“Keeping Moving”: factors associated with sedentary behaviour among older people recruited to an exercise promotion trial in general practice

Sedentary behaviour (SB), conventionally defined as low energy-expenditure activity undertaken in a sitting or reclining position [1], is associated with adverse physical and mental health outcomes [2]. Sedentary behaviour appears to have deleterious health effects even where physical activity recommendations are met [3], and so sitting time is now recognised as a health risk factor independent of physical activity [4‐6]. Older adults are most likely to be sedentary [7, 8].

Long periods of sitting are associated with a bigger waist circumference, depression and social isolation, and an increased risk of death [2]. Sedentary older adults are more likely to have the metabolic syndrome [9, 10], type 2 diabetes [3, 9, 11], cardiovascular disease [3, 9, 12], depression [13], lower bone mineral density [14], greater co-morbidity [13] and higher all-cause mortality [3, 15] than less sedentary older adults. Increased sedentary behaviour is further associated with functional limitations [11, 13, 16], falls [13], poorer quality of life [17], experiencing severe pain [16] and lower likelihood of successful aging, measured across both physical and psychological domains [17]. Since the health risks are significant and far reaching, understanding the characteristics of sedentary individuals is potentially important in targeting health interventions to reduce sedentary behaviour.

Epidemiological studies have described the characteristics of sedentary older people. Increasing sedentary behaviour is associated with older age [11, 16, 18], abnormal BMI [9, 12, 16, 18‐20], higher waist circumference [11], smoking [11, 12], living alone [13, 19], being unmarried [11, 12], lack of full-time employment [19] and lower levels of social support [16]. Occasionally the associations are conflicting. Sedentary behaviour has been shown to be more prevalent in women [9, 16], men [11, 12], neither sex [18], in those with lower education [9, 11, 12, 16, 18, 19], higher education [13], lower income [18] and higher income [13].

A small qualitative study by Chastin et al. [21] sheds light on the determinants, motivators and barriers older women express in relation to reducing sitting time. They attributed their sedentary behaviour to pain (predominantly musculo-skeletal), variable daily energy levels, external pressure from family and friends to undertake sitting activities and societal stereotypes of older people [21]. They also felt an entitlement to sit in older age, failed to recognise its objective harms and felt a sense of wellbeing from social sedentary activities [21]. Motivators to activity included pain relief (after sedentary periods), the necessity of household chores, in order to be useful to those around them and to relieve boredom & depression [21]. They also identified environmental barriers to increasing activity including lack of standing activities for older people, poor weather, and lack of public resting places outside the home [21]. However, they felt that more community-based opportunities to be active would help them reduce their sedentary behaviour [21]. Whilst sometimes perceived as a hard to reach group, promotion of appropriately tailored exercise in older adults may prove both acceptable and effective in reducing sedentary behaviour.

To reduce sedentary behaviour we must first quantify it. Although challenging, several studies have quantified SB in older people [19, 22‐24]. Whilst younger adults are engaged in SB during 60 % [25] of their waking hours, older adults have been shown objectively (using accelerometry) to be sedentary more than 70 % of the time [22, 24], for around 8–10 h of the waking day, and this increases linearly with age [22, 26]. Conversely, self-reported SB is typically underestimated by as much as 50 % [23, 27]. Espana-Romero et al. [28] showed that older people both overestimate their physical activity and underestimate their sedentary behaviour; men by 26 % and women by 34 % amounting to a difference of 4–6 h/day. Sedentary behaviour is thus commonplace in older adults and under-estimated by self-report.

This discrepancy between objective & self-reported measures is explained by Van Uffelen et al. who demonstrate that older adults made judgements and generalisations when answering physical activity questionnaires [29]. Older people had difficulty in generating examples of sedentary activities beyond those explicitly listed. They were uncertain whether non-leisure sedentary activities should be included as sedentary (eating, driving etc.). They also generalised to a ‘typical day’ rather than giving a contemporaneous report of the day’s activities [29]. Nonetheless, six activities have been shown to correlate best with SB; napping, reading, listening to music, watching TV, having a hobby and talking to friends [30]. These can be used to estimate total sedentary time. The underestimate in self-reporting appears to correlate in a linear fashion with objectively measured sedentary behaviour [30]. Therefore, it is reasonable to measure SB using self-reported questionnaires like PASE (Physical Activity Scale for the Elderly) and adjust for under-reporting.

Understanding the volume of sedentary behaviour in older people and the negative associations with health leads to questions over the validity of physical activity targets. Guidelines focus on the attainment of moderately-vigorous physical activity (MVPA) [4], even though older adults spend as little as 1 % of their waking day in MVPA [22]. Exercise promotion for older adults should perhaps also aim at reducing SB [31] and displacing inactivity into light physical activity such as household chores, slow walking or light gardening [6]. These changes can increase the metabolic rate and energy expenditure markedly [32] and are associated with better physical health in adults aged 65+ [31]. Breaks to sedentary time are independently & beneficially associated with lower waist circumference, BMI, triglyceride concentration and 2-h plasma glucose [33]. Reduction & displacement of SB could be a useful target for older people’s health promotion.

However, the literature on modifying sedentary behaviour is limited. Fitzsimons et al. [34] demonstrated an objective reduction in sedentary activity and an increase in activity in a small group of older adults following a motivational interview about reducing sedentary behaviour. Gardiner et al. found an objective reduction in sedentary time after a single-session of goal setting [35]. Magistro et al. [36] found that functional fitness improved in a group of sedentary older adults who undertook a 4 month small-group walking exercise programme. These smaller exploratory studies suggest that change is possible. However, when Stevens et al. [37] conducted a meta-analysis of activity-based interventions in general practice, only 6 suitable studies were found which were “heterogeneous and difficult to replicate or standardise”. Further work is required to establish the effectiveness of interventions to reduce sedentary behaviour in older adults.

This study is novel because it explores the extent of sedentary behaviour in participants in an exercise intervention trial aimed at older people (65 and over) and carried out in general practice, and describes the characteristics associated with sedentary behaviour. The research questions were:1) Do sedentary older people join an exercise study? 2) What demographic, functional and health factors are associated with sedentary behaviour in this self-selected population of older people?

Three hundred eighty seven of the 1104 participants (35 %) were sedentary at baseline. Table 3 shows the associations between participant characteristics and sedentary behaviour, with each association presented unadjusted and adjusted for all other characteristics.

In this study sample, sedentary behaviour was not significantly different between men and women, and was not more common amongst those aged 75 and over than those 65–74. SB was associated with having an abnormal BMI (<18.5 or >25) and 64 % of the sample had abnormal BMIs. Only 18 of the sample had BMI values below 18.5, but 44 % were sedentary; 42 % of the 625 participants with BMI >25 kg/m² were sedentary.

Univariable analyses showed that those who were sedentary were more likely to: ever have smoked, have more comorbidities, take more medications, have difficulty using public transport, use a walking aid, not live in a couple, have informal home help, and describe themselves as inactive. The sedentary reported greater activity limitation, poorer quality of life and poorer physical health as well as performing less well on some functional tests; timed up and go & chair stand. There were no statistically significant associations between sedentary behaviour and educational attainment, household income, employment status, falls in the last year, falls risk (Falls Risk Assessment Tool, FRAT), functional reach, balance (FICSIT), self-reported mental health (SF12-mental component score (Mental Component Score, MCS) and social isolation (Lubben social network score). These variables were excluded from further analysis. Table 3 shows the associations between sedentariness and continuous variables, and dichotomised variables.

Logistic regression analyses showed that only 4 covariates remained independently significantly associated with sedentary behaviour:

For each additional medication the odds of being sedentary increased and for each additional point on the SF-12 PCS (indicating better self-rated health) the odds of being sedentary decreased.