Identifying Acute Neuropsychiatric Events in Children and Adolescents

Friday, August 26, 2022

5:30 PM – 5:45 PM CEST

Location: C1M0

Publication Number: 91

Background: Neuropsychiatric events (NPEs) encompass a range of signs, symptoms, and diagnoses, some of which are life-threatening (e.g., psychosis, seizures, suicidal or homicidal ideation) and require emergency healthcare utilization. The association between use of certain medications and NPEs in children remains unclear, in part because clinical trials were not powered to identify these uncommon yet potentially serious outcomes. Furthermore, while several observational studies have used discharge diagnosis codes to identify pediatric NPEs, the validity of this approach is unknown.

Objectives: To develop and validate approaches to accurately identify incident pediatric NPEs requiring hospitalization using administrative data.

Methods: We performed a cross-sectional, multicenter study of children 5-17 years of age hospitalized at 2 U.S. children's hospitals with an NPE between April 1, 2016 and March 30, 2020. NPE codes were categorized by clinical significance (e.g., critical or important or unclear) and subtype (e.g., neurologic or psychiatric) based on expert opinion. The Critical category included Homicidal, Suicidal or Self Harm Ideation/Attempts; the Important category includes mood disorders, psychosis, hallucination, altered mental status, movement disorders, encephalitis/encephalopathy and seizures; and, the Unclear Significance category consisted of dizziness, headache, sleeping disorders and vision changes.
We developed and evaluated 3 NPE identification algorithms: 1) Primary or secondary NPE ICD-10 diagnosis alone; 2) NPE diagnosis plus indication that the NPE was present on admission plus non-malignancy or surgery-related primary diagnosis; and, 3) Algorithm 2 but without requiring the NPE to be present on admission. Using chart reviews as reference, we computed the positive predictive value (PPV) for each algorithm overall, and by diagnosis field position (primary or any secondary) and clinical significance.

Results: There were 1098 NPE hospitalizations included in the study. A total of 857 confirmed NPEs were identified for Algorithm 1, yielding a PPV of 0.78 (95% CI 0.76-0.80). The overall PPV was 0.89 (95% CI 0.87-0.91) for algorithm 2 (n = 846) and 0.86 (95% CI 0.83-0.88) for Algorithm 3 (n = 938). PPVs varied by diagnosis order and clinical significance. The PPV for "critical" NPEs was 0.99 (0.97-0.99) for all 3 algorithms.

Conclusions: We identified highly accurate approaches to identify NPEs in children and adolescents. Algorithm 2, utilizing specific diagnosis codes and present on admission flags, retained the highest PPV. Use of these approaches can reduce outcome misclassification in future evaluations of medication use and the risk of NPE.

Presenting Author(s)

James W. Antoon, MD, PhD, MPH

Assistant Professor of Pediatrics
Department of Pediatrics, Vanderbilt University Medical Center, Nashville, TN
Vanderbilt University Medical Center
Nashville, Tennessee, United States

Slides

Co-Author(s)

JF

James Feinstein, MD MPH

Associate Professor of Pediatrics
University of Colorado
Aurora, United States
CG

Carlos Grijalva, MD, MPH

Professor, Health Policy
Vanderbilt University Medical Center, TN, United States
YZ

Yuwei Zhu, MD, MS

Senior Associate in Biostatisitcs
Vanderbilt University Medical Center
Nashville, Tennessee, United States
ED

Emily Dickenson

Colorado Children's Hospital, United States
JS

Justine Stassun

Vanderbilt University Medical Center, United States
JJ

Jakobi Johnson

Vanderbilt University Medical Center, United States
MS

Mert Sekmen

Vanderbilt University Medical Center, United States
YT

Yasas Tanguturi, MD

Vanderbilt University Medical Center, United States
JG

James Gay, MD, MMHC

Professor of Pediatricw
Vanderbilt University Medical Center
Nashville, Tennessee, United States
DW

Derek Williams, MD, MPH

Vanderbilt University Medical Center, United States