Table 2:
Results from 100 repeated holdouts testing the association between the phthalate WQS index and PMA at first growth spurt (latency) and slope of the first growth spurt (growth rate), controlling for sex, SGA, NICU morbidity score, and year of birth.
Outcome | Parameter | Mean | Median | 2.5th percentile | 97.5th percentile |
---|---|---|---|---|---|
Latencya | WQS | 0.17 | 0.21 | −0.39 | 0.64 |
WQS2 | −0.045 | −0.049 | −0.103 | 0.030 | |
Growth rateb | WQS | 0.13 | 0.13 | 0.05 | 0.21 |
N=101 (training n~40, validation n~61); a nonlinear effect was assessed by including a WQS*WQS term
N=96 (training n~38, validation ~58); βs were constrained in the positive direction; log transformed growth rate to meet assumptions of linear regression