ORIGINAL ARTICLE Calibration of Accelerometer Output for Ambulatory Adults With Multiple Sclerosis Robert W. Motl, PhD, Erin M. Snook, PhD, Stamatis Agiovlasitis, PhD, Yoojin Suh, MS ABSTRACT. Motl RW, Snook EM, Agiovlasitis S, Suh Y. Calibration of accelerometer output for ambulatory adults with multiple sclerosis. Arch Phys Med Rehabil 2009;90:1778-84. Objectives: To examine the association between the rates of accelerometer activity counts and energy expenditure during walking in persons with multiple sclerosis (MS) versus controls and then to calibrate the output of accelerometers for comput- ing time spent in light, moderate, and vigorous physical activity based on common metabolic equivalent unit categories in per- sons with MS. Design: Mixed-model design. Setting: Laboratory. Participants: People with MS (n=24) and people without MS (n=24) who were similar in age, sex, height, and weight. Interventions: The participants undertook three 6-minute periods of walking at 3.2, 4.8, and 6.4km·h -1 on a motor- driven treadmill. Main Outcome Measures: Activity counts and energy ex- penditure were measured with an accelerometer worn on the right hip and open-circuit spirometry, respectively. Results: The results indicated that (1) persons with MS had greater energy expenditure, but not activity counts, during walking on a treadmill than did controls; (2) there was a strong linear relationship between activity counts and energy expen- diture during treadmill walking, but the slope of the relation- ship was steeper in persons with MS than in controls; and (3) the cut-points for light, moderate, and vigorous physical activ- ity were lower in persons with MS than in controls. Conclusions: Such findings provide evidence for a strong linear relationship between activity counts and energy expen- diture during walking in persons with MS and cut-points based on counts per minute for quantifying time spent in light, moderate, and vigorous physical activity using accelerometers in this population. Key Words: Exercise; Multiple sclerosis; Oxygen consump- tion; Rehabilitation; Walking. © 2009 by the American Congress of Rehabilitation Medicine T HERE HAS BEEN AN INCREASED interest in studying physical activity behavior among people with MS because of the beneficial relationship of physical activity with indices of quality of life, symptom management, and disease progres- sion. 1-3 Of central importance in studying physical activity among persons with MS is its measurement. Accurate mea- surement is critical when examining the relationship between physical activity and outcomes linked with disease progression, symptom management, and rehabilitation. If the measurement of physical activity does not precisely characterize the behavior itself, then the strength of the association between physical activity and outcomes is likely underestimated or eliminated. 4 Motion sensors such as accelerometers might represent an ideal method of measuring physical activity in persons with MS. Accelerometers in physical activity monitors provide a raw output of activity counts. The activity counts represent a digital integration of a bidirectional, positive and negative acceleration signal over a given time window. This accelera- tion signal is proportional to the net external force generated during bodily movement associated with physical activity be- havior. By extension, the acceleration signal should directly reflect the energy costs of physical activity. Activity counts from an accelerometer seemingly provide a valid, objective measurement of physical activity and avoid the problems of recall and self-reporting biases that can plague surveys. Indeed, several teams of investigators have examined the validity of activity counts from an accelerometer as a measure of physical activity in persons with MS. 5-9 Those studies have either examined the association between activity counts from an accelerometer, step counts from a pedometer, and scores from self-reported surveys or compared mean ac- tivity counts from accelerometers between individuals with MS and controls. For example, 2 studies 7,8 reported a large corre- lation between activity counts from an accelerometer and step counts from a pedometer, and moderate to large correlations between activity counts from an accelerometer and scores from 2 self-report surveys in individuals with MS. Another study reported that activity counts from an accelerometer, but not scores from a self-report survey, were sensitive in detecting differences in physical activity among persons with MS com- pared with sedentary and active controls. 9 That body of evi- dence provides a promising, albeit preliminary, empirical basis for studying physical activity using accelerometers in persons with MS. There is a major limitation of that body of research. Re- searchers have not examined the primary assumption 10,11 of a strong association between accelerometer activity counts and energy expenditure during dynamic physical activity in persons From the Department of Kinesiology and Community Health, University of Illinois at Urbana-Champaign, Urbana, IL. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organi- zation with which the authors are associated. Reprint requests to Robert W. Motl, PhD, Dept of Kinesiology, The University of Illinois Urbana-Champaign, 350 Louise Freer Hall, 906 S Goodwin Ave, Urbana, IL 61820-3895, e-mail: robmotl@uiuc.edu. 0003-9993/09/9010-00897$36.00/0 doi:10.1016/j.apmr.2009.03.020 List of Abbreviations ANOVA analysis of variance CI confidence interval EDSS Expanded Disability Status Scale MET metabolic equivalent unit MS multiple sclerosis PDDS Patient-Determined Disease Steps SEE standard error of the estimate SEM standard error of the mean V ˙ O 2 oxygen consumption 1778 Arch Phys Med Rehabil Vol 90, October 2009