Adherence to physical activity guidelines in older adults, using objectively measured physical activity in a population-based study

Sample

7735 men participating in the British Regional Heart Study, an on-going prospective, population-based cohort study were recruited from primary care centres in 24 British towns in 1978–80 and followed up repeatedly [18]. In 1999–2001, a parallel cohort study (British Women’s Heart Health Study) of 4286 women of the same age and in the same primary care centres was established, omitting two study towns (Dewsbury and Maidstone) and adding another (Bristol), with a response rate 60% [19]. In 2010–2012, 6529 survivors from these two cohorts (3292 men and 3237 women) were invited to participate in a study of objectively measured PA by post. Men were also asked to wear accelerometers on other occasions, here we use the postal survey that occurred concurrently with the women’s study and using the same study protocol. The National Research Ethics Service (NRES) Committee for London provided ethical approval. Participants provided informed written consent to the investigation, which was performed in accordance with the Declaration of Helsinki.

Measures

Statistical methods

The total number of minutes spent in MVPA in each of the valid days (range 3–8 days) was summed. For participants who did not have 7 valid days, the total MVPA minutes were scaled to the number of valid days. Attaining ≥150 minutes of MVPA in 7 days was calculated for the total minutes of MVPA accumulated in bouts lasting (i) ≥5 minutes (MVPA5+) and (ii) ≥10 minutes (MVPA10+).

Descriptive statistics of demographic, social, and environmental variables stratified by gender were calculated and adjusted mean PA levels were estimated. Summary measures of demographic, social and environmental variables were calculated according to achieving ≥150 minutes MVPA5+ and MVPA10+. ANOVA and chi square tests were used to compare continuous and categorical variables, respectively.

Logistic regression models were used to compare participants who achieved guidelines to those who did not. Model 1 was adjusted for age, season, wear time (average minutes/day on valid days) and region of residence. Potential mediators were added one at a time to model 1, to evaluate the role of each. These included physical and mental health and wellbeing, perceptions of local environment and behavioural factors. Model 2 was additionally adjusted for mental and physical health (depression score and number of chronic conditions). Complete case analysis was used. As sensitivity analyses, the main analyses were repeated using a cut-point of >1952 cpm to define MVPA. Statistical analyses were run in R version 2.15.3 and Stata version 12.

Prevalence and correlates of attaining PA recommendations in bouts of 10 minutes or more

Prevalence and correlates of attaining PA recommendations in bouts of 5 minutes or more

28% (n = 452) men and 21% (n = 177) achieved 150 minutes of MVPA defined as >1040 cpm in bouts of 5 minutes or more; (Additional file 1: Table S1). Correlates of achieving MVPA5+ were mostly similar to achieving MVPA10+, but also included social isolation, feeling safe in the daytime and doing shopping within walking distance of home. These associations were mostly confirmed in logistic regression models; (Additional file 1: Table S2, Model 1 and with additional adjustment for depression and number of chronic conditions (Model 2), stronger associations were seen for regularly walking a dog, leaving the house on more than 5 days per week, having no mobility limitations, not being depressed and, particularly for women, using active transport (walking or cycling).

Attainment of PA recommendations using 1952 cpm cut-point

In further analyses using >1952 cpm to define MVPA, the prevalence of adherence to MVPA guidelines MVPA5+ was for 12% men and 8% for women, and MVPA10+ was 7% for men and 3% for women (Table 1). Correlates of adherence to MVPA10+ guidelines in men and women are presented in Additional file 1: Table S3. In logistic regression models, similar variables predicted adherence to MVPA10+ guidelines in men for 1952 cpm cut-point as for 1040 cpm cut-point and effects were in the same direction (Additional file 1: Table S4). The physical health (chronic conditions, mobility limitations), depression and the behavioural variables were all consistent correlates of achieving MVPA. Logistic regression models were not presented for women due to low numbers, only 3% (n = 26), attaining guidelines with the higher cut-point for MVPA (Additional file 1: Table S3).

Levels of PA in older people in the UK

Adherence to the PA guidelines was low using either intensity cut-point to define MVPA. Our estimates of 15% men and 10% women attaining guidelines based on the age appropriate cut-point of activity >1040 cpm to define MVPA were higher than previous studies have reported. However our estimates of 7% and 3% using the cut-point of >1952 cpm are more in line with estimates from other UK-based studies of older adults which have used cut-points nearer to 2000 cpm [2, 15, 17]. In the Health Survey for England subsample (approximately 400 adults with valid accelerometer data aged over 65), 5% men and 0% of women adhered to the guidelines [2]. In the OPAL study, among 230 community-dwelling adults in UK aged over 70 years, only 1.3% met the MVPA10+ (>1952 cpm) guidelines [15]. Another study (n = 240) of older adults recruited from UK primary care centres reported that 2.5% of men and women achieved 150 minutes or more per week of MVPA10+ (defined as >1999 cpm), while 62% did not do any minutes of MVPA in a week [17]. Data from National Health And Nutrition Examination Survey in the USA estimated that 6.3% of over 70 year olds met guidelines, counting activity in bouts of ≥10 minutes and 10.4% not considering bouts [27]. Whilst half of 70–74 year olds in the OPAL study accumulated 150 minutes/week of MVPA in bouts of one minute or more, this prevalence declined steeply with age to 7% of over 85 year olds [15]. These studies used the same brand of accelerometer to measure activity as we did, although there were differences in the data processing protocols and the cut-points to define MVPA (1952 cpm [15], 1999 cpm [17] and 2020 cpm [2]) whereas we used 1040 cpm or 1952 cpm.

Importance of MVPA bout length

When 5 minute bouts were considered instead of 10 minute bouts, the prevalence of attaining 150 minutes/week of MVPA was doubled. The 10 minute bouts stipulated by guidelines were based on trial data for cardiometabolic risk factors only, not clinical end points [12]. Newer data indicate that accumulating 10 minute bouts of PA is associated with reduced risks of adiposity and intermediary markers of disease in older adults [28]. However, we lack data showing whether health benefits for hard clinical endpoints accrue from bouts of activity shorter than 10 minutes, or whether it is the total volume of PA which is important, irrespective of the pattern in which is accumulated. Until evidence from studies linking accelerometer data to incident morbidity and mortality data are available, it remains unclear whether older adults would benefit from higher cumulated total MVPA, or from increasing the duration of shorter spells of MVPA. Data from studies of high intensity training suggest that short high intensity interval training can have beneficial effects on CVD risk factors [29] but we do not know if such training regimes are sustainable over the long term. It is unlikely that high intensity training would be suitable for older adults, especially in unsupervised settings, particularly given the higher levels of co-morbidities, mobility limitations and balance impairments in that age-group.

Correlates of meeting MVPA guidelines

The directions of associations between the correlates and adherence to MVPA guidelines fitted with expectations from studies with self-reported PA and previous smaller studies using accelerometers. As expected [17, 30, 31], we found that poor physical health (chronic health conditions and mobility limitations) was an important barrier to achieving MVPA guidelines. We found that associations with poor physical health remained after adjusting for depression. Depression itself was an important barrier to achieving MVPA guidelines, after accounting for presence of chronic physical conditions. Other modifiable factors which could be potential levers for PA interventions in older adults were identified; as seen elsewhere higher levels of exercise self- efficacy [14, 17] and exercise outcomes expectations were associated with greater adherence to guidelines. Men left the house on more days than women and also reported walking or cycling for transport more often than women, and each of these behaviours, as well as dog-walking (as seen elsewhere [17]) consistently predicted achieving guidelines, suggesting that encouraging older adults to leave the house each day and do walking based activities could improve adherence to guidelines. Most participants had positive perceptions of their neighbourhood, and these were not strongly related to achieving guidelines, with the exception of feeling safe walking alone after dark, but very few women felt safe walking after dark. Other studies report that associations between neighbourhood deprivation measures and PA are mediated by personal factors [16].

Strengths and limitations

To our knowledge, this is the largest study to have estimated adherence to national guidelines using accelerometer-measured PA in older British men and women including the oldest old. Previous studies included fewer participants and fewer correlates. We study community dwelling adults, not special high risk clinical populations, hence results should be generalizable to this age group. We used the GT3x, a widely used accelerometer. Given the lack of consensus about which cut-points to use to define MVPA in older adults, we tested different, validated, cut-points [22, 23] to define MVPA levels. Our data are cross-sectional, so we cannot assess causality among the correlates of MVPA adherence, longitudinal studies will be necessary. Loss to follow-up may be greater among the more ill and infirm who are likely to be less active, resulting in under-representation of the less active. However, our response rates to the postal accelerometer study (51% in men and 29% in women) are line with other studies of older adults: 21% [15], 43% [17] and in the Health Survey for England 37% women and 48% men over 75 years had 4 or more days with valid data [2].