Page 73 - Safe Wrestling

Models of Injury Occurrence, Prevention and Prediction

• Intervention: introducing concrete measures, preventive exercises,
or programmes and monitoring their effectiveness.
A good-practice example of integrating such models into the field is
the Injury-Prevention Warm-up (IPW) in grappling sports (Bell et al.,
2024), where the TIP model was used to collate the information required
for programme development. This evidence-based warm-up protocol is
presented in the following section.
Models for Predicting Injury Risk in Athletes
In addition to conceptual models of injury occurrence and prevention,
sports science has also developed quantitative models that attempt to
predict the probability that an athlete will sustain an injury within a given
period. These models are based on statistical methods, most commonly
logistic regression, and enable the calculation of individual injury risk on
the basis of selected variables (e.g., anthropometric characteristics, test
performance, injury history, and training and competition loads).

Logistic Regression Models in Elite Athletes
Several attempts to develop logistic regression equations for predicting
injuries in elite athletes have been described in the literature (Sham-
baugh et al., 1991; Serfontein, 2009). In this approach, researchers use
a range of potential predictor variables (e.g., age, body height, perfor-
mance in physical tests, muscular asymmetries, previous injuries, and
exposure time to training and competition) to construct a mathematical
model that estimates, for each athlete, the probability of sustaining an
injury within a specified period (e.g., during a season).
Some models have achieved very high predictive accuracy, particularly
in cohorts of elite athletes in which injury incidence is relatively high and
well documented (Leckey et al., 2024; Yuan et al., 2025). Such models can
be used to:
• identify athletes at above-average risk (Bahr, 2016; Bahr & Holme,
2003);
• adjust the training process (e.g., through additional preventive ex-
ercises or closer monitoring of training load) (Bourdon et al., 2017;
Gabbett, 2016);
• support decision-making regarding return to play following injury
(Creighton et al., 2010).
However, these models are generally population- and sport-specific,
as they have been developed within a particular sport, at a particular

73