The importance of physical activity assessment
The American College of Sports Medicine and the American Heart Association recently issued updated health-based physical activity recommendations (1).  It is indisputable that an association between physical activity and health exists and the measurement of physical activity behavior is central to defining this relationship and clarifying the nature of the dose-response relationship.  Additionally, physical activity assessment is useful for evaluating the effectiveness of programs and interventions designed to increase physical activity as well as for monitoring population trends and providing individuals with feedback on their current level of physical activity.

Measuring physical activity behavior, where are we and what are some of the challenges?
While the accurate assessment of physical activity is possible in the laboratory, the measurement of physical activity in a free-living environment over a period of several days poses challenges and it is in that context that the measurement of physical activity is essential. 

Subjective methods such as interview or questionnaire based recall have widespread use for estimating physical activity in a free-living environment.  From small studies to large epidemiological studies and national surveys such as the National Health Interview Survey and the National Health and Nutrition Examination Survey (NHANES), physical activity assessment instruments are used to gather information about self-reported physical activity behavior.  For example, the NHANES includes questions about physical activities over the past 30 days.  On an international level, the International Physical Activity Questionnaires provide a means of obtaining comparable estimates of physical activity behavior across nations.  The validity of these instruments is limited however, as self-report methods often suffer from errors associated with respondent recall, cognitive development (in young children) and social desirability bias.  These errors can be reduced or eliminated by the use of objective measures, but these methods are subject to other limitations.

Objective methods of assessment include direct and indirect calorimetry, doubly-labeled water, accelerometers, heart rate monitors, pedometers and more recently multi-sensor devices, which measure two or more variables (e.g., heart rate and acceleration).  Accelerometer-based methods of physical activity assessment are generally preferred to other established objective methods of assessment in a free-living environment as accelerometers provide information about the pattern, amount and intensity of activity and unlike heart rate, they are not influenced by factors other than activity such as stress or environmental conditions.  Moreover, accelerometers are reasonably priced and can be used over an extended period with minimal subject burden and reactivity.  In fact, accelerometers were used in the NHANES 2003-2006 data collection cycle to obtain population-based objective data on physical activity behavior. 

In spite of their extensive use, accelerometer-based methods of physical activity assessment have inherent shortcomings.  Accelerometers are generally worn on the hip, which limits their ability to detect upper body movements and during walking they can not detect grades or whether an individual is carrying a load.  Additionally, in the original and widely used method of data analysis, one regression equation is used to estimate energy expenditure from activity monitor output.  The prediction of energy expenditure for different kinds of activities using an equation developed on specific activities such as walking and running often results in the misclassification of intensity level.      

Comparisons among studies that use different activity monitors can also be challenging or impossible due to the use of proprietary software and lack of disclosure about monitor specifications.  Additionally, inconsistent practices among researchers can thwart comparisons among studies and the end points selected as boundaries between intensities can influence the interpretation of the data.  

Opportunities for improving the assessment of physical activity behavior
To address some of the challenges associated with current methods, best practices and research recommendations on the use of accelerometers to measure physical activity have been published (2).  The areas addressed included monitor selection and use protocols, monitor calibration, analysis of accelerometer data and the integration of accelerometry with other data sources.  As self-report methods have widespread use, the standardization of procedures for these methods is important and should not be overlooked. 

Attempts to improve the original regression approach have resulted in more equations but no single equation is accurate across all activities.  The use of a two-equation regression model that utilizes variability in counts to determine whether to employ the locomotion or lifestyle equation should improve the accuracy of energy expenditure estimates (3).

While the regression model approach has been used extensively to characterize and quantify physical activity, the emergence of new methods made feasible by technological advances offer exciting possibilities to advance the field of physical activity assessment.  One of the newer approaches is the integration of accelerometers with other data sources.  The addition of global positioning systems (GPS) provides information about the context of physical activity and the integration of heart rate and accelerometry has been shown to improve the prediction of energy expenditure (4, 5). 
   
The accuracy of physical activity estimates may also be enhanced by employing a classification approach that focuses on determining the mode of physical activity.  Intelligent data processing methods similar to those used for speech recognition could be used to determine mode of activity (6).  The use of statistical techniques to model and correct for measurement error is also under investigation and could lead to improved estimates of physical activity at the population level using self-report instruments.

An important consideration in the development of new physical activity assessment methods is the promotion of open-source technology.  The use of “off-the-shelf” software and disclosure of instrument specifications would facilitate comparisons and could promote rapid growth of an extensive database. 

Self-monitoring of physical activity behavior
An additional area where assessing physical activity can be is useful is self-monitoring.  Previously the interest in self-monitoring was primarily limited to the athlete who was interested in quantifying activity during exercise training sessions.  Today many individuals use self-monitoring to establish activity goals and evaluate success in attaining those goals.  The advent of the relatively inexpensive pedometer and the slogan “10,000 steps/day” have played a major role in promoting self-monitoring and motivating individuals to increase physical activity.

Summary
Physical activity behavior is assessed with many different types of instruments for a variety of reasons. Assessment of whether public health recommendations for physical activity are being met and to clarify how the amount and intensity of physical activity influence health outcomes have driven efforts to improve upon existing objective methods for assessing physical activity behavior in a free-living environment at the population level.  Monitoring physical activity level also has broad appeal in the health and wellness arena for individuals who use devices such as pedometers to provide immediate feedback, which can be used to establish and evaluate personal physical activity goals.

References
1.  Haskell WL, Lee IM, Pate RR, Powell KE, Blair SN, Franklin BA, Macera CA, Heath GW, Thompson PD, and Bauman A. Physical Activity and Public Health. Updated Recommendation for Adults From the American College of Sports Medicine and the American Heart Association. Circulation, 2007.
2.  Ward DS, Evenson KR, Vaughn A, Rodgers AB, and Troiano RP. Accelerometer use in physical activity: best practices and research recommendations. Med Sci Sports Exerc 37: S582-588, 2005.
3.  Crouter, SE, Clowers, KG, Bassett, DR Jr.  A novel method for using accelerometer data to predict energy expenditure.  J. Appl. Physiol. 100: 324-331, 2006
4.  Brage S, Brage N, Franks PW, Ekelund U, and Wareham NJ. Reliability and validity of the combined heart rate and movement sensor Actiheart. Eur J Clin Nutr 59: 561-570, 2005.
5.  Crouter SE, Churilla JR, and Bassett DR, Jr. Accuracy of the Actiheart for the assessment of energy expenditure in adults. Eur J Clin Nutr (epub), April 18: 1-8, 2007.
6.  Pober DM, Staudenmayer J, Raphael C, and Freedson PS. Development of novel techniques to classify physical activity mode using accelerometers. Med Sci Sports Exerc 38: 1626-1634, 2006.