P2A228: Comparing Rates of Delirium Documentation for Medical versus Surgical Patients: An Observational Study at a Tertiary Care Pediatric Hospital

Background: Delirium is a common complication of critical illness. Diagnosing and documenting delirium are important given its associations with prolonged hospital stays, long-term cognitive impairment, mood disturbances and increased mortality. Despite well-validated screening tools, such as the Cornell Assessment of Pediatric Delirium (CAPD), delirium remains under-reported. This study compares rates of delirium documentation between medical and surgical patients during transitions of care within the hospital setting.

Methods: At a single, tertiary care, pediatric hospital, we identified patients ages 0-21 years old admitted to the pediatric ICU (PICU) from January 2016-August 2022 who screened positive for delirium (CAPD >9). Patients with a maximum score of 9 and who screened positive for < 2 days were excluded. We considered patients to be “medical” if they transferred from the PICU to the pediatric hospital medicine service, while surgical patients were those managed by a surgical team in the PICU and then also on the floor following transfer. We collected demographic information and characteristics of the hospital stay. Four physicians on our study team reviewed patient charts for the word “delirium” in transfer and discharge notes, as well as for an ICD-10 delirium code. Univariate analysis, using chi-square and t-tests, compared patients on medical versus surgical primary teams.

Results: We identified 772 medical patients and 467 surgical patients. Documentation of delirium was low in both groups, with 10.1% of surgical and 14.8% of medical patients having delirium documented in discharge notes (p=0.017). The same trend was observed for documentation on transfer and for ICD-10 coding (Figure 1). On average, surgical patients screening positive for delirium were older, less likely to have a developmental delay, have a lower maximum CAPD score, screen positive for delirium for fewer days and have a shorter PICU length of stay (LOS) compared to medical patients (Table 1). There was no difference in percentage of days screening positive for delirium out of total PICU LOS (67% vs 69%, p=0.51).

Conclusion: We found that medical patients were more likely than surgical patients to have delirium documented at transitions of care and recorded via ICD-10 codes. Overall, the rates of documented delirium in our study were well below prior research estimating the prevalence of delirium ranging from 20% within the general medicine PICU to as high as 65% among postoperative pediatric patients. This may reflect under-recognition and/or under-documentation of this condition. Given the long-term sequelae of delirium and risk for recurrence in future hospitalizations, more work is needed to understand the basis of these findings and further explore communication during transitions of care. Study limitations include focus on documentation only and not verbal communication, variable note templates between services, and lack of control for comorbidities that may limit the reliability of CAPD scoring.