O2059: Neighborhood Opportunity and Risk of Hospital-acquired Venous Thromboembolism Among Critically Ill Children: A Multicenter, Retrospective Study

Background: One in 50 critically ill children will develop a hospital-acquired venous thromboembolism (HA-VTE). Acquired morbidity during hospitalization for acute and chronic medical conditions has been associated with social determinants of health including neighborhood-level determinants such as access to education, community resources, affordable housing, and the presence of food insecurity. We aim to characterize and identify potential associations between the Childhood Opportunity Index and HA-VTE among critically ill children.

Methods: We performed a retrospective cohort study assessing encounters from 47 United States hospital centers participating in the Pediatric Health Information Systems (PHIS) registry. We included children < 18 years of age hospitalized in a pediatric intensive care unit (PICU) from January – December 2022 and excluded for VTE presence at time of admission. Our primary outcome was the overall Childhood Opportunity Index scores (an ordinal, parametric scale ranging from: very low, low, moderate, high, and very high) among cohorts defined by the presence of a HA-VTE. Additional HA-VTE risk factors were characterized and assessed as covariates for HA-VTE risk including the presence of a central venous catheter (CVC), invasive mechanical ventilation (IMV), treatment for an infectious disease, and patient age. Qualitative proportional data were analyzed using Chi-squared tests. Quantitative data were assessed using student’s t and Wilcoxon rank-sum tests. Candidate HA-VTE risk factors were assessed using univariate and adjusted logistic regression.

Results: Of the 112,068 PICU encounters included for study, 2,551 (2.3%) developed a HA-VTE (78.6% of which represented limb/neck deep venous thromboses and 16.5% were pulmonary emboli). The overall median COI Level was moderate (interquartile range (IQR): low to high) and rates were evenly distributed among critically ill children: very low COI – 22%, Low COI – 20%, Moderate COI – 20.5%, High COI – 19%, and Very High COI – 18.5%. There were no detectable differences in COI levels for those with and without HA-VTE. Children with a HA-VTE were older (median 6.2 [IQR:0.6,14.6] versus 3.7 [IQR:0.8,10.8] years), more commonly had a CVC (45.3% versus 9.2%), received MV (63.6% versus 30.2%), had a greater PHIS pediatric medical complexity algorithm category (very high: 88.5% versus 55.8%), and more commonly were technology dependent (46.9% versus 25.4%) (all P< 0.001). In an adjusted logistic model, greater overall COI was not associated with HA-VTE. Factors associated with HA-VTE included IMV (adjusted odds ratio [aOR]: 2.8, 95% CI: 2.5-3) CVC (aOR: 5.5, 95% CI: 5-5.9), and infectious diagnoses (aOR: 1.7, 95% CI: 1.6-1.9).

Conclusion: In this multicenter, retrospective cohort study, we found no detectable relationship between COI and acquired HA-VTE during pediatric critical illness. Future studies could consider the impact of neighborhood opportunity of post-discharge outcomes related to HA-VTE.