Discussion
With the high rate of LPIs and ETIs (8.3% and 17.5%, respectively) and the costs incurred with such deliveries in the United States, decisions about the management of LPIs and ETIs are a significant part of obstetric practice. Therefore, physicians and policymakers need accurate estimates of the risk of morbidity and mortality associated with LPIs and ETIs to make informed decisions concerning the infants' and mothers' care. A strength of the current study is the amount of maternal data available, both demographic and clinical. More comprehensive maternal data allowed for greater statistical control of factors related to both infant morbidity and early delivery, thus providing more accurate estimates of the effect of early delivery itself. Because of the large number of infants born late preterm or early term, any increased morbidity could potentially have a very large public health impact, especially if effects of these morbidities persist beyond the birth hospitalization.
This study found rates of ADHD at 2.8% for term and ETIs, and 3.6% for LPIs. This is similar to recent Center for Disease Control (CDC) rates (2007 to 2011) that ranged from 1.1 to 2.5% across the nation. The CDC data present the rate for ages 3 to 5 and does not break down the percentages for gestational age at birth. Although there is no way to account for different practice patterns, the consistency of the rates would imply either similar patterns or no clinically significant difference among practice patterns.
Our study shows significant differences in several demographic- and delivery-related characteristics between the study groups; however, not all statistically significant differences are clinically significant. Specific values for these characteristics are available online as Supplementary Material http://www.nature.com/jp/journal/v35/n8/suppinfo/jp201528s1.html. The infant weight and size characteristics are consistent with what would be expected for deliveries at their respective gestational ages. The difference in infant race, maternal age, education and characteristics are not clinically significant. The delivery characteristics support the current trend of an increased cesarean delivery rate at earlier gestational ages. The number of prenatal visits is skewed by the fact that preterm deliveries lose several of their visits, depending on how early the delivery occurs. The inter-pregnancy interval, labor and delivery complications, and maternal comorbidities are important clinically to help with counseling and pregnancy management. The hospital characteristics are notable for confirming what is already known: those facilities that see the highest volume of patients and have the highest level of neonatal intensive care unit facilities will see an increased proportion of late preterm deliveries.
Several studies have examined long-term morbidities of LPIs. Shapiro-Mendoza et al. found that rates of enrollment in early intervention programs (surrogate for neurodevelopment) increased as gestational age at delivery decreased. Woythaler et al. retrospectively compared the mental developmental index and psychomotor developmental index between LPIs and term infants and found that LPIs were more likely to have lower scores and/or scores in the significant developmental delay range. Peacock et al. compared LPIs to term infants' success rates of a reading/writing/math assessment of 5 to 7 year old children and found that the success rates were significantly different at 71% for LPIs and 79% for term infants. Samra et al. performed a thorough review of the literature and found that there is both conflicting data and knowledge deficits in long-term morbidities of LPIs, suggesting that more research is needed. Our study reinforces the current knowledge that there are increased risks of long-term neurodevelopmental problems with late preterm deliveries, but it also demonstrates the conflicting nature of the data. We show an increase in developmental speech/language disorder, which is supported by other studies. However, Morag et al. published a prospective study comparing LPIs and term infants using the Alberta Infant Motor Scale and Griffiths Mental Development Scales. They found a significant decrease in scores, indicating impaired neurodevelopment, among LPIs; however, when the scores were corrected for post-conception age, no difference was observed.
Our study has weaknesses which further exhibit the need for more research in this area. The data set came from Medicaid claims and birth certificates that were not designed for research purposes. We based outcomes on the use of diagnosis and procedure codes, which are subject to error and bias. Additionally, the population was limited to Medicaid patients, who are associated with a lower socioeconomic class and traditionally have poorer obstetric outcomes. However, Medicaid is the primary payer for 48% of deliveries in the United States, making the study relevant to a large portion of the United States population. Another weakness of the study is the significant drop off over time in Medicaid enrollment. Early childhood outcome data were only available for approximately half of the deliveries in the study. Although appropriate econometric techniques were used in an attempt to address the lack of outcome data for a significant proportion of the study population, statistical techniques cannot completely correct for the potential bias introduced from the lack of data.
It is difficult to get consistent data in this area because studies focus on various aspects of neurodevelopment or on various types of neurodevelopmental issues. Our study used ICD-9 codes as an identifier of morbidity, while other studies use mental development index and physical development index, academic assessments, early intervention enrollments, and other surrogates for neurodevelopmental delay. In addition, although our study and others agree that there are potential morbidities with late preterm deliveries, we cannot ignore the fact that the majority of these deliveries are medically indicated. Therefore, while we should strive to decrease non-medically indicated late preterm deliveries, we can also use this information to better counsel our patients who will undergo LPI deliveries that are indicated for maternal and/or fetal conditions.