Validity and reliability of an inertial device (WIMU PROTM) to quantify physical activity level through steps measurement

U.S. adults may have lower levels of ambulatory physical activity compared with adults living in other countries. The purpose of this study was to provide descriptive, epidemiological data on the average number of steps per day estimated to be taken by U.S. adults and to identify predictors of pedometer-measured physical activity on the basis of demographic characteristics and self-reported behavioral characteristics. The America On the Move study was conducted in 2003. Individuals (N = 2522) aged 13 yr and older consented to fill out a survey, including 1921 adults aged 18 yr and older. Valid pedometer data were collected on 1136 adults with Accusplit AE120 pedometers. Data were weighted to reflect the general U.S. population according to several variables (age, gender, race/ethnicity, education, income, level of physical activity, and number of 5- to 17-yr-old children in the household). Differences in steps per day between subgroups were analyzed using unpaired t-tests when only two subgroups were involved or one-way ANOVA if multiple subgroups were involved. Adults reported taking an average of 5117 steps per day. Male gender, younger age, higher education level, single marital status, and lower body mass index were all positively associated with steps per day. Steps per day were positively related to other self-reported measures of physical activity and negatively related to self-reported measures on physical inactivity. Living environment (urban, suburban, or rural) and eating habits were not associated with steps per day. In the current study, men and women living in the United States took fewer steps per day than those living in Switzerland, Australia, and Japan. We conclude that low levels of ambulatory physical activity are contributing to the high prevalence of adult obesity in the United States.

Step counting has long been used as a method of measuring distance. Starting in the mid-1900s, researchers became interested in using steps per day to quantify ambulatory physical activity. This line of research gained momentum after 1995, with the introduction of reasonably accurate spring-levered pedometers with digital displays. Since 2010, the use of accelerometer-based “activity trackers” by private citizens has skyrocketed. Steps have several advantages as a metric for assessing physical activity: they are intuitive, easy to measure, objective, and they represent a fundamental unit of human ambulatory activity. However, since they measure a human behavior, they have inherent biological variability; this means that measurements must be made over 3–7 days to attain valid and reliable estimates. There are many different kinds of step counters, designed to be worn on various sites on the body; all of these devices have strengths and limitations. In cross-sectional studies, strong associations between steps per day and health variables have been documented. Currently, at least eight prospective, longitudinal studies using accelerometers are being conducted that may help to establish dose–response relationships between steps/day and health outcomes. Longitudinal interventions using step counters have shown that they can help inactive individuals to increase by 2500 steps per day. Step counting is useful for surveillance, and studies have been conducted in a number of countries around the world. Future challenges include the need to establish testing protocols and accuracy standards, and to decide upon the best placement sites. These challenges should be addressed in order to achieve harmonization between studies, and to accurately quantify dose–response relationships.

Large physical activity surveillance projects such as the UK Biobank and NHANES are using wrist-worn accelerometer-based activity monitors that collect raw data. The goal is to increase wear time by asking subjects to wear the monitors on the wrist instead of the hip, and then to use information in the raw signal to improve activity type and intensity estimation. The purposes of this work was to obtain an algorithm to process wrist and ankle raw data and to classify behavior into four broad activity classes: ambulation, cycling, sedentary, and other activities. Participants (N = 33) wearing accelerometers on the wrist and ankle performed 26 daily activities. The accelerometer data were collected, cleaned, and preprocessed to extract features that characterize 2-, 4-, and 12.8-s data windows. Feature vectors encoding information about frequency and intensity of motion extracted from analysis of the raw signal were used with a support vector machine classifier to identify a subject's activity. Results were compared with categories classified by a human observer. Algorithms were validated using a leave-one-subject-out strategy. The computational complexity of each processing step was also evaluated. With 12.8-s windows, the proposed strategy showed high classification accuracies for ankle data (95.0%) that decreased to 84.7% for wrist data. Shorter (4 s) windows only minimally decreased performances of the algorithm on the wrist to 84.2%. A classification algorithm using 13 features shows good classification into the four classes given the complexity of the activities in the original data set. The algorithm is computationally efficient and could be implemented in real time on mobile devices with only 4-s latency.