Fig 1.
Flowchart of recruitment of study participants at Hoa Binh Province General Hospital, Vietnam.
Table 1.
Demographic and clinical characteristics of the study participants (N = 485).
Fig 2.
Receiver operating characteristic (ROC) curves for anthropometric measures as diagnostic tools for predicting each outcome, and areas under each curve (AUC).
Top row: ROC curves for anthropometric measures as predictors of low birth weight (<2500g). Middle row: ROC curves for anthropometric measures as predictors of prematurity (<37 weeks according to the New Ballard Score). Bottom row: ROC curves for anthropometric measures as predictors of low birth weight and prematurity. The solid diagonal line represents a theoretical ROC curve that is no better than random as a predictor of the outcome; the dashed horizontal line represents the required threshold sensitivity of 0.8.
Table 2.
Sensitivity, specificity, and positive and negative predictive values of optimal cut-points1 for each outcome and anthropometric measure.
Fig 3.
Scatter plots of anthropometric measurements taken at day 1 of life against measurements taken at day 5 of life (n = 200).
The line of best fit and the correlation coefficient (r) are also shown. All measurements are in centimetres.
Table 3.
Comparison of anthropometric measures at days 1 and 5 of life, in a subset of newborns remeasured on day 5 (N = 200).