This study investigated age and sex specific differences in time use in three active daily domains (leisure, transport, home-based) and examined associations between time allocation in these domains and objectively measured PA with different health outcomes (self-rated health, BMI, handgrip strength, ADL). Furthermore, we explored differences in time allocation between working and non-working older adults. Results showed age and sex differences in total PA as well as in time use in the active daily domains. Participants with a paid occupation spent less time in almost every active daily domain than participants without an occupation. Furthermore, the results showed several associations between active daily domains and health indicators. Higher PA was associated with a lower risk of a bad self-rated health, less limitations when doing moderate activities or when bending, kneeling, and stooping, as well as a lower risk of being overweight, or obese. Time spent on leisure activities seemed to lower the risk of bad self-rated health and having limitations in ADL. Having an occupation was associated with better self-rated health. Time spent on active transport showed a lower risk for having limitations when walking more than one kilometer and being overweight, or obese. Home-based activities showed no significant associations with any health indicators.
This study showed that participants older than 70 years were significantly less physically active than those under 70 years. Yet, they spent more time in several SLOTH domains with increasing age. One reason could be the decreasing proportion of people having a paid occupation or doing volunteer work as they get older (women: 63.3% compared to 51.4%, men: 57.3% compared to 53.9%), which leads them to allocate their time differently. Another possibility why they spent more time in several SLOTH domains is that with increasing age, more time might be required to perform certain tasks, because of the ageing-related decline in health [27]. Both men and women in the older age group (age 70 + years) more often reported less good or bad health than the younger ones. Since our data only provides information about the time allocation to active domains and cannot provide information about the intensity, we can only speculate that although the time spent in the SLOTH domains increases with higher age, the intensity might decrease. Spinney et al. [28], for example, found decreasing rates of older Canadians meeting PA recommendations with increasing age, when looking at the active domains.
Furthermore, this study showed significant differences in objectively measured PA between men and women. These results are in contrast to several other studies objectively measuring PA in older adults. The systematic review by Sun et al. [3], for instance, found older men to be more physically active than older women. These differences could arise from differences in PA measurements. In this study the accelerometer was worn on the non-dominant wrist. However, most studies use PA monitors on the hip, which results in differences in the measured movements, especially regarding the upper body [3]. Since the women in our sample devote a lot of time to housework, which includes lots of upper body movements, it is possible that the higher amount of PA stems from this.
The sex differences in time allocation regarding home-based activities are in line with previous research by Adjei and Brand that was conducted in Germany [9], as well as Sprod et al. that was conducted in Australia [11, 29], who also found that women spend more time on housework but less on gardening than men. Gauthier and Smeeding [12] reported the same findings for overall home-based activities in nine different countries, but, contrasting our results, saw a decrease with age in women devoting time to housework and an age-related increase in men. In terms of the amount of time spent doing home-based activities, the same study found on average three hours per day being devoted to this domain in Germany [12]. Our results differ a lot with 80.2 ± 69.4 minutes per day in men and 118.5 ± 87.8 minutes per day in women. These differences could be a result of measuring home-based activities, since our study used a questionnaire asking for weekly hours carrying out the activity, which can lead to recall bias and reporting errors. The review only included studies using time use diaries, that can deliver more accurate results [7]. Another reason could be different interpretations of housework.
Adjei and Brand [9] found positive associations between time spent doing household chores and self-reported health, in men and women. Our results, however, did not match these findings, which could also be a result of different operationalisations or assessment of time spent in housework.
In line with our results, a systematic review by Gauthier and Smeeding [12] reported average leisure time PA of 0.5 to 1 hour per day, with men devoting more time to it than women and Krantz-Kent and Stewart [30] found similar results in their American study.
While our findings suggest that active transport is the domain older adults devote the least amount of time on, several other studies reported leisure PA to be the domain the least time is spent on [11, 12, 29, 30]. Sprod and colleagues [11, 29] also made a distinction between active and inactive travel and reported almost three times more minutes per day being spent on it than our study. These differences could be a result of us only using everyday destinations, whereas Sprod et al. included every destination. Another explanation could be the difference in infrastructure between the study countries (Australia vs. Germany), that leads to the participants having to cover longer routes.
Our findings further indicated that the time spent in leisure activities is associated with better self-rated health, which is in line with the studies by Abu-Omar and Rotten [8], as well as Kaleta and colleagues [31]. Having an occupation seemed to increase the likelihood of a better self-rated health, which does not coincide with the existing literature, that found contrasting results [32]. The different results could stem from varying definitions of occupation and retirement, for example if having a part-time job whilst being retired counts as having an occupation.
Time spent in active transport was associated with a lower risk of overweight and obesity, and no limitations in walking more than 1 km. A comparison with existing research is difficult, since other time use studies often do not distinguish between active and inactive transport or define it as general PA [33] or commuting [8]. A study by Foley et al. [34] could, however, associate time in active transport with spending more time on healthy behaviours.
Our results showed that participants with higher total PA were more likely to rate their health as good, which is in accordance with prior research [35, 36]. In line with our results, a study by Riebe et al. [37] found an association between higher PA and a lower risk of obesity. Higher total PA was also associated with having no limitations doing moderate activities as well as bending, kneeling, and stooping. Yorston and colleagues [38], who focused on the associations of PA and physical function in older adults, found similar results. They reported people with higher levels of PA having a lower risk of functional limitations.
We found no associations of home-based activities and health outcomes. This is in contrast to the results from Adjei and Brand [9], that found older adults who spent more time doing home-based activities to have higher odds of good self-reported health. Other studies, however, could not find any effects of household chores on either lower odds of being overweight [39] or on having a better health status [40].
Furthermore, our results showed no associations between time use and PA with handgrip strength. This is in contrast to a study by Spartano et al. [41], that found associations between PA and better handgrip strength in middle aged and older adults. They did not assess the time being physically active but time in MVPA and used handgrip strength in kilogram, whereas we used reference ranges. These differences in measurements could be a reason for the different results.
The study has a few limitations that need to be addressed. The questionnaire used in the OUTDOOR ACTIVE study was not initially designed for time use analyses, thus some domains of the SLOTH model were not fully assessed. Furthermore, the assessment of housework and gardening might be biased, since it was not clarified which tasks account to these domains. It is, for example, unknown whether the participants included cooking to housework or only referred to cleaning. Moreover, the activity “riding a bike” was excluded from the leisure domain, since it is possible that participants included time spent on riding a bike for transport in their answer. This could lead to underestimated time in this domain. Since the participants had to estimate their weekly time spent in leisure and household activities, recall bias and reporting errors could be an issue. The use of a proper time use diary could reduce these risks. Additionally, they would deliver data for full 24-hour days. However, the results of the present study are still a good indicator for the time allocation to active domains in older adults.
The assessment of transport only included everyday destinations, leaving out travelling to work, social events, or other obligations, which could lead to an underestimation of time allocation.
Another limitation of this study is the cross-sectional design. Therefore, no statements regarding causation can be made and associations because of possible reverse causation (e.g. between PA and self-rated health or between occupation and self-rated health) cannot be determined. Thus, future research should look at time allocation and PA, and its effect on health outcomes in a longitudinal study.
Despite the limitations, this study provided insights in time allocation of older adults to active daily domains and PA, which are important information for developing PA programs. Additionally, associations of time use and several health indicators were presented. One strength of this study is the representativeness of the sample for the population of Bremen-Hemelingen, when comparing social demographic factors. Furthermore, the PA data was assessed objectively using accelerometers, which is a reliable measurement for PA in older adults [42]. Furthermore, the SLOTH model is a well-known and fitting time budget model to analyse time allocation regarding PA [43].