Journal Logo


Objectively Measured Physical Activity in a Diverse Sample of Older Urban UK Adults


Author Information
Medicine & Science in Sports & Exercise: April 2011 - Volume 43 - Issue 4 - p 647-654
doi: 10.1249/MSS.0b013e3181f36196
  • Free


The benefits of physical activity (PA) for the prevention of cardiovascular disease, stroke, diabetes, and some cancers are now well established regardless of the age of the adult population (5,25). Evidence has also shown the preventive and therapeutic effects of PA on dementia (8) and depression (21). Further, PA is related to independence and social well-being (6) in older people because it provides daily access to communities. The potential for PA therefore to support a higher quality of life is great.

Little is known about the levels and patterns of PA of older adults. Activity surveys, which suffer from recall bias, have relied on self-reported activity (16). Daily activity for older people is predominantly incidental walking, which is difficult to estimate and recall accurately (19). Studies with diverse samples of healthy older adults using accelerometry are rare. We have located only four published studies other than ours (4), which are based in Japan (26), Hong Kong (2), England (9), and Canada (3).

The Older People and Active Living Project was designed to recruit a representative sample of older adults aged at least 70 yr from neighborhoods differing in level of deprivation and availability of local amenities in the United Kingdom. The aim was to provide a rich descriptive profile of daily PA and its determinants and consequences for health and quality of life. This article summarizes through accelerometry the daily and the weekly PA levels and associations with PA by age, gender, weight status, frequency of daily journeys, and seasonality.


Sampling and recruitment.

Participants were recruited through medical practices distributed within the city of Bristol, UK. Practices were stratified by amenity access (the proximity of the nearest shop) according to their patient list profile and by the Index of Multiple Deprivation (IMD). The IMD combines 38 economic, social, and housing indicators into a single deprivation score for each locality (20). A 3 × 2 sampling matrix on the basis of the tertiles of IMD and the top and bottom 10% of amenity access was used to select 12 practices distributed across Bristol. Sample size was based on detecting a 20-min difference between three groups (n = 70) in moderate-intensity PA per day (equivalent to approximately 0.5 SD) with 80% power and two-sided pairwise alpha of 2.5%.

Participants of at least 70 yr old were randomly selected from practice patient lists. General practitioners (GP) screened for exclusion according to 1) recent bereavement, 2) terminal illness, 3) debilitating mental illness, 4) inability to complete a questionnaire, and 5) any other illness preventing participation. GP were advised to be as inclusive as possible (including those using mobility aids) to achieve the broadest range of physical function and PA levels. Those not excluded by their GP were mailed an invitation, an information pamphlet, and a consent form. The study was approved by the Bristol Southmead NHS Research Ethics Committee (reference No. 06/Q2002/127).


PA was assessed by ActiGraph GT1Ms (Firmware Version 02.03.01; ActiGraph, Pensacola, FL) programmed to record activity in 10-s epochs, producing both count and step-count data. Participants wore the ActiGraph (supplied in a custom Velcro™ pouch) on a belt for waking hours for 7 d and kept a daily log. The log was used to record details of the days and times when the ActiGraph was worn and any journeys made away from the home. For each journey, participants recorded the time, purpose (shopping, personal business (e.g., banking or posting letters), visiting friends or family, sport or exercise, day trip or excursion, going for a walk or walking the dog, escorting a friend or a relative, work or volunteer activity, entertainment or going out to eat or drink, or "other"), and mode of transport (walking, cycling, driving, car passenger, bus, train, or "other") for each journey, ticking the appropriate checkbox or specifying an unlisted alternative. Height and weight were measured using portable scales and a stadiometer. Demographic and health (diagnosed physical and mental conditions) data were collected by an interview-administered questionnaire. The participant's residential postcode was used to derive the relevant IMD score. Seasonal, local, and monthly weather data were obtained from the Meteorological Office ( and daily sunset times from the US Naval Meteorology and Oceanography ( These data were used to provide date-matched weather (mean maximum temperature, mean minimum temperature, days of frost, millimeters of rainfall, and hours of sunshine per month) and available daylight data for the last day of data recorded.

Data reduction and analyses.

ActiGraph data were reduced (in 2009) using the MAH/UFFE Analyzer Version (MRC Epidemiology Unit, Cambridge, UK) set to ignore runs of 100 min of zeros. Long zero-count periods are not uncommon in older adults (4). Files failing to meet the inclusion criteria (10 h of monitoring on at least 5 d) were excluded from analysis. The summary variables used in analysis were as follows: daily steps; registered accelerometer (wear) time; counts per (registered) minute (CPM) time spent at sedentary (0-99 CPM-calculated by subtracting active time (>99 CPM) from registered time); active (>99 CPM), light (100-1951 CPM, <3 METs), and at least moderate-to-vigorous PA (MVPA) (≥1952 CPM, ≥3 METs) activity levels; and 10-min bouts of MVPA (MVPA10+). Ten-minute bouts are the smallest unit considered adequate to contribute to the recommended 30 min on most days (25). To reduce the stringency of criteria for MVPA10+ (60 continuous 10-s epochs), we reduced the data using 60-s epochs (revised criteria ten 60-s epochs). In the absence of widely accepted cut points for older people, we adopted cut points for adults (7,13). The summarized data were further reduced using Microsoft Excel 2002 and SPSS Version 16.0 (SPSS Inc., Chicago, IL) to derive diurnal patterns of activity, day-of-the-week summaries. The date of data collection was used to establish seasonal effects.

Nonnormally distributed data were log transformed (log [x + 1]). Independent t-tests or one-way ANOVA using covariance where appropriate were used to determine differences between age, gender, and body mass index (BMI) groups. Repeated-measures ANOVAs were used to determine differences in PA levels by day of the week. Pearson correlations were used to determine relationships between the number of journeys performed, the seasonal and meteorological variables, and the PA levels.


Of the 1328 older adults randomly selected from GP patient lists, 156 were excluded and 1170 were sent invitations. From 725 respondents, 244 gave informed consent to participate and 240 completed the study. Overall recruitment rate from those invited to take part was 20.8%. Incomplete or faulty data were provided by 10 participants. Of the 230 valid cases, 8 participants' data were recorded without step-count data. The majority of participants (n = 196, 85%) provided 7 d of data (6 d, n = 27; 5 d, n = 7).

Participant Characteristics

Age, gender, and BMI data for the 230 participants with valid accelerometry data are shown in Table 1. Walking aids (e.g., walking stick, walking frame, wheeled walker, or shopping trolley) were used by 60 participants (26.1%), and mobility aids (e.g., mobility scooter) were used by 6 participants (2.6%) in conjunction with walking aids. Approximately 72.2% (n = 166) reported having been diagnosed with a chronic physical health condition, 6.1% (n = 16) reported having been diagnosed with a mental health condition, and 45.2% reported current treatment for a health condition. Although recruitment of members of this age group for PA studies is challenging (10,11), a reasonably representative sample for gender, age, and BMI was achieved. The age and the gender of the sample differed minimally from the patient lists from which they were selected (differences in proportions between pooled practice lists and recruits in each age and gender group: 70-74 yr, −6.5% men, −4.4% women; 75-84 yr, +4.4% men, −7.0% women; 85-89 yr, +0.9% men, + 7.8% women; 90+ yr, +1.2% men, +3.6% women). The residential IMD scores of participants were representative of the quartiles for the IMD distribution in England.

Participants characteristics.

PA Profiles

Activity levels among participants varied widely (9.2-1006.8 CPM per day, 184.1-16,427 steps per day, 0.0-169.9 min of MVPA per day, and 98.7%-48.1% registered accelerometer time spent in sedentary activity). Journeys per day and PA measures were moderately related (steps per day, r = 0.55, P < 0.001; CPM per day, r = 0.56, P < 0.001; MVPA per day, r = 0.55, P < 0.001; CPM and MVPA were not normally distributed, and log-transformed values were used in the analyses). There was also a weak inverse relationship between daily journeys and sedentary time (r = −0.19, P = 0.005).

Age Effects

There were no significant differences in amount of registered accelerometer time between age groups. Expected age effects of lower levels of PA and more sedentary time in older adults were found after adjusting for BMI (Table 2). Half (47%) of the sample spent more than 80% of their time being sedentary. There was a steep decline in the proportion of active time spent in MVPA, with the 70- to 75-yr age group spending more than double the proportion of active time in MVPA (13.4%) compared with the ≥85-yr age group (3.1%).

Comparisons of key PA measures by age and gender.

Half (50.6%) of the 70- to 74.9-yr age group accumulated the weekly equivalent of 150 min or more per week of MVPA; this proportion declined steeply with age (31.7% of 75-79.9 yr, 15.3% of 80-84.9 yr, and 7.1% of ≥85 yr). However, only three (1.3%) of the participants met adult PA recommendations for health (equivalent of ≥150 min in ≥10-min duration bouts of MVPA per week). Further, only half (50.6%) of the 70- to 74.9-yr age group performed any MVPA10+ per 7-d period, and this proportion declined with age (45.0% of 75-79.9 yr, 28.8% of 80-84.9 yr, and 14.3% of ≥85 yr). Younger participants performed significantly more weekly journeys than older participants (mean ± SD: 70-74.9 yr = 11.5 ± 5.3 journeys per week, 75-79.9 yr = 10.1 ± 4.7 journeys per week, 80-84.9 yr = 8.3 ± 3.5 journeys per week, and ≥85 yr = 6.4 ± 4.2 journeys per week, F(3,218) = 10.5, P < 0.001).

Although this expected age decline was strong, the range within age groups was also wide. In the 70- to 79.9-yr age group, the range of mean steps per day was 690.3-16,427, and in those ≥80 yr, it was 184-12,322. Similarly, for the proportion of time spent being sedentary, the range was broad (70-79 yr = 48.1%-92.4%, ≥80 yr = 64.0%-98.7%).

Gender Effects

Men accumulated 30 min more registered accelerometer time per day than women (mean ± SE: 14.4 ± 0.1 vs 13.9 ± 0.01 h·d−1, P = 0.011). Men spent significantly more time in MVPA (22.6 ± 18.3 vs 14.3 ± 18.3 min, P = 0.001) (Table 2; results-adjusted registered accelerometer time) and reported significantly more journeys from the home (10.6 ± 5.5 vs 8.8 ± 3.9 journeys per week, F(1,217) = 7.1, P = 0.008). Women spent significantly (P = 0.017) more time in light activity (2.9 h·d−1) than men (2.6 h·d−1) (Table 2). Only 20.4% of women and 41.9% of men accumulated the equivalent of 150 min of MVPA per week, and 25.6% of men and 14.2% of women performed at least one MVPA10+ per week.

BMI Effects

There were no significant differences in registered accelerometer time between BMI groups. After adjusting for age, there were significantly (P < 0.05) lower levels of PA in overweight and obese compared with normal weight participants (Table 2). Similarly, sedentary time was higher in overweight (P = 0.030) and obese (P = 0.075) compared with normal weight participants (Table 2).

Walking and mobility aid effects.

Those who used walking aids and mobility aids were significantly older (F = 13.6, P < 0.001), had higher BMI (F = 5.2, P = 0.006), and recorded less registered accelerometer time (F = 5.6, P = 0.004) than those who did not. After adjusting for these factors, those who used a walking aid (mean ± SE: 3597.0 ± 283.5 steps per day, 150.5 ± 13.6 CPM per day, 13.6 ± 2.4 min of MVPA per day, and 11.4 ± 0.1 sedentary hours per day) or mobility aid (2520.9 ± 889.1 steps per day, 96.3 ± 40.9 CPM per day, 9.0 ± 7.3 min of MVPA per day, and 12.1 ± 0.4 sedentary hours per day) were significantly less active and more sedentary than those who did not (4806.8 ± 158.8 steps per day, F (2,216) = 8.3, P < 0.001; 195.6 ± 7.9 CPM per day, F(2,224) = 19.6, P < 0.001; 20.6 ± 1.4 min of MVPA per day, F(2.224) = 24.5, P < 0.001; 11.0 ± 0.1 sedentary hours per day, F(2,224) = 5.8, P = 0.003). Those who used walking aids and mobility aids made fewer journeys per week (mean ± SD = 7.8 ± 4.9 and 5.2 ± 2.1, respectively) than those who did not (10.6 ± 4.7 journeys per week, F(2,219) = 10.3, P < 0.001).

Patterns of Activity

Journey frequency effects.

After adjustment for age and registered accelerometer time, those who made more journeys from the home had significantly (P < 0.05) higher levels of PA and were significantly (P < 0.05) less sedentary (Table 3).

Comparisons of PA measures by level of journey frequency.

Diurnal effects.

Participants were significantly (P < 0.001) more active in the mornings (8:00 a.m. to 1:00 p.m.) than during afternoons (1:00 to 6:00 p.m.) and significantly (P < 0.001) less active in the evenings than both afternoons and mornings (mean ± SD: mornings = 259.3 ± 175.9 CPM, afternoons = 181.8 ± 128.0 CPM, evenings = 102.5 ± 82.7 CPM, F(2, 229) = 1.4, P < 0.001). Peak CPM for men and women was recorded between 10:00 and 11:00 a.m. Men were more active than women throughout the day, but this was only significant in the morning (mean ± SD: men = 289.5 ± 207.2 CPM vs women = 228.0 ± 129.7 CPM, F(1,229) = 4.5, P = 0.034). Younger participants (70-74.9 yr old) were significantly (P ≤ 0.005) more active than older participants (≥85 yr old) across the day (morning = 334.3 ± 210.5 vs 114.2 ± 71.2 CPM, afternoon = 230.8 ± 148.8 vs 93.0 ± 80.4 CPM, and evening = 125.0 ± 116.7 vs 59.6 ± 34.5 CPM for 70-74.7 and ≥85 yr old, respectively). The proportion of activity (steps) in each segment of the day was as follows: morning, 46.9%; afternoon, 37.4%; and evening, 15.7% (mean % of steps for all, N = 220). There was no significant age effect for these proportions (morning, F(3,216) = 1.4, P = 0.245; afternoon, F(3,216) = 0.2, P = 0.891; and evening, F(3,216) = 2.4, P = 0.066).

Day of the week and seasonal effects.

There were no significant gender or age group differences for activity pattern across the days of the week or differences for PA measures between individual weekdays (so weekdays were collapsed for comparison with weekend days). Sundays had significantly less registered accelerometer time and were significantly less active than Saturdays and weekdays (Table 4). Only time spent in light PA was significantly (P = 0.005) higher in spring (March to May) than that in winter (December to February) (winter, n = 52, 2.7 ± 1.0; spring, n = 46, 3.4 ± 1.3; summer, n = 89, 3.1 ± 1.3; and autumn, n = 43, 3.1 ± 1.2 h·d−1). Seasonal differences were hypothesized to be related to available daylight or prevailing weather conditions. Weak but significant relationships were found between time spent in light PA and 1) time from 08:00 a.m. (a time before which little PA occurred) to sunset (r = 0.183, P = 0.005), 2) hours of sunshine per month (r = 0.146, P = 0.027), 3) days per month with frost (r = −0.146, P = 0.026), and 4) mean maximum monthly temperature (r = 0.134, P = 0.042). No such relationships were found for daily journeys.

Comparisons of PA measures by day of the week.


The strength of this study is the use of accelerometry estimates of daily PA in a diverse group of 230 older adults living in neighborhoods with a wide range of deprivation and access to amenities. The importance of the collection of objective data on PA in this population is demonstrated by the large differences between self-report data and accelerometer data (compliance with PA recommendations for health for the ≥65-yr age group: men 16% vs 5% and women 12% vs 0% for self-report and accelerometry, respectively) in a recent nationally representative sample (age = 16-75+ yr, N = 2066) (22). The challenges of recruitment for this population (10) suggest that volunteers were likely to be healthier and less infirm than UK population norms. Even so, data indicated that PA was less than half of that found in younger (age = 26.9 ± 4.1 yr) adults (men = 404.3 ± 134.0 CPM, women = 370.0 ± 81.1 CPM) participating in Better Ageing (4). In the United States, a recent large-scale (N = 6329, age = 6-70+ yr) investigation of PA using accelerometry confirmed that CPM levels of participants 20-59 yr old (mean ± SEM: men = 385.7 ± 8.0 CPM, women = 309.7 ± 5.0 CPM) were more than double than that of participants ≥70 yr old (mean ± SEM: men = 188.9 ± 5.4 CPM, women = 169.3 ± 3.0 CPM) (23).

In our sample, half (47%) of the participants spent more than 80% of their time being sedentary. The Health Survey for England 2008 (22) reported average proportions of sedentary time (0-199 CPM) in those ≥75 yr of 79% for men and 84% for women compared with 67% for younger adults (45-54 yr, 67% sedentary time). These results are not directly comparable with the current study because they have been derived using a different cut point to define sedentary time (0-199 vs 0-100 CPM used in the current study), which would inflate the volume of sedentary time. Nevertheless, age is still associated with higher volumes of sedentary time and is accompanied by lower levels of PA. The steep age-related decline in PA seems to be characterized by lower activity volume, less higher-intensity PA, and lower frequency of "getting out and about." Age differences were apparent across the day with those who were aged ≥85 yr performing just a third of the activity of those aged 70-74.9 yr at peak activity times.

However, for each age group, a wide range of activity levels was recorded, with some recording very little movement (184.1 steps per day) and who were sedentary for almost all (98%) of their time. There may be concerns that such a low value may be erroneous, and it is possible that walking aids may distort the data. However, in this case, journey log and researcher notes supported this evidence. The participant was a 93-yr-old man living in sheltered accommodation who used a wheeled walking aid to get about his accommodation. He left his home only once during the 7-d monitoring period, when he recorded his highest daily step count. Conversely, there were exceptionally active people who accumulated nearly 3 h of MVPA per day accrued through work-related activity or leisure walking. Clearly, low activity is not characteristic of all older people. Indeed, those who were overweight or obese were less active and more sedentary than normal weight adults regardless of age.

Although many older adults achieve 150 min of MVPA per week, they fail to meet national physical recommendations because they do not sustain MVPA continuously for 10 min. The Health Survey for England 2008 (22) reported just 5% of men and no women ≥65 yr old achieved 30 min or more of MVPA on at least 5 d·wk−1. Even in younger individuals (16- to 34-yr-olds), prevalence was just 11% of men and 8% of women. Although this study used the same device (GT1M), it should be noted that a different epoch (60 s) and MVPA cut point (2020 CPM) was used. These factors might, if anything, be expected to result in reduced prevalence of achieving PA recommendations. However, considering the lower age of the sample (≥65 vs ≥70 yr), these results do not appear to be inconsistent with the data presented here and highlight both low and declining prevalence with age of achieving these recommendations. Idiosyncrasies of accelerometry methods such as measurement epoch length do not explain these deficits. It seems likely that the gentler and intermittent nature of older adults' activity may be normal even for those who achieve high volumes of activity. The recent recommendations for activity in the United States (25) have adopted a pragmatic approach for older adults (currently in review in the United Kingdom). They suggest using moderate PA as a relative rather than an absolute term to accommodate the lower levels of cardiorespiratory fitness and physical function in this group. Adoption of this approach would suggest that more older adults would achieve more qualifying ≥10-min bouts.

Women recorded significantly less registered accelerometer time than men. The reason for this is not clear but may be accounted for by women choosing to retire to bed earlier than men (15). After adjusting for registered accelerometer time, men were more active than women. Additionally, men were more active at higher intensities than women and recorded more MVPA, with women recording greater volumes of light activities. The volumes of MPVA and gender differences found (mean ± SE: men = 23.1 ± 1.8, women = 13.8 ± 1.8) are similar to those of the Health Survey for England 2008 (22) (65-74 yr old, men = 23 min, women = 16 min; ≥75 yr old, men = 12 min, women = 9 min MVPA). However, data collected using the same epoch (60 s) and MVPA cut point (2020 CPM) from US adults (NHANES) (23) ≥70 yr old were 8.7 min (SEM = 0.7 min) MVPA per day for men and 5.4 min (SEM = 0.3 min) MVPA per day for women. These lower values in the US sample suggest that device (ActiGraph 7164) and international or cultural differences may account for the lower MVPA levels reported.

Our explanation for these confirmed gender differences is the maintenance of traditional gender roles, with men engaging in heavier activities and making more daily journeys as "bread winners" of the family while women are more involved in lighter domestic activities, such as home management, care, and cooking. Daily journey logs confirmed that those making most journeys away from their home were more likely to be the more active men.

The diurnal pattern of activity shows that most activity in older adults, regardless of age group, is performed in daylight hours with a peak volume in the late morning. Very little activity takes place beyond 6:00 p.m. This confirms our earlier work (4) and that of others (3) who have broken the day into morning, afternoon, and evening and also found that more longer-duration bouts of activity occurred in the mornings.

The morning peak in diurnal activity is related to activities that take the individual out of the house (e.g., shopping). Survey data (14) have indicated that for adults ≥70 yr old, the largest single purpose for journeys (41%) was shopping. Journey frequency has been associated in previous studies with increased physical function in older adults (17). Careful consideration of these peak activity times and the purposes of the activity will be required when designing PA promotion programs. Late morning is a key time for activity, but it is also a time when shopping or other purposeful journeys take place. It may be possible to capitalize on these outings without interfering with them.

It might be expected that, for a retired population without weekday work commitments, PA levels would differ little across the days of the week. Although there were no differences among individual weekdays, there were differences between weekdays and each weekend day. Saturday and Sunday PA were 10% and 20% lower, respectively, than weekday levels. There was 20 min less registered accelerometry time on Sundays than other days, and this might be partially explained by longer sleeping hours. There was a greater reduction (just less than 30%) in MVPA and steps per day on Sundays, indicating fewer journeys occurring on this day. It is not clear whether the reduction in activity on Sundays is due to reduced local shopping and service availability or a reflection of the sustained tradition of Sunday being a day of rest. Seasonal differences were likely to be a result of available daylight or prevailing weather conditions. However, regular patterns of daily journeys appeared to be largely uninterrupted.


Although these data provide useful insight into the patterns of PA in older adults for age and gender groups, the cross-sectional nature of the study limits implications for causal mechanisms producing group and individual differences. There are many determinants of low activity, most obviously pain, illness, and loss of physical function. However, psychosocial factors such as isolation, acceptance of inactivity as the norm, and challenges provided by the local neighborhood may also be critical. Tracking these factors over time within individuals, and examining their interaction is needed to fully understand age-related declines in PA.

According to the limited evidence, accelerometers record higher step-count scores than pedometers. Furthermore, accelerometer-derived step-count data have been reported, and their precision with older adults is not fully established. Even less is known of the validity of step counts when walking aids are used. The differences between pedometer- and accelerometer-derived step counts are probably because accelerometers have a lower sensitivity threshold than pedometers (24). The GT1M used here has a higher sensitivity threshold (24) than previous models and thus may attenuate this difference. Further, pedometer comparisons may be invalid as they have been found to underrecord steps by as much as 34% (18). With older adults having a less dynamic gait (1), accelerometers may be able to detect steps that pedometers do not and thus provide a more appropriate measure of steps in this population. More research with specific instruments is required to fully enlighten on these differences and the mechanisms underpinning them.


This study provides objective data for PA in a representative sample of older adults. The data show that few meet current PA for health recommendations. Increasing age brings about lower volumes and intensity of PA and greater amounts of time spent sedentary. However, there is a wide range of PA levels among older adults from each age group, indicating that given the right conditions, higher activity levels may be sustainable for longer than is currently observed. Data confirm gender and weight status as determinants of PA and suggest opportunities for focused intervention. Diurnal, weekly, and seasonal patterns of PA indicate when peak activity takes place and show that journeys out of the house for shopping and personal business are important in their contribution to PA levels. These patterns of activity can help inform the nature and timing of delivery of PA promotion strategies among older adults.

This work was supported by the National Prevention Research Initiative (grant No. G0501312) ( whose funding partners are as follows: British Heart Foundation; Cancer Research UK; Department of Health; Diabetes UK; Economic and Social Research Council; Medical Research Council; Research and Development Office for the Northern Ireland Health and Social Services; Chief Scientist Office, Scottish Executive Health Department; Welsh Assembly Government; and World Cancer Research Fund.

The authors are grateful to the participants and general medical practices who gave their time to the study. The authors also acknowledge all of the investigators and the team members of the Older People and Active Living Project not listed as coauthors.

The results of the present study do not constitute endorsement by the American College of Sports Medicine.


1. Baloh RW. Disequilibrium and gait disorders in older people. Rev Clin Gerontol. 2002;12(1):21-30.
2. Cheng YH, Choui KL, MacFarlane DJ, Chi I. Patterns of physical exercise and contributing factors among Hong Kong older adults. Hong Kong Med J. 2007;13(4 suppl):S7-12.
3. Copeland JL, Esliger DW. Accelerometer assessment of physical activity in active, healthy older adults. J Aging Phys Act. 2009;17(1):17-30.
4. Davis MG, Fox KR. Physical activity patterns assessed by accelerometry in older people. Eur J Appl Physiol. 2007;100(5):581-9.
5. Department of Health Physical Activity Health Improvement and Prevention. At Least Five a Week: Evidence on the Impact of Physical Activity and its Relationship to Health. A report from the Chief Medical Officer. London (UK): Department of Health; 2004. p. 1-9.
6. Elavsky S, McAuley E, Motl R, et al. Physical activity enhances long-term quality of life in older adults: efficacy, esteem, and affective influences. Ann Behav Med. 2005;30(2):138-45.
7. Freedson P, Melanson E, Sirard J. Calibration of the Computer Science and Applications, Inc. accelerometer. Med Sci Sports Exerc. 1998;30(5):777-81.
8. Hamer M, Chida Y. Physical activity and risk of neurodegenerative disease: a systematic review of prospective evidence. Psychol Med. 2009;39(1):3-11.
9. Harris TJ, Owen CG, Victor CR, Adams R, Ekelund ULF, Cook DG. A comparison of questionnaire, accelerometer, and pedometer: measures in older people. Med Sci Sports Exerc. 2009;41(7):1392-402.
10. Harris TJ, Victor CR, Carey IM, Adams R, Cook DG. Less healthy, but more active: opposing selection biases when recruiting older people to a physical activity study through primary care. BMC Public Health. 2008;8(182):1-6.
11. Hildebrand M, Neufeld P. Recruiting older adults into a physical activity promotion program: active living every day offered in a naturally occurring retirement community. Gerontologist. 2009;49(5):702-10.
12. Deleted in press.
    13. Matthews CE, Chen KY, Freedson PS, et al. Amount of time spent in sedentary behaviors in the United States, 2003-2004. Am J Epidemiol. 2008;167(7):875-81.
    14. National Statistics. National Travel Survey: 2006. London (UK): National Statistics; 2007. p. 25-32.
    15. Reyner A, Horne JA. Gender- and age-related differences in sleep determined by home-recorded sleep logs and actimetry from 400 adults. Sleep. 1995;18(2):127-34.
    16. Shephard RJ, Vuillemin A. Limits to the measurement of habitual physical activity by questionnaires. Br J Sports Med. 2003;37(3):197-206.
    17. Shimada H, Ishizaki T, Kato M, et al. How often and how far do frail elderly people need to go outdoors to maintain functional capacity? Arch Gerontol Geriatr. 2010;50(2):140-6.
    18. Silva M, Shepherd EF, Jackson WO, Dorey FJ, Schmalzried TP. Average patient walking activity approaches 2 million cycles per year: pedometers under-record walking activity. J Arthroplasty. 2002;17(6):693-7.
    19. Sims J, Smith F, Duffy A, Hilton S. The vagaries of self-reports of physical activity: a problem revisited and addressed in a study of exercise promotion in the over 65s in general practice. Fam Pract. 1999;16(2):152-7.
    20. Social Disadvantage Research Centre. The English Indices of Deprivation 2004. Oxford (UK): Department of Social Policy and Social Research, Communities and Local Government; 2004. p. 14-42.
    21. Teychenne M, Ball K, Salmon J. Physical activity and likelihood of depression in adults: a review. Prev Med. 2008;46(5):397-411.
    22. The Health and Social Care Information Centre. Health Survey for England 2008: Physical Activity and Fitness. London: The NHS Information Centre for Health and Social Care; 2009. p. 59-88.
    23. Troiano RP, Berrigan D, Dodd KW, Masse LC, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc. 2008;40(1):181-8.
    24. Tudor-Locke C, Johnson WD, Katzmarzyk PT. Accelerometer-determined steps per day in US adults. Med Sci Sports Exerc. 2009;41(7):1384-91.
    25. US Department of Health and Human Services. Physical Activity Guidelines Advisory Committee Report, 2008. Washington (DC): US Department of Health and Human Services; 2008. p. A1-10.
    26. Yoshiuchi K, Nakahara R, Kumano H, et al. Yearlong physical activity and depressive symptoms in older Japanese adults: cross-sectional data from the Nakanojo study. Am J Geriatr Psychiatry. 2006;14(7):621-4.


    ©2011The American College of Sports Medicine