A new tool – a clinical risk score – may help identify which children and adolescents who recently sustained a head injury are at risk for persistent postconcussion symptoms, according to a report published online March 7 in JAMA.
Approximately one-third of pediatric patients with concussion will have ongoing somatic, cognitive, psychological, and/or behavioral symptoms at 28 days, and at present, there are no tools to help predict which patients will be affected. The 5P (Preventing Postconcussive Problems in Pediatrics) study was performed to develop and validate a clinical risk score for this purpose, said Dr. Roger Zemek of Children’s Hospital of Eastern Ontario Research Institute, Ottawa, and his associates ( JAMA 2016 Mar 8;315:1014-25 ).
This prospective cohort study involved patients aged 5-17 years (median age, 12 years) who presented to one of nine Canadian pediatric emergency departments within 48 hours of sustaining a concussion.
In the derivation cohort, 510 of 1,701 participants (30%) met the criteria for persistent postconcussion symptoms (PPCS). A total of 47 possible predictive variables were assessed for their usefulness in predicting PPCS in this cohort. They were collected from demographic data, patient history, injury characteristics, physical examination, results on the Acute Concussion Evaluation Inventory and the Postconcussion Symptom Inventory , and patient/parent responses to weekly follow-up surveys during the month following the injury.
The investigators devised a clinical risk score using the nine predictors they found to be most accurate: patient age, patient gender, the presence or absence of prior concussion, migraine history, the presence or absence of current headache, sensitivity to noise, fatigue, slow responses to questions, and an abnormal tandem stance. They then selected three cutoff points to delineate PPCS risk: 0-3 points indicated low risk, 4-8 points indicated intermediate risk, and 9 or more points indicated high risk.
Treating physicians also were asked to predict the likelihood of PPCS.
In the validation cohort, 291 of 883 participants (33%) met the criteria for PPCS.
For low-risk patients, the sensitivity of the clinical risk score was 94%, the specificity was 18%, the negative predictive value was 85%, and the positive predictive value was 36%. For high-risk patients, the sensitivity of the clinical risk score was 20%, the specificity was 94%, the negative predictive value was 70%, and the positive predictive value was 60%.
In both sets of patients, the clinical risk score was significantly better than physician judgment in predicting PPCS. However, in its present form, it is only modestly accurate at distinguishing who will and who will not have the disorder. This tool could be further refined, perhaps by adding information regarding biomarkers, genetic susceptibility, or advanced neuroimaging, Dr. Zemek and his associates wrote.
“Before this score is adopted in clinical practice, further research is needed for external validation, assessment of accuracy in an office setting, and determination of clinical utility,” they concluded.
This work was supported by the Canadian Institutes of Health Research, the Ontario Neurotrauma Foundation Mild Traumatic Brain Injury Team, and the Alberta Children’s Hospital Foundation. Dr. Zemek and his associates reported having no relevant financial disclosures.