• Fitness & Health
• Sport & Exercise Science
• Physical Education
• Strength & Conditioning
• Sports Medicine
• Sport Management
• Dance

Electronic devices used in physical activity surveillance and public health research

This is an excerpt from Foundations of Physical Activity and Public Health 3rd Edition With HKPropel Access by Harold W. Kohl III,Tinker D. Murray & Deborah Salvo.

Early studies in the field have relied on self-report instruments to assess physical activity. We will learn more about self-report measures further along in this chapter. Advances in technology, however, have resulted in the development of devices that can objectively measure movement.

Accelerometers

Accelerometers are small piezoelectric devices that measure the magnitude and direction of acceleration. Accelerometers of many varieties are used for applications as varied as oil and gas drilling, navigation, and transportation. For human movement and physical activity, accelerometers have proven very useful in determining total physical activity and in estimating energy expenditure. However, although accelerometers have advanced our ability to measure physical activity, they provide no information regarding the type of activity being performed. For example, a period of brisk walking is indistinguishable from a tennis match to an accelerometer.

Accelerometers measure movement and physical activity by measuring acceleration forces in one, two, or three planes (or axes) as a result of a change in the velocity of the body. Semiconductors translate the processing of movement to acceleration. The following simple equation shows the mathematical relationship between acceleration (a), change in velocity (∆v), and change in time (∆t). As velocity increases for a given time, so does acceleration. When the amount of time decreases (at a constant velocity), acceleration also increases.

a = ∆v / ∆t

Using this relationship, an accelerometer worn against a person’s body measures changes in velocity and time. Accelerometers record these measures as counts—numerical values given to the acceleration recorded at a given point in time. Counts alone are not very meaningful, because the scale is an arbitrary one that varies depending on the make and model of the accelerometer. Modern-day research-grade accelerometers record between 30 and 120 data points (counts) per second. The number of counts collected per second by the accelerometer is referred to as the sampling rate. Because such sampling rates produce a very large amount of information, researchers usually export the data by aggregating them to the 1-second, 10-second, 30-second, or 60-second level. This aggregated time segment is known as an epoch. For instance, if an accelerometer uses a sampling rate of 30 counts per second, and we use a 60-second epoch to summarize the data, each data point (cell) in our accelerometer dataset would represent the average value of 1,800 counts (30 counts per second × 60 seconds in the epoch). Counts per epoch can then be converted into estimates of energy expenditure.

Several researchers have conducted laboratory-based studies employing different models and brands of research-grade accelerometers to develop specific energy expenditure formulas. Most accelerometer energy expenditure formulas are based on counts per minutes; that is, they assume a 60-second epoch. These formulas are useful because they provide cut points for different intensities of physical activity. These cut points tell us the minimum number of counts per minute that need to be recorded by an accelerometer to indicate light-, moderate-, or vigorous-intensity physical activity. Something critical to keep in mind when using accelerometers is that physical activity intensity cut points are age- and brand-specific. One must become familiar with all the models, brands, and energy expenditure formulas (cut points) available before collecting and processing accelerometer data in a given study.

Accelerometers are very useful for measuring physical activity because they take human recall out of the equation. Participants can attach an accelerometer unit to a waistband and forget it throughout the course of the day—they are free to go about their usual activities of daily living without having to remember anything. In addition to estimating time spent at specific intensities of physical activity, accelerometers can also determine time spent in continuous periods of physical activity above a certain time threshold (e.g., 10-minute bouts). Continued technological advances allow for multiday data storage (sometimes more than 20 days) for long-term behavior monitoring.

Accelerometers can be relatively expensive (the better ones exceeded $325 USD in 2018) and may not be accurate for all kinds of activities. Most accelerometers are not waterproof; thus, they cannot be used to measure physical activity in the pool, ocean, or anywhere it is wet. Most accelerometers accurately record physical activity levels when worn on the waist, making them an inadequate instrument for measuring bicycling activity. Because downloading, processing, and analyzing accelerometer data is a complex task requiring high-level training, accelerometers are best suited for rigorous research studies.

Geographic Positioning Systems (GPS) Monitors

Investigators use wearable GPS monitors to measure and characterize physical activity spatial patterns (figure 4.3). Research-grade GPS devices collect accurate geographic location data every 5 to 60 seconds. These instances, referred to as waypoints, represent each geolocation where a participant is detected via satellite systems every x seconds. In addition to recording locational characteristics (latitude and longitude), each waypoint contains additional critical information: a date stamp, a time stamp, and a measure of altitude. By using these variables together, we can derive the velocity at which a person is traveling in space (from waypoint A to waypoint B to waypoint C and so on). Investigators employ algorithms to categorize trips as car-based, bicycle-based, or walking based on velocity data. This travel mode detection and the total minutes per day spent in active travel are the typical physical activity indicators that can be obtained from GPS monitors. However, a small number of researcher groups around the world simultaneously collect GPS and accelerometer data. Using complex time-matching algorithms, the combined use of GPS and accelerometers allows investigators to map moderate- and vigorous-intensity physical activity in space (e.g., to identify the places within a city where most physical activity occurs), as well as to detect and assess duration and intensity of bicycling behaviors (which are difficult to assess with accelerometers alone).

Like accelerometers, GPS monitors remove human recall from the equation, making them a powerful assessment tool. Also like accelerometers, GPS devices can be attached to a waistband; are relatively expensive (high-quality models exceeded $100 USD per unit in 2024); and require highly trained individuals for data download, processing, and analysis. Because of the large amount of spatial data being collected by GPS monitors, battery life continues to be a significant drawback for these instruments. Most research-grade GPS devices have to be recharged by the participant nightly or every few days to ensure sufficient battery life for another full day of data collection. It is expected that with time these technologies will improve, allowing for longer battery lives and continuous days or weeks of data collection. One limitation of GPS monitors, when used alone (as opposed to simultaneously with accelerometers), is that they only capture physical activity with spatial displacements (moving from point A to point B in space). GPS monitors alone would not be useful to assess physical activity occurring in a static point in space (e.g., treadmill running). Ethical concerns related to participants’ privacy are also a consideration when dealing with GPS monitors for physical activity assessment. Some people are hesitant to participate in studies that will collect continuous data on their whereabouts. Rigorous protocols that protect participants’ right to privacy, and that safeguard their geolocational data, must be followed and further enhanced as these instruments become increasingly popular for physical activity research.

Pedometers

Pedometers (or step counters) are another kind of electronic monitoring device that can be used to take the recall bias out of physical activity assessment. They are usually most useful for measuring walking, running or jogging, or any other type of physical activity that involves the lower body. Many kinds of pedometers exist, using several types of mechanisms, although they all fundamentally measure total steps (see figure 4.4). Some rely on a spring or a spring lever to record the movement, others use a strain gauge, and still others use a magnetic switch.

A key strength of pedometers is that they are fairly inexpensive and thus can be used by many people (good pedometers could be purchased for $12 to $40 USD in 2024). They are fairly simple and straightforward to use and seem to accurately measure the number of steps taken. Studies have shown a range of accuracies among various brands, with the less expensive pedometers generally being less accurate than more expensive ones. Pedometers help people become more active by reporting the total number of steps they have taken (e.g., toward a preset goal), thereby reminding them to be active.

When collecting pedometer-based physical activity data, the least common denominator is steps taken during the observation period. With additional data collection, such as a diary or questionnaire, or with better pedometer models, one can obtain information on steps taken per day or steps within different periods of the day, weekdays versus weekends, and so on. A substantial drawback to using pedometers for measuring physical activity is that they do not directly measure velocity or time and thus cannot estimate acceleration. This means that time spent in different intensities of physical activity cannot be derived simply by having information on total steps taken over a period of time. A pedometer weights each recorded step with the same value, whether that step was taken while strolling slowly, during a sprint run, or during a soccer game. Clearly, the dose of physical activity differs in these various situations, but a pedometer would not be able to determine that without additional data collection.

Consumer-Based Wearable Tracking Devices and Apps

There has been a rapid surge in the availability of smartphone applications (apps) for health-monitoring purposes. This includes apps to monitor one’s physical activity levels (figure 4.5). While many apps simply offer monitoring services, others also include interactive features such as competition with other app users, or goal-setting, to help the user achieve a better health status. In addition to smartphone apps, many people also own wearable wristband monitors (e.g., Fitbit) that link to one’s smartphone via Wi-Fi or Bluetooth and that track physical activity patterns throughout the day. These apps and tracking devices rely on technologies similar to those previously described, including built-in pedometers and GPS receivers. However, the technology in these consumer-based products is usually not as advanced as those found in research-grade devices, and the algorithms used to derive acceleration and velocity, and thus energy expenditure, are mostly unknown because these are proprietary products.

While some research has emerged in an attempt to determine if these new technologies are good at accurately measuring physical activity in individuals and populations, to date, these consumer-based tracking devices are not considered ideal for measurement purposes in research. Some of their limitations include unstandardized wearing positions (some people carry their smartphone in their pocket, others put it in a backpack or purse), low validity when compared to research-grade accelerometers, and high potential for reactivity due to the readily available information about the user’s activity patterns. Reactivity is the tendency to modify one’s habitual behavior after becoming aware of being measured or observed. The phenomenon of reactivity can be very difficult to control in studies of physical activity behavior, but it becomes particularly problematic when participants have direct and real-time access to the results of the measures being performed, as is the case with consumer-based tracking devices and apps.

More Excerpts From Foundations of Physical Activity and Public Health 3rd Edition With HKPropel Access

---