Epidemiological Analysis of Shoulder Injuries Among Greek CrossFit Participants and Predictive Modeling for Shoulder Injury Incidence.

CrossFit is prescribed as "a constantly varied, high-intensity, functional movement". It is a highly motivational way of training including a wide variety of different exercises modalities, including calisthenics, gymnastics, metabolic conditioning, and weightlifting, which includes both Olympic and powerlifting movements. This training model has been demonstrated to improve 10 physical skills: cardiovascular and respiratory endurance, stamina, strength, flexibility, power, speed, coordination, agility, balance, and accuracy. Its popularity has increased in recent years with more than 200,000 athletes competing worldwide and many more participants of a lower level. These characteristics led to concerns about CrossFit's safety, accordingly many researchers began to investigate whether the injury incidence rates are higher than other sports with retrospective studies, which were conducted via online questionnaires. The purpose of this prospective cohort study is to investigate shoulder injury rate among CrossFit participants in Greece and to design and examine the efficacy of a sport-specific injury prevention program in these kinds of injuries.

One hundred and eleven CrossFit participants will be surveyed. After reading and signing informed consent form, they will be interviewed for personal data, previous injuries history and general health. Shoulders and core range of motion, shoulders and hips muscle strength, shoulder stabilizers endurance will be measured with valid and reliable instruments. Also, functional assessment will be carry out using a novel instrument which has been developed in previous stage via a pilot study. CrossFit Screening Tool had been designed as a functional assessment instrument and applied as a pilot study to 20 CrossFit athletes. Based on the results, we have selected the functional tasks which will implement the baseline measurements process.

The sample will be monitored for any shoulder injury by the researcher for the following 12 months. In the term of shoulder injury is included any new or aggravated old injury which costs the athlete's absence or urges performance modifications for tissue protection. When an injury occurs, the researcher will contact the athlete, assess him using special questionnaires and clinical examination, including valid and reliable special clinical tests. She will document the kind of injury and give him or her rehabilitation and return-to-sport guidelines.

Baseline data of the participants who experienced injury will be correlated with the corresponding data of those who did not, using specific statistical analysis methods to disclose the aetiologic parameters. According to the results, at the next stage, a predictive model for shoulder injury incidence will be created using machine learning. The researchers of the present study aim to generate a model which be able to prognosticate the athletes who are going to be injured in whichever shoulder presenting high accuracy and low overfitting.