Technology and Health Care 22 (2014) 199–208 199 DOI 10.3233/THC-140817 IOS Press Evaluation of active video games intensity: Comparison between accelerometer-based predictions and indirect calorimetric measurements Julien Tripette a,∗ , Takafumi Ando b , Haruka Murakami a , Kenta Yamamoto c , Kazunori Ohkawara d , Shigeho Tanaka b and Motohiko Miyachi a a Department of Health Promotion and Exercise, National Institute of Health and Nutrition, Tokyo, Japan b Department of Nutritional Sciences, National Institute of Health and Nutrition, Tokyo, Japan c Department of Integrative Physiology, University of North Texas Health Science Centre, TX, USA d Faculty of Informatics and Engineering, University of Electro-Communications, Tokyo, Japan Received 6 March 2014 Accepted 14 April 2014 Abstract. BACKGROUND: Several active video game (AVG) intervention studies failed in showing an increase in physical activity by using accelerometry measurements. OBJECTIVE: To test the validity of accelerometry for monitoring AVG playing intensity. METHOD: Twenty-two adults performed 80 activities included in the Wii Sports and Wii Fit Plus series. The energy expen- diture (EE) and subsequent MET values were measured by indirect calorimetry using metabolic chambers. Subjects wore an accelerometer-based monitor displaying MET values. For each activity, METs values obtained from indirect calorimetry and accelerometry were compared. Each activity was classified as light or moderate to vigorous physical activity (LPA: < 3METs or MVPA:  3METs) for the two methods. RESULTS: AVG intensities have been slightly but significantly underestimated by the acceleromater-based monitor compared to the indirect calorimetry (2.5 ± 1.0 instead of 2.7 ± 0.9 METs). Fourty percent of activities have been significantly misesti- mated, and 20% have been misclassified. CONCLUSION: Those results point out the potential bias of accelerometry measurements for evaluating AVG intensities. Because average AVG intensity lays at the boundary between LPA and MVPA classes, misclassifications can frequently occur. Accelerometry data should be interpreted with caution in intervention studies using AVG. Keywords: Exergame, accelerometry, metabolic chamber, energy expenditure, metabolic equivalent * Corresponding author: Julien Tripette, Department of Health Promotion and Exercise, National Institute of Health and Nutrition, Tokyo, Japan. Tel.: +81 3 3203 8061; Fax: +81 3 3203 1731; E-mail: tripette@nih.go.jp. 0928-7329/14/$27.50 c  2014 – IOS Press and the authors. All rights reserved AUTHOR COPY