The incidence of musculoskeletal injuries in elite football (soccer) is high. Injury prevention strategies have potential benefits to teams, such as improving player availability and therefore team performance, and reducing demand on medical and financial resources. Periodic health examination (PHE) is perceived to be key to injury prevention. Used by the majority of elite football teams, PHE consists of medical, musculoskeletal, functional and performance tests, typically evaluated during preseason and in-season periods. PHE is used to detect pathology that may be dangerous or performance limiting, to monitor rehabilitation and performance, and to measure factors thought to increase, or predict, an athlete's future injury risk. However, this latter function of PHE, and its use in informing individual injury prevention strategies, is contentious. Through eight papers, this thesis explores the value of PHE for injury prediction in elite football. The first paper describes the conceptual differences between aetiology and injury prediction, and argues that the related concept of prognosis is more relevant to PHE in practice. The second paper is a systematic review which demonstrates the serious limitations of the current evidence. Only age and a history of a previous hamstring injury have a confirmed association with future hamstring injuries, and no previous prognostic prediction models have been developed for any spinal or lower extremity injuries in elite players. The third paper is an audit of routinely collected injury and PHE data available at Manchester United Football Club, where index indirect muscle injury (I-IMI) outcomes affecting the lower extremity and sixty candidate prognostic factors are shown to be eligible for inclusion in prognosis research. The fourth and fifth papers investigate whether data recorded with inertial measurement units (IMUs) during two routinely used PHE tests at the Club could provide eligible candidate prognostic factors. These papers demonstrate that kinematic evaluation of single leg squats and kinetic analysis of running using these IMU systems are insufficiently reliable to be of value in prognostic research or for rehabilitation monitoring. The sixth paper is a protocol for a study to: 1) develop and internally validate a multivariable prognostic prediction model for individual injury risk prediction, and; 2) to explore the associations between additional PHE-derived factors and injury outcomes during a season in elite football players. The seventh and eighth papers contain the results of this study. The seventh paper shows that it is currently only possible to develop a modestly performing multivariable prognostic model with low potential for clinical utility. The eighth paper shows that while there may be a prognostic association between I-IMI occurrence and a hamstring injury greater than 12 months prior to PHE, no other factors have clear prognostic value over and above a player's age. In conclusion, PHE has no current value for injury prediction purposes and should only be considered useful for rehabilitation or performance monitoring purposes in elite football medical practice at this time.