754 - Validating a <25 weeks Gestational Age Neonatal Cohort using the Epic Cosmos Database: Continued Increased Active Treatment and Improved Survival in 21- and 22-week Infants

Publication Number: 2735.754

Lindsey A. Knake, University of Iowa Roy J. and Lucille A. Carver College of Medicine, Iowa City, IA, United States; Evan Miller, Nemours Children's Hospital, Newark, DE, United States; Sara Handley, Childrens Hospital of Philadelphia, Philadelphia, PA, United States; Heather Burris, Children's Hospital of Philadelphia, Philadelphia, PA, United States; Thomas Nienaber, Cincinnati Childrens Hospital Medical Center, Cincinnati, OH, United States; Jay Greenspan, Sidney Kimmel Medical College at Thomas Jefferson University, Philadelphia, PA, United States; Edward F. Bell, University of Iowa, Iowa City, IA, United States; Kevin Dysart, Children's Hospital of Philadelphia, Philadelphia, PA, United States

Background: A recent NICHD Neonatal Research Network (NRN) study (Bell et al, JAMA 2022) found that survival of infants born at < 28 weeks gestational age (GA) is improving. There is significant treatment variability for infants born at 21-24 weeks GA in the US, thus evaluating trends beyond the NRN academic centers warrants examination.

Objective: To 1) compare trends in active treatment and mortality among infants < 25 weeks' GA in the 2022 NRN study to those in a national US database and 2) describe contemporary national trends in treatment and outcomes as well as assess long-term outcome data availability.

Design/Methods: This retrospective cohort study used data in the Epic Systems Cosmos research platform. Participating hospitals submit a defined limited dataset, which currently includes over 1,809 hospitals and 300 million patients. Included infants were born at < 25 weeks' GA and had a neonatal birth encounter. Infants were divided into two cohorts: 2013-2018 to match the NRN study and 2019-2024 to capture contemporary trends. Active treatment was defined as a documented medication or surfactant administration. We used a hospital system fixed-effects logistic regression model with standard errors clustered by hospital system to adjust for sex, birth weight, major anomalies, and antenatal corticosteroids.

Results: From 223 hospital systems, 11,019 infants were included in the 2013-2018 cohort and 18,209 infants in the 2019-2024 cohort. The 2013-2018 data show similar 22-week GA trends compared to those reported by the NRN with 11% of all infants (Cosmos 8%) and 30% of infants with active treatment (Cosmos 30%) survived to discharge home (Table 1). Compared to 2013-2018, the 22-week GA contemporary cohort had higher odds of active treatment (aOR 2.5, 95% CI 1.8-3.4) and survival to discharge home (aOR 1.5, 95% CI 1.15-1.96). The 21-week GA contemporary cohort, had higher odds of antenatal corticosteroids exposure (aOR 14.3, 95% CI 6.5-31.8). For the 2013-2018 infants who survived or transferred at discharge we found 60-80% had clinical data available in Cosmos at 1 and 5 years (Table 1) with more limited diagnosis code-based neurodevelopmental outcome data at 5 years (Table 2).

Conclusion(s): In the contemporary cohort, active treatment of 21-week and 22-week infants continues to increase. Our findings are comparable to results from the NRN study and thus support the use of the Epic Cosmos database for neonatal research in the periviable population. This data source may provide advantages of broader representation and accessibility to monitor evolving current practices and outcomes, including long-term outcomes.

Figure 1: Active Treatment and Survival curves over study years
21-week GA curves were excluded due to low numbers

Table 1: Infant Characteristics and Long-Term Follow Up data
%Missing data due to unassigned or unknown Sex #Antenatal Corticosteroids data based on maternal linked data available
*Any Presence in cosmos based on the Survived at discharge denominator

Table 2: 5-year Neurodevelopmental Outcomes Available in Cosmos
Data were excluded due to low numbers (ie. 21 week outcomes and Bilateral and Unilateral Blindness)