Longitudinal studies provide an opportunity to study prospectively or retrospectively physical activity over time and whether physical activity already in youth leads to a certain pathway with consequences for later activity and well-being. These studies are relatively scarce because they require a large financial effort and the long-term commitment of subjects, and encounter certain problems such as changing measurement techniques, a changing concept of physical activity, and confounding effects that occur in repeated measurements [35]. The concept of physical activity may have varied; those who, nowadays, know that when they walk somewhere (shopping, work), they are taking exercise might not recognize this earlier in life (e.g., 30 years previously). Socioeconomic status may be a confounding effect that occurs in repeated measurements in different ways; those with higher status continue participating more often in later phases of life.
Most commonly, physical activity is studied across two or three time points, for example, using Spearman’s rank order correlations in tracking studies; across childhood into adolescence, from adolescence into young adulthood, and during adulthood. Other methods for considering activity stability or change in activity include using baseline characteristics to predict physical activity and fitness several years later and relating current activity with retrospective recall of earlier activity levels.
From childhood to adolescence and adulthood
Early physical activity in childhood and adolescence predicts a physically active lifestyle in adulthood [9, 12, 13, 36, 37]. However, the tracking of physical activity during transitions from childhood to adulthood and adolescence to adulthood has been shown to be quite low, while the tracking of physical activity during adult life is somewhat higher [8, 30]. From early ages (7–11 years) to middle adulthood, Spearman’s rank order correlations have been found to be from 0.03 to .035 and from adolescence to adulthood from 0.15 to 0.44 for men and for 0.09 to 0.34 women [8]. The low tracking correlation indicates that many other factors influence physical activity in adulthood. Low tracking correlations may also be due to measurement errors. Education, occupation, living environment, marital status, having children, and health attitudes have been found to be important factors for later activity [9, 12, 15].
As early as 1992, Kuh and Cooper [9] examined the influence of childhood activity on adult activity patterns in work and leisure in a national sample in England, Scotland, and Wales of over 3,300 men and women aged 36 years from a birth cohort of the year 1946. Those who were most active in sport in adulthood had been above average at sports in school, more outgoing socially in adolescence, had fewer health problems in childhood, were better educated, and had more mothers with a secondary education than those who were less active. Two decades later, Mäkinen et al. [10] studied a representative sample of Finns aged ≥30 years in 2000–2001 and found that childhood socioeconomic conditions had an effect in leisure-time physical activity. The direct effect of childhood socioeconomic conditions on educational differences in inactivity was stronger than its indirect effects through adulthood socioeconomic conditions and other health behaviors. Tammelin et al. [12] studied a Finnish cohort of 8,000 men and women at ages 14 and 31 in follow-up surveys of the Northern Finland 1966 birth cohort. Their results suggest an association between father’s occupation and type of sport. Participants in downhill skiing, dancing, orienteering, and riding came from families of the highest social class, and participants in strength sports, walking, cycling, and soccer came more commonly from families of low social class. They also found associations between participation in some sports with high grades in school sports and membership of a sports club that carried over into adult years. In addition, Cleland et al. [15] found that a persistently high socioeconomic position and upward social mobility were associated with increases in self-reported physical activity and objectively measured fitness from childhood to adulthood among over 2,000 Australian adults aged 26–36 who had first been examined between the ages of 7 and 15. Upward social mobility, defined as reaching a higher level of educational attainment than one’s parents, was associated with a greater likelihood of high physical activity in adulthood and a greater likelihood of a high fitness level. The results of secular trends in the Finnish Adolescent Health and Lifestyle Survey from 1977 to 2005 [16] showed that participation in youth sport organized by sport clubs was strongly associated with parental socioeconomic status over a period of 28 years in both sexes. Social inequality was smaller in sport organized by schools and in young people’s spontaneous physical activities. Participation in physical activity is unevenly distributed across population groups. Various writers have discussed [12, 15, 16] that associations between childhood socioeconomic position, social mobility, and physical activity and fitness may run through the mechanisms, such as improved socioeconomic circumstances leading to greater social or economic and psychological support for physical activity for example through more flexible work hours and higher income; a low social class may weaken parents’ ability to transport children to organized sports, as well as to afford the fees and equipment required for the young person to participate.
Biological and psychological determinants and correlates are also crucial to the stability of physical activity. Allender et al. [38] found in their review of qualitative studies of motives and barriers that body shape and weight management are the important reasons behind the participation of girls in adolescence. Those who participated from youth to adulthood recalled the importance of positive influences at school in becoming and staying physically active (good teachers, friends). Instead, competitive classes, a lack of teacher support, negative experiences at school, and peer pressure and identity conflict were the major barriers to physical activity. The findings of Lunn [39] suggest that there is a gender gap in team sports but not in individual sports; boys are many times more active in team sports. Participation in sports and exercise activities moves closer to gender equality once individuals enter adulthood, where presumably they have greater choice and autonomy with respect to leisure-time activities.
It is important to investigate also the stability of various physical activities from childhood and adolescence to adulthood, and analyze how well adult physical activity can be predicted by physical activities measured in childhood and adolescence to obtain information for prevention and physical education about factors that may be relevant to later activity. In the study of cardiovascular risk in Young Finns [13], a random sample of 3,500 children, surveyed in 1980 and 2001, is analyzed. The aim was to reveal the influence of the type of sport practiced in youth on adult physical activity. The results indicated that a high level of physical activity at ages 9–18, especially when it was continuous, significantly predicted a high level of adult physical activity and that the level of activity in adulthood did not depend on the type of physical activity at a young age. It seems that intensive participation in general in physical activity and sports, and continuous participation at school age in particular, were more important than participation in specific sports. In addition, when Tammelin et al. [12] evaluated the association between participation in different types of adolescent sports and physical activity in adulthood, they found that very different types of participation in adolescence were beneficial with respect to the enhancement of adult physical activity. Adolescent participation in intensive endurance sports, such as cross-country skiing, running, orienteering, as well as certain sports that require and encourage diversified sports skills, such as ice hockey, soccer, volleyball, track and field, gymnastics, and riding and combat sports, was associated with a high or very high level of adult activity. Matton et al. [14] found that fitness characteristics demonstrated higher levels of stability from adolescence to middle adulthood than physical activity, and inactivity rather than activity tended to continue from youth to adulthood when they studied the stability of the physical fitness and physical activity of 138 women from adolescence (17 years of age) into adulthood (40 years of age) in Flemish women in the Leuven Longitudinal Study. Similar results were reported in the Amsterdam Growth and Health Study. The tracking of daily physical activity and physical fitness along with cardiovascular disease risk factors was investigated by means of six repeated measurements of 5,100 subjects over a period from 13 to 27 years of age [11]. The total time spent on all habitual physical activities in relation to school, work, sports, and on other leisure time activities was measured and calculated with the intensity to energy expiatory MET scores. Low to moderate tracking was observed for daily physical activity and VO2 max, whereas good tracking was observed for neuromotor fitness. In contrast, daily physical activity seemed to predict activity clearly from early years to adulthood—an odds ratio of 3.6 was found—indicating that those in the lowest quartile for physical activity at the age of 13 were 3.6 times more likely to remain in the lowest quartile along the measurement period of 15 years than those in the upper three quartiles at the age of 13.
In adulthood
The tracking coefficients during adulthood have been moderate, from 0.35 to 0.65, and seem to be smaller when the follow-up period has been longer [8]. The stability in a Finnish study over 5 years in 18–64-year-old male and for female subjects was 0.46 and 0.34, and stability over 10 years was 0.25 and 0.29, respectively [18]. The study examined both the tracking of physical activity and the effect of earlier physical activity and other lifestyle and socio-demographic determinants of sport, exercise, recreation, and housework physical activity. The level of physical activity at baseline was a good predictor of activity during the follow-up stages. Changes along the follow-up reflect a polarization of the distributions of physical activity within the sample. Low activity, old age, smoking, a blue-collar occupation, female gender, a high body mass index, and poor perceived health were significant predictors of inactivity in the 5-year follow-up in 1978, whereas old age, smoking, and low activity were predictors in the 10-year follow-up in 1983. In the model for the 28-year follow-up, in 2000, the dominant factors were age, along with female gender, a blue-collar occupation, smoking, poor perceived health, and low initial activity. Age and an initially low level of activity were the strongest predictors of inactivity. A similar gender difference was found in the study of Fortier et al. [17] with male subjects showing greater stability and higher physical activity levels than with female subjects. The study compared the tracking of physical activity levels between male and female subjects during middle adulthood. The Canadian sample comprised 1,900 subjects, aged 11–69 years, in the 1981 Canada Fitness Survey and its 7-year follow-up, the Campbell’s Survey. The measures of physical activity were estimated activity energy expenditure and time spent on activity. Significant tracking correlations for physical activity levels were limited to adulthood, whereas significant tracking of musculoskeletal fitness was observed at all ages. The results suggest that the level of physical activity has greater plasticity than musculoskeletal fitness and is thus subject to greater change.
Individual determinants such as personality have been recognized as an important correlate of physical activity [40], e.g., in the Terman Life-Cycle Study, a lifespan longitudinal cohort study, Friedman et al. [19] found that childhood energy and activity ratings correlated with adult activity. They examined the activity levels of 723 men and 554 women. The sample originally consisted of middle class boys and girls. Childhood activity and interests were assessed in1922 by parents, who evaluated their children’s “hobbies or enterprises.” Additionally, participants indicated how much they enjoyed playing games that required lots of exercise. These ratings were averaged to indicate childhood interest in being physically active. Adult physical activity was assessed in 1936, 1940, 1950, 1960, and 1972. In 1977, when the participants were, on average, 67 years old, individuals were asked about their level of participation in various activities. The results showed activity levels were somewhat stable from childhood through middle and late adulthood. Active, energetic children tended to become active, energetic adults, and in turn tended to remain active.
These studies, the Kirjonen et al. study [18] and the Friedman et al. [19] study spanning many decades, and also the Frontier study [17] sample including subjects aged 11–69 years suggest that there are very long-term consistencies in levels of physical activity and that fitness and inactivity tends to track better than activity. On the other hand, the level of physical activity has greater plasticity than fitness and is thus subject to greater change [17]. If interest in voluntary physical activity is lacking during early adulthood, it does not easily become a part of life later on, particularly among blue-collar workers, women, and people with initially poor perceived health.
In old age
There are only a few tracking studies that extend into old age [17–19], and it is difficult to see any systematic age differences in the tracking coefficients reported in the different phases of life, the coefficients being quite low from childhood to adulthood and at most moderate in adulthood [8]. When the subjects have already reached older age, the activity profiles have been shown to be quite stable. The Nottingham Longitudinal Study found that both the activity profiles and the time spent on various activities were quite stable over an 8-year follow-up in subjects aged 65–74-years at the baseline. The 8-year stability correlations in outdoor activities and men’s walking and leisure activities were moderate (0.45–0.58) [21]. Within the cohorts of older people, traditional gender roles continue to exert an influence with women showing higher levels of activity participation indoors and men showing higher levels of activity participation outdoors. In levels of walking/shopping activity, however, there was evidence of gender differences narrowing over time [22]. The major sources of change might plausibly include declining health with reduced physical ability, widowhood, and adaptive moves into specialized, but less spacious accommodation.
The Evergreen study in Finland [23] found that men’s physical activity level decreased less than women’s level over an 8-year follow-up with non-institutionalized seniors aged 65–84 years at baseline in 1988. The study examined changes in subjects’ involvement in various types of physical exercise, and the motives for and obstacles to participation. Men took part in a wider range of leisure time physical activities and sports than women, and their participation in supervised exercise classes and doing calisthenics exercises at home increased over the follow-up, while in the case of women, the general trend was that physical exercise declined during the follow-up. In the same Evergreen study project, the continuity of lifespan physical activity was studied [24] by examining the predictors of maintaining a high level of physical activity over 8 years of follow-up. For men and women, self-reported competitive sport participation from as early as 10–19 years of age was a significant predictor of maintaining physical activity in old age regardless of chronic conditions. Women’s participation in recreational sports at the age of 40–64 years also predicted activity. Bonsdorf et al. [25] studied the same participants in their 16-year follow-up in 2004. The association of self-reported physical activity from midlife to old age and all-cause hospital and long-term care in the last year of life was studied. Men who reported consistent physical activity from midlife onward needed fewer days of hospital care in their last year of life than those who reported less physical activity. For women who died before 90 years of age, the same was true for long-term care.
In the longitudinal study in Gothenburg, Sweden [20] of 293 women and 233 men, the relation between the level of physical activity in earlier life and both physical activity and performance at age 76 was analyzed. The lifespan was divided into five age periods from the age of 10 and described as competitive sport, recreational sport, occupational physical load, household work, and means of transportation. The Swedish results revealed that the earliest periods could not explain activity patterns or performance late in life when all fields of activity were combined into a total measure of physical activity. After the age of 35 years in men and 50 years in women, however, the associations grew stronger, and the activity level of the last life period (66–76) was strongly correlated with physical activity in old age. When, on the other hand, each field of activity was treated alone, above all recreational sport turned out to be positively related to walking speed and physical activity at age 76 in both women and men for most life periods.
Earlier physical activity seems to be the most important determinant, along with gender, in old age. Probably, the cumulative effects of different determinants still affect physical activity in old age. This may be due, at least in part, to gender-role expectations that have been assimilated over the years and that still apply in old age, namely, that it is more appropriate for men to be active than it is for the women. O’Brien Cousins and Vertinsky [41] conducted life-course interviews with women born in or before 1890, and Kluge et al. [4] explored the nature and meaning of being physically active in women aged 65 and older. Both studies found that the late life exercise patterns of elderly women appear rooted in their pasts. These physically active women hung on to a concept of themselves as physically active. However, findings suggest that the continuity of a physically active lifestyle was affected by gender socialization, ageist attitudes, and physical challenges. In addition, they found that social empowerment during adulthood and a commitment to exercise through later life were important to remaining physically active. Older women judge their efficacy for exercise based on their understanding of perceived health, chronological age, and previous skills [42]. The findings support the self-determination theory, which argues that experiences about physical competence might have roots in early mastery experiences. In most cases, positive social support from parents, peers, and teachers was a feature of the participants’ early experiences of physical activity.
The attitudes of older people toward physical activity may sometimes be undermined by prevailing attitudes that older people are passive members of society who need to rest. Long working hours and other responsibilities may have prevented people from participating in physical activity in earlier life, and thus, exercise skills have not developed. In addition, physical activity behaviour is susceptible to changes in environmental factors, such as latent age discrimination in exercise clubs, the accessibility of exercise facilities, leadership, and role models. For example, in a systematic review of qualitative research into barriers and motivations to physical activity [43], older adults identified cost and time barriers as the main problems. The lack of realistic role models in the community was a deterrent. Exercise prescriptions were perceived as targeted at young people and not relevant to older groups. Some older adults were unsure about the “right amount” of physical activity for someone of their age. The health benefits of physical activity were emphasized in terms of reducing the effects of aging and being fit and able to play with the grandchildren. The studies reviewed also highlighted the importance of social networks in maintaining participation.