A Short List About What I Do and Don't Know

INTRODUCTION Objective monitoring of physical activity using body-worn pedometers and accelerometers has become a mainstay measure in many activity-related research protocols and practical applications, including clinical practice. For example, the U.S. National Health and Nutrition Examination Survey (NHANES), the primary source of our nation's health surveillance data, uses an accelerometer to objectively capture free-living physical activity behaviors. Although accelerometers are indisputably important in terms of studying physical activity intensity and pattern (for example, moderate-to-vigorous physical activity [MVPA], a key constituent of public health recommendations), pedometers generally are considered more practical for clinical applications, largely because they are relatively inexpensive and feasible in terms of implementation and data collection and management. Recent metaanalyses have documented the ability of pedometers to increase physical activity when they are used as part of a behavior change program. Interventions have produced incremental increases in physical activity on the order of 1800-4500 steps·d−1, accompanied by modest weight loss and improvements in waist circumference and blood pressure. Buoyed by such consistently positive reports, clinicians are beginning to use both pedometers and accelerometers to describe usual activity levels in children, older adults, and special populations, including individuals with heart and vascular diseases, chronic obstructive lung disease, diabetes, neuromuscular disease, arthritis, and intellectual disability. Objective monitors of physical activity have potential utility for screening, prescription, monitoring, feedback, and evaluation. And there are more and more review articles being published to help guide clinicians on related methodological issues and other considerations. Before jumping on the bandwagon, however, clinicians need to carefully consider their personal clinical needs and resources (including personnel expertise and time); therefore, I have compiled a short list reflecting on what I do and don‘t know about objective monitoring of physical activity. WHAT I KNOW Objective monitors are commercial items. New ones appear on the market at regular intervals; the number of options and their various features are staggering. Most objective monitors are worn at the waist attached to clothing, but some are designed to be worn in a pocket, in a shoe, or around an ankle. They also vary widely in cost. Data management demands vary greatly. Some require additional docking stations and software to access data, some provide a colorful array of graphics, and some simply provide an onboard digital screen display. Clinicians need to consider carefully what, exactly, they want and the time and expertise they have to invest. They change. Just like the particular running shoe model that seems to perform best for you or the computer operating system you have finally become comfortable with, the specific objective monitoring instrument that you love most can disappear or reappear with changed features. This can be frustrating when you want to compare data collected using different models of the same instrument. Although manufacturers claim new generation models are fully backwards-translatable, independent evidence suggests that this might not be the case. They are protected by patents. This means that one has a measurement mechanism driven by a coil spring, another has a hair spring, another has a piezoelectric mechanism, and so forth. The ramification is that a step measured by one isn't necessarily a step measured by another. And this goes for minutes in MVPA or sedentary behaviors as well: MVPA time detected by one doesn't agree with another, and one infers sedentary behavior from inclinometers, another by interpreting acceleration cut points, and so on. Increased sensitivity comes with a specificity tradeoff. In a dogged pursuit of ever lower force accelerations (e.g., very slow walking), manufacturers are serving up more and more sensitive instruments. The flip side is that these same instruments also pick up a lot of nonambulatory ``noise'' from riding in a car, fidgeting, or shifting in a chair. There is no easy answer to this conundrum. But clinicians and others need to be aware of the inevitable tradeoff. For example, if they select a sensitive instrument to detect gait patterns in mobility-compromised elderly individuals, the output detected cannot be interpreted using a scale developed for use in community-dwelling elderly people who walk at relatively more forceful paces, even during activities of daily living. We are in a buyer beware situation. There are a lot of options for instrumentation. And there is no authoritative body that governs or otherwise oversees the measurement definition of a step or time spent in MVPA, or even sedentary behavior for that matter. Again, this makes it difficult to compare across studies using different instruments. Price isn't necessarily indicative of an instrument that would work best in any given clinical application. And sometimes a low-technology, reasonably priced device (that also is valid and reliable) is both a sufficient and appropriate choice. WHAT I DON'T KNOW The best instrument for all situations. A technology that is perfect for a specific research question may be too cumbersome and require too much data distillery to be useful in a clinical application. One instrument might be perfect for use in children but inappropriate for use with older adults. There is no simple answer. The best defense is educating yourself before diving in. How well similarly-named outputs translate between instruments. This is unfortunate and curtails advancement in our ability to say how many steps are enough, how much moderate-to-vigorous physical activity is enough, and how much sedentary behavior is too much. It would be nice to have some way to convert the output of one instrument to another, but we are not there yet (and, dare I say, may never be, given innate differences in measurement mechanisms). About obsolescence. This scares me. In 50 yr, when we go to measure physical activity objectively, and we say that the population is taking X number of steps·d−1 (or minutes in MVPA or sedentary behavior), will we really be able to say it has changed from today if instrumentation has changed? What the future will bring. Well, I do know that commercially-driven technology will continue to evolve. This is the proverbial double-edged sword. Don't get me wrong: I am excited about new advances in technology, its application, and its evaluation. However, I fret that the "ruler" keeps changing, and that is why I place "what the future will bring" under the "What I don't know" category. BABIES AND BATHWATER Doom and gloom aside, objective monitoring of physical activity has transformed researchers' and practitioners' ability to conduct surveillance, quantify dose-response relationships, screen for low activity, communicate specific prescriptions, motivate patients through self-monitoring and personalized feedback, and evaluate progress. These tasks previously were informed primarily by self-report questionnaires, and the limitations of those are well known. Selected with care and used thoughtfully and appropriately, objective monitoring of physical activity can be a powerful clinical tool. Clinicians should take advantage of this technology, while cognizant of the caveats I have listed in this article.