A life-course perspective on physical functioning in women
Geeske Peeters a, Annette J Dobson b, Dorly JH Deeg c & Wendy J Brown d
a. Schools of Human Movement Studies and Population Health, The University of Queensland, Bldg 26b Blair Drive, St Lucia, Queensland 4072, Australia.
b. School of Population Health, The University of Queensland, Brisbane, Australia.
c. Department of Epidemiology and Biostatistics, VU University Medical Center, Amsterdam, Netherlands.
d. School of Human Movement Studies, The University of Queensland, Brisbane, Australia.
Correspondence to Geeske Peeters (e-mail: firstname.lastname@example.org).
(Submitted: 11 April 2013 – Accepted: 23 July 2013.)
Bulletin of the World Health Organization 2013;91:661-670. doi: http://dx.doi.org/10.2471/BLT.13.123075
In 1997, Kalache & Kickbusch, then of the World Health Organization’s Ageing and Health programme, published a report1 containing a figure (Fig. 1) that summarized a concept unchallenged until now: that functional capacities – i.e. respiratory capacity, muscular strength and cardiovascular performance – increase and peak during early adulthood and then decline linearly with advancing age. The two diverging lines in the graphic, which show how widely physical capacity can theoretically vary over the lifespan across the population, suggest that in early adulthood everyone has similar functional capacity but that the subsequent rate of decline depends on lifestyle and environmental factors. As people age, this creates an increasing gap in capacity across the population – the so-called fitness gap. Consequently, people with lower peak functional capacity reach the “disability threshold” – i.e. need help with daily activities and care – at a younger age than those who start off with higher peak functional capacity. The shaded boxes indicate the interventions applicable at different stages in life as dictated by this model.
Fig. 1. Theoretical framework illustrating changes in functional capacity across the life course and potential interventions for maintaining the highest possible level of functional capacity
The model represented by the figure described above has traditionally inspired health promotion initiatives worldwide.2,3 However, no study that we know of has tested the validity of the model, specifically: (i) the shape describing the relationship between age and functioning; (ii) the rate of decline as a function of peak capacity level and the consequential fitness gap; and (iii) the moment in life in which the disability threshold is reached. Generating better evidence surrounding all of these aspects is important for planning interventions.
The figure developed by Kalache & Kickbusch focused on “functional capacity”, a term synonymous with “fitness”.1 These terms, together with “physical functioning”, denote the individual’s capacity to undertake everyday tasks.4 A different but related concept is that of physical activity, defined as “any bodily movement produced by skeletal muscles that results in energy expenditure”.5 “Physical activity” refers, in other words, not to the capacity to do something, but to what one actually does. Previous studies of mid-age and older adults show that more active people have better levels of physical functioning.6–13 The proposed mechanism behind this association is that physical activity helps to maintain muscle and cardiac function and subsequently prevents functional decline.14,15 If this is true, patterns of functional capacity may differ for people with low and high levels of physical activity.
The aim of this study was to verify and quantify the theoretical model suggested by Kalache & Kickbusch by answering the following research questions:
- What is the rate of decline in physical functioning at different stages of the adult life course?
- What is the average age at which women reach the disability threshold?
Since data on functional capacities were not available for this study, we focused on physical functioning as measured with the Short Form (36) Health Survey (SF-36). To quantify the fitness gap, we stratified the analyses by baseline level of physical functioning. To determine whether physical activity influences the rate of decline in physical functioning and hence the age when the disability threshold is reached, we also stratified the analyses for baseline level of physical activity.
Data were obtained from the Australian Longitudinal Study on Women’s Health, a large population-based study of factors affecting health and well-being in three cohorts of women: a younger cohort (birth in 1973–78); a mid-age cohort (birth in 1946–51) and an older cohort (birth in 1921–26).16 The study methods were approved by the human research ethics committees of the Universities of Newcastle and Queensland and all participants signed informed consent. Detailed information on design, recruitment and attrition can be found elsewhere.16,17 Briefly, in 1996 women aged 18–23, 45–50 and 70–75 were randomly selected from the Medicare database, which includes all Australian citizens and permanent residents, and were mailed information packs, an invitation to participate in the study and survey materials, including the SF-36.18,19 No selection or exclusion criteria were applied. Women from rural and remote areas were oversampled to capture the experiences of women outside the metropolitan area. After 1996 follow-up surveys were completed on a rolling basis at 2–3-year intervals until 2011: four in the younger cohort and five in the mid-age and older cohorts. The 1996 survey included 14 247 younger, 13 715 mid-age and 12 432 older women, respectively. Table 1 shows participant retention rates.
Table 1. Retention rates in the three cohorts of the Australian Longitudinal Study on Women’s Health
As we were interested in functioning over the life course and in the general population, we did not select participants based on disease status. If women had any diseases that limited their functioning, they were included in the study nonetheless.
Physical functioning and disability threshold
The SF-36 subscale18,19 explores, through 10 items, whether any health-related conditions limit the respondent’s ability to perform a range of daily tasks, some involving vigorous and moderate activity, or to climb stairs, move about and engage in self-care. The score ranges from 0 to 100 and higher scores indicate better physical functioning. In each cohort, we classified baseline physical functioning into quintiles to represent physical functioning as a continuum while also capturing the groups with the best and poorest levels of functioning.
To quantify the disability threshold, we calculated the mean physical functioning score obtained in the fourth survey (2005) in women in the oldest cohort who had reached the disability threshold – i.e. those who reported needing help from another person to carry out any of several activities. These included grooming, eating, bathing/showering, dressing, getting up from a chair, walking inside the house, using the toilet, shopping for groceries, doing light or heavy housework, managing money, preparing meals, taking medications, using the telephone and engaging in leisure activities or hobbies. Questions and response options for these items were based on Gill et al. (1998).20
The baseline level of physical activity was assessed from responses to questions about the frequency of vigorous exercise (e.g. sports, aerobics) and moderate exercise (e.g. walking, swimming) in a usual week.21 Participants were classified as engaging in the following levels of physical activity: none (moderate activity 0–1 times per week); low (moderate activity 2–4 times or vigorous activity 1–2 times per week); moderate (moderate activity 5–8 times or vigorous activity 3–5 times per week); or high (moderate activity ≥ 8 times or vigorous activity ≥ 5 times per week, or equivalent combination).
We explored demographic variables and a few lifestyle variables besides physical activity for descriptive purposes. Area of residence was classified as urban, rural or remote. Response options for the highest educational qualification completed were collapsed into high school or lower and post-high school. Adequacy of the respondent’s income was assessed by asking: “How do you manage on the income you have available?” Response options were collapsed into “impossible/difficult” or “not too bad/easy”. Those reporting occasional or regular smoking at the time of the survey were classified as current smokers. Risky level of alcohol consumption was defined as 15 glasses or more per week or 3 or more per day.22 (Further details can be obtained from the corresponding author on request.)
We explored the baseline characteristics of each cohort separately. We compared women with complete data on physical functioning throughout all follow-up surveys with women with incomplete data using t-tests for age and χ2 tests for categorical variables.
To quantify the rate of decline in physical functioning in each cohort, we used spline regression and assumed linear associations between knots. The first knot was placed at the upper end of the baseline age interval of each cohort (i.e. at ages 23, 50 and 75) to visualize potential regression-to-the-mean effects. We selected additional knots for each cohort separately by starting with a model that included knots at three-year intervals (e.g. younger cohort: ages 20, 23, 26, 29, 32 and 35) and subsequently removing knots that gave no statistically significant (P ≥ 0.05) difference in regression coefficient (slope) with respect to the preceding age interval. We used robust standard errors to account for within-subject correlation. We conducted the main analyses only with cases having complete data but performed sensitivity analysis with all participants who had provided any data at any point. Using the knots as selected in the total sample, we plotted physical functioning against age for (i) each quintile of baseline physical functioning and (ii) each level of baseline physical activity. We used graphs to estimate the mean age when the disability threshold was reached in each subgroup.
Of the 14 247, 13 715 and 12 432 younger, mid-age and older women who returned the first survey in 1996, 5635, 8092 and 2999 provided complete data on physical functioning over all five (younger women) or six (mid-age and older women) surveys. The women with complete data were on average 20.8 (standard deviation, SD: 1.5), 47.6 (SD: 1.5) and 72.3 (SD: 1.4) years old at baseline, respectively. Women with complete data were the same age as those with incomplete data but were better educated, less likely to report being smokers or having difficulty managing on their income and more likely to engage in moderate physical activity, and they had higher levels of baseline physical functioning (P < 0.001) (Table 2).
Table 2. Baseline (1996) characteristics of the younger, mid-age and older women in the Australian Longitudinal Study on Women’s Health
Overall, physical functioning declined with age and more rapid declines were observed at higher ages (Fig. 2). The decline ranged from none in women aged 18–23 years (slope: 0.05; 95% confidence interval, CI: −0.13 to 0.22) to more than 2 units annually in women aged 82–90 (slope −2.43; 95% CI: −2.64 to −2.23) (Table 3). The disability threshold – i.e. mean score for physical functioning in older women who reported that they needed help with daily activities – was 62.8. Across the three cohorts, the average age when the disability threshold was reached was 79 years. In the sensitivity analyses with participants with any data (nyounger = 14 243; nmid-age = 13 709; nolder = 12 409), a similar, slightly attenuated decline was observed (Table 3, Fig. 2) and the disability threshold was reached at an average age of 75 years.
Fig. 2. Decline in physical functioning with age
Table 3. Rate of decline in physical functioning in each of the three cohorts in the Australian Longitudinal Study on Women’s Health
In all three birth cohorts, the patterns of physical functioning over time clearly differed as a function of baseline physical functioning quintile (Fig. 3). In the first years of follow-up, women with high baseline physical functioning showed a sharp decline, whereas those with poor baseline functioning showed a sharp increase. Towards the end of the follow-up, however, the rate of decline was greater in quintiles with lower baseline physical functioning in the younger and mid-age cohorts. (This was evidenced by greater negative slopes that fell outside the 95% CIs of the higher quintiles, Table 4). In contrast, in older women the rate of decline during the later years of follow-up was greater among those with higher baseline physical functioning (Table 4).
Fig. 3. Decline in physical functioning with age, by baseline quintile of physical functioning
Table 4. Rate of decline in physical functioning in each of the three cohorts in the Australian Longitudinal Study on Women’s Health, by quintile of baseline physical functioning
No younger or mid-age women who were in the four highest quintiles of baseline physical functioning reached the disability threshold during follow-up (Fig. 3). In younger women, the lowest quintile started at the disability threshold but improved rapidly (possibly due to regression-to-the-mean) and remained above this threshold during follow-up. In mid-age women, the lowest quintile started at the disability threshold and remained below it during follow-up. In the oldest women, the lowest two quintiles started at the threshold and remained below it throughout follow-up, whereas in the three highest quintiles, women started above the threshold but ended below it. In these three quintiles, women reached the disability threshold at an average age of 77.5, 83.5 and 88 years, respectively.
When the relationship was stratified by baseline level of physical activity, those with higher levels of activity started at higher levels of functioning, but rates of decline were similar in all activity groups (Fig. 4). In the younger and mid-age cohorts, the levels of physical functioning remained above the disability threshold, regardless of level of physical activity. In the oldest cohort, the disability threshold was reached at an average age of 70, 78, 80.5 and 84 years in the groups that reported no physical activity, low activity, moderate activity and high activity, respectively.
Fig. 4. Decline in physical functioning with age, by baseline level of physical activity
The aim of this study was to verify and quantify the theoretical physical capacity model suggested by Kalache & Kickbusch.1 Clear differences were observed between the theoretical model (Fig. 1) and the empirical data in various respects: (i) the shape of the relationship between age and functioning (Fig. 2); (ii) the rate of decline as a function of physical capacity levels at baseline (Fig. 3); (iii) the spread of physical functioning scores (Fig. 3); and (iv) the age when the disability threshold is reached. The most important findings of this study are that the rate of decline of physical functioning is not constant, as postulated in the original paper, but rather, that it increases with age. Furthermore, there is considerable spread in physical functioning (fitness gap) throughout the life course.
The first difference between the theoretical and empirical models lies in the shape of the modelled relationship between age and functioning. This is not linear, as in the original model, but curvilinear, with decline occurring more rapidly at older ages. To our knowledge, only one other study has published life-course data on physical functioning and that study showed a similar curvilinear pattern.23 Furthermore, the theoretical model suggested that functioning improves in childhood and adolescence but undergoes a sharp decline when a person moves into adulthood.1 Our results have shown that functioning levels are stable between the ages of 18 and 23 years (Table 2), but that a small but statistically significant decline ensues. The rate of decline was greater at younger ages in the mid-age cohort, although throughout the mid-age period the average decline seemed to slow down rather than to accelerate. One explanation for the higher rate of decline observed in women aged 45–50 years when compared with women aged 51–64 may be that in the former age group the menopause causes symptoms that lead women to perceive their physical functioning as declining more rapidly.24,25 In women moving from middle into early old age, the rate of decline was in approximately the same range, but in old age it increased from 0.5 to 2.4 units annually (on a 0–100 scale).
The second difference between the theoretical model and our findings is that, even at younger ages, physical functioning varies widely (Fig. 3). This finding is clinically important because it means that interventions to help those with poorer physical functioning at baseline need to be implemented at a much younger age. Furthermore, in all three cohorts, a sharp decrease in physical functioning was observed in women in the highest quintile of baseline physical functioning, whereas a sharp increase was observed in women in the lowest quintile. This suggests a regression to the mean between the first and second surveys.26 After the first knot in the spline models (and the fading out of the regression-to-the-mean effect), the rates of decline were greater in younger and mid-age women with lower baseline physical functioning than in those with higher baseline physical functioning. These patterns were in line with those of two studies in which cluster analyses were performed to identify trajectories of disabilities.27,28 According to these studies, which were conducted in adults aged 15–74 and 55–85 years at baseline who were followed up for 5 and 6 years, respectively, trajectories characterized by more disabilities at baseline had more rapid rates of decline over time.27,28 In the current study, however, the opposite was found among the oldest women, in whom a faster rate of decline was observed in those with higher baseline functioning. Again, this deviates from the pattern depicted by Kalache & Kickbusch but is in line with the findings of a study in which trajectory modelling was used to identify patterns of disability in men and women 70 years of age and older who were followed up for 10 years.29 It may be that the levels of baseline physical functioning in the lowest quintiles were so low, particularly in the older women, that there was little room for further decline on the scale used.
The third noteworthy limitation of the theoretical model is its assumption that disability occurs only in older ages, whereas the empirical model shows that disability occurs at all ages. In our study, the average functioning levels of the lowest 20% of the sample of mid-aged women fell below the disability threshold and stayed below it during follow-up. Other studies have also quantified disability in middle adulthood. For example, in a representative sample of Swedish adults aged 18–75 years, mobility began to decline around the age of 40.30 In a similar study in Belgium, moderate activity disabilities arose around the age of 25–34, while prevalence of severe activity disabilities increased at higher ages.31 In a Danish cohort of men and women aged 40 and 50 at baseline, the prevalence of difficulty climbing stairs increased from 4% to 32% during 7 years of follow-up.32 In the US Behavioural Risk Factor Surveillance System, 29% of 50–65 year old men and women reported having limitations in activity or requiring aids or adaptations.33 Interventions for improving physical functioning and quality of life should therefore not be restricted to older women, but should include women of all ages with disabilities.
When we stratified physical functioning by baseline level of physical activity, we found that activity was associated with the baseline level of physical functioning but not with the course of physical functioning over time. A similar finding was reported in a study of 1297 people aged 55–85 years in the Netherlands. In this cohort, being physically inactive at older ages was not associated with a greater risk of physical decline. In contrast, both this Dutch study and a British study of mid-age adults showed that level of physical activity throughout adulthood was associated with physical functioning later in life.6,34 As physical activity levels tend to fluctuate over time,35 cumulative activity may be more important for functioning over time than activity at any given point. Also, high physical activity before the decline begins is important, since it is associated with higher baseline physical functioning and thus with later onset of disability.
Among the strengths of this study is the large population-based sample composed of three birth cohorts who provided 15 years of follow-up. A large proportion of the original sample (60%, 41% and 76% in the younger, mid-age and older cohorts, respectively) had missing data or had died or dropped out. Women with incomplete data were less educated, were more likely to be smokers and had lower levels of baseline physical functioning and physical activity than women with complete data, which suggests a healthy survivor bias. Repeating the analyses with all women who provided any data at any survey resulted in similar patterns, but with slightly attenuated rates of decline (Table 3; Fig. 2). Also, linear associations were assumed between knots in the spline regression, even though the change in physical functioning over time was nonlinear, particularly in later life. In preliminary analyses, more complex models were examined, including higher-order polynomials and exponential functions. Since these functions added little to the explained variance of the models, the use of a simplified piecewise linear model that was easier to interpret was justified.
The rate of functional decline increased with age. Depending on baseline level of physical functioning, the average age when the disability threshold was reached (i.e. when a woman needs assistance with daily activities) ranged from 45 to 88. Physical activity was associated with baseline level of physical functioning but not with the rate of decline. The current results improve on and quantify the theoretical model proposed by Kalache & Kickbusch and have implications for the timing of interventions to maintain optimal levels of physical functioning in the population.
The authors thank the Australian Government’s Department of Health and Ageing and the women who provided the survey data.
The research on which this paper is based was conducted as part of the Australian Longitudinal Study on Women's Health, the University of Newcastle and the University of Queensland. GP was supported by a program grant from the (Australian) National Health and Medical Research Council (NHMRC grant 569940) and the National Health and Medical Research Council Centre of Research Excellence (grant number APP1000986). The funding sources had no involvement in the research presented in this manuscript.
- Kalache A, Kickbusch I. A global strategy for healthy ageing. World Health 1997; 50: 2.
- Saunders N, Broe T, Brown W, Earle L, Gregory B, Head R et al. Promoting healthy ageing in Australia. Canberra: Prime Minister’s Science, Engineering and Innovation Council; 2003. Available from: http://www.innovation.gov.au/Science/PMSEIC/Documents/PromotingHealthyAgeing.pdf [accessed 24 July 2013].
- Active ageing: a policy framework. Geneva: World Health Organization; 2002. Available from: http://whqlibdoc.who.int/hq/2002/who_nmh_nph_02.8.pdf [accessed 24 July 2013].
- Cooper R, Kuh D, Cooper C, Gale CR, Lawlor DA, Matthews F, et al., et al. Objective measures of physical capability and subsequent health: a systematic review. Age Ageing 2011; 40: 14-23.
- Caspersen CJ, Powell KE, Christenson GM. Physical activity, exercise, and physical fitness: definitions and distinctions for health-related research. Public Health Rep 1985; 100: 126-31.
- Cooper R, Mishra GD, Kuh D. Physical activity across adulthood and physical performance in midlife: findings from a British birth cohort. Am J Prev Med 2011; 41: 376-84.
- Blair SN, Wei M. Sedentary habits, health, and function in older women and men. Am J Health Promot 2000; 15: 1-8.
- Balboa-Castillo T, Leon-Munoz LM, Graciani A, Rodriguez-Artalejo F, Guallar-Castillon P. Longitudinal association of physical activity and sedentary behavior during leisure time with health-related quality of life in community-dwelling older adults. Health Qual Life Outcomes 2011; 9: 47.
- LaCroix AZ, Guralnik JM, Berkman LF, Wallace RB, Satterfield S. Maintaining mobility in late life. II. Smoking, alcohol consumption, physical activity, and body mass index. Am J Epidemiol 1993; 137: 858-69.
- Seeman TE, Berkman LF, Charpentier PA, Blazer DG, Albert MS, Tinetti ME. Behavioral and psychosocial predictors of physical performance: MacArthur studies of successful aging. J Gerontol A Biol Sci Med Sci 1995; 50: M177-83.
- Visser M, Pluijm SM, Stel VS, Bosscher RJ, Deeg DJ. Longitudinal Aging Study A. Physical activity as a determinant of change in mobility performance: the Longitudinal Aging Study Amsterdam. J Am Geriatr Soc 2002; 50: 1774-81.
- Simonsick EM, Lafferty ME, Phillips CL, Mendes de Leon CF, Kasl SV, Seeman TE, et al., et al. Risk due to inactivity in physically capable older adults. Am J Public Health 1993; 83: 1443-50.
- van Oostrom SH, Smit HA, Wendel-Vos GC, Visser M, Verschuren WM, Picavet HS. Adopting an active lifestyle during adulthood and health-related quality of life: the Doetinchem Cohort Study. Am J Public Health 2012; 102: e62-8.
- Keysor JJ. Does late-life physical activity or exercise prevent or minimize disablement? A critical review of the scientific evidence. Am J Prev Med 2003; 25: 129-36.
- Rantanen T, Guralnik JM, Sakari-Rantala R, Leveille S, Simonsick EM, Ling S, et al., et al. Disability, physical activity, and muscle strength in older women: the Women's Health and Aging Study. Arch Phys Med Rehabil 1999; 80: 130-5.
- Lee C, Dobson AJ, Brown WJ, Bryson L, Byles J, Warner-Smith P, et al., et al. Cohort profile: the Australian Longitudinal Study on Women's Health. Int J Epidemiol 2005; 34: 987-91.
- Brown WJ, Bryson L, Byles JE, Dobson AJ, Lee C, Mishra G, et al., et al. Women's Health Australia: recruitment for a national longitudinal cohort study. Women Health 1998; 28: 23-40.
- Bohannon RW, DePasquale L. Physical Functioning Scale of the Short-Form (SF) 36: internal consistency and validity with older adults. J Geriatr Phys Ther 2010; 33: 16-8.
- Ware JE, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 1992; 30: 473-83.
- Gill TM, Robison JT, Tinetti ME. Difficulty and dependence: two components of the disability continuum among community-living older persons. Ann Intern Med 1998; 128: 96-101.
- Brown WJ, Mishra G, Lee C, Bauman A. Leisure time physical activity in Australian women: relationship with well being and symptoms. Res Q Exerc Sport 2000; 71: 206-16.
- Australian Institute of Health and Welfare. Australian alcohol guidelines: health risks and benefits. Canberra: National Health and Medical Research Council; 2001.
- Long JS, Pavalko EK. The life course of activity limitations: exploring indicators of functional limitations over time. J Aging Health 2004; 16: 490-516.
- Dennerstein L, Lehert P, Guthrie JR, Burger HG. Modeling women's health during the menopausal transition: a longitudinal analysis. Menopause 2007; 14: 53-62.
- Kumari M, Stafford M, Marmot M. The menopausal transition was associated in a prospective study with decreased health functioning in women who report menopausal symptoms. J Clin Epidemiol 2005; 58: 719-27.
- Davis CE. The effect of regression to the mean in epidemiologic and clinical studies. Am J Epidemiol 1976; 104: 493-8.
- Deeg DJ. Longitudinal characterization of course types of functional limitations. Disabil Rehabil 2005; 27: 253-61.
- Nusselder WJ, Looman CW, Mackenbach JP. The level and time course of disability: trajectories of disability in adults and young elderly. Disabil Rehabil 2006; 28: 1015-26.
- Gill TM, Gahbauer EA, Han L, Allore HG. Trajectories of disability in the last year of life. N Engl J Med 2010; 362: 1173-80.
- Ahacic K, Parker MG, Thorslund M. Mobility limitations in the Swedish population from 1968 to 1992: age, gender and social class differences. Aging 2000; 12: 190-8.
- Ethgen O, Gillain D, Gillet P, Gosset C, Jousten A, Reginster JY. Age- and sex-stratified prevalence of physical disabilities and handicap in the general population. Aging Clin Exp Res 2004; 16: 389-97.
- Nilsson CJ, Avlund K, Lund R. Social inequality in onset of mobility limitations in midlife: a longitudinal study in Denmark. Eur J Ageing 2011; 8: 255-69.
- Thompson WW, Zack MM, Krahn GL, Andresen EM, Barile JP. Health-related quality of life among older adults with and without functional limitations. Am J Public Health 2012; 102: 496-502.
- Pluijm SM, Visser M, Puts MT, Dik MG, Schalk BW, van Schoor NM, et al., et al. Unhealthy lifestyles during the life course: association with physical decline in late life. Aging Clin Exp Res 2007; 19: 75-83.
- Dobson A, Byles J, Brown W, Mishra G, Loxton D, Hockey R, et al. Adherence to health guidelines: findings from the Australian Longitudinal Study on Women's Health. Canberra: Australian Government Department of Health and Ageing; 2012. Available from: http://www.alswh.org.au/images/content/pdf/major_reports/2012ALSWHMajorReportG.pdf [accessed 24 July 2013].