644 - Association of Preferred Language With In-NICU Mortality in a Level IV NICU

Publication Number: 4630.644

Sally A. Muskett, Cincinnati Children's Hospital, Cincinnati, OH, United States; John Feister, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, United States; Sara Bechtel, Cincinnati Children's Hospital Medical Center, Newport, KY, United States

Background: Language-based disparities in health outcomes are well documented. Less is known about variation in hospital outcomes based on preferred language. Communication between the medical team and patient families plays complex roles in patient care and is limited in the setting of discordant language. This barrier to communication may influence mortality in the NICU through multifactorial pathways.

Objective: To determine if preferred language other than English (LOE) is associated with in-hospital mortality in a large, quaternary NICU.

Design/Methods: Retrospective analysis of all infants admitted to the Cincinnati Children's NICU in a 3 year period with length of stay >1 day. The primary outcome considered was death during the initial hospitalization. Family preferred language and sociodemographic/medical characteristics that plausibly influence mortality were extracted from the electronic medical record. Univariable and multivariable logistic regression (adjusting for significantly different covariates) was performed to calculate odds ratios (OR) of LOE versus English for in-hospital mortality.

Results: 2528 infants met inclusion criteria. The overall mortality rate was 4.2%. Mortality rate for infants of English-speaking families was 3.8% and for those of LOE families was 8.8% [unadjusted odds ratio 2.42 (1.29-4.21)] . Compared to infants of English families, infants of LOE families had similar birth weights, were of slightly older gestational age, were more likely to get surgery, more likely to have received prenatal care through a specialized fetal care center, and more likely to have a complex diagnosis (Table 1). Among infants who died, infants of LOE families were more likely to die of sepsis (Table 2). In multivariable logistic regression analysis, association between language and mortality remained statistically significant [adjusted OR 3.06 (1.32, 6.89)] (table 3).

Conclusion(s): LOE is independently associated with in-hospital mortality in infants receiving care at a large quaternary NICU. These findings warrant further a) replication at other NICUs, and b.) investigation of the potential mechanisms of language discordancy on mortality in the NICU. Efforts to improve care team communication with LOE families represent a potential intervention to address this disparity in mortality.

Table 1


Table 2


Table 3