Parent-child agreement on postconcussion symptoms in the acute postinjury period

OBJECTIVES: To evaluate parent-child agreement on postconcussion symptom severity within 48 hours of injury and examine the comparative predictive power of a clinical prediction rule when using parent or child symptom reporting. METHODS: Both patients and parents quantified preinjury and current symptoms using the Postconcussion Symptom Inventory (PCSI) in the pediatric emergency department. Two-way mixed, absolute measure intraclass correlation coefficients were calculated to evaluate the agreement between patient and parent reports. A multiple logistic regression was run with 9 items to determine the predictive power of the Predicting and Preventing Postconcussive Problems in Pediatrics clinical prediction rule when using the child-reported PCSI. Delong’s receiver operating characteristic curve analysis was used to compare the area under the curve (AUC) for the child-report models versus previously published parent-report models. RESULTS: Overall parent-child agreement for the total PCSI score was fair (intraclass correlation coefficient = 0.66). Parent-child agreement was greater for (1) postinjury (versus preinjury) ratings, (2) physical (versus emotional) symptoms, and (3) older (versus younger) children. Applying the clinical prediction rule by using the child-reported PCSI maintained similar predictive power to parent-reported PCSI (child AUC = 0.70 [95% confidence interval: 0.67–0.72]; parent AUC = 0.71 [95% confidence interval: 0.68–0.74]; P = .23). CONCLUSIONS: Overall parent-child agreement on postconcussion symptoms is fair but varies according to several factors. The findings for physical symptoms and the clinical prediction rule have high agreement; information in these domains are likely to be similar regardless of whether they are provided by either the parent or child. Younger children and emotional symptoms show poorer agreement; interviewing both the child and the parent would provide more comprehensive information in these instances.