384 - Who's in the NICU? A Population-level Descriptive Analysis

Poster Number: 384
Publication Number: 384.133

Carolyn Smith Hughes, University of California, San Francisco, Department of Epidemiology and Biostatistics, San Francisco, CA, United States; Susan K. Schmitt, Department of Veterans Affairs, Grand Junction, CO, United States; Molly Passarella, Childrens Hospital of Philadelphia, Philadelphia, PA, United States; Scott A. Lorch, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, United States; Ciaran S. Phibbs, Stanford University School of Medicine, Palo Alto, CA, United States

Background: At least 10% infants born in the US receive some level of neonatal intensive care unit (NICU) care. However, characteristics of infants admitted to NICUs, their conditions, and their lengths of stay (LOS) have not been recently documented longitudinally across a range of NICUs in a specific state.

Objective: To describe the characteristics and disposition of infants admitted to NICUs in a population-level analysis, including gestational age (GA), birthweight (BW), reason for admission, and relative contribution to total LOS.

Design/Methods: We performed a descriptive analysis of liveborn infants in South Carolina from 1996-2015 using birth certificates linked to maternal and infant hospital discharge records. Infants were excluded (0.28%) if they were born outside the analysis period, if their birth certificate was not linked to discharge records, or if they were not considered viable (GA < 22 weeks and/or BW < 400 g). Missing or implausible GA or BW were accounted for using ICD-9 codes or imputation. NICU admission status was based on hospital revenue codes, as previously documented. Descriptive statistics included characteristics of mothers, births, and infants. We also assessed total LOS (NICU and non-NICU days) and variation by NICU level and year. We categorized infants by highest level of condition severity based on morbidity, LOS, and GA.

Results: Of 1,035,236 live births, 11.5% (118,809) were admitted to a NICU; those admitted had mean LOS of 17 days (SD: 25) and had >2 million days of care. 55.9% were preterm and/or low BW and accounted for nearly 80% of total LOS.
 
64.8% of admissions were near-term or term (GA >34 weeks) infants, who made up 30.6% of total LOS. Of these infants, 54.2% had an LOS < 6 days. In terms of condition severity, 27.8% of infants with GA >34 weeks had only mild conditions; 37.6% had moderate conditions; and 12.7% had severe conditions. Among those with mild conditions, 76.7% had only one mild condition; hyperbilirubinemia (48.2%) and observation for suspected infection (39.8%) were the most frequent single mild conditions. 
 
The share of admissions and LOS was generally constant over time by GA and among those with mild conditions.  Conclusion(s): More half of infants admitted to NICUs are near-term/term, and account for a lower relative proportion of total LOS. More than a quarter of near-term/term infants were admitted for mild conditions, the majority of which were a single condition. Understanding short-stay NICU admissions, and whether their care can be accommodated in other settings, is important for optimizing the use of NICU care.
Table 1
Figure 1: Composition of infants in NICUs by gestational age and birthweight