Predictive Models for Musculoskeletal Injury Risk: Why statistical Approach Makes All the Difference — BMJ Journal

Rhon D, Teyhen D, Collins G, Bullock G

Abstract submitted to The World Congress of Sports Physical Therapy 2022

Using the same data, compare performance between an injury prediction model categorizing predictors and one that did not, and compare selection of predictors based on univariate significance versus assessing non-linear relationships. Validation and replication of a previously developed injury prediction model in a cohort of 1466 healthy military service members followed for one year after physical performance, medical history, and sociodemographic variables were collected. The original model dichotomized 8 predictors. The second model (M2) kept predictors continuous but assumed linearity, the third model (M3) conducted non-linear transformations. The fourth model (M4) chose predictors the proper way (clinical reasoning and supporting evidence) which led to an addition of 7 additional predictors (15 predictors total), but still kept predictors dichotomized. Model performance was assessed with R2, calibration in the large, calibration slope, and discrimination. Decision curve analyses were performed with risk thresholds from 0.25 to 0.50.

Rhon, D. I., Teyhen, D. S., Collins, G. S., & Bullock, G. S. (2022). Predictive Models for Musculoskeletal Injury Risk: Why statistical Approach Makes All the Difference. BMJ Open Sport & Exercise Medicine. https://doi.org/10.1136/bmjsem-2022-001388