Predicting Concussion Outcomes

Childhood and injury go hand-in-hand. When these injuries take the form of concussions, they can have serious effects including problems with balance, feelings of fogginess, and confusion. Usually, symptoms resolve within a few weeks, but in some patients the effects of a concussion can last for months. A risk score that helps clinicians predict which of their patients are most likely to experience the long-term effects of a concussion – known as persistent postconcussion symptoms, or PPCS – could facilitate more individualized injury management. Researchers of a recently published study in JAMA set out to create just such a score.

The study enrolled over 2,500 children aged 5-18 years to develop and validate their risk prediction tool. To be included, a study participant had to present with a head injury to one of nine participating emergency departments and meet the criteria for a concussion diagnosis. A participant was categorized as having PPCS if they had at least three concussion symptoms for 28 days or more when compared to their pre-injury state.

To create their risk score, the researchers analyzed 47 variables thought to be associated with postconcussion symptoms; nine of them (age, sex, prior concussion with symptoms lasting over one week, migraine history, headache, noise sensitivity, fatigue, slow to answer questions, and difficulty standing with one foot in front of the other) were included in their final tool. When the researchers validated the tool in a separate group of patients, they found that it only did moderately well at distinguishing patients likely to experience PPCS. A measure known as the area under the curve was used to evaluate the risk score: the researchers’ tool scored a .7 (a perfect prediction tool would have scored a 1, while a tool that was no better at prediction than random guesses would score a .5). The researchers’ tool, however, was better than clinicians at predicting which patients would experience PPCS (the physicians’ prediction had an area under the curve score of .55).

More research is needed before the risk score can be used to personalize management of concussion symptoms. It’s possible that the tool’s prediction capabilities could be improved with the addition of biomarkers or neurological imaging.  Also, as the study authors note, determining whether their tool is clinically meaningful will require new studies. Though still in the formative stage, this work demonstrates the potential of using accessible data to develop tools that may ultimately tailor patient care.     


*To learn more about how Hopkins inHealth develops new clinical measurements, visit our Measurement Corps page.

*To learn about concussion treatment at Johns Hopkins, visit here.

Editor’s Note: We welcome comments, but ask that they remain civil and on-topic. We reserve the right to remove any inappropriate comments.

Join Our Mailing List:

Please enter the word you see in the image below: