Adiposity and changes in movement-related behaviors in older adult women in the context of the built environment: a protocol for a prospective cohort study

The world’s population is aging [1], and this process is accompanied by the loss or reduction of health-enhancing physical activity (PA) and the concurrent rise of sedentary time [2]. At the same time, the proportion of older adults who are overweight or obese increases [3]. These issues have resulted in an urgent need for effective social, educational, medical, environmental and economic solutions to mitigate these negative trends.

In the older adult population, a moderate amount of regular PA can minimize the physiological effects of a lifestyle that is generally sedentary and increase active life expectancy by limiting the development and progression of chronic diseases and disabling conditions [4], thus reducing medical and social services costs [5].

Obesity has been found to be significantly inversely associated with PA in numerous studies [6]. Moreover, other studies in which participants were segmented into multiple groups according to the volume or intensity of their PA and incremental differences in BMI or adiposity parameters supported a dose-response relationship [7, 8]. However, these cross-sectional studies did not allow for conclusions about causality to be drawn.

A theory-based study in an adult population using the theory of planned behavior [7] confirmed a one-directional relationship and suggested that obesity could be a direct predictor of future PA. A number of prospectively designed studies concluded that obesity is a risk factor for a decrease in PA [9‐11], an increase in physical inactivity [9, 12], and a sedentary lifestyle [13, 14] in the future. However, the majority of the abovementioned studies focused on middle-aged populations, with the rare inclusion of older adults [14], and have often been limited by the use of self-reported data of MRBs and obesity status. The most recent findings [15], which indicate approximately twofold stronger associations between PA and adiposity as measured with objective measurements than with self-reported PA, emphasize the need to incorporate objective measures in future studies.

These assessment methods, while typical for large-scale studies, are prone to increasing measurement error and misclassification [16, 17]. Thus, a prospectively designed investigation examining the association between obesity and PA level using device-based methods, including accelerometer-based measures of PA and a well-validated body composition assessment in an old-aged population, which would likely yield valuable valid data, might help to provide more valuable insights.

Social ecological models [18, 19] suggested that specific health-related behaviors (e.g., regular moderate PA) are influenced at multiple levels, ranging from policy and the built environment to psychosocial and individual factors. The influence of environmental exposures on individual health behaviors may increase with age, as older adults spend longer periods of time in or near residential areas. As adults age and increasingly lose control of their functioning, they become more vulnerable to environmental challenges and sensitive to environmental constraints [20‐22]. The majority of previous studies have indicated a positive association between the quality of the neighborhood environment and PA in older adults [22, 23]. In the Central European population aged 50+ years, neighborhood aesthetics, land use mix-proximity, street connectivity and neighborhood safety were found to be associated with higher odds of achieving health-enhancing PA [24, 25]. The built environment-obesity relationship is not as clear in older adults [23]. Some studies have indicated that neighborhood walkability (i.e., describing the extent to which the environment is conducive to walking and an active lifestyle) and land use mix are negatively associated with obesity [22, 26]. However, some studies have indicated a null association between certain characteristics of the built environment and obesity [27, 28]. In Central Europe, obesity was found to be lower in older women living in highly walkable areas [29].

Obesity in older age could be a possible barrier to involvement in regular PA [30]. Examining the contribution of obesity to changes in future activity levels is desirable and could lead to cues for predicting PA decreases in old age years before this behavior manifests. Such findings may also serve to create more specific programs and health promotion strategies to prevent weight gain and to avoid decline in habitual PA late in life. Elderly women are an important part of the target group for PA promotion, since the menopausal transition years place them at high risk of abdominal obesity [31], which is a widely recognized risk factor for cardiovascular and metabolic diseases [32]. Since multilevel interventions may be the most effective way to increase PA in older adults [33], studies should shift from simplistic to more complex models that take potential moderators and mediators into account [34]. Accounting for the moderating effect of certain environments might be appropriate, since more walkable environments may help older adults with obesity to engage in more PA. It is necessary to describe the role of the built environment in the long-term influence of obesity on PA in older adults to guide the development of environmental and policy initiatives that might lead to the development of both physical-activity-friendly and aging-friendly neighborhoods.

Recent epidemiological studies highlight the fact that behavior should be seen as a time-use variable in the finite number of discrete behavior modes that are changing. In this approach, any behavior can be understood as a varying time use in concrete patterns of behavior. Consequently, movement-related behavior (MRB) data should be considered compositional, and an appropriate methodological approach has to be applied in corresponding studies [35, 36]. In this regard, based on energy consumption, body position, and consciousness, studies commonly differentiate among four discrete composite parts of 24-h units: time spent in sedentary behavior (SB), time spent in light PA (LPA), time spent in moderate-to-vigorous PA (MVPA), and time spent in sleep [36].

Recently, Pedišić, Dumuid and Olds [37] noted that the application of inappropriate methods in compositional data can lead to incorrect conclusions. Above all, they described the risks related to the application of regression analysis in previous epidemiological studies in which the compositional character of data is not respected. Consequently, they also presented a proper statistical data procedure applicable in MRB data. Their approach is based on a statistical compositional data analysis [38]. We can therefore register a number of epidemiological studies in the last few years applying this new statistical approach that respects the compositional character of MRB data. We assume that this integrated time-use approach [37] has the potential to provide new insight into knowledge about the risk of obesity with respect to MRBs.

During the aging process, older adults become less active and more sedentary, which are considered to be risk factors for mortality. The time spent in different MRBs throughout the day might be associated with body composition. An inverse association between obesity and PA volume and positive associations between obesity and SB were found in cross-sectional [6] and prospective studies [12‐14]. However, these studies focused mainly on middle-aged populations with the inclusion of mainly self-reported assessment methods. Therefore, our study will mainly focus on older adults, including adults in the transition from age 65+ to age 70 + .

Few studies have suggested that obesity may lead to a subsequent increase in SB among middle-aged and older adults [14, 52]. However, these studies considered SB to be an independent predictor of obesity without conceptualizing it as a part of the time-use composition.

Studies using an integrated time-use approach to explore the associations between device-measured MRBs and obesity among older adults have, to our knowledge, all been cross-sectional [52‐56]. Evidence from these studies appears to consistently suggest that reallocating sedentary time to PA may be associated with reduced adiposity, with the magnitude of associations being greater for higher intensities of PA [56]. However, there is a lack of evidence linking changes in obesity and MRBs over time, although such evidence could underpin obesity interventions and public health policy for older people.

In general, it seems that the strongest association with obesity is observed when time is reallocated from SB to MVPA [55], but reverse causation is also plausible. Studies using an integrated time-use approach to explore the associations between device-measured MRB and obesity among older adults have, to our knowledge, all been cross-sectional [52‐56]. Evidence from these studies appears to consistently suggest that reallocating the time spent in SB to the time spent in PA may be associated with reduced adiposity, with the magnitude of associations being greater for higher intensities of PA [56]. However, there is a lack of evidence linking changes in obesity and MRBs over time, although such evidence could underpin obesity interventions and public health policy for older people.

In general, it seems that the strongest association with obesity is observed when time is reallocated from SB to MVPA [55], but reverse causation is also plausible. Especially in older adults, increasing adiposity might lead to PA reduction with the concurrent rise in sedentary time. However, evidence about the influence of obesity and changes in obesity on the later composition of MRBs in older age is missing. Therefore, the major strength of this study is that it will use an integrated time-use approach to investigate how obesity is prospectively associated with longitudinal changes in the time spent in various MRBs.

The built environment might have a powerful impact on the health of older people and on their ability to successfully age in place [57]. Although obesity cannot be directly affected by the built environment, the built environment clearly has a relationship with obesity as a consequence of PA. A recent review on obesity and the built environment highlighted the lack of longitudinal research on this topic [58]. Hence, another strength of this study is the employment of the built environment as a potential moderating variable affecting longitudinal associations between obesity and MRBs.

The main limitation of the study is that the baseline sample cannot be considered fully representative of older adults. In addition, the sample will include only women; hence, the results cannot be extrapolated to men without caution. Due to the recruitment of volunteers, we may expect selection bias (healthier and more educated sample with fewer smokers compared to the general population) [59]. Moreover, we expect a relatively high dropout of participants aged 70+ years in the 7-year follow-up, which can lead to unspecific changes in the baseline profile of the population.

With an integrated time-use approach, this study is the first longitudinal study investigating the prospective association of obesity and MRBs in older women. The findings from this study might help to clarify the causality of obesity—MRB association, which is necessary for successful weight gain interventions. Moreover, the insights provided into the potential moderating effects of the built environment may help to develop multilevel interventional models in which physical activity in older adults is enhanced.