Background
Regular physical activity in older adults is critically important for healthy aging [
1]. The link between regular physical activity and disease prevention, maintenance of independence and improved quality of life is supported by strong evidence [
2,
3]. However, there is a lack of knowledge on the physical activity levels and sedentary behavior among older people. Current knowledge is primarily based on studies using subjective assessment methods (e.g. questionnaires). Recalling physical activity is a complex cognitive task, and old adults are likely to have particular memory and recall skill limitations [
4‐
6].
The introduction of accelerometers for objective assessment of physical activity allows for valid and reliable assessments of activity intensity, frequency, and duration [
7,
8]. Accelerometry is less prone to the recall and social desirability biases associated with self-report instruments [
9]. Objective information on the physical activity levels and sedentary behavior has the potential to increase our understanding of physical activity in old age [
3].
There are only a limited number of studies that have assessed physical activity using accelerometers in older adults. Most of these studies were completed in the USA [
10‐
12], Canada [
13] and the United Kingdom [
14,
15] and relatively few studies are anchored in the northern European countries [
16‐
18]. Additionally, there is a lack of knowledge regarding physical activity levels in adults over 79 years of age [
11,
13,
18].
The World Health Organization recommends that information on how individuals perceive their own health should be collected in population-based studies including older individuals [
19]. Self-reported health status is considered as a sensitive measure of overall health in older adults, influenced by physical function, the presence of disease, the existence of disabilities, functional limitations, and the rate of aging [
20]. It is viewed as a holistic measurement of health, reflecting both physical and mental health as well as well-being [
21]. At present, few studies have examined physical activity level measured objectively in the elderly in combination with self-report instruments including simple measures of health [
22].
The aims of the present study were therefore to describe the level of accelerometer-determined physical activity in a random national sample of Norwegian older adults (65–85 years), and secondary to investigate the associations between physical activity level and self-reported health.
Discussion
The main findings of the present study were that objectively-measured physical activity level significantly differed by age in a national sample of older adults. There were no sex differences in physical activity level within each age group. In the age groups 65–69 years and 70–74 years, men had higher levels of sedentary minutes than women, whilst men in the age group 75–79 years achieved more minutes of MVPA than women. In all age groups, except for the oldest one, women spent significantly more minutes of low-intensity PA than men. Also, overall physical activity was associated with self-reported health.
We found that accelerometer-determined physical activity significantly differed between the different age groups, with the oldest age group having substantially lower mean physical activity levels than the youngest age group. This is in accordance with other cross-sectional studies using the same objective method [
10‐
17]. Our population appeared to have somewhat higher overall physical activity level than what has been reported in other studies [
12,
16]. While Norwegian men and women in age group 75–79 years had a mean cpm of 256 and 215, respectively, data from this age group in Iceland showed lower physical activity levels (mean cpm 150 and 139 for men and women, respectively) [
16]. Our mean physical activity levels in individuals aged 65–74 years are higher than what has been reported among Americans [
12]. However, the activity levels in Norway are similar to what has been reported in Sweden [
17]. This might be due to differences in socioeconomic status, cultural differences with respect to retirement age, infrastructure and degree of environmental security among the populations studied.
We did not find significant sex differences in physical activity level within each age group, which is in contrast with similar studies from other countries usually showing a higher mean physical activity level among men than among women [
10,
11,
13‐
16]. This discrepancy might be connected to cultural differences as described above. Also, the lack of a difference in PA level between sexes in the present study is also in contrast to earlier Norwegian studies using self-reported measures of PA [
29]. Women may spend more time doing low and lifestyle intensity activities, such as walking, household chores, and gardening [
14]. Subjectively-assessed PA have limited accuracy at capturing activities that are unstructured and of low intensity [
4], which have a tendency to be performed more often in older populations and in particular among older women [
30‐
32]. This is supported by the fact that Norwegian women spent more time in low-intensity PA and have less sedentary time compared to their male counterparts.
The participants spent the majority of the day being sedentary (66% of the total wear time), and this was followed by low-intensity PA (24%), lifestyle PA (7.1%) and MVPA (3.0%). These findings are comparable to what has been reported among older adults in Iceland [
16], Great Britain [
14], and Canada [
13]. Resent research has also shown dose–response associations between sitting time and mortality from all causes, independent of leisure time physical activity [
33]. The large proportion of sedentary time and increased sitting-time is worrying as it might lead to substantial health problems for older people and as a consequence, reduced quality of life and need for assistance. It is therefore important to develop and initiate interventions where the goal is to increase physical activity levels and reduce sedentary time among older adults. In addition to the PA promotion, physicians should also discourage sitting time for extended periods.
When looking at sex- and age trends, Norwegian women are spending less time being sedentary and more time in low-intensity PA per day compared to men at the same age as mentioned above, while men (75-79-yr age group) accumulate more minutes of MVPA than women. In comparison, older men in the UK performed significantly more minutes of MVPA per day than women (23.1 vs. 13.8 min) [
14]. Furthermore, the British older adults had a steep decline in the proportion of active time spent in MVPA with increasing age [
14], which is in accordance with our results. Similar patterns are also observed among US older adults [
10] and among Canadians aged 20–79 years [
13], where MVPA decreased across increasing age [
10].
The age group 65–69 years averaged 5525 steps more per day than the individuals in age group 80–85 year (p = 0.02), a relative difference of 59%. This is in accordance to what has been found in two other studies [
14,
15] including older adults, both using accelerometer to assess PA levels. Davis et al. [
14] found that younger participants (70–75 years) averaged significantly more steps per day (5661 steps per day) than participants aged 80+ years (3410 steps per day). Harries et al. [
15] also showed that step-count declined steadily with age. In the latter study, however, sex differences in step counts were also reported and men achieved 754 more steps daily than women. This is in contrast to the result of the present study where no sex differences in step counts were reported.
Overall, 21% of the participants (women and men: 18% and 22%, respectively) fulfilled the current Norwegian PA recommendations. Data from the United Kingdom shows a similar prevalence among older men (25.6% met national recommendations), but a lower prevalence among older women (14.2%) [
14]. In the oldest age group, we found that only 6% reached the national physical activity recommendations. This is a higher percentage compared with a study conducted in the United Kingdom by Harris et al. [
15], showing that only 2.5% of the participants 65 years and older met the PA recommendations. On the other hand, looking at the Icelandic oldest (85 years and older), as much as 25% of the men and 9% of the women fulfilled the recommendations, defined as having at least one ≥10 minutes MVPA boats [
16]. However, comparability between the current study and the Iceland study [
16] is hampered by the use of different physical activity recommendation criteria and differences in data reduction strategies.
In Norway, mean physical activity level declines by approximately 30% between the ages of 9 and 15 years [
34]. A further decline of 30% for women and 35% for men have been observed when going from 15 years into adulthood, followed by a stable level of activity until retirement age [
23]. Following retirement to 80–85 years, a further decline of 47% in women and 53% in mean PA level was observed in the present study. The causes for these age-related changes in physical activity level are not fully known, although the overall decline of 50% observed during the age of being 65 years to entering 85 years, might be caused by changes in health status and of course the aging process in itself [
35]. The higher mean physical activity level in the youngest age group might also be explained by higher prevalence of participants in this age group reporting part- or full time employment than participants in the oldest age group (23% versus 4%). 23% of the youngest age group still reported the fact to be employed. For example, if their work involves a lot of walking and their physical activity measurement period includes only working days then their measured activity level may be higher compared to someone whose measurement period includes non-working days where they may be less active. This will overall affect their computed average activity levels, and has to be taken into consideration.
In the present study significant differences in the overall level of PA were observed between all self-reported health groups, except between those who perceived their health as “either good or bad” and “poor/very poor health”. One of few available studies mentioned above is targeting community-dwelling people in the U.K. from 65 years and older showed that those with poor health took fewer steps compared to those with better health [
15]. This difference (p > 0.05) was not found in the current study (data not shown). The latter study used a different method (Health Survey form England, 1988: questions related to general health, disability, long-standing illness, pain, medication use, chronic disease, falls, and walking aid use) to register self-reported health compared to the this study and therefore, the degree of comparability is rather limited. The associations between physical activity level and perceived health are strong, but due to the study design we cannot determine causality.
The major strength of this study is the use of accelerometers to assess physical activity in a relatively large sample of older adults. The participants showed good compliance with the protocol and few data were lost because of insufficient wearing time or defect monitors. Objectively-measured physical activity in combination with self-reported health in older adults, is rather novel. These variables are often presented separately in other studies [
11,
14,
21], and few studies [
15] have objectively measured physical activity levels and its association with multiple health factors (e.g. general health).
We acknowledge some limitations to our study. One limitation is the relatively low participation rate. A drop-out analysis performed via registry linkage showed that the responses varied according to socio-demographic variables [
23], which is consistent with other population-based studies conducted in Western countries [
36].
Furthermore, there are limitations worth noting when interpreting accelerometry data [
11]. Accelerometers do not provide qualitative information on the type of physical activities being performed, and hip-mounted accelerometers underestimate upper body movements and activities such as carrying heavy loads, weight training, swimming, and cycling [
11]. Nevertheless, accelerometers are sensitive to ambulatory activities such as walking. The participants reported walking as the most frequently performed activity during the measuring period, which decreases the possibility that physical activity level was underestimated [
23]. Walking technique must also be taken into consideration because it can affect the validity of accelerometer counts, especially in older individuals [
11]. It seems that some accelerometers can undercount activity in individuals with a non-standard gait, e.g. upper body leaned forward and bended knees during walking, thereby underestimate the activity level in these individuals [
37]. Furthermore, when interpreting accelerometer data, there is a possibility that the observed differences in physical activity may simply reflect differences in accelerometer wear time between groups. However, there were no significant differences between sexes and between age groups in minutes of daily accelerometer wear time and the sample were compliant to the accelerometer protocol with a mean wear time of 14.0 hours per day.
In the past, methods based on self-ratings of health have been questioned because of their obvious subjective bias [
5,
6]. Self-reported height and body mass is therefore considered as a limitation to our study. However, several studies have shown that self-report instruments concluding simple measures of health and self-reported functioning in old persons have acceptable reliability and validity [
38,
39]. Furthermore, because it is inexpensive and easy to administer and interpret, self-reported health is a practical tool suitable for the clinical environment [
40] and has become an important variable to assess the state of health in the older population [
20,
41].
Our findings help to better understand older peoples’ rate of physical activity and thereby help guide the development of needed physical activity interventions targeted at older adults in Norway. The link between PA and prevention of disease, maintenance of independence and improved quality of life in older adults is supported by strong evidence [
2,
3], and therefore it is of great importance to maintain PA levels as long as possible. Implementation of PA among community-dwelling older adults should therefore be prioritized in the future, with a special focus on the least physically active and the oldest individuals, especially in those with low levels of self-reported health.
Competing interests
The authors declare that they have no competing interests. The results of the present study do not constitute endorsement by the BMC Public Health.
Authors’ contributions
SAA contributed to the conception and design of the study. BHH was responsible for the collection of the KAN data in corporations with colleagues at nine other test centers throughout Norway. BHH provided the data for analysis. HLS undertook the data analysis and drafted the manuscript. All authors provided critical insight, and revisions to the manuscript. All authors read and approved the final version of the manuscript submitted for publication.