Accelerometry Accuracy In Differentiating Between Sedentary Behavior And Light Physical Activity With Overweight Pregnant Females

PURPOSE: Physical activity is an important aspect of health in pregnant females, however little is known regarding how to best assess activity behaviors in this population. The purpose of this study was to compare the validity of ActivPAL (AP) versus ActiGraph (AG) in measuring and differentiating between sedentary behavior (SB) and light physical activity (LPA) in overweight pregnant females. METHODS:13 pregnant overweight (BMI > 25 kg/m2) females (32.46 + 4.68 years, BMI: 28.46 + 2.10 kg/m2) in their 2nd or 3rd trimester wore AGs on their wrist and waist and APs on both thighs while performing 3 SBs and 8 LPAs in a randomized order for 7 minutes each. An Oxycon Mobile measured oxygen consumption (VO2) as the criterion. SBs included lying, sitting watching TV, and sitting typing. LPAs included standing/reading, standing/folding laundry, washing dishes, walking, stretching, setting a table, making a bed, and vacuuming. Mean percentage error (MPE) was used to assess the error for each of the wearable monitors relative to VO2. We dichotomized each device measurement as either SB (<1.5 METs) or LPA (1.5 - 3.0 METs). Kappa statistics assessed the agreement among the dichotomized measures and VO2. Kappa ≥0.6 were considered acceptable. Bland-Altman (BA) plots were used to visually inspect device agreement. RESULTS: For the AP, MPEs for SBs ranged from 20.2 to 38.7% (1.29+ 0.07 METs), and MPEs for LPAs ranged from -17.6 to 32.1% (1.75 + 0.54 METs). For the ankle AG, MPEs for SBs ranged from -7.5 to 11.3% (1.0 + 0.01 METs) and MPEs for LPAs ranged from -32.1 to 179.8% (2.24 + 1.87 METs). For the waist AG, MPEs for SBs ranged from -7.5 to 11% (1.0 + 0.02 METs) and MPEs for LPAs ranged from -44.3 to -23.8% (1.12 + 0.31 METs). Kappa was >0.6 for all measures except waist AG. Inspection of BA plots indicated that, in terms of alignment with VO2, the most optimal performance for SB was ankle AG, and for LPA was AP. CONCLUSIONS: AP was able to acceptably measure and differentiate LPA, but overestimated METs for SB. Compared to AP, AG METs were closer to the criterion for SBs, but worse at capturing LPA METs. Additionally, waist-worn AG should be avoided for assessment of and differentiation between SBs and LPA in overweight pregnant females. Compared to AG, AP may be better able to better measure and differentiate between SBs and LPA in overweight pregnant females.