American Journal of Epidemiology Copyright 2007 by the Johns Hopkins Bloomberg School of Public Health All rights reserved;

; printed in U.S.A.

Vol. 165, No. 12 DOI: 10.1093/aje/kwm053 Advance Access publication April 9, 2007

Original Contribution Aging, Retirement, and Changes in Physical Activity: Prospective Cohort Findings from the GLOBE Study

Annabelle S. Slingerland1,2,3, Frank J. van Lenthe3, J. Wouter Jukema2, Carlijn B. M. Kamphuis3, Caspar Looman3, Katrina Giskes3,4, Martijn Huisman3, K. M. Venkat Narayan5, Johan P. Mackenbach3, and Johannes Brug3,6
1 2

Institute of Biomedical Sciences, Peninsula Medical School, Exeter, United Kingdom. Department of Cardiology, Leiden University Medical Center, Leiden, the Netherlands. 3 Department of Public Health, Erasmus-MC, University Medical Center Rotterdam, Rotterdam, the Netherlands. 4 Center for Health Research, School of Public Health, Queensland University of Technology, Brisbane, Australia. 5 Hubert Department of Global Health, The Rollins School of Public Health, Emory University, Atlanta, GA. 6 EMGO Institute, VU University Medical Center, Amsterdam, the Netherlands. Received for publication October 29, 2006; accepted for publication January 4, 2007.

Downloaded from http://aje.oxfordjournals.org/ by guest on February 20, 2013

There is increased recognition that determinants of health should be investigated in a life-course perspective. Retirement is a major transition in the life course and offers opportunities for changes in physical activity that may improve health in the aging population. The authors examined the effect of retirement on changes in physical activity in the GLOBE Study, a prospective cohort study known by the Dutch acronym for Health and Living Conditions of the Population of Eindhoven and surroundings, 19912004. They followed respondents (n 971) by postal questionnaire who were employed and aged 4065 years in 1991 for 13 years, after which they were still employed (n 287) or had retired (n 684). Physical activity included 1) work-related transportation, 2) sports participation, and 3) nonsports leisure-time physical activity. Multinomial logistic regression analyses indicated that retirement was associated with a signicantly higher odds for a decline in physical activity from work-related transportation (odds ratio (OR) 3.03, 95% condence interval (CI): 1.97, 4.65), adjusted for sex, age, marital status, chronic diseases, and education, compared with remaining employed. Retirement was not associated with an increase in sports participation (OR 1.12, 95% CI: 0.71, 1.75) or nonsports leisure-time physical activity (OR 0.80, 95% CI: 0.54, 1.19). In conclusion, retirement introduces a reduction in physical activity from work-related transportation that is not compensated for by an increase in sports participation or an increase in nonsports leisure-time physical activity. aging; cohort studies; leisure activities; motor activity; prospective studies; retirement; sports

Abbreviations: CI, condence interval; OR, odds ratio.

Currently 600 million people worldwide are aged 60 years or more, and this number will grow to 1.2 billion by the year 2025 according to the World Health Organization (1). Prevalence of major chronic diseases increases with age. For example, diabetes is found in one in ten individuals. By the time a person reaches 65 years of age (retirement), this increases to one in ve according to the US Centers for Dis-

ease Control and Prevention. People aged over 65 years account for almost 40 percent of the population with diabetes, and diabetes is the leading underlying cause of death in this age group (24). Not surprisingly, the Organization for Economic Cooperation and Development (5) reports that the medical burden and rise in public spending (from 6.7 to 13 percent of the gross domestic product by 2050) are mainly

Correspondence to Annabelle S. Slingerland, C. van Rijplantsoen 28, 1063 MC Amsterdam, the Netherlands (e-mail: annabelleslingerland@ hetnet.nl).

1356

Am J Epidemiol 2007;165:13561363

Changes in Physical Activity after Retirement

1357

due to the health-care demands of people this age. Interventions to improve physical activity and diet can decrease diabetes incidence by 3258 percent (4). There is strong evidence that, in the retired age group, physical activity can also decrease other diseases, falls, disability, and immobility (68). Physical activity can improve independence, since it contributes to maintenance of basic activities (such as climbing stairs and personal hygiene) and psychological functioning (by preventing isolation, depression, and, possibly, progression of dementia) (68). Before retirement age, work-related physical activity (e.g., manual work and transportation) contributes substantially to total physical activity while, at the same time, work-related time pressures and constraints are reported as main barriers to leisure-time physical activity (sports and nonsports) (9, 10). Therefore, although retirement is expected to result in a decline in physical activity from work and work-related transportation, it may also reduce time-related barriers to sports and nonsports leisure-time physical activities. Little is known about changes in physical activity after retirement. Studies are limited by their cross-sectional design, short follow-up, and/or restrictions on the scope of physical activities addressed (10, 11). Investigating specic aspects of physical activities (such as work-related transportation, sports, and nonsports leisure-time physical activity) is of utmost importance in directing possible ways of intervention. We examined the effect of retirement on changes in the three major aspects of physical activity (work-related transportation, sports, nonsports leisure time) over 13 years follow-up among employees aged 4065 years who participated in the GLOBE Study, a prospective cohort study known by the Dutch acronym for Health and Living Conditions of the Population of Eindhoven and surroundings, 19912004. Specically, we hypothesized that people who retired during follow-up would be more likely to have reduced their work-related physical activities but increased their sports and nonsports leisure-time physical activities.
MATERIALS AND METHODS Population

We sent participants from these three samples and who were still alive in 2004 a questionnaire to which 4,323 responded (response rate: 66.3 percent). For our analyses, we selected participants aged 4065 years who were employed in 1991 and responded in 2004 (n 1,246). We excluded participants with missing values for employment status in 2004 (n 34). On the basis of their employment status in 2004, we classied participants as retired (n 703, 58 percent), still employed (n 295, 24 percent), homemaker (n 74, 6 percent), unable to work (n 59, 5 percent), unemployed (n 27, 2 percent), owner of a company (n 12, 1 percent), other employment status (n 9, 1 percent), or nonresponder to this particular question (n 33, 3 percent). We limited subsequent analyses to the population who were either retired or employed after 13 years follow-up (n 998 (703 295)). Owner of a company was not included in analyses as participants in this category had no clear point of retirement. We performed analyses on retired and employed participants for whom complete data on age, sex, marital status, chronic disease, and education were available at baseline (n 971 (684 287)). Our study sample had more chronic disease than did the total population from which it was obtained (respondents of baseline postal questionnaire limited to those that were employed and aged 4065 years in 1991). There were no differences in terms of sex, age, marital status, or education (table 1).
Measurements Employment status and physical activity. We assessed employment status by a closed question, similar for 1991 and 2004, with the main categories being retired or employed as described above. Physical activity measures differed for 1991 and 2004. In 1991, participants reported their 1) work-related transportation physical activity, that is, time spent on walking and cycling to work (in minutes/day); 2) sports participation, that is, time per week spent on sport activities (hardly ever, <1 hour, 12 hours, >2 hours); and 3) nonsports leisure-time physical activity, that is, time per week spent on walking, cycling, and gardening during leisure time (hardly ever, <1 hour, 12 hours, >2 hours) (appendix 1). Work-related transportation time was categorized in the same way as were sports participation and nonsports leisure-time physical activity (appendix 1). In 2004, we used the standardized and validated questionnaire known as the SQUASH questionnaire (Short QUestionnaire to ASsess Health-enhancing physical activity) (13). For work-related transportation physical activity, participants reported frequency (days/week) and duration (minutes/day) of walking to work and cycling to work (appendix 2). Minutes per week of walking and cycling to work were summarized and recategorized (hardly ever, <1 hour, 12 hours, >2 hours). For sports participation and for nonsports leisure-time physical activity, the same procedure was followed. The latter comprised separate questions for walking, cycling, and gardening in leisure time, which aspects were summarized before being recategorized into the aforementioned four categories. We assessed a change in physical activity over

Downloaded from http://aje.oxfordjournals.org/ by guest on February 20, 2013

The GLOBE Study aimed to explain inequalities in health and living conditions in a cohort of 27,070 noninstitutionalized Dutch inhabitants, aged 1574 years in 1991 and living in the city of Eindhoven and surroundings (the southern part of the Netherlands). A detailed description of the purpose and design of the GLOBE Study is presented elsewhere (12). We collected, with a response rate of 70.1 percent, baseline information from 18,973 inhabitants using a postal questionnaire in 1991. Nonrespondents showed no differences in age, gender, or socioeconomic position (12). We sent a follow-up postal questionnaire in 2004 to three samples taken from the respondent population in 1991. The rst sample was a random sample comprising 2,800 respondents. The second was a sample with an overrepresentation of participants with chronic diseases compared with respondents, comprising 2,867 individuals. The third sample was of respondents who resided in Eindhoven and surroundings in 1991 and who still resided there in 2004 (n 2,190).
Am J Epidemiol 2007;165:13561363

1358 Slingerland et al.

TABLE 1. Baseline characteristics of the total study population and stratied according to employment status after 13 years follow-up, the GLOBE* Study, 19912004
Total population that responded to the baseline postal questionnaire, being employed, and aged 4065 years in 1991 No. % Total study population of the three samples at baseline, being employed, and aged 4065 years in 1991 No. % Total study population, restricted to employment status in 2004 (employed or retired), excluding those missing Total No. % Stratied according to employment status after 13 years follow-up Employed No. % Retired p valuey No. %

No. of participants Sex Male Female Age (years)z Marital status Married Single/divorced/widowed Missing No. of current chronic diseases 0 1 2 Education 1 (low) 2 3 4 (high) Missing

4,257 2,980 1,277 49.5 (5.4) 3,572 652 33 84 15 1 70 30

1,246 857 389 49.6 (5.3) 1,048 189 9 84 15 1 69 31

971 729 242 50.0 (5.3) 820 151 84 16 75 25

287 204 83 44.4 (3.0) 246 41 86 14 71 29

684 0.07 525 159 52.4 (4.3) 574 110 84 16 77 23 <0.001 0.44

Downloaded from http://aje.oxfordjournals.org/ by guest on February 20, 2013

<0.001 2,428 1,208 621 594 1,673 870 1,023 97 57 28 15 14 39 20 24 2 593 437 216 159 469 270 324 24 48 35 17 13 38 22 26 2 469 342 160 113 348 216 294 48 35 17 12 36 22 30 155 98 34 25 98 63 100 54 34 12 9 34 22 35 314 244 126 88 249 153 194 46 36 18 0.01 13 37 22 28

* GLOBE, Dutch acronym for Health and Living Conditions of the Population of Eindhoven and surroundings. y Two-sided p value of baseline values according to employed versus retired status after 13 years of follow-up; the v2 test was used for categorical variables, and the independent t test was used for the continuous variable age. z Age is expressed as mean (standard deviation). Education refers to the highest educational level attained, with 1 elementary education (primary school), 2 lower secondary (lower vocational) education, 3 higher secondary (intermediate vocational) education, and 4 tertiary (higher vocational) education.

the 13 years of follow-up by subtracting categories of physical activity in 1991 from the physical activity in 2004 for each of the aspects of physical activity. Subsequently, we expressed a downward change in categories as decline in physical activity and an upward change as an increase in physical activity. Covariates. Participants reported baseline covariates in the postal questionnaire of 1991 that included questions on sex (male/female), age (years), marital status, having chronic diseases, and education. We adjusted for these factors in subsequent analysis. Questions on chronic disease were part of a standard Dutch 24-item checklist that is used for national health statistics research (Statistics Netherlands general survey), comprising heart disease, pulmonary disease, stroke, peptic ulcer, kidney disorders, diabetes, rheumatism or arthritis, illness of the nervous system, and cancer. From this checklist, we calculated the number of chronic diseases and categorized participants as having no chronic disease, having one chronic disease, or having two or more chronic

diseases. The question on education asked for the highest educational level attained and had eight categories that were recategorized into four: 1) elementary education (primary school), 2) lower secondary education (lower vocational), 3) higher secondary education (intermediate vocational), and 4) tertiary education (higher vocational).

Analyses

We applied weighing factors to make the sample representative of the original source population, taking into account the sampling design and nonresponse. We described the continuous variable age by mean (standard deviation) and categorical variables by frequency (percent) and odds ratio (95 percent condence intervals). We compared retired participants with participants who were still employed in 2004 by using independent t tests for the normally distributed continuous variable age and v2 tests for comparing
Am J Epidemiol 2007;165:13561363

Changes in Physical Activity after Retirement

1359

TABLE 2. Physical activity of the study population at baseline and after 13-year follow-up, stratied by employment status after 13 years, the GLOBE* Study, 19912004
Employed after follow-up Physical activity in 1991 No. % Physical activity in 2004 No. % Retired after follow-up Physical activity in 1991 No. % Physical activity in 2004 No. %

Work-related transport physical activitytime spent on walking and cycling to work Hardly ever <1 hour/week 12 hours/week >2 hours/week Total (all categories) Sports participationtime spent on sports Hardly ever <1 hour/week 12 hours/week >2 hours/week Total (all categories) Nonsports leisure-time physical activitytime spent on walking, cycling, and gardening in leisure time Hardly ever <1 hour/week 12 hours/week >2 hours/week Total (all categories) 35 45 78 126 284 12 16 28 44 32 23 24 205 284 11 8 9 72 69 106 195 312 682 10 15 29 46 57 18 26 581 682 8 3 4 85 116 23 79 65 283 41 8 28 23 139 7 39 98 283 49 2 14 35 361 53 114 138 666 54 8 17 21 341 15 80 230 666 51 2 12 35 76 48 41 115 280 27 17 15 41 155 15 28 82 280 55 5 10 29 158 102 95 295 650 24 16 15 45 587 7 4 52 650 90 1 1 8

Downloaded from http://aje.oxfordjournals.org/ by guest on February 20, 2013

* GLOBE, Dutch acronym for Health and Living Conditions of the Population of Eindhoven and surroundings.

baseline categorical variables and the categorical variables of physical activity. To explore the effect of retirement on the change in physical activity between baseline and follow-up, we used multinomial logistic regression. Full models without missing values were used for each of the aspects of physical activity (transportation, leisure time, and sports). We calculated odds ratios and 95 percent condence intervals for the decline and increase in physical activity with no change as the referent of retired compared with employed participants 13 years after follow-up. SPSS, version 10, software (SPSS, Inc., Chicago, Illinois) was used to analyze the data. All tests were two sided, and signicance was dened as p < 0.05 or if the condence interval did not contain the value of unity.
RESULTS

Among the employed participants at baseline (4065 years of age), 70 percent had retired during 13 years of follow-up. As expected, retired participants were older at baseline than participants still employed at follow-up. Furthermore, retirees reported chronic disease more frequently and were generally less educated (table 1).
Am J Epidemiol 2007;165:13561363

There were small baseline differences between participants who had retired and those still employed after 13 years of follow-up (table 2). The proportion of participants that hardly ever spent time on sports participation was lower at baseline in those that were still employed after 13 years of follow-up than in those who had retired. At the end of follow-up, the proportion of retirees spending time on workrelated transportation physical activity had dropped substantially. The proportion of participants that hardly ever spent time on work-related physical activity rose from 55 percent to 90 percent. The reduction among those still employed was much smaller (from 27 percent to 24 percent). Those still employed had increased their sports participation while the retirees had not, and the distribution of respondents over the three sports participation categories was similar between retirees and those still employed. The proportion of retirees that participated in nonsports leisure-time physical activity was slightly higher than the proportion among those still employed. The difference was most apparent in the highest category of nonsports leisure-time activity (>2 hours/week). Multinomial logistic regression analyses indicated that retirement was associated with a signicantly higher odds for a decline in work-related transportation physical activity

1360 Slingerland et al.

TABLE 3. Change in physical activity and odds ratios (condence interval) of the change in physical activity over 13 years for the retired versus employed after a 13-year follow-up adjusted for age, sex, marital status, chronic disease, and education at baseline, the GLOBE* Study, 19912004
Employed Change in physical activity over 13 years No. % p valuey No. % Retired Odds ratio Retired vs. employed 95% condence interval p valuez

Work-related transport physical activitytime spent on walking and cycling to work Decline No change Increase Total (all categories) Sports participationtime spent on sports Decline No change Increase Total (all categories) Nonsports leisure-time physical activitytime spent on walking, cycling, and gardening in leisure time Decline No change Increase Total (all categories) 49 119 116 284 17 42 41 0.002 66 300 316 682 10 44 46 0.36 1.00 0.80 0.54, 1.19 0.27 0.19, 0.68 0.001 64 147 72 283 23 52 25 0.001 101 368 197 666 15 55 30 0.66 1.00 1.12 0.39, 1.10 0.71, 1.75 0.11 119 122 39 280 42 44 14 <0.001 445 177 28 650 69 27 4 3.03 1.00 0.51 0.22, 1.16 0.11 1.97, 4.65 0.001

Downloaded from http://aje.oxfordjournals.org/ by guest on February 20, 2013

0.63

* GLOBE, Dutch acronym for Health and Living Conditions of the Population of Eindhoven and surroundings. y v2 tests were used for each of the categorical variables of physical activity: transport, sports, and nonsports leisure-time physical activity. z Two-sided p value of employed versus retired status after 13 years of follow-up for change in physical activity over these 13 years. Full models without missing values on physical activity were used in multinomial logistic regression to analyze adjusted odds ratios for a decline and increase in physical activity.

(odds ratio (OR) 3.03, 95 percent condence interval (CI): 1.97, 4.65, adjusted for sex, age, marital status, chronic disease, and education) (table 3). Remarkably, retirement was not signicantly associated with a decline (OR 0.66, 95 percent CI: 0.39, 1.10) or an increase (OR 1.12, 95 percent CI: 0.71, 1.75) in sports participation. Retirement was associated with a signicantly lower odds for a decline in nonsports leisure-time physical activity (OR 0.35, 95 percent CI: 0.19, 0.68) compared with those remaining employed but not with an increase in nonsports leisure-time physical activity (OR 0.80, 95 percent CI: 0.19, 0.68).
DISCUSSION

In a 13-year prospective follow-up into retirement of employees aged 4065 years at baseline, work-related transportation physical activity greatly reduced. This reduction was not compensated for by increased sports participation or nonsports leisure-time physical activity, with retirement associated with a lower odds for a decline in nonsports leisure-time physical activity only. These results indicate

that retirement results in a net reduction in physical activity rates. Our study is, to our knowledge, the rst longer-term prospective cohort study investigating the effect of retirement on physical activity. It is the rst to separately address workrelated transportation physical activity, sports participation, and nonsports leisure-time physical activity; the results conrm the importance of making this distinction. A number of limitations should be acknowledged. One potential limitation of the study was the use of self-reported, single-item assessments for the different categories of physical activity. We have, however, no reason to expect that the vulnerability to the potential bias of giving socially desirable answers would be different according to employment status. At baseline, no validated physical activity questionnaire was available and, therefore, at follow-up the standardized and validated SQUASH questionnaire was available and adopted. Using two different physical activity questionnaires may have biased our results, but both questionnaires were quite similar in format and content and allowed categorization in identical categories of physical activity. There is no reason
Am J Epidemiol 2007;165:13561363

Changes in Physical Activity after Retirement

1361

to believe that differential misclassication was stronger in one of the questionnaires by employment status. Another potential limitation could be the external validity of our ndings. Our study was carried out in a medium-sized city in the Netherlands (190,000 inhabitants) and, although it seems reasonable to generalize our ndings to other cities of similar sizes, ndings might have been different if the study had been conducted in more rural or larger city environments. A possible selection bias could have resulted if participants of ill health had retired earlier because of their ill health and if their ill health rather than their retirement per se negatively inuenced their physical activity. We adjusted for chronic disease and have, therefore, minimized the possibility of this selection bias. Selection bias could have also resulted if the retirees with ill health had died without our knowing, resulting in the retired participants being healthier (the healthy worker effect). In that case, our results are an underestimation of the true effects, and the absence of compensatory sports participation and nonsports leisuretime physical activity cannot be ascribed to health issues. Our results are, to some degree, consistent with those from a study that showed that people who retired were more likely to maintain leisure-time physical activity than were those who continued working (11). The authors of that study had suggested extending their follow-up of 6 years, as they were uncertain that their results were maintained. We extended that study and showed that their ndings of a protective effect of retirement on a decline in physical activity is not maintained 13 years after follow-up. In addition, by including work-related transportation physical activity, we showed that losses in this area would not be compensated for by just maintenance of preretirement physical activity. Currently, the average home-to-work distance in the Netherlands is 18.2 km. Twenty-ve percent commute by bike, while 3 percent walk to work, taking 15 minutes and 10 minutes on average (Statistics Netherlands). The net decline in physical activity would be even larger if loss in physical activity due to loss in manual work would have been taken into account (14). There are several explanations for the failure to compensate for a loss in work-related physical activity after retirement. Time pressures and work-constraints before retirement might be perceived rather than actual barriers to physical activity. Alternatively, time pressures and constraints might continue from other sources, but our study lacks such data. Aging might be another explanation. However, our data were adjusted for this, as well as for possible increased chronic disease. Moreover, according to a recent study (15), increased age is no limitation for physical activity and, hence, one should not infer with expectations that the extra time after retirement would be used for extra physical activity. This implies that retirement, seen as a hallmark in the transition to old age, should still open excellent opportunities to increase physical activity (16). Our results are reason for concern, since retirement appears to be associated with a relative reduction in physical activity, while the objective opportunities for being more physically active should be better after retirement (16). Remaining physically active at older age has great health benets, and it is therefore important that physical activity
Am J Epidemiol 2007;165:13561363

promotion should be targeted toward people who are preparing for retirement and who have recently retired (14, 1723). Further research should be aimed at replication of the present study ndings in other study populations and at investigations of the underlying social, physical, environmental, and personal factors that may encourage or discourage leisure-time physical activity after retirement in order to inform tailored physical-activity-promotion interventions (24). Future such studies also need to investigate the impact of a net decrease in physical activity during retirement on outcome measures, such as depression, disability, quality of life, and health-care costs.

ACKNOWLEDGMENTS

The Prevention Fund and the Dutch Ministry of Public Health, Welfare, and Sports have nancially supported the GLOBE Study. The study was conducted by the Department of Public Health of the Erasmus University Rotterdam, in collaboration with the Public Health Services of the city of Eindhoven and the region of South-East Brabant. The authors wish to thank all collaborators without whose help this study would have been impossible, as well as Dr. Andrew T. Hattersley and Lorna W. Harries, Institute of Biomedical and Clinical Sciences, Peninsula Medical School, Exeter, United Kingdom. They also would like to acknowledge the organizers and participants of Diabetes, Exercise, and Sports Association 2006; the Cambridge Diabetes Seminar 2006; and Dr. Peter H. Bennett, Epidemiology and Clinical Research Branch, Phoenix, Arizona, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, for suggestions and helpful comments. This paper has been dedicated to Dr. G. Jan (Mu) Bruining and Dr. Peter H. Bennett, professors of diabetes who refused to retire, and to J. M. Slingerland and A. M. Slingerland-Wennekers, who actively integrated physical activity into retirement planning. Conict of interest: none declared.

Downloaded from http://aje.oxfordjournals.org/ by guest on February 20, 2013

REFERENCES 1. World Health Organization. Health topics. Ageing and life course. Geneva, Switzerland: World Health Organization, 2006. (http://www.who.int/ageing/en). 2. Narayan KM, Boyle JP, Geiss LS, et al. Impact of recent increase in incidence on future diabetes burden: U.S., 2005 2050. Diabetes Care 2006;29:211416. 3. Narayan KM, Boyle JP, Thompson TJ, et al. Lifetime risk for diabetes mellitus in the United States. JAMA 2003;290: 188490. 4. Centers for Disease Control and Prevention. Diabetes public health resource. Atlanta, GA: Division of Diabetes Translation, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, 2006. (www.cdc.gov/diabetes/). 5. Organization for Economic Cooperation and Development. Study projects growing pressure on public health spending

1362 Slingerland et al.

6. 7. 8.

9.

10. 11. 12.

13. 14.

over and above effects of ageing society. Paris, France: Organization for Economic Cooperation and Development, 2006. (http://www.oecd.org/topic/). Mensink GB, Ziese T, Kok FJ. Benets of leisure-time physical activity on the cardiovascular risk prole at older age. Int J Epidemiol 1999;28:65966. Young A, Dinan S. Activity in later life. BMJ 2005;330: 18991. Haight T, Tager I, Sternfeld B, et al. Effects of body composition and leisure-time physical activity on transitions in physical functioning in the elderly. Am J Epidemiol 2005;162: 60717. Popham F, Mitchell R. Leisure time exercise and personal circumstances in the working age population: longitudinal analysis of the British household panel survey. J Epidemiol Community Health 2006;60:2704. Mein GK, Shipley MJ, Hillsdon M, et al. Work, retirement and physical activity: cross-sectional analyses from the Whitehall II study. Eur J Public Health 2005;15:31722. Evenson KR, Rosamond WD, Cai J, et al. Inuence of retirement on leisure-time physical activity: the Atherosclerosis Risk in Communities Study. Am J Epidemiol 2002;155:6929. Mackenbach JP, van de Mheen H, Stronks K. A prospective cohort study investigating the explanation of socio-economic inequalities in health in the Netherlands. Soc Sci Med 1994; 38:299308. Wendel-Vos GC, Schuit AJ, Saris WH, et al. Reproducibility and relative validity of the short questionnaire to assess healthenhancing physical activity. J Clin Epidemiol 2003;56:11639. Nooyens AC, Visscher TL, Schuit AJ, et al. Effects of retirement on lifestyle in relation to changes in weight and waist

15.

16. 17. 18.

19. 20. 21. 22. 23. 24.

circumference in Dutch men: a prospective study. Public Health Nutr 2005;8:126674. Aijanseppa S, Notkola IL, Tijhuis M, et al. Physical functioning in elderly Europeans: 10 year changes in the North and South: the HALE project. J Epidemiol Community Health 2005;59:41319. Keller KB, Lemberg L. Retirement is no excuse for physical inactivity or isolation. Am J Crit Care 2002;11:2702. Caspersen CJ, Pereira MA, Curran KM. Changes in physical activity patterns in the United States, by sex and crosssectional age. Med Sci Sports Exerc 2000;32:16019. Harland J, White M, Drinkwater C, et al. The Newcastle exercise project: a randomised controlled trial of methods to promote physical activity in primary care. BMJ 1999;319: 82832. Elley CR, Kerse N, Arroll B, et al. Effectiveness of counselling patients on physical activity in general practice: cluster randomised controlled trial. BMJ 2003;326:7938. Davies BL, Matte-Lewis L, OConnor AM, et al. Evaluation of the Time to Quit self-help smoking cessation program. Can J Public Health 1992;83:1923. Davis AL. Fighting obesity: physical activity has major role. BMJ 2004;329:53. Siegenthaler JK, Brenner AM. Flexible work schedules, older workers, and retirement. J Aging Soc Policy 2000;12:1934. Nothwehr F, Stump T. Health-promoting behaviors among adults with type 2 diabetes: ndings from the Health and Retirement Study. Prev Med 2000;30:40714. Brug J, Oenema A, Ferreira I. Theory, evidence and intervention mapping to improve behavior nutrition and physical activity interventions. Int J Behav Nutr Phys Act 2005;2:28.

Downloaded from http://aje.oxfordjournals.org/ by guest on February 20, 2013

APPENDIX 1 Postal Questionnaire 1991: Questions Used to Assess Self-reported Physical Activity Work-related transport physical activity. How many minutes do you spend on walking or cycling to work and/or shops on an

average day? About . . . minutes a day?

Leisure time physical activity. How many hours of your leisure time a week do you spend on walking, cycling, gardening,

letting the dog out?     hardly ever less than 1 hour 12 hours more than 2 hours
Sports participation. Are you participating in sports?

   

hardly ever less than 1 hour 12 hours more than 2 hours

APPENDIX 2 Postal Questionnaire 2004: Questions Used to Assess Self-reported Physical Activity

Imagine a normal week of the past few months. Could you state how many days a week you participated in physical activity and how many hours and minutes you spent on an average day?
Am J Epidemiol 2007;165:13561363

Changes in Physical Activity after Retirement

1363

Work-related transportation physical activity.

Walking to/from work or school Cycling to/from work or school

Leisure-time physical activity.

No. of days a week h days h days

Average time a day h hours h minutes h hours h minutes

Walking Cycling Gardening

Sports participation.

No. of days a week h days h days h days

Average time a day h hours h minutes h hours h minutes h hours h minutes

Sport Sport Sport Sport

1 2 3 4

No. of days a week h days h days h days h days

Average time h hours h h hours h h hours h h hours h

a day minutes minutes minutes minutes

Downloaded from http://aje.oxfordjournals.org/ by guest on February 20, 2013

Am J Epidemiol 2007;165:13561363