WIP 105 - Optimizing genetic diagnosis in the NICU: a pooled analysis

Poster Number: WIP 105
Publication Number: WIP 105.2425

Background: Infants with genetic disorders account for a large proportion of neonatal intensive care unit (NICU) admissions, morbidity, and mortality. Accurate and timely diagnosis is important in order to aid in medical decision-making and impact clinical outcomes. Although rapid genomic sequencing (rGS, including exome or genome sequencing) has a high diagnostic yield in the NICU, questions regarding its optimal implementation and clinical utility compared to alternate approaches limit its broad application in clinical practice.

Objective: To identify the impact of rGS on diagnostic and clinical outcomes for NICU infants in a large, population-based cohort, in addition to evaluating the effect of phenotype, gestational age, and illness severity on these outcomes.

Design/Methods: This is an IRB-approved retrospective analysis of observational data pooled across five level IV NICUs, comprising all infants in these NICUs who have had genetic diagnostic evaluations in the past five to ten years. We have already collected data from two of five NICUs, comprising 2833 infants. We are in the process of assembling the final cohort that consists of clinical data including demographics, phenotypic information, and markers of illness severity. To evaluate the causal effect of rGS on our specified outcomes, we will utilize an instrumental variable approach, with year of NICU admission as the instrument, in order to minimize bias by confounding. The primary outcome is NICU length of stay and secondary outcomes include: identification of a molecular genetic diagnosis within the first year of life, age at diagnosis, total hospital length of stay, and mortality. Other covariates include specific phenotypic categories and markers of illness severity. We will also utilize binomial linear regression to analyze the effect of gestational age on the likelihood of genetic diagnosis. We anticipate that we will have completed data collection and curation from all sites by December 2023 and have completed data analysis by February 2024.