Validation and human factor analysis study of an infant weight estimation device.

Journal Article (Journal Article)

BACKGROUND: Weight is critical for the medical management of infants; however, scales can be unavailable or inaccessible in some practice settings. We recently developed and validated a robust infant weight estimation method based on chest circumference (CC) and head circumference (HC). This study was designed to determine the human factors (HF) experience with, and predictive performance of, an infant weight estimation device that implements this method. METHODS: Prospective, multi-center, observational, masked study of 486 preterm and term infants (0-90 days) assessed by 15 raters. Raters measured the infant using calibrated scales/measures and masked versions of the device. Raters also evaluated critical tasks associated with device use. Mean error (ME) and mean percentage error (MPE) were used to assess predictive performance. RESULT: Among 486 infants enrolled (36.8 ± 4.0 weeks gestational age, 31.5 ± 28.6 days postnatal age), predicted weight correlated highly with actual weight (r = 0.97, ME: - 69 ± 257 g, MPE: - 1.3 ± 6.9%). Predicted weight was within 10 and 15% of actual weight in 86 and 99%, of infants. HF errors were low, 0.1-0.8% depending on task. In all cases raters were confident or very confident in their measurements. CONCLUSION: The device was statistically equivalent to the method on which it was based and approximated weight with acceptable variance from the true weight. HF data suggest the device is easy to use. This device can be used to estimate weight in infants when calibrated scales are impractical or unavailable.

Full Text

Duke Authors

Cited Authors

  • Abdel-Rahman, SM; Paul, IM; Delmore, P; Chen, J-Y; Mills, M; Greenberg, RG; Best Pharmaceuticals for Children Act – Pediatric Trials Network,

Published Date

  • January 22, 2020

Published In

Volume / Issue

  • 20 / 1

Start / End Page

  • 30 -

PubMed ID

  • 31969129

Pubmed Central ID

  • PMC6977278

Electronic International Standard Serial Number (EISSN)

  • 1471-2431

Digital Object Identifier (DOI)

  • 10.1186/s12887-020-1933-5

Language

  • eng

Conference Location

  • England