In Reply Ambrose et al raise 2 important points. First, we did not calculate gestational age (GA) from the last menstrual period (LMP) but assumed that the provided estimate on the state birth certificates was derived following the Centers for Disease Control and Prevention algorithm that is used for the US natality file and that defaults to LMP. We contacted both Florida and Texas Vital Statistics offices and verified that recorded (GA) estimates are indeed the clinical estimate and not solely calculated from the LMP. To explore the effect of the various GA estimates on our analysis, we recalculated age thresholds at which moderate premature infants reach similar respiratory syncytial virus (RSV) hospitalization risk as their preterm counterparts at 1 month of age using (1) the clinical GA estimate as done in our original analysis (age = 4.2 months [95% CI, 2.5-5.7] and 4.5 months [95% CI, 2.8-6.4] for Florida and Texas, respectively),1 (2) GA calculated from the LMP (age = 2.8 months [95% CI, 1.8-4.5] and 2.6 months [95% CI, 1.8-4.3]), and (3) the Centers for Disease Control and Prevention “combined GA estimate,” which selects the GA estimate based on compatibility with birth weight and defaults to the LMP-based estimate if both estimates are compatible (age = 2.9 months [95% CI, 1.9-4.5] and 2.7 months [95% CI, 1.9-4.1]).2 Ambrose et al are correct in their assumption that LMP-based estimates establish comparison groups that are more similar in RSV hospitalization risk, likely because of greater misclassification of term infants as preterm. Based on LMP-based determination of term status, moderate preterm infants approximated term infants’ RSV infection risk at a slightly younger age than estimated in our original analysis where clinical GA estimates were used.
Winterstein AG, Knox C, Kubilis P. Methods to Estimate Gestational Age Can Significantly Affect Study Results—Reply. JAMA Pediatr. 2014;168(4):388-389. doi:10.1001/jamapediatrics.2013.5178