Background/Purpose: People living with inflammatory arthritis are likely to be less physically active due to pain and fatigue.  To understand the relationship between physical inactivity and poor health outcomes in inflammatory arthritis, it is necessary to accurately identify and quantify lower levels of physical activity. Accelerometers are small activity monitors used to measure moderate and higher levels of physical activity.  However, it is unclear how accurately different monitors capture sedentary and light physical activities. The purpose of this study was to compare the accuracy of the SenseWear Mini^TM (SW) and Actigraph GT3X^TM (AG) accelerometers for differentiating between sedentary and non-sedentary/ light physical activities in a controlled laboratory setting.

Methods: 22 adults (F=15, M=7, mean age 35.7, range 19 to 72) volunteered to participate.  Both monitors were synchronized with Greenwich Mean Time and were set to collect data at 1-minute intervals. Participants were randomly assigned to perform 9 activities for 5 minutes each.  All activities were observed and timed.  The activities were performed in: 1) standing (**slow treadmill, **wash dishes, *text messaging), 2) sitting (**slow cycling, *watch television, **computer typing), and 3) lying down (*rest with music, *read book, *knee ROM exercise in lying). According to the 2011 Compendium of Physical Activities (CPA), 5 activities were defined as *sedentary (<1.6 METs) and 4 were defined as **non-sedentary/ light (1.6 to 3 METs).  The middle 3-minute interval from each 5-minute time period was analyzed. Each SW activity was categorized as sedentary (< 1.6 METs) or non-sedentary (> 1.6 METs) based on the SW proprietary MET data.  Each AG activity was categorized as sedentary or non-sedentary, based on 2 different sedentary cut-points (< 100 and < 589 activity counts/ minute) for the AG vertical axis data. 

Percentage of activities correctly identified as sedentary (sensitivity) and non-sedentary (specificity), proportion of all activities identified as sedentary or non-sedentary that were in fact sedentary (positive predictive value – PPV) or non-sedentary (negative predictive value – NPV),  as well as, the positive (LR+) and negative (LR-) likelihood ratios were calculated for both accelerometers.

Results: Compared against derived values from the 2011 Compendium of Physical Activities, SW performed notably better than AG in differentiating between sedentary and non-sedentary/ light physical activities under controlled laboratory settings (Table 1).  Accuracy of AG was not affected by the cut-point chosen.

Conclusion:  Our results suggest that SenseWear Mini may be a better accelerometer for objective measurements of low physical activity in people with inflammatory arthritis.  Further studies are needed to examine the accuracy of SenseWear Mini for measuring physical activity under free-living conditions.