Validation of digital anthropometric measurements during pregnancy: a longitudinal study

Abstract

Maternal anthropometric changes are rapid during pregnancy and reflect increments in maternal and fetal tissues. These dynamic changes in body composition and shape during pregnancy are associated with maternal and fetal outcomes and are often monitored with simple tools such as a flexible tape for quantifying selected circumferences. The current study aim was to evaluate the hypothesis that circumferential measures of maternal body size and shape acquired with a 3D-optical imaging system will correlate, be accurate, and agree closely with ground-truth estimates made with a flexible tape by trained staff. 3D-optical scans and flexible tape measurements were acquired at 15-, 25-, and 35-weeks of gestation in 57, 41, and 35 participants, respectively. 3D avatars obtained at each time point were analyzed for waist (two sites), hip, mid-upper arm, mid-thigh, and calf circumferences. 3D-optical and flexible tape measurements were compared using linear regression analyses (r, r 2s), mean absolute errors (MAEs), root-mean square errors (RMSEs), concordance correlation coefficients (CCCs), and Bland-Altman plots. Overall, associations between 3D and conventional anthropometric measurements were strong (univariate linear regression r values, 0.80–0.98; p’s all <0.001) with good corresponding accuracy and agreement at all five anatomic sites (MAEs, –8.5 to 0.8 cm; RMSEs, 0.61–9.94 cm; CCCs 0.6–1.0; small significant (p < 0.05) bias was present at some sites/timepoints for some measures). Post hoc analyses revealed a potential basis for impact of advancing pregnancy on between-method differences. Strong associations, good accuracy, and close agreement with flexible tape circumference measurements, combined with feasible implementation, supports further development of 3D-optical technology as an alternative to conventional anthropometry for evaluating and monitoring body size, shape, and composition over the course of pregnancy.