A new statistical methodology to investigate post intensive care syndrome in pediatric survivors and their families

Abstract

BACKGROUND AND AIM: After PICU discharge, survivors and their families not only struggle with recovery from initial critical condition, but also have to deal with acquired comorbidities or altered functioning due to hospitalization that can persist with adverse consequences. Recently PICS-p framework has been developed in children and their family offering a new lens for post-PICU outcome research. However, evidence on this syndrome remains limited. As part of a multicenter, national, longitudinal study aiming to describe PICS syndrome in PICU survivors, and their families, we will in this presentation describe advanced statistical methods to explore PICS-p with the new lens and to investigate associated factors related to child, caregiver, family and society based on Bioecological theory of Human Development. METHOD(S): This statistical methods uses multicenter, longitudinal data from PICU survivors, their main family caregivers and siblings. Statistical methods will include descriptive statistics and spaghetti plots, Structural Equation and Growth Mixture Modeling (GMM) frameworks. RESULT(S): In this presentation we discuss descriptive statistics and spaghetti plots to describe and visualize PICS trajectories in four domains of physical, cognitive, emotional, and social health, Structural Equation and Growth Mixture Modelings to investigate associated factors with different outcome trajectories of child-family dyads. CONCLUSION(S): These advanced modelings besides statistical advantages and robustness, can accommodate the complex nature of PICS-p and to find associated factors of the four outcome domains. It also handle the dyadic influence of child and their family members on each other while looking at changes over time and teasing out potential heterogeneity among the outcome trajectories.