LAUSR

Unauthorized reproduction of this article is prohibited 676 www.pccmjournal.org July 2019 • Volume 20 • Number 7 The search for optimal care practices in neonatology is ongoing. This is particularly true for our most vulnerable patients, those born at the edge of viability. Despite recent advances, this population continues to face high mortality, substantial morbidity, and an array of long-term complications (1). Certain practices like continuous positive airway pressure and early surfactant administration are known to improve some of these outcomes (2, 3). However, the search for “good ideas” to improve the care of the extremely low birthweight infant continues. To improve outcomes, it is critical first to identify the most critical causes of failure. Therefore, we are grateful to Park et al (4), in this issue of Pediatric Critical Care Medicine, for their study examining the mortality rate-dependent variations in the timing and cause of death of infants born at 23–24 weeks’ gestation in the Korean Neonatal Network (KNN) during a recent 3-year time period. The hospitals in the KNN were arbitrarily categorized to low-mortality (< 50%) and high-mortality (> 50%) groups, based on comprehensive institutional statistics. The primary findings were that infants in the low-mortality group had significantly reduced mortality in all subgroups and at all time periods less than 28 days of life and that reductions in cardiorespiratory, infectious, and gastrointestinal causes were responsible for much of these differences. Importantly, modifiable factors such as antenatal steroid use and body temperature at time of admission were associated with reduced mortality (4). One unique aspect the study by Park et al (4) is that withdrawal of care, which may confound mortality results, is prohibited by law in Korea. We are aware that some of these causes are already known to be independent risk factors for morbidity and mortality in premature neonates (5). The most notable finding, however, is that there was a large discrepancy in patient volumes between the two groups. The low-mortality group had an average of 19 patients per hospital, versus only six patients per hospital for the high-mortality group. Although the authors themselves briefly touch on this finding, we emphasize that regionalization, large patient volumes, and familiarity/expertise in managing vulnerable patient populations have a profound impact on outcomes. Regionalization is not a new concept, as guidelines for perinatal regionalization have existed for decades that have strongly recommended that very-low-birth-weight infants be born at institutions that provide highly specialized maternal and neonatal care. We also know that vulnerable infants born at low-volume, low-acuity centers face an increased likelihood of death (6). The Committee on the Fetus and Newborn of the American Academy of Pediatrics has also endorsed this view, with their most recent policy statement in 2015 recognizing the critical need to regionalize perinatal care (7). Regionalization seems also to reduce other severe morbidities in this high-risk population, including necrotizing enterocolitis and bronchopulmonary dysplasia (8, 9). Clearly, expertise in the care of the vulnerable premature infant is enhanced and perpetuated by large patient volumes (10, 11). Although the need for continued perinatal regionalization is widely recognized in the United States, we continue to fall short of this goal. A recent geospatial analysis of advanced obstetric and neonatal care centers demonstrated areas of wide gaps in access, as well as large discrepancies in the availability of suitably joined centers for both the mother and infant (12). As we continue to grapple with reducing maternal and neonatal mortality, we must find ways to improve delivery of proven treatments to the target populations. In many instances it appears that as our knowledge of taking care of high-risk maternal-fetal dyads has increased, so too have the barriers and limitations in administering that care—a frustrating dichotomy to be sure. As Park et al (4) have demonstrated, there are several modifiable risk factors for mortality in this population of extremely preterm infants, especially antenatal steroid administration and optimal thermoregulation. Clearly, improvements in these care practices are required in high-mortality centers everywhere; however, it is likely that these differences may be influenced, in part, by patient volume as well. In addition, we believe that the reduction in deaths due to cardiorespiratory, infectious, and gastrointestinal causes noted in the study by Park et al (4) are likely due to improved care practices and should be emulated by high-mortality centers. As a field, we are right to emphasize the benefits of properly administered antenatal steroids to the mother and the importance of optimally performed neonatal resuscitation and close attention to admission body temperatures. Unfortunately, we do not always insist that such interventions occur reliably and effectively. Regionalization of both maternal and newborn care to high-volume, high-acuity centers, can help ensure that these life-saving interventions occur as intended. *See also p. 630.