Validity of Postconcussion Only Algorithms in Collegiate Athletes Following Sports-Related Concussion

Comparing preinjury baseline neuropsychological testing with postconcussion testing has become the standard in sports concussion management. However, limitations to this model have led to calls for the development of approaches that only require postconcussion testing. The present study tested evidence-based postconcussion algorithms using a hybrid neuropsychological test battery. "Recovered " and "Not Recovered " groups of concussed collegiate athletes were identified using base rate of impairment algorithms (Arnett et al., 2016). This yielded 145 Recovered and 27 Not Recovered athletes in each of 2 algorithms, and 140 Recovered and 32 Not Recovered athletes based on the combined algorithm. Outcome variables included postconcussion symptom factor scores (ImPACT Post-Concussion Symptom Scale (PCSS)), and indices of cognitive variability (Intra-Individual Standard Deviation and Maximum Discrepancy scores) across the 17 indices. Across algorithms, results consistently showed that, compared with the Recovered group, the Not Recovered group reported significantly higher Headache, Sleep, and Cognitive scores on PCSS factor scores; they also showed significantly greater cognitive variability. Inconsistent with predictions, the groups did not differ significantly on the Affective PCSS factor; results for the Physical factor were mixed. Sex differences were also observed, with more than twice the proportion of females falling in the Not Recovered compared with the Recovered groups; these sex differences did not result in any changes in the group results when controlled for statistically. Our study provides evidence for the validity of base rate of impairment algorithms for collegiate athletes postconcussion who have not had baseline testing.What is the significance of this article for the general public?Nearly 4 million traumatic brain injuries (TBI) are estimated to occur annually (Langlois et al., 2006), with direct and indirect costs estimated at $60 billion (Daneshvar et al., 2011), highlighting that TBI is a significant public health concern. The majority of TBIs are mild (also known as concussions), and within this, many occur in the context of sports. Determining whether or not someone has recovered from a sports-related concussion is critical, as returning someone to play too soon can exacerbate existing symptoms and increase the risk of a second concussion. Many current models favor costly and time-consuming neurocognitive and symptom report baseline testing for purposes of comparing athletes' preinjury reports and performance with postinjury variables. Beyond cost and time, baseline models have also been criticized for being susceptible to changes in motivation from pre- to postinjury, practice effects, and test-retest reliability issues, among others. In the present study, we test an algorithm derived from base rates of impairment in healthy athletes to determine whether just testing athletes postconcussion and applying an existing algorithm (without benefit of baseline) is a feasible way of identifying those who have recovered versus not recovered. We found that, compared with those athletes identified as "recovered " by our algorithm, "not recovered " athletes showed more indicators that they were in fact not recovered, supporting the validity of the algorithm. Our results suggest that application of such an algorithm could result in significant time and cost savings over existing baseline models, while still being able to identify athletes who have not sufficiently recovered from their concussions to return to play.