P4A098: Delirium Documentation During Transitions of Care: A Single Site Observational Study

Background: Critically ill pediatric patients ​are at high risk for delirium, and those who experience delirium are at risk for worse outcomes during and after hospitalization. Despite this knowledge, studies show pediatric providers frequently under-diagnose and under-document delirium. This study describes frequency of and factors associated with delirium documentation for pediatric patients at both transfer from the PICU and at hospital discharge at a single institution.

Methods: In this cross-sectional study, we identified patients ages 0-21 years-old admitted to a 30-bed PICU from 2016-2022 who screened positive for delirium via Cornell Assessment of Pediatric Delirium (CAPD) (score>9) and then transferred to the hospital medicine service at a tertiary care pediatric institution. We excluded patients with developmental delay, intractable epilepsy, and other severe psychiatric or neurological conditions. Four physicians reviewed data on demographic factors, specialty consultation, and severity and duration of delirium. Primary outcomes were frequency of delirium ICD-10 coding as well as written documentation of delirium on transfer and discharge. We used descriptive statistics and univariate analysis with chi-square and t-tests as well as Fleiss' kappa.

Results: A total of 255 patient encounters met inclusion criteria. Delirium was rarely communicated at transfer (23%, N = 59) or discharge (21%, N = 53); even fewer encounters (19%, N=49) had relevant ICD codes for delirium.

Factors associated with documenting delirium at transfer included delirium ICD code for hospitalization (41% vs 13%, p= < 0.0001), older age (9.3 vs 4.0 years, p< 0.0001), and psychiatry consult during PICU stay (78% vs 11%, p= < 0.0001). Maximum CAPD score, final CAPD score prior to transfer, and days of positive screen were not associated with documentation. Discharge summaries were more likely to document delirium if the patients’ transfer notes also mentioned delirium compared to patients whose transfer notes did not mention delirium (75% vs 5%, p< 0.0001).

Conclusion: Our findings suggest pediatric intensivists and hospitalists continue to under-recognize and -document delirium during transitions of care, despite a validated screening tool. This supports findings by previous studies. Psychiatry consultation was significantly associated with increased rates of communicating delirium, but severity and duration of delirium symptoms were not.

Limitations include sampling from a single institution, affecting generalizability. The CAPD is a screening test, thus our sample may include false positives. To address this, we excluded patients with conditions that may lower the CAPD's specificity. Our study could not assess verbal handoffs.

Further education for pediatric providers on delirium’s clinical signs and impacts may help them better identify this high risk population and ensure comprehensive care and longitudinal follow-up.