Milken Institute School of Public Health Poster Presentations (Marvin Center & Video)
Evaluating methods for utilizing time loss data in sports settings using a sample of U.S. collegiate soccer-related injury observations
Poster Number
74
Document Type
Poster
Status
Graduate Student - Doctoral
Abstract Category
Exercise and Nutrition Sciences
Keywords
Injury Epidemiology, Time loss, Severity, Random Effects
Publication Date
Spring 2018
Abstract
Background: Time loss has featured heavily in assessments of sports-related injury severity. Typically, it is measured as a count of days lost to injury and analyzed using ordinal cut points. We argue that a refinement of methods for the analysis of time loss which acknowledges the role of severity, is advantageous. We propose to instead model time loss with count or survival regression and adopt the view that it is a manifestation of injury severity, which is a latent variable. Inclusion of a random intercept in the model enables representation of latent injury severity as an unobservable predictor of time loss and admits an interesting, clinically relevant interpretation of observable covariate effects as being ‘severity-adjusted.’
Methods: Using a sample of U.S. collegiate soccer-related injury observations, we fit random effects Poisson and Weibull Regression models to perform ‘severity-adjusted’ evaluations of time loss.
Results: Injury site, injury mechanism and injury history emerged as the strongest predictors in our sample. In comparing random effects and fixed effects models, we noted that the incorporation of the random effect attenuated associations between most observed covariates and time loss, and model fit statistics revealed that the random effects models improved model fit over the fixed effects models.
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.
Open Access
1
Evaluating methods for utilizing time loss data in sports settings using a sample of U.S. collegiate soccer-related injury observations
Background: Time loss has featured heavily in assessments of sports-related injury severity. Typically, it is measured as a count of days lost to injury and analyzed using ordinal cut points. We argue that a refinement of methods for the analysis of time loss which acknowledges the role of severity, is advantageous. We propose to instead model time loss with count or survival regression and adopt the view that it is a manifestation of injury severity, which is a latent variable. Inclusion of a random intercept in the model enables representation of latent injury severity as an unobservable predictor of time loss and admits an interesting, clinically relevant interpretation of observable covariate effects as being ‘severity-adjusted.’
Methods: Using a sample of U.S. collegiate soccer-related injury observations, we fit random effects Poisson and Weibull Regression models to perform ‘severity-adjusted’ evaluations of time loss.
Results: Injury site, injury mechanism and injury history emerged as the strongest predictors in our sample. In comparing random effects and fixed effects models, we noted that the incorporation of the random effect attenuated associations between most observed covariates and time loss, and model fit statistics revealed that the random effects models improved model fit over the fixed effects models.