P3.I179: Characteristics of Code Status Discordant Resuscitation in a Pediatric Quaternary Care Center

Background: While the majority of pediatric deaths occur in hospitals after withdrawal of life-sustaining therapy, roughly 1 in 5 die of failed resuscitation in what is often perceived by families to be a “bad death.” Code status discussions are important for providing goal-concordant end-of-life care and preventing undesired resuscitation, but occur infrequently in the pediatric population. We sought to investigate the incidence of discordance in code status and attempted resuscitation as well as describe the clinical and socio-demographic factors associated with code status discordant resuscitation.

Methods: This is a retrospective cross-sectional study of pediatric patients (ages 0-26 years) at a single institution who died between 2013 and 2021. Socio-demographic factors, illness characteristics, and end-of-life care features including location of final admission, location of death, consultation with palliative care and/or ethics, technology dependence, expectation of death, and attempted resuscitation were compared by documented code status and mode of death. The primary outcome of interest was code status discordant resuscitation. Code status discordance was defined as lack of resuscitation in the presence of full code status or attempted resuscitation in the presence of do-not-resuscitate (DNR) code status. Multivariable logistic regression assessed for clinical and socio-demographic factors associated with code status discordant resuscitation.

Results: Of 857 total included individuals, the majority of patients had chronic conditions (78%), were admitted to acute care (77%), and ultimately died in an ICU (84%) with technology dependence (78%). The most common mode of death was withdrawal of life-sustaining therapy (67%), followed by unsuccessful resuscitation (19%) and non-initiation/non-escalation of life-sustaining therapy (14%). Of 473 individuals with DNR code status, 19 (4%) had attempted resuscitation, and of 384 individuals with full code status, 235 (61%) were not resuscitated, for a total of 254 individuals (30%) with code status discordant resuscitation. Technology dependence and expectation of death were strongly associated with code-status discordance (Odds Ratio [OR] 3.41, 95% Confidence Interval [CI] 2.00-5.80, p< 0.0001 and OR 3.55, 95% CI 2.15-5.85, p< 0.0001 respectively) while palliative care consultation was strongly protective against code-status discordance (OR 0.45, 95% CI 0.31-0.64, p< 0.0001). Discordance occurred less frequently among individuals of color (OR 0.68, 95% CI 0.44-1.06, n.s), Hispanic ethnicity (OR 0.75, 95% CI 0.48-1.17, n.s.), and individuals with preferred language other than English (OR 0.76, 95% CI 0.49-1.20, n.s.).

Conclusion: Code status discordant resuscitation occurred in 30% of pediatric deaths in a single center. Technology dependence and expectation that death would occur were associated with more discordant care while palliative care consultation was strongly protective. Non-White/Hispanic race/ethnicity and non-English language preference were less likely to experience code status discordance. Further investigation to understand the etiology of code status discordant deaths in this population may prevent unwarranted resuscitation attempts in the future.