775 MEP064 – Injury prediction with artificial intelligence: preliminary results of a prospective cohort study over an athletics season

AbstractBackground During an athletics (track and field) season, about 70% of athletes experienced at least one injury complaint justifying their prevention. One of these prevention measures could be to provide athletes with individualised feedback on their health status using e-health possibilities. Prognostic modelling approaches using machine learning techniques allow to estimate the individual athlete's injury risk. Before analysing its potential efficacy to reduce injuries, it is of interest to investigate if athletes use it. Objective To analyse the frequency of view and the level of use of the injury risk estimation feedback (I-REF) over one athletics season.Design Prospective cohort-study.Setting One athletics season of 34 weeks follow-up.Participants 117 competing athletics athletes licensed with the French Federation of Athletics.Main Outcomes Measurements The frequency of I-REF view was calculated by dividing the total number of days those athletes checked their I-REF by the overall number of days that an I-REF was generated. The level of I-REF use by athletes was calculated by the mean score over the study follow-up of the self-declared level of consideration in I-REF, ranging from 0 ('not at all') to 100 ('totally'), that athletes give to their I-REF each day. Results 117 athletes (37% women and 63% men, mean age: 34±15 years, weight: 67±11kg, height: 176±8cm, 13±11 years of athletics practice, 7±3 hours of athletics training per week) were included in this analysis. The mean frequency of I-REF generated view was 46±29%, ranging from 0 to 100. The mean I level of I-REF use was 31±32, ranging from 0 to 100.Conclusions This study reports that an individual athlete's injury risk estimation developed using machine learning is consulted by athletes on average in 50% of its availability, and its consideration to adapt the practice is on average 30/100 with variability according to athletes.