Baron, R.M. and Kenny, D.A. (1986) The moderator–mediator variable distinction in social psychological research: conceptual, strategic and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182.[CrossRef][Web of Science][Medline]
Berkman, L.F., Seeman, T.E., Albert, M., Blazer, D., Kahn, R., Mohs, R., Finch, C., Schneider, E., Cotman, C., McClearn, G., Nesselroade, J., Featherman, D., Garmezy, N., McKhann, G., Brim, G., Prager, D. and Rowe, J. (1993) High, usual and impaired functioning in community-dwelling older men and women: findings from the MacArthur Foundation Research Network on Successful Aging. Journal of Clinical Epidemiology, 46, 1129–1140.[CrossRef][Web of Science][Medline]
Brodin, E., Ljungman, S., Hedberg, M. and Sunnerhagen, K.S. (2001) Physical activity, muscle performance and quality of life in patients treated with chronic peritoneal dialysis. Scandinavian Journal of Urology and Nephrology, 35, 71–78.[CrossRef][Web of Science][Medline]
Brown, D.W., Balluz, L.S., Heath, G.W., Moriarty, D.G., Ford, E.S., Giles, W.H. and Mokdad, A.H. (2003) Associations between recommended levels of physical activity and health-related quality of life. Findings from the 2001 Behavioral Risk Factor Surveillance System (BRFSS) survey. Preventive Medicine, 37, 520–528.[CrossRef][Web of Science][Medline]
Brummett, B.H., Barefoot, J.C., Vitaliano, P.P. and Siegler, I.C. (2003) Associations among social support, income and symptoms of depression in an educated sample: the UNC Alumni Heart Study. International Journal of Behavioral Medicine, 10, 239–250.[CrossRef][Web of Science][Medline]
Daly, M.C., Duncan, G.J., McDonough, P. and Williams, D.R. (2002) Optimal indicators of socioeconomic status for health research. American Journal of Public Health, 92, 1151–1157.
Dias, R.C., Dias, J.M. and Ramos, L.R. (2003) Impact of an exercise and walking protocol on quality of life for elderly people with OA of the knee. Physiotherapy Research International, 8, 121–130.
Ellingson, T. and Conn, V.S. (2000) Exercise and quality of life in elderly individuals. Journal of Gerontological Nursing, 26, 17–25.
Ford, E.S., Moriarty, D.G., Zack, M.M., Mokdad, A.H. and Chapman, D.P. (2001) Self-reported body mass index and health-related quality of life: findings from the Behavioral Risk Factor Surveillance System. Obesity Research, 9, 21–31.[Web of Science][Medline]
Gallant, M.P. and Dorn, G.P. (2001) Gender and race differences in the predictors of daily health practices among older adults. Health Education Research, 16, 21–31.
Groessl, E.J., Kaplan, R.M., Barrett-Connor, E. and Ganiats, T.G. (2004) Body mass index and quality of well-being in a community of older adults. American Journal of Preventive Medicine, 26, 126–129.[CrossRef][Web of Science][Medline]
Hage, C., Mattsson, E. and Stahle, A. (2003) Long-term effects of exercise training on physical activity level and quality of life in elderly coronary patients—a three- to six-year follow-up. Physiotherapy Research International, 8, 13–22.
Hellhammer, D.H., Buchtal, J., Gutberlet, I. and Kirschbaum, C. (1997) Social hierarchy and adrenocortical stress reactivity in men. Psychoneuroendocrinology, 22, 643–650.[CrossRef][Web of Science][Medline]
Hemingway, H., Nicholson, A., Stafford, M., Roberts, R. and Marmot, M. (1997) The impact of socioeconomic status on health functioning as assessed by the SF-36 questionnaire: the Whitehall II study. American Journal of Public Health, 87, 1484–1490.
Hickman, I.J., Jonsson, J.R., Prins, J.B., Ash, S., Purdie, D.M., Clouston, A.D. and Powell, E.E. (2004) Modest weight loss and physical activity in overweight patients with chronic liver disease results in sustained improvements in alanine aminotransferase, fasting insulin and quality of life. Gut, 53, 413–419.
Jacobs, D.R., Ainsworth, B.E., Hartman, T.J. and Leon, A.S. (1993) A simultaneous evaluation of ten commonly used physical activity questionnaires. Medicine and Science in Sports and Exercise, 25, 81–91.
Jenkinson, C., Coulter, A. and Wright, L. (1993) Short form 36 (SF36) health survey questionnaire: normative data for adults of working age. British Medical Journal, 306, 1437–1440.
Kempen, G.I., Brilman, E.I., Ranchor, A.V. and Ormel, J. (1999) Morbidity and quality of life and the moderating effects of level of education in the elderly. Social Science and Medicine, 49, 143–149.
Kriska, A.M. and Caspersen, C.J. (1997) Introduction to a collection of physical activity questionnaires. Medicine and Science in Sports and Exercise, 29, S5–S9.
Kristenson, M., Orth-Gomeer, K., Kucinskiene, Z., Bergdahl, B., Calkauskas, H., Balinkyniene, I. and Olsson, A.G. (1998) Attenuated cortisol response to a standardized stress test in Lithuanian versus Swedish men: the LiVicordia study. International Journal of Behavioral Medicine, 5, 17–30.[CrossRef][Web of Science][Medline]
Kristenson, M., Eriksen, H.R., Sluiter, J.K., Starke, D. and Ursin, H. (2004) Psychobiological mechanisms of socioeconomic differences in health. Social Science and Medicine, 58, 1511–1522.
Lachman, M.E. and Weaver, S.L. (1998) The sense of control as a moderator of social class differences in health and well-being. Journal of Personality and Social Psychology, 74, 763–773.[CrossRef][Web of Science][Medline]
Lantz, P.M., House, J.S., Lepkowski, J.M., Williams, D.R., Mero, R.P. and Chen, J. (1998) Socioeconomic factors, health behaviors and mortality: results from a nationally representative prospective study of US adults. Journal of the American Medical Association, 279, 1703–1708.
Lantz, P.M., Lynch, J.W., House, J.S., Lepkowski, J.M., Mero, R.P., Musick, M.A. and Williams, D.R. (2001) Socioeconomic disparities in health change in a longitudinal study of US adults: the role of health-risk behaviors. Social Science and Medicine, 53, 29–40.
Lavizzo-Mourey, R. and McGinnis, J.M. (2003) Making the case for active living communities. American Journal of Public Health, 93, 1386–1388.
Leon, A.S. and Connett, J. (1991) Physical activity and 10.5 year mortality in the Multiple Risk Factor Intervention Trial (MRFIT). International Journal of Epidemiology, 20, 690–697.
Lynch, J.W., Kaplan, G.A. and Salonen, J.T. (1997) Why do poor people behave poorly? Variation in adult health behaviours and psychosocial characteristics by stages of the socioeconomic lifecourse. Social Science and Medicine, 44, 809–819.
Marmot, M.G. and Shipley, M.J. (1996) Do socioeconomic differences in mortality persist after retirement? 25 year follow up of civil servants from the first Whitehall study. British Medical Journal, 313, 1177–1180.
Marmot, M.G., Shipley, M.J. and Rose, G. (1984) Inequalities in death—specific explanations of a general pattern? Lancet, i, 1003–1006.
Marmot, M.G., Smith, G.D., Stansfeld, S., Patel, C., North, F., Head, J., White, I., Brunner, E. and Feeney, A. (1991) Health inequalities among British civil servants: the Whitehall II study. Lancet, 337, 1387–1393.[CrossRef][Web of Science][Medline]
Paffenbarger, R.S., Hyde, R.T., Wing, A.L. and Hsieh, C.C. (1986) Physical activity, all-cause mortality and longevity of college alumni. New England Journal of Medicine, 314, 605–613.[Abstract]
Peterson, N.A. and Hughey, J. (2002) Tailoring organizational characteristics for empowerment: accommodating individual economic resources. Journal of Community Practice, 10, 41–60.[CrossRef]
Peterson, N.A. and Hughey, J. (2004) Social cohesion and intrpersonal empowerment: gender as moderator. Health Education Research, 19, 533–542.
Pinsky, J.L., Leaverton, P.E. and Stokes, J. (1987) Predictors of good function: the Framingham study. Journal of Chronic Diseases, 40, 159S–167S.
Pols, M.A., Peeters, P.H., Bueno-De-Mesquita, H.B., Ocke, M.C., Wentink, C.A., Kemper, H.C. and Collette, H.J. (1995) Validity and repeatability of a modified Baecke questionnaire on physical activity. International Journal of Epidemiology, 24, 381–388.
Powell, K.E., Thompson, P.D., Caspersen, C.J. and Kendrick, J.S. (1987) Physical activity and the incidence of coronary heart disease. Annual Review of Public Health, 8, 253–287.[CrossRef][Web of Science][Medline]
Pruessner, J.C., Hellhammer, D.H. and Kirschbaum, C. (1999) Burnout, perceived stress and cortisol responses to awakening. Psychosomatic Medicine, 61, 197–204.
Rejeski, W.J. and Mihalko, S.L. (2001) Physical activity and quality of life in older adults. Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 56, 23–35.
Rejeski, W.J., Brawley, L.R. and Shumaker, S.A. (1996) Physical activity and health-related quality of life. Exercise and Sport Sciences Reviews, 24, 71–108.[Medline]
Rose, G. and Marmot, M.G. (1981) Social class and coronary heart disease. British Heart Journal, 45, 13–19.
Sallis, J.F., Linton, L. and Kraft, M.K. (2005) The First Active Living Research Conference: growth of a transdisciplinary field. American Journal of Preventive Medicine, 28, 93–95.[CrossRef][Web of Science][Medline]
Sapolsky, R.M. and Mott, G.E. (1987) Social subordinance in wild baboons is associated with suppressed high density lipoprotein–cholesterol concentrations: the possible role of chronic social stress. Endocrinology, 121, 1605–1610.
Sapolsky, R.M., Alberts, S.C. and Altmann, J. (1997) Hypercortisolism associated with social subordinance or social isolation among wild baboons. Archives of General Psychiatry, 54, 1137–1143.
Seeman, T.E., Charpentier, P.A., Berkman, L.F., Tinetti, M.E., Guralnik, J.M., Albert, M., Blazer, D. and Rowe, J.W. (1994) Predicting changes in physical performance in a high-functioning elderly cohort: MacArthur studies of successful aging. Journal of Gerontology, 49, M97–108.
Ware, J.E. and Kosinski, M. (2001) SF-36 Physical and Mental Health Summary Scales: A Manual for Users of Version 1, 2nd edn. QualityMetric, Lincoln, RI.
Ware, J.E. and Sherbourne, C.D. (1992) The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Medical Care, 30, 473–483.[Web of Science][Medline]
Ware, J.E., Kosinski, M. and Gandek, B. (2002) SF-36 Health Survey Manual and Interpretation Guide. QualityMetric, Lincoln, RI.
Received on September 20, 2004; accepted on July 8, 2005
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||