Yongquan Dong

Data Acquisition and Preprocessing in Studies on Humans: What is Not Taught in Statistics Classes?

The aim of this article is to address issues in research that may be missing from statistics classes and important for (bio-) statistics students. In the context of a case study, we discuss data acquisition and preprocessing steps that fill the gap between research questions posed by subject matter scientists and statistical methodology for formal inference. Issues include participant recruitment, data collection training and standardization, variable coding, data review and verification, data cleaning and editing, and documentation. Despite the critical importance of these details in research, most of these issues are rarely discussed in an applied statistics program. One reason for the lack of more formal training is the difficulty in addressing the many challenges that can possibly arise in the course of a study in a systematic way. This article can help to bridge the gap between research questions and formal statistical inference by using an illustrative case study for a discussion. We hope that reading and discussing this article and practicing data preprocessing exercises will sensitize statistics students to these important issues and achieve optimal conduct, quality control, analysis, and interpretation of a study.

Longer breastfeeding duration reduces the positive relationships among gestational weight gain, birth weight and childhood anthropometrics

Background The relationship between gestational weight gain (GWG) and childhood growth remains controversial. An examination on whether infant feeding practices mediate this relationship may improve our understanding of it. Methods We investigated whether the relationships among GWG, birth weight and childhood anthropometrics were mediated through infant feeding practices (breastfeeding duration and age at introduction of solid foods) in a cross-sectional multiethnic study of 1387 mothers and their children aged 0–5.9 years in the USA (2011–2012). Child anthropometrics included age-specific and sex-specific z-scores for weight-for-age (WAZ), height/length-for-age (HAZ), weight-for-height/length (WHZ) and body mass index-for-age (BMIZ); and ulnar length, a marker for limb growth. We used structural equation modelling to calculate standardised path coefficients and total, direct and indirect associations of GWG, birth weight and infant feeding practices with child anthropometrics. Results Maternal GWG had a positive indirect association with all anthropometrics mediated via birth weight, whereas longer breastfeeding duration reduced the positive associations of GWG and birth weight with WAZ, WHZ and BMIZ in non-Hispanics (β=−0.077, −0.064 and −0.106, respectively). Longer breastfeeding duration and introducing solid foods at a later age were positively associated with ulnar length (β=0.023 and 0.030, respectively) but not HAZ, suggesting a distinct association, for the first time, with limb growth. Conclusions Findings suggest that promoting longer breastfeeding duration among women with excessive GWG who had high birthweight newborns may mitigate the potential for their offspring to develop obesity. In addition, findings reinforce the importance of promoting appropriate GWG and preventing high birth weight, which are positively associated with childhood anthropometrics.

Arm span and ulnar length are reliable and accurate estimates of recumbent length and height in a multiethnic population of infants and children under 6 years of age.

Surrogate measures are needed when recumbent length or height is unobtainable or unreliable. Arm span has been used as a surrogate but is not feasible in children with shoulder or arm contractures. Ulnar length is not usually impaired by joint deformities, yet its utility as a surrogate has not been adequately studied. In this cross-sectional study, we aimed to examine the accuracy and reliability of ulnar length measured by different tools as a surrogate measure of recumbent length and height. Anthropometrics [recumbent length, height, arm span, and ulnar length by caliper (ULC), ruler (ULR), and grid (ULG)] were measured in 1479 healthy infants and children aged <6 y across 8 study centers in the United States. Multivariate mixed-effects linear regression models for recumbent length and height were developed by using ulnar length and arm span as surrogate measures. The agreement between the measured length or height and the predicted values by ULC, ULR, ULG, and arm span were examined by Bland-Altman plots. All 3 measures of ulnar length and arm span were highly correlated with length and height. The degree of precision of prediction equations for length by ULC, ULR, and ULG (R(2) = 0.95, 0.95, and 0.92, respectively) was comparable with that by arm span (R(2) = 0.97) using age, sex, and ethnicity as covariates; however, height prediction by ULC (R(2) = 0.87), ULR (R(2) = 0.85), and ULG (R(2) = 0.88) was less comparable with arm span (R(2) = 0.94). Our study demonstrates that arm span and ULC, ULR, or ULG can serve as accurate and reliable surrogate measures of recumbent length and height in healthy children; however, ULC, ULR, and ULG tend to slightly overestimate length and height in young infants and children. Further testing of ulnar length as a surrogate is warranted in physically impaired or nonambulatory children.