Regular engagement in physical activity is an essential behavior to promote health and prevent chronic disease for persons of all ages and those with and without functional limitations. The ability to accurately measure physical activity in a free-living environment is crucial to any investigation in which physical activity is an outcome or exposure variable of interest. However, physical activity is a complex behavior with high levels of interindividual and intraindividual variation, making it a difficult construct to measure. Techniques for measuring physical activity need to be able to distinguish different characteristics of activity, namely, the frequency, duration, intensity, and type, to further our understanding of population levels of physical activity. Measurement techniques also are needed to determine the relationship between physical activity and disease/disability and to be able to document the efficacy of interventions designed to increase physical activity participation.
Despite advances in the use and development of accelerometers to measure physical activity, careful consideration is needed when deciding to apply objective monitoring methods to certain populations, especially children, older adults, and adults with functional limitations, because methods are not “one-size-fits-all.” Factors to consider include 1) the physical activity indicators of interest, 2) the application of objective physical activity monitoring in these populations, and 3) the choice of whether calibration and validation studies support the application in the population of interest.
This article addresses each of these considerations with an aim toward describing the current use of accelerometers with children, older adults, and adults with functional limitations; describing best practices for future applications; and recommending future research to advance the use of objective monitoring technology. For the purposes of this article, discussion is limited to children age 3–18 yr, older adults (≥65 yr), and adults with functional limitations.
DERIVING INDICATORS OF PHYSICAL ACTIVITY FROM OBJECTIVE MONITORS
Understanding the basic terminology is important to understand the applicability of accelerometers for selected study designs and measures of interest within selected populations. Physical activity is conventionally defined as any bodily movement produced by the contraction of the skeletal muscle that results in energy expenditure (10). It is important to recognize that physical activity is a behavior that results in energy expenditure. Whereas physical activity is quantified by characteristics of frequency and duration, energy expenditure reflects the sum of all metabolic processes involved to support the physical activity in question. Several factors influence energy expenditure, including age and functional status.
If one assumes human mechanical efficiency to perform an activity to be similar across populations, then energy expenditure would generally be constant, and volume indicators such as kilocalories per day, MET-minutes per day, or MET-hours per day would be comparable across different populations. If one cannot assume a similar mechanical efficiency across populations, however, an unwanted source of variation in the measurement of physical activity is introduced, obscuring meaningful interpretation and comparison of findings. Many disabilities (stroke, spinal cord injury, Parkinson disease) affect neuromuscular and movement-related functions that collectively affect movement itself. The ability to engage in physical tasks can therefore become mechanically altered and more physiologically demanding to an individual. When considering such facts, many practical issues arise when considering the use of objective physical activity monitoring in this population subgroup. For example, using a commercially available algorithm to extrapolate accelerometer counts per day into kilocalories per day for a stroke survivor with gait abnormalities is unlikely to yield precise estimates for total or physical activity–related energy expenditure for that individual.
Furthermore, additional unwanted variation is introduced when one does not consider absolute versus relative intensity differences for individuals. Figure 1 highlights differences between absolute and relative intensity for selected walking activities across different levels of physical fitness. Absolute intensity may assume a similar mechanical efficiency (e.g., walking at 3 mph represents a similar activity irrespective of fitness level), but relatively speaking, this could constitute a light-intensity activity for an individual with a 12-MET capacity and a vigorous-intensity activity for an individual with a 6-MET capacity. Age-related decreases in the average fitness level (21) and absolute-versus-relative physical activity intensity differences are amplified and necessitate consideration when evaluating indicators of physical activity derived from objective monitors. Further, this issue becomes particularly important when examining the child and adolescent population, given that growth and maturation can affect several factors that may be related to the ability of accelerometers to detect movement (e.g., body size, relative length of body segments, fitness).
Such factors as noted above are essential to consider when using objective monitors to assess physical activity behavior in different subpopulations. The next section describes the current application of accelerometers in children age 3–18 yr, older adults, and adults with functional limitations. We focus on observational, intervention, and determinant studies, with further categorization by school level (preschool, elementary, middle, and high school) for studies of children. It is important to note that the available body of literature regarding children and adolescents is more extensive than the available body of literature for older adults and adults with functional limitations; thus, this review is not intended to be comprehensive for the child and adolescent population.
CURRENT APPLICATION OF OBJECTIVE PHYSICAL ACTIVITY MONITORING
Due partly to the growing public health concern about childhood obesity, the number of investigations examining physical activity in children has increased for the past decade. Evidence from most observational studies of preschool children indicates that accelerometer-assessed levels of moderate-to-vigorous physical activity tend to be low, while the amount of time spent in sedentary and/or light activity is high (8,57). Population-based accelerometry studies of elementary school children have shown varied results regarding the prevalence of meeting physical activity recommendations. In the United States, data from the 2003–2004 National Health and Nutritional Examination Survey (NHANES) indicated that 42% of children age 6–11 yr met physical activity recommendations. Studies of children in the United Kingdom show extremes in the prevalence of meeting physical activity recommendations, from 7% in the northeast of England (41) to almost 70% in the middle east section of England (72). Another recent investigation from England showed that, on average, children were achieving 60 min of physical activity per day (60). This same investigation also examined bouts of activity and the effects of seasonality, which are constructs deserving more attention in future research. Regarding the middle and high school populations, 2003–2004 NHANES data showed 8% of youth age 12–19 yr met physical activity recommendations. Taken together, the results corroborate the long-standing belief, first observed in self-report studies, that physical activity declines with age. Although accelerometry has confirmed the self-report results, discrepancies exist between self-report and accelerometer-based estimations of physical activity.
Population-based accelerometry studies provide essential statistics for public health planning and intervention, but Gidlow et al. (26) demonstrated that prevalence estimates change drastically when different cut points to define physical activity intensity levels are used. Comparison among studies would be easier if researchers adopted standard methods for population assessment by accelerometer. Despite the cut point issue, accelerometers allow investigators to examine facets of physical activity not previously measurable by self-report instruments, and they offer an additional degree of accuracy. Accelerometer data from the Study of Early Child Care and Youth Development showed a decline of 38 min·yr−1 in weekday and 41 min·yr−1 in weekend moderate-to-vigorous physical activity (46). The authors estimated that age 12.8 yr (girls) and 14 yr (boys), averaged across weekdays and weekend days, to be the ages at which adolescents dropped below the recommended levels. Before the use of accelerometers, accurate estimates such as these were not possible.
Accelerometers have been used in several physical activity intervention investigations involving children and youth, and they are emerging as useful tools for comparing effectiveness across interventions.
Only a few studies of interventions in the preschool population exist in the published literature, but it seems accelerometers can be used to assess changes in physical activity in preschool children. Specker et al. (63) found children who participated in the gross motor portion of an intervention spent a higher percentage of time after intervention in vigorous physical activity than those who participated in the fine motor portion. Cardon et al. (9) discovered that supplying physical activity equipment and creating playground markings did not affect moderate-to-vigorous physical activity during recess. However, the addition of playground equipment positively affected light physical activity and moderate-to-vigorous physical activity in another investigation (30). Increases in preschooler physical activity also have been detected for interventions that implemented curricula to increase classroom time spent in moderate-to-vigorous physical activity (68).
Intervention strategies with elementary school children have ranged from multicomponent school-based approaches (including physical education, classroom-based, and extracurricular physical activities and playground adjustments) (74) to individual approaches, such as rewarding physical activity participation with time for television viewing (28). In addition, an innovative investigation aimed at creating an “activity-permissive” environment during the school day used accelerometry for the main outcome measure (38). Similar to the interventions in younger populations, accelerometers have been used in several investigations involving middle school and high school children and youth. Some interventions have been clinically based (52), even including behavioral economics strategies (16). Others have included newer, innovative strategies such as use of the Internet and Boy Scout troop badges (34) and active video games (47). Accelerometers have been demonstrated to be reliable and feasible for use in these types of studies and have allowed investigators to examine more complex variables such as in-school versus after-school physical activity in a more reliable manner than previous recall-based investigations.
The body of literature including correlates and/or determinants of accelerometer-assessed physical activity is too vast to be thoroughly discussed here. Most investigations have shown that demographic variables such as age, sex, race, and preschool attended are associated with preschoolers’ physical activity (53), as well as weight status (70), and psychosocial correlates such as parent perception of child’s athletic competence (53). A few studies have examined other correlates, such as bone area and mass (76). Aside from demographic variables, the literature examining determinants of physical activity has not consistently shown many factors to be related to physical activity in preschool children, warranting more work in this area in future investigations.
One of the first investigations of correlates of accelerometer-assessed physical activity in elementary school children published in 1991 showed a relationship between family members’ physical activity and child physical activity (23). Many biological variables related to physical activity in elementary school children have been examined, including weight status (1), visceral fat (14), insulin (1), and appendicular bone mineral density (32). Psychosocial variables seem to be less studied in this age group. However, Roemmich et al. (59) showed that liking and relative reinforcing values were associated with accelerometer-assessed physical activity in children age 8–12 yr, and Shoup et al. (62) demonstrated that less active children had lower quality-of-life scores. Psychosocial factors related to accelerometer-assessed physical activity in elementary school-aged children are less clear than biological variables.
In contrast to the elementary school group, more psychosocial factors have been investigated in middle-school- and high-school-age groups. Self-efficacy for physical activity (69), participation with community organizations (69), parent physical activity (69), and active transportation to school (71) are all positively related to accelerometer-assessed physical activity in older children. Perceived access and proximity to commercial physical activity facilities also have been associated with physical activity, particularly in adolescent girls (61). Biological variables such as maturation also potentially seem to be related to physical activity in this age group (2).
Despite currently published data examining determinants and correlates of accelerometer-assessed physical activity, a core group of variables on which to intervene has yet to be identified. In part, this may be due to the lack of a standard method for collecting and analyzing accelerometer data for children.
Knowing that the number of older adults in the US population is rapidly increasing and that advancing age is accompanied with a greater risk of chronic disease and disability, monitoring free-living physical activity behavior among older adults is crucial. Accelerometer data from the 2003–2004 NHANES revealed that older adults are the least active segment of the population, with adults age 60–69 yr accumulating 12–17 min of moderate-to-vigorous physical activity per day and adults age 70 yr and older accumulating 5–9 min of moderate-to-vigorous physical activity per day (67). Other studies using accelerometers have documented an age-associated decrease in physical activity (35), but the absolute volume of physical activity reported by studies varies depending on the methods for analyzing accelerometer data. The NHANES data were analyzed using a cut point of 2020 counts per minute (CPM) to delineate moderate physical activity, whereas Johannsen et al. (35) used a lower 574 CPM and reported that US adults age 60–74 yr engaged in 126 min of moderate-to-vigorous physical activity per day. In addition to discrepancies introduced by use of various intensity level cut points, analyzing accelerometer data in “accumulated bouts” results in drastically reduced population estimates of physical activity (13). Not surprisingly, given the low level of activity performed in bouts, one report suggests than fewer than 3% of US older adults are currently attaining sufficient physical activity to meet public health recommendations (67).
Very few intervention studies using objective monitoring have been conducted with older adults. Pruitt et al. (55) explored the ability of accelerometry to differentiate between intervention groups using a measure of physical activity relative to individual functional capacity. Individually determined accelerometer CPM thresholds were derived from performance on a 400-m walk test, and this relative measure of physical activity was able to distinguish that the physical activity intervention increased activity levels above those obtained by a nonexercise health education program. These early data are promising, both in their ability to detect intervention differences and in their overall practicality and application to design and implementation.
Recent studies have shown many different variables to be related to accelerometer-assessed physical activity in older adults. Studies consistently show demographic variables of age and sex to be associated with physical activity (35,67), as well as functional variables such as balance and gait speed (25,35) and psychosocial variables—notably quality of life and general mental health indices (50). Investigators also have begun to document associations between outdoor temperature and daily activity levels (5). Other literature examining biological factors such as bone mineral density (25) have not shown consistent relationships. Despite the growth of work published thus far pertinent to older adults, more data are needed to establish core intervening variables for this group of the population that has been shown to be the least active in the nation.
Adults with Functional Limitations
Studies have documented the feasibility and use of objective monitoring to assess physical activity levels in many different populations with functional limitations, including those with multiple sclerosis (29), osteoarthritis (19,45), peripheral arterial occlusive disease (24), heart transplant (17), and stroke (56). This is a rapidly expanding literature set, all using different activity monitors, different protocols for wearing the activity monitors, and different physical activity indicators.
In a recent study by Farr et al. (19), a hip-worn accelerometer was used to assess the activity levels of patients with early knee osteoarthritis. Using a cut point value averaged from the calibration literature to demarcate physical activity intensity levels, they found individuals engaged in approximately 24 min·d−1 of moderate physical activity, approximately 1 min·d−1 of vigorous physical activity, and that 70% of those sampled did not meet physical activity recommendations. Studies with hip-worn uniaxial and triaxial accelerometers have also shown that adults with kidney disease (36), multiple sclerosis (29), stroke (56), and Parkinson disease (29) are considerably less active compared with healthy adults. Other comparison studies have been conducted with wrist-worn accelerometers. For instance, Kop et al. (37) compared individuals with fibromyalgia and chronic fatigue syndrome with healthy controls. Similar to that previously reported by others in different populations (29), activity levels in these patient populations were upward of 50% lower than the healthy controls. Comparisons across populations and studies are cumbersome because of differences in monitors and site placements. It is not clear whether the same cut points generated on healthy adults should apply to populations with functional limitations.
Similar to other populations, intervention studies that use objective monitors to measure physical activity are lacking in adults with functional limitations, but evaluation of treatment therapies, rehabilitation effectiveness, and other potential intervention effects is possible with objective assessment. The application of objective monitoring in therapy and rehabilitation could be particularly helpful to document time to recover, diurnal fluctuations, and nonimpaired limb compensation. In a recent study by Reiterer et al. (58), a wrist-worn accelerometer was used to objectively examine the recovery process of 38 patients during stroke rehabilitation. Activity levels of both arms were assessed at four different time points after stroke—24–36 h, 5–7 d, 3 months, and 6 months. Wrist activity levels at the impaired side significantly increased during rehabilitation. A growing body of literature in individuals with lower extremity osteoarthritis is beginning to document the association between pacing as a therapy to reduce pain and its association with accumulated levels of physical activity as documented by wrist-worn actigraphy (45). Objective measurement of physical activity would aid evaluation of intervention components that are presently difficult to assess with self-report methods.
As in studies of the other populations, investigations have shown several demographic variables, such as sex, age, and disease/disability type to be associated with accelerometer-assessed physical activity (29,56). A recent investigation by Sumukadas et al. (65) also showed day length, mean maximal outdoor temperature, and sunshine duration to independently predict physical activity engagement. Investigators also have begun to document associations between physical activity and functional correlates in individuals with fibromyalgia and chronic fatigue syndrome (37). Psychosocial correlates, such as overall quality of life and self-efficacy, have been observed in individuals who have undergone heart transplant (17). Evangelista et al. (17) documented inverse associations between hypertension, hyperlipidemia, and obesity and overall physical activity levels in individuals having undergone heart transplant. More work is warranted in this area to corroborate findings from other literature sets and document the relationship of physical activity to aspects of metabolic, vascular, bone health, and other known health correlates, as well as the effect of broader macro-level influences, such as the built and social environment.
VALIDATION AND CALIBRATION STUDIES
The goal of validation and calibration studies is to establish the relationship between the signals generated from objective monitors and the actual physical activity performed. Articles in the current supplement discuss different types of validation and calibration principles and discuss the need to cross-validate established equations and prediction parameters in independent samples. For the purpose of the current article, we will primarily focus on the review of the literature pertaining to the criterion-referenced validity of accelerometers for assessing physical activity among children, older adults, and adults with functional limitations.
Many different calibration and validation studies have been conducted across the full age span of childhood and adolescence, using different criterion measures (indirect calorimetry, doubly labeled water, direct observation) and different accelerometer brands. Since the 2004 accelerometry meeting in Chapel Hill, NC (75), calibration studies with indirect calorimetry as the criterion measure have been published for the preschool population for the ActiGraph (LLC, Fort Walton Beach, FL) (51) and Actical (Mini Mitter Co., Inc., Bend, OR) (54). For elementary children, Evenson et al. (18) created ActiGraph and Actical physical activity count cut points, and other investigators, such as Hussey et al. (33), have provided calibration and/or validation studies for the RT3 (StayHealthy, Inc., Monrovia, CA; or other triaxial monitor). The study of Hussey et al. (33) also applied to middle school ages.
Although indirect calorimetry has primarily been used as a criterion measure, some studies since 2004 have used doubly labeled water (39) and direct observation (40) to evaluate the criterion validity of objective monitors in younger children. Studies comparing accelerometry to pedometry (7), parent proxy (4), and HR (31) have provided construct validation data. In addition, reliability (66) and stability (49) of accelerometry have been examined. Thus, many investigations have contributed to the body of knowledge regarding calibration, validation, and reliability of accelerometry as a measure of physical activity for children.
Criterion-referenced validity studies with older adults have been distinctly lacking. Although recent reviews exist on the use of accelerometry in older adults (44), the lack of data specific to this population means that conclusions have been drawn largely from the general adult literature. One laboratory-generated calibration study of older adults reported strong correlations between treadmill walking and measured oxygen consumption (r = 0.6) (11); other studies compared objective monitors to self-report measures, providing information about construct validity (50,55).
In 1992, Nichols et al. (48) examined the validity of the Caltrac (Muscle Dynamics, Torrence, CA) accelerometer in 28 young (mean age ∼26 yr) and 28 older (mean age ∼65 yr) men and women and determined that treadmill walking assessed by an accelerometer worn on the upper back was significantly correlated to measured net kilocalories obtained by indirect calorimetry. Reported relationships, although all significant, were weaker in the older group versus the younger group and were weaker for women than for men. In a follow-up study conducted by Fehling et al. (20), the criterion-referenced validity of two waist-mounted accelerometers (Caltrac and Tritrac [StayHealthy, Inc., Monrovia, CA]) to measure energy expenditure was assessed against indirect calorimetry in a group of 86 older adults (mean age ∼71 yr). Both accelerometers significantly misestimated energy expenditure.
Miller et al. (43) examined the criterion-referenced validity of the hip-worn ActiGraph to measure treadmill walking and running for adults age 20–29 yr, 40–49 yr, and 60–69 yr and found no differences in accelerometer output (CPM) across age groups for a given walking or running speed. These results highlight that age may not be an important factor in using objective monitoring with healthy older adults when evaluating accelerometer data output. No consensus exists on applicable cut points to delineate physical activity intensity for older adults.
Other criterion-referenced validation studies exist among older adults that have used doubly labeled water. Starling et al. evaluated the Caltrac in a sample of women up to age 84 yr (64) and showed the Caltrac significantly underestimated PAEE under free-living conditions. Meijer et al. (42) evaluated a triaxial low-back-worn accelerometer and found strong correlations with doubly labeled water (r = 0.78).
Adults with Functional Limitations
Studies using indirect calorimetry as a criterion-referenced calibration standard against objective monitoring are fairly scarce in this population. Ekelund et al. (15) evaluated an ActiGraph worn on the lower back in a group of patients with coronary artery disease during level walking. Accelerometer counts were significantly correlated with speed (r = 0.92), measured oxygen consumption (r = 0.87), and energy expenditure (r = 0.85). Equations to derive energy expenditure, using body weight and accelerometer counts, explained 75% of the variance in measured energy expenditure, and mean differences (−0.2 kcal·min−1) between measured and predicted energy expenditure values were not significant. The individual limits of agreement were greater, a finding generally consistent with the overall accelerometry literature in any population, which suggests that estimates are better for group or pooled data than they are for individuals. A follow-up study by Focht et al. (22) also examined the validity of a hip-worn ActiGraph among older adults with a documented chronic disease. Results confirmed those of previous studies, reporting strong significant correlations between accelerometer data and indirect calorimetry. Free-living investigations comparing objective measures to doubly labeled water also have found strong correlations between energy expenditure and data obtained from the Caltrac in older adults with peripheral arterial occlusive disease (r = 0.8) (24) and data obtained from a triaxial accelerometer in adults with chronic low back pain (r = 0.7) (73).
Comparison of objective monitors with direct observation is another approach that has been used to establish criterion validity, including visual analysis of simultaneous video recordings (6) and motion capture systems in a laboratory environment (27). This latter approach is particularly useful if raw movement or motion is the objective monitoring outcome desired. Gironda et al. (27) compared motion and acceleration scores from the Actiwatch placed at three sites (wrist, waist, and ankle) with an optical three-dimensional motion capture system for typical physical therapy exercises and a walking trial. Overall, all accelerometer site placements were significantly associated with the motion capture system, but different placements performed better depending on the task undertaken.
Studies such as these continue to contribute to the scientific knowledge in this area. Other studies have reported on aspects of reliability adding much needed content to this growing literature set (27).
Strengths and Weaknesses of Validation and Calibration Studies
Existing validation and calibration studies have strengths and weaknesses that are important to consider. A large validation and calibration literature base exists, but this literature base is considerably more developed for children than in older adults or adults with functional limitations. A strength of the existing literature is that many studies have used rigorous designs with appropriate criterion-reference standard measures. Although the validation and calibration literature is expanding, it is questionable whether the creation of additional single-regression prediction equations can assist with moving the field forward. Having too many equations creates a “cut point conundrum” and leaves researchers and practitioners wondering which set is “correct,” particularly when they yield vastly different answers when classifying physical activity intensity. Unfortunately, this issue has no one good solution. Predicting physical activity behavior from single-regression equations may be too simple an approach to use in examining complex physical activity behaviors. Walking calibration studies conducted in the laboratory work well for walking and locomotion but may misclassify other types of activity (12).
Although it is commonly recognized that validation and calibration studies should be population specific, it is not known how specific they truly need to be. The answer depends on the outcome variable of interest. For instance, Miller et al. (43) demonstrated that age is not a classifying variable for calibration because they saw no differences in raw accelerometer output across multiple walking speeds spanning different ages in healthy adults. However, the role of age is not clear when attempting to classify physical activity levels or derive energy expenditure and the issue is considerably more complex when considering individuals with functional limitations and children and adolescents who are undergoing growth and maturation. Knowing that the mechanical efficiency of an activity can be markedly changed across different populations, using prediction equations generated on one population and applying to another is likely to produce erroneous estimates of activity.
Consensus on site placement also is needed. Although there seems to be a good agreement within the child and adult literature that waist-mounted devices are the site placement of choice, this does not extend to all populations. Specific site placement studies in individuals with functional limitations are lacking, thereby limiting conclusions that can be drawn. A good example of this pertains to recent studies published in Arthritis & Rheumatism. In 2008, two articles were published on activity levels in individuals with osteoarthritis (19,45). Both studies used different activity monitors, used different site placements (one on the wrist, the other on the waist), and reported different outcomes (raw movement and time spent in physical activity intensity levels). Results of these studies are not comparable, limiting the ability to pool studies to draw conclusions about activity levels in this population.
Along similar lines, standard time sampling intervals (epoch lengths) are not uniformly applied. Researchers believe that young children are active in shorter, potentially more intense bouts of activity than are older children and adolescents. Part of the problem is that the effects of growth and maturation have not previously been accounted for in existing studies. Maturation is subject to timing (when it occurs) and tempo (rate at which it occurs) (3). Thus, variance in variables such as limb segment length, muscle activity, and neural development could all play significant roles in locomotor patterns at different ages.
In light of the strengths and weaknesses of the use of objective monitoring in youth and underserved populations, we recommend several best practices to move this field of research forward.
- Always clearly define the outcome of interest when determining how to use objective measures of activity.
- Discontinue the creation of single linear regression calibration equations and count cut points. The addition and repetition of studies in this area presents confusion.
- Establish and use standard methods for obtaining, cleaning, and analyzing data for all activity monitors. Clearly outline this in all disseminated work.
- Use shorter time sampling intervals (epoch length). To date, data obtained by 1-s epochs may not be informative on their own but can be summed to create 10-, 15-, or 30-s increments until the field develops a better understanding of data obtained from short sampling intervals.
Here we suggest avenues that could be pursued to advance future research in physical activity monitoring among these and other understudied populations.
- Adopt more complex mathematical modeling strategies for detecting patterns of movement, such as hidden Markov models and artificial neural networking. In addition to improving prediction of outcomes such as energy expenditure, these models also have the capability to identify specific activities. The type of activity may be a more appropriate outcome for some studies, rather than the amount of energy expended.
- Gain understanding about how transitions between activities affect accelerometer output and the agreement between device output and criterion measures of energy expenditure.
- Direct efforts to study children and adolescents with mental or physical disabilities, pregnant women, toddlers, obese individuals, and adults with chronic medical conditions and how differences in these population subgroups affect monitor output.
- Explore the effects of growth and maturation on monitor output.
- Explore classifying groups where appropriate. For instance, gait speed could help classify many different functional limitations. Such an approach, or better methods for classifying individuals, could remove the need for multiple population-specific algorithms.
- Account for relative and absolute intensity differences in studies. Given the differences in physical fitness level in the population, a difference typically highlighted with aging or functional limitations/disability, this effort becomes paramount.
- Encourage research groups to work in unison rather than in isolation to foster progress and optimize the use of objective monitoring in diverse populations.
This work was partially supported by a Career Development Award (Strath) from the National Institute on Aging (K01AG025962).
The authors would like to thank the organizing committee and sponsors for this conference, and Dr. Heather Bowles and Anne Rodgers for their helpful edits and suggestions while preparing this article.
The authors have no financial conflict of interest with any of the monitor manufacturers and have received no research funding from these companies.
The results of the present study do not constitute endorsement of the products described in this article by the authors or the American College of Sports Medicine.
1. Alhassan S, Robinson T. Objectively measured physical activity and cardiovascular disease risk factors in African American girls. Ethn Dis. 2008; 18 (14): 421–6.
2. Baker B, Birch L, Trost S, Davison K. Advanced pubertal status at age 11 and lower physical activity in adolescent girls. J Pediatr. 2007; 151 (5): 488–93.
3. Baxter-Jones ADG, Eisenmann JC, Shearer LB. Controlling for maturation in pediatric exercise
science. Ped Exerc Sci. 2005; 17 (1): 18–30.
4. Bender J, Brownson R, Elliott M, Haire-Joshu D. Children
’s physical activity: using accelerometers to validate a parent proxy record. Med Sci Sports Exerc. 2005; 37 (8): 1409–13.
5. Brandon CA, Gill DP, Speechley M, Gilliland J, Jones GR. Physical activity levels of older community-dwelling adults are influenced by summer weather variables. Appl Physiol Nutr Metab. 2009; 34 (2): 182–90.
6. Bussmann JB, van de Laar YM, Neeleman MP, Stam HJ. Ambulatory accelerometry to quantify motor behaviour in patients after failed back surgery: a validation study. Pain. 1998; 74 (2–3): 153–61.
7. Cardon G, De Bourdeaudhuij I. Comparison of pedometer and accelerometer
measures of physical activity in preschool children
. Pediatr Exerc Sci. 2007; 19 (2): 204–14.
8. Cardon G, De Bourdeaudhuij I. Are preschool children
active enough? Objectively measured physical activity levels. Res Q Exerc Sport. 2008; 79 (3): 326–32.
9. Cardon G, Labarque V, Smits D, De Bourdeaudhuij I. Promoting physical activity at the pre-school playground: the effects of providing markings and play equipment. Prev Med. 2009; 48 (3): 335–40.
10. Casperson CJ, Powell KE, Christenson GM. Physical activity, exercise
, and physical fitness: definitions and distinctions for health-related research. Public Health Report. 1985; 100 (2): 126–31.
11. Copeland J, Esliger D. Accelerometer
assessment of physical activity in active, healthy older adults
. J Aging Phys Act. 2009; 17 (1): 17–30.
12. Crouter SE, Churilla JR, Bassett DR Jr. Estimating energy expenditure using accelerometers. Eur J Appl Physiol. 2006; 98 (6): 601–12.
13. Davis MG, Fox KR. Physical activity patterns assessed by accelerometry in older people. Eur J Appl Physiol. 2007; 100 (5): 581–9.
14. Dencker M, Andersen L. Health-related aspects of objectively measured daily physical activity in children
. Clin Physiol Funct Imaging. 2008; 28 (3): 133–44.
15. Ekelund U, Tingstrom P, Kamwendo K, et al.. The validity of the Computer Science and Applications activity monitor for use in coronary artery disease patients during level walking. Clin Physiol Funct Imaging. 2002; 22 (4): 248–53.
16. Epstein L, Roemmich J, Paluch R, Raynor H. Physical activity as a substitute for sedentary behavior
in youth. Ann Behav Med. 2005; 293: 200–9.
17. Evangelista LS, Dracup K, Doering L, Moser DK, Kobashigawa J. Physical activity patterns in heart transplant women. J Cardiovasc Nurs. 2005; 20 (5): 334–9.
18. Evenson K, Catellier D, Gill K, Ondrak K, McMurray R. Calibration of two objective measures of physical activity for children
. J Sports Sci. 2008; 26 (14): 1557–65.
19. Farr JN, Going SB, Lohman TG, et al.. Physical activity levels in patients with early knee osteoarthritis measured by accelerometry. Arthritis Rheum. 2008; 59 (9): 1229–36.
20. Fehling PC, Smith DL, Warner SE, Dalsky GP. Comparison of accelerometers with oxygen consumption in older adults
. Med Sci Sports Exerc. 1999; 31 (1): 171–5.
21. Fitzgerald M, Tanaka H, Tran Z, Seals D. Age-related declines in maximal aerobic capacity in regularly exercising vs. sedentary women: a meta-analysis. J Appl Physiol. 1997; 83 (1): 160–5.
22. Focht BC, Sanders WM, Brubaker PH, Rejeski WJ. Initial validation of the CSA activity monitor during rehabilitative exercise
among older adults
with chronic disease. J Aging Phys Act. 2003; 11 (3): 293–304.
23. Freedson P, Evenson S. Familial aggregation in physical activity. Res Q Exerc Sport. 1991; 62 (4): 384–9.
24. Gardner AW, Poehlman ET. Assessment of free-living daily physical activity in older claudicants: validation against the doubly labeled water technique. J Gerontol A Biol Sci Med Sci. 1998; 53 (4): M275–80.
25. Gerdhem P, Dencker M, Ringsberg K, Akesson K. Accelerometer
-measured daily physical activity among octogenerians: results and associations to other indices of physical performance and bone density. Eur J Appl Physiol. 2008; 102 (2): 173–80.
26. Gidlow C, Cochrane T, Davey R, Smith H. In-school and out-of-school physical activity in primary and secondary school children
. J Sports Sci. 2008; 26 (13): 1411–9.
27. Gironda RJ, Lloyd J, Clark ME, Walker RL. Preliminary evaluation of reliability and criterion validity of Actiwatch-Score. J Rehabil Res Dev. 2007; 44 (2): 223–30.
28. Goldfield G, Mallory R, Prud’homme D, Adamo K. Gender differences in response to a physical activity intervention in overweight and obese children
. J Phys Act Health. 2008; 5 (4): 592–606.
29. Hale LA, Pal J, Becker I. Measuring free-living physical activity in adults with and without neurologic dysfunction with a triaxial accelerometer
. Arch Phys Med Rehabil. 2008; 89 (9): 1765–71.
30. Hannon J, Brown B. Increasing preschoolers’ physical activity intensities: an activity-friendly preschool playground intervention. Prev Med. 2008; 46 (6): 532–6.
31. Harro M. Validation of a questionnaire to assess physical activity of children
ages 4–8 years. Res Q Exerc Sport. 1997; 68 (4): 259–68.
32. Hasselstrom H, Karlsson M, Hansen S, Gronfeldt V, Froberg K, Andersen L. A 3-year physical activity intervention program increases the gain in bone mineral and bone width in prepubertal girls but not boys: the Prospective Copenhagen School Child Interventions Study (CoSCIS). Calcif Tissue Int. 2008; 83 (4): 243–50.
33. Hussey J, Bennett K, Dwyer J, Langford S, Bell C, Gormley J. Validation of the RT3 in the measurement of physical activity in children
. J Sci Med Sport. 2009; 12 (1): 130–3.
34. Jago R, Baranowski T, Baranowski J, et al.. Fit for Life Boy Scout badge: outcome evaluation of a troop and Internet intervention. Prev Med. 2006; 42 (3): 181–7.
35. Johannsen DL, DeLany JP, Frisard MI, et al.. Physical activity in aging: comparison among young, aged, and nonagenarian individuals. J Appl Physiol. 2008; 105 (2): 495–501.
36. Johansen KL, Chertow GM, Ng AV, et al.. Physical activity levels in patients on hemodialysis and healthy sedentary controls. Kidney Int. 2000; 57 (6): 2564–70.
37. Kop WJ, Lyden A, Berlin AA, et al.. Ambulatory monitoring of physical activity and symptoms in fibromyalgia and chronic fatigue syndrome. Arthritis Rheum. 2005; 52 (1): 296–303.
38. Lanningham-Foster L, Foster R, McCrady S, et al.. Changing the school environment to increase physical activity in children
. Obesity (Silver Spring). 2008; 16 (8): 1849–53.
39. Lopez-Alarcon M, Merrifield J, Fields D, et al.. Ability of the Actiwatch accelerometer
to predict free-living energy expenditure in young children
. Obes Res. 2004; 12 (11): 1859–65.
40. McClain JJ, Abraham TL, Brusseau TA Jr, Tudor-Locke C. Epoch length and accelerometer
outputs in children
: comparison to direct observation. Med Sci Sports Exerc. 2008; 40 (12): 2080–7.
41. McLure S, Summerbell C, Reilly J. Objectively measured habitual physical activity in a highly obesogenic environment. Child Care Health Dev. 2009; 35 (3): 369–75.
42. Meijer EP, Goris AH, Wouters L, Westerterp KR. Physical inactivity as a determinant of the physical activity level in the elderly
. Int J Obes Relat Metab Disord. 2001; 25 (7): 935–9.
43. Miller N, Strath S, Swartz A. Estimating absolute and relative physical activity intensity across age via accelerometry in adults. J Aging Phys Act. 2010; 18 (2): 158–70.
44. Murphy SL. Review of physical activity measurement using accelerometer
in older adults
; Considerations for research design and conduct. Prev Med. 2009; 48 (2): 108–14.
45. Murphy SL, Smith DM, Clauw DJ, Alexander NB. The impact of momentary pain and fatigue on physical activity in women with osteoarthritis. Arthritis Rheum. 2008; 59 (6): 849–56.
46. Nader P, Bradley R, Houts R, McRitchie S, O’Brien M. Moderate-to-vigorous physical activity from ages 9 to 15 years. JAMA. 2008; 300 (3): 295–305.
47. Ni Mhurchu C, Maddison R, Jiang Y, Jull A, Prapavessis H, Rodgers A. Couch potatoes to jumping beans: a pilot study of the effect of active video games on physical activity in children
. Int J Behav Nutr Phys Act 5. 2008; 5: 8.
48. Nichols JF, Patterson P, Early T. A validation of a physical activity monitor for young and older adults
. Can J Sport Sci. 1992; 17 (4): 299–303.
49. Nyberg G, Ekelund U, Marcus C. Physical activity in children
measured by accelerometry: stability over time. Scand J Med Sci Sports. 2009; 19 (1): 30–5.
50. Parker SJ, Strath SJ, Swartz AM. Physical activity measurement in older adults
: relationships with mental health. J Aging Phys Act. 2008; 16 (4): 369–80.
51. Pate R, Almeida M, McIver K, Pfeiffer K, Dowda M. Validation and calibration of an accelerometer
in preschool children
. Obesity (Silver Spring). 2006; 14 (11): 2000–6.
52. Patrick K, Calfas KJ, Norman GJ, et al.. Randomized controlled trial of a primary care and home-based intervention for physical activity and nutrition behaviors: PACE+ for adolescents. Arch Pediatr Adolesc Med. 2006; 160 (2): 128–36.
53. Pfeiffer KA, Dowda M, McIver KL, Pate RR. Factors related to objectively measured physical activity in preschool children
. Pediatr Exerc Sci. 2009; 21 (2): 196–208.
54. Pfeiffer KA, McIver KL, Dowda M, Almeida MJ, Pate RR. Validation and calibration of the Actical accelerometer
in preschool children
. Med Sci Sports Exerc. 2006; 38 (1): 152–7.
55. Pruitt LA, Glynn NW, King AC, et al.. Use of accelerometry to measure physical activity in older adults
at risk for mobility disability
. J Aging Phys Act. 2008; 16 (4): 416–34.
56. Rand D, Eng JJ, Tang PF, Jeng JS, Hung C. How active are people with stroke?: use of accelerometers to assess physical activity. Stroke. 2009; 40 (1): 163–8.
57. Reilly JJ. Physical activity, sedentary behaviour and energy balance in the preschool child: opportunities for early obesity prevention. Proc Nutr Soc. 2008; 67 (3): 317–25.
58. Reiterer V, Sauter C, Klösch G, Lalouschek W, Zeitlhofer J. Actigraphy—a useful tool for motor activity monitoring in stroke patients. Eur Neurol. 2008; 60 (6): 285–91.
59. Roemmich J, Barkley J, Lobarinas C, Foster J, White T, Epstein L. Association of liking and reinforcing value with children
’s physical activity. Physiol Behav. 2008; 93 (4–5): 1011–18.
60. Rowlands A, Pilgrim E, Eston R. Seasonal changes in children
’s physical activity: an examination of group changes, intra-individual variability and consistency in activity pattern across season. Ann Hum Biol. 2009; 36 (4): 363–78.
61. Scott M, Evenson K, Cohen D, Cox C. Comparing perceived and objectively measured access to recreational facilities as predictors of physical activity in adolescent girls. J Urban Health. 2007; 84 (3): 346–59.
62. Shoup J, Gattshall M, Dandamudi P, Estabrooks P. Physical activity, quality of life, and weight status in overweight children
. Qual Life Res. 2008; 17 (3): 407–12.
63. Specker B, Binkley T, Fahrenwald N. Increased periosteal circumference remains present 12 months after an exercise
intervention in preschool children
. Bone. 2004; 35 (6): 1383–8.
64. Starling RD, Matthews DE, Ades PA, Poehlman ET. Assessment of physical activity in older individuals; a doubly labeled water study. J Appl Physiol. 1999; 86 (6): 2090–6.
65. Sumukadas D, Witham M, Struthers A, McMurdo M. Day length and weather conditions profoundly affect physical activity levels in older functionally impaired people. J Epidemiol Community Health. 2009; 63 (4): 305–9.
66. Toschke JA, von Kries R, Rosenfeld E, Toschke AM. Reliability of physical activity measures from accelerometry among preschoolers in free-living conditions. Clin Nutr. 2007; 26 (4): 416–20.
67. Troiano RP, Berrigan D, Dodd KW, Mâsse LC, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer
. Med Sci Sports Exerc. 2008; 40 (1): 181–8.
68. Trost SG, Fees B, Dzewaltowski D. Feasibility and efficacy of a “move and learn” physical activity curriculum in preschool children
. J Phys Act Health. 2008; 5 (1): 88–103.
69. Trost SG, Kerr LM, Ward DS, Pate RR. Physical activity and determinants of physical activity in obese and non-obese children
. Int J Obes Relat Metab Disord. 2001; 25 (6): 822–9.
70. Trost SG, Sirard JR, Dowda M, Pfeiffer KA, Pate RR. Physical activity in overweight and nonoverweight preschool children
. Int J Obes (Lond). 2003; 27 (7): 834–9.
71. van Sluijs EM, Fearne VA, Mattocks C, Riddoch C, Griffin SJ, Ness A. The contribution of active travel to children
’s physical activity levels: cross-sectional results from the ALSPAC study. Prev Med. 2009; 48 (6): 519–24.
72. van Sluijs EM, Skidmore PM, Mwanza K, et al.. Physical activity and dietary behaviour in a population-based sample of British 10-year old children
: the SPEEDY study (Sport, Physical activity and Eating behaviour: Environmental Determinants in Young people). BMC Public Health. 2008; 8: 388.
73. Verbunt JA, Westerterp KR, van der Heijden GJ, Seelen HA, Vlaeyen JW, Knottnerus JA. Physical activity in daily life in patients with chronic low back pain. Arch Phys Med Rehabil. 2001; 82 (6): 726–30.
74. Verstraete SJ, Cardon GM, De Clercq DL, De Bourdeaudhuij IM. A comprehensive physical activity promotion programme at elementary school: the effects on physical activity, physical fitness and psychosocial correlates of physical activity. Public Health Nutr. 2007; 10 (5): 477–84.
75. Ward D. Objective monitoring of physical activity: closing the gaps in the science of accelerometry. Med Sci Sports Exerc. 2005; 37 (11 suppl): S487–588.
76. Wosje KS, Khoury PR, Claytor RP, Copeland KA, Kalkwarf HJ, Daniels SR. Adiposity and TV viewing are related to less bone accrual in young children
. J Pediatr. 2009; 154 (1): 79–85.