PURPOSE: Researchers collecting accelerometry data of physical activity habits often need to compare, or even collate, data acquired from different versions of a device that have come from the same manufacturer. Since few independent results exist, the purpose of this study was to determine if the outputs of three generations of accelerometer from the same manufacturer are similar and interchangeable.

METHODS: Ten volunteers simultaneously wore three generations of Actigraph accelerometers (1D-Actitrainer, GT1M, and 7164) under two conditions: (i) during a single normal working day, and (ii) during a progressive exercise protocol on a treadmill (7 incremental steps lasting 3 minutes, progressing from slow walking at 1.5mph to running at 5mph).

RESULTS: Over a single day some variation in the total number of steps taken was evident across the 3 accelerometers (10611, 10824, 12103 steps) for the Actitrainer, GTIM and 7164 respectively. However the respective time (minutes) spent in light activity (913, 913, 913 min), moderate (78, 78, 76 min), heavy (7, 8, 10 min), and very heavy activity (0, 1, 1 min), were in very good agreement. The respective mean accelerometer outputs (counts/min) for the Actitrainer, GTIM and 7164 during the treadmill test showed acceptable agreement at 1.5mph (541, 613, 732), 2.5mph (2321, 2377, 2439), 3.5mph (4324, 4333, 4455), 4.0mph (6630, 6709, 6937), 4.5mph (7583, 7641, 7795), 5.0mph (8002, 8064, 8335), and 5.5mph (8622, 8691, 8948).

CONCLUSIONS: Other than the accrued step counts, the outputs from the three generations of Actigraph accelerometer appear to have adequately low inter-machine variability across a wide range of treadmill speeds and over a normal day.
Acknowledgement: This was part of a larger series of studies funded by the Research Grants Council of Hong Kong (#747807H) and the HKU URC Strategic Research Theme (Public Health).