Impact of different evaluation methods on the dropout rate of objectively detected physical activity on age specific subgroups: Results from the Motorik-Modul-Study (MoMo)

Introduction: In order to determine physical activity in free-living conditions the most common used objective method is accelerometry. This study analyzed the different technical methods and their outcomes evaluating accelerometer data in the large-scale epidemiological MoMo study. MoMo marks the first nationwide objective collection of data on physical activity and sedentary behavior of German children, adolescents and young adults. Method: The accelerometer sample size in MoMo from the latest survey from 2014-2017 is n = 1.971 (4-18 years). Wearing ActiGraph accelerometers GT3X+/wGT3X-BT movement behavior was objectively registered for one week. Effects on the dropout rate were analyzed with different epoch lengths (EL), non-wear-time (NWT) definitions, different valid day/week criteria and intensity-classification-algorithms. Results: Recording is in EL of 1s with the possibility to convert into 5s, 10s, 15s, 30s and 60s for future analysis. For the NWT Calculation the Choi-Algorithm was chosen because of the constancy in detecting NWT whereas Troiano shows variance up to 3% in NWT with different EL. Valid day criteria is 8h of recordings (+13% valid data compared to 10h) on four weekdays and one further weekend day when wearing the device for 7d. ... mehrFour different intensity classification algorithms (Evenson, Hängii, Romanzini, Sasaki) were validated for the different age groups (5-8, 7-11, 12-18, adults). Conclusion: Because there is no general consensus, these results will help researchers to make more suitable decisions for data processing criteria before and after data collection using the GT3X/+ accelerometer in largescale epidemiological studies. External funding details: Federal Ministry of Education and Research, funding reference number: 01ER1503