Machine Learning in Modeling High School Sport Concussion Symptom Resolve

Abstract

Introduction Concussion prevalence in sport is well recognized, so too is the challenge of clinical and return-to-play management for an injury with an inherent indeterminant time course of resolve. A clear, valid insight into the anticipated resolution time could assist in planning treatment intervention. Purpose This study implemented a supervised machine learning–based approach in modeling estimated symptom resolve time in high school athletes who incurred a concussion during sport activity. Methods We examined the efficacy of 10 classification algorithms using machine learning for the prediction of symptom resolution time (within 7, 14, or 28 d), with a data set representing 3 yr of concussions suffered by high school student-athletes in football (most concussion incidents) and other contact sports. Results The most prevalent sport-related concussion reported symptom was headache (94.9%), followed by dizziness (74.3%) and difficulty concentrating (61.1%). For all three category thresholds of predicted symptom resolution time, single-factor ANOVA revealed statistically significant performance differences across the 10 classification models for all learners at a 95% confidence interval (P = 0.000). Naïve Bayes and Random Forest with either 100 or 500 trees were the top-performing learners with an area under the receiver operating characteristic curve performance ranging between 0.656 and 0.742 (0.0–1.0 scale). Conclusions Considering the limitations of these data specific to symptom presentation and resolve, supervised machine learning demonstrated efficacy, while warranting further exploration, in developing symptom-based prediction models for practical estimation of sport-related concussion recovery in enhancing clinical decision support.

Use of artificial intelligence in sports medicine: a report of 5 fictional cases

“…Hence, a guideline-corresponding followup by AI algorithms after minutes to hours could be built into the app to enable a fast implementation from the recommendations [20]. Other studies have already shown significant differences in diagnostic capacities among different algorithms in the context of concussion [21], but data also suggested a great potential of AI diagnostic support as assisting tool to clinicians [22].…”

“…Various aspects must be critically discussed in the context of this report. One major limitation is that this case report only used one of various existing apps on the market, and the efficacy will likely vary between different algorithms [21]. In general, it has to be acknowledged that nowadays the purpose of AI-based chatbots cannot yet be seen as diagnosing complex clinical injuries or pathologies, instead is intended to give patients useful insight before getting a chance to meet or talk with a medical professional.…”

A systematic review of the effectiveness of machine learning for predicting psychosocial outcomes in acquired brain injury: Which algorithms are used and why?

“…Outcomes included post-concussive symptoms (Bergeron et al, 2019;Cnossen et al, 2017), functional outcome (Gupta et al, 2017;Nishi et al, 2019;Walker et al, 2018), indicators of mood and psychological symptoms (Hirata et al, 2016;Huttunen et al, 2016;Shafiei et al, 2017), and employment (Stromberg et al, 2019).…”

Utility of MemTrax and Machine Learning Modeling in Classification of Mild Cognitive Impairment

“…For our preliminary analysis, we employed 10 commonly used learning algorithms: 5-Nearest Neighbors, two versions of C4.5 decision tree, Logistic Regression, Multilayer Perceptron, Naïve Bayes, two versions of Random Forest, Radial Basis Function Network, and Support Vector Machine. Key attributes and contrasts of these algorithms have been described elsewhere [21] (see respective Appendix). These were chosen because they represent a variety of different types of learners and because we have demonstrated success using them in previous analyses on similar data.…”