D25-03 Routine Delirium Screening Fails to Predict Cognitive Outcomes

Introduction Delirium is common in the ICU and is associated with adverse outcomes, including increased mortality, morbidity, and long-term cognitive deficits. It remains the strongest independent risk factor for post-ICU cognitive decline, yet it is frequently missed in routine clinical care. In our prior prospective cohort of English- and Spanish-speaking ICU patients, routine delirium screening using the Confusion Assessment Method for the ICU (CAM-ICU) demonstrated limited accuracy, particularly among Spanish-speaking patients. Because unrecognized delirium may obscure which patients are at risk for long-term cognitive decline, we sought to evaluate whether delirium identified by routine care predicts post-ICU cognitive outcomes. Methods We originally conducted a prospective cohort study in the ICU designed to evaluate the diagnostic accuracy of Usual Care delirium screening (CAM-ICU completed by bedside providers) compared to a Reference-standard (CAM-ICU completed by bilingual research staff) across language groups. We then extended this cohort to examine post-ICU cognitive outcomes 9 months after discharge. Cognitive outcomes were assessed using the Telephone Interview for Cognitive Status (TICS; range 0–41, higher = better cognition). We examined whether delirium identified by Usual Care predicted post-ICU cognitive performance using linear regression, with language-stratified analyses (English vs. Spanish) to assess differential associations by language group. Results Among 142 ICU patients enrolled in the original study (71 English-speakers, 71 Spanish-speakers), in-hospital mortality was 33%, increasing to 45% by 9 months. Of the 78 survivors, 86% (n = 66; 34 English-speakers, 32 Spanish-speakers) completed TICS evaluations. Usual Care was not associated with post-ICU cognitive outcomes (β = -1.84 [95% CI -4.6 to 0.9]; p = 0.19). In contrast, Reference-Standard was associated with significantly lower TICS scores, indicating worse cognitive outcomes as expected (β = -5.30 [95% CI -7.7 to -2.9]; p < 0.01). When examined by language group, the association between Usual Care and cognition remained non-significant in both English (β = -1.18 [95% CI -4.27 to 1.90]; p = 0.44) and Spanish-speaking patients (β = -2.39 [95% CI -6.72 to 1.94]; p = 0.27), though the effect size was larger among Spanish speakers. Conclusion Usual Care delirium screening did not predict post-ICU cognitive outcomes, with the weakest performance among Spanish-speaking patients. Missed delirium diagnoses in this group may contribute to unrecognized post-ICU cognitive impairment. These findings highlight the need for more accurate, linguistically appropriate, and objective delirium screening tools to ensure equitable detection and follow-up for all ICU survivors. This abstract is funded by: None