Table 6.

Summary of considerations and recommendations for future research.

Recommendation	Rationale
Developing preventative frameworks focusing on movement quality in risky movements	Of the 29 studies included, only four evaluated the effect of any exercise-based intervention on movement quality during potentially risky movements associated with the common ACL injury mechanism in football [59,66,70,76], while only one study has been carried out with a cost-effective tool which could be easily implemented in the football context [66]. Even though the effect of several interventions on movement quality has been performed with different team sports including football players [37], the specific nature of some exercise-based adaptations and the uniquely context in which football occur [119] could justify the need of developing more football-specific preventative programs. Furthermore, only one NRCT study incorporated TM as a strategy to mitigate risk factors of ACL injury [66] despite it is widely known the contribution of biomechanical determinants in increasing ACL loads [120], and knowing its promising results in other sports [121] and in young female players [122]. On the other hand, given the influence that neurocognition may have in ACL injuries [123], open skills tasks that evaluate athletes under football-specific neurocognitive demands (e.g., unanticipated COD) should be included in screening tests assessing effectiveness of intervention programs [123].
Improving quality of the interventions	None of the included studies had pre-registered the study protocol before to its execution, by which it may be speculated that this is not common trend in Sports Sciences. Since this pre-registration would allow to compare evaluation and data analysis finally carried out with those initially intended, it would be easier to detect risk of bias, especially that related to bias in the se-lection of reported outcomes [49]. Although problems arise from the inability of blinding athletes and care providers are sometimes unavoidable in the context of interventions carried out in football teams, others such us bias due to confounding variables in NRCT or bias due to deviations from intended interventions and unequal training volumes in RCT can be prevented with appropriate analysis performed (i.e., pre-registration and overall transparency with the research process) [50]. Randomization and concealment of allocation sequence processes should be improved and explicitly reported. By doing this, the number of low risks of bias studies (only one in the present review) would be higher and, therefore, findings more reliable [49]. Additionally, by increasing samples and/or the follow-up times, greater statistical power would be reached in the associations, especially those with ACL injury incidence [21].
Appropriate reporting of important features of the program: reliability of outcomes, SWC, compliance and supervisor	It is also suggested that reliability and SWC data are directly measured so that practical relevance of the results obtained because of an exercise-based intervention could be determined [90], and to ensure that training induced changes exceed the measurement error to increase the certainty improvements are “real”. This is extremely pertinent when researchers do not have the opportunity to utilize a CG and therefore adopt a single-arm design. In the present review, only 5/24 and 3/24 of the studies evaluating risk of injury variables have reported directly measured reliability and SWC data, respectively. Of note, it is suggested that supervisor of the interventions is specified, as it is known the potential positive influence of the quality of the feedback provided (i.e., through verbal, auditory and visual cues) [124,125,126] in the reinforcement of proper technique during anterior cruciate ligament injury prevention exercises [127]. 17/29 included studies reported who were the intervention supervisors. Additionally, despite the positive relationship that has been shown between compliance and effectiveness of exercise-based interventions targeting injury reduction [57,58], only 8/29 studies reported compliance rates. Indeed, compliance to the intervention has been shown to be a critical component of prevention programmes, as it highly determines its effectiveness [87]. Therefore, going forward, before concluding a training modality as potentially ineffective, it is central to consider the training compliance which, unfortunately, 72% of studies in this review failed to report. Thus, it is suggested to incorporating such data in future research to confirm the efficacy of injury mitigation training interventions.

ACL anterior cruciate ligament, CG control group, NRCT nonrandomized studies, RCT randomized controlled trial, SWC smallest worthwhile change.