It was one of the indelible sights of the 2023 Champions League final. With a noticeable limp, Kevin De Bruyne, Manchester City’s star midfielder, crossed the touchline and into the arms of Pep Guardiola, who hugged him tight.
Unbeknown to De Bruyne, that was one of the last times he’d wear the City shirt in 2023. Two months later, in the opening match of the 2023-24 Premier League season, the Belgian reinjured his hamstring and was sidelined for the next five months.
De Bruyne returned to the pitch in January, but he highlights the ongoing crisis that has left everyone – teams, coaches, players, medical staff – searching for answers. How about fewer games? Probably not, despite the pleas of players. A modified schedule? Unlikely. What about expanding medical support teams? Sure, but that’s an option only for the wealthiest few – and a beefier staff wouldn’t necessarily mean a reduction in injuries.
The industry continues to turn towards technology for answers. In the modern era, that means one thing: AI. Clubs and national teams are turning to machine learning to help with early detection and injury management to try to get a grip on the injury crisis – and to help boost performance.
“Sport by nature is unpredictable,” Dr Brian Moore, CEO of Orreco, a bio-analytics company, told the Guardian. “New AI tools can help the world’s greatest athletes reduce their risk of illness and injury, and play for longer.”
Why is there an increase in injuries?
Premier League clubs have been hit hard this season. Multiple teams have had double-digit players miss time. Before the turn of the year, the average number of injuries per club that forced a player to miss a game was 9.8, according to data tracking from Premier Injuries, and overall injuries were up 15% compared to the last four seasons. Through their first 20 games, Newcastle lost 16 players to injuries. Fourteen of the league’s 20 clubs are dealing with six injuries or more.
At the onset of the league season, rule changes saw more accurate added time awarded for goals and goal celebrations, on-field injuries, substitutions, and time wasting, bringing the league in line with reforms across the footballing world. While the English football community was initially concerned the increased playing time would lead to more injuries, Dr Lee Herrington, athlete health lead at the UK Sports Institute and a sports injury rehabilitation specialist, says rule changes and playing time don’t necessarily correlate with the injury rates. Which is good news, given the changes are here to stay for good.
The problems date back to the start of the pandemic. During the UK’s lockdown, the interrupted schedule and at-home training deconditioned players. After the lockdowns, athletes returned to their domestic clubs and many participated in the European Championships, which extended the season to a record rate. Soon after, the following campaign started early due to the winter World Cup, so players haven’t had the opportunity to rest or rebuild their fitness.
Herrington, however, believes the increased football calendar is “an easy out”. In reality, recovery time and preparedness are the main factors.
“From a scientific perspective, the physical efforts across the whole aren’t massive so it’s about what happens between games,” Herrington said. “One consequence is players not fully recovering from their efforts leading to fatigue. But also, perhaps players are underprepared for games because there’s an overemphasis on recovery. Maybe they aren’t training long enough in preseason or with appropriate intensity in season to sufficiently cope with the schedule.”
Injury from cognitive load, or mental fatigue, is another problem. Herrington explains that the speed of the game causes players to think faster for extended periods of time, alongside pressure from the media and fan expectations. The accumulation of cognitive load and fatigue perhaps leaves players more vulnerable to injury.
Still, there’s a degree of luck. In a contact sport, injuries happen. All the training and planning in the world cannot prevent from feeling the aftershocks of a dodgy tackle.
What are the game’s most common injuries?
Hamstrings are now the league’s chief concern. According to data supplied by Premier Injuries, last season, there were 118 hamstring strains, eight hamstring surgeries and 19 incidents of tight hamstrings. This does not include injuries that went unreported. This season, those figures are climbing. “Hamstring injuries are up massively, in about the 90% range,” Jason McKenna, head of content for Premier Injuries, told the Guardian.
Adds Harrington, “other than hamstrings, the most common injuries in football are general muscle injuries, groin issues, and problems with the knees and ankles.”
In the women’s game, ACL tears are a particular problem. Star names like Leah Williamson, Beth Mead, Alexia Putellas, Ellie Carpenter and Sam Kerr have sustained ACL injuries in the past year.
Overall, women are two to eight times more likely to sustain ACL injuries than men. While alignment and anatomy might contribute to the problem, Powell says there are countless other internal and external risk factors, a combination that likely holds the key to understanding injury rates. The lack of investment in women’s sport – and specific research – remains a concern. Male athletes typically get more access to performance monitoring and injury management strategies than women do, which could contribute to higher ACL rates.
“Participation, training loads and schedules within women’s football has been given considerable focus, but it’s also important to improve our understanding of rehabilitation and recovery from injury,” says Dr Dylan Powell, assistant professor in public health and innovation at the University of Stirling.
Powell, whose research focuses on digital applications in healthcare, believes there is opportunity for more personalised monitoring and return to play strategies for both men and women.
“There is limited evidence to accurately or reliably predict injuries. Some advances have been made in stratifying those at potentially higher risk and identifying opportunities to deploy preventive strategies to minimise their risk. But, the causes of injury are complex and can encompass medical history, biomechanics, physiology, and disparities in access to training. So we are definitely still early on in our understanding,” Powell said.
Herrington, though, is encouraged that ACL rates in the women’s game will improve. “We’re talking about great footballers with great footballing brains, but they haven’t had the exposure to appropriate physical development in a way that’s prepared them for the intensity of the women’s game these days … I’m optimistic this will change with the professionalism of women’s football and not make ACL injuries so epidemic.”
How can AI help?
Every industry is grappling with the reality of the impact of AI now and in the future, and football is no exception. In general, AI has been used to gather and interpret player data, analyze performance in real-time, and has the potential to assist in athlete recovery and injury management.
AI can be used by clubs to help managers and medical staff make more informed decisions when it comes to training and matchday selection. But can the technology predict and prevent injuries?
“You can probably never predict injury, but how you use the data to understand risk and make decisions is important,” Herrington said. “There’s no such thing as absolute, so the technology can help the coaches and players make more informed choices on player selection based on the risk and the context.”
For example, AI can quickly analyze blood, urine, sweat, sleep, diet and hydration to quickly tell backroom staff the overall condition of a player. During practice or matches, AI technology worn by players can reveal how much a player has run, how fatigued their muscles are, their biomechanical movements and how much oxygen is pumping through their blood. If this information is not to the team’s liking, a player can be pulled or benched before an injury gets worse. AI can help a player on their path to recovery by providing more accurate markers on when they’re able to return to the pitch.
Virtual reality, or computer-generated simulation of a three-dimensional environment, is a technology that can also aid in player return. In football, players can replicate or practice specific movements in a controlled environment, thus testing their availability without putting them in a high-risk scenario.
“VR is a developing area that could augment our current approaches to training and recovery for players and clinicians. While there is no substitute for real world or authentic training, using VR may provide a reliable, convenient solution to replicating complex scenarios athletes typically face,” Powell said.
Orreco is one of the leaders in the space, combining AI and bio-analytics to optimize performance and accelerate recovery of elite athletes. Orreco’s statisticians and sport scientists use AI to identify risk profiles from athlete training sessions, physiological data, match performance, travel schedules, styles of play, environmental conditions and recovery times.
“There are a vast array of tools that are available to help understand how an athlete is responding to the training or games,” Moore said. “We consider internal load (blood biomarkers), GPS or camera-derived external load (distance covered, accelerations and decelerations from camera and GPS), and a player’s kinetics and kinematics (how they move).
Orreco has worked with England’s women’s national team, the Lionesses, to help the FA monitor and understand the link between menstrual cycles and player performance. Orreco also worked with the team to provide action plans and individualized reports so the athletes can make adjustments to their diet, sleep and training schedules throughout the month.
“Our team, led by Dr Georgie Bruinvels, helped support the [Lionesses’] high performance team decision making … a key focus was menstrual cycle symptom reduction. Data was monitored using our Fitrwoman app and Fitrcoach platform and personalized further with biomarker analysis,” Moore said.
That said, not all clubs, teams and players, regardless of gender or location, are open to such monitoring. Considering how new and rapid the technology is developing, there is still a level of uncertainty.
What’s the future of AI in management?
Injuries, no matter how mild or severe, have repercussions. While lost playing time is a concern, there is also an impact on an athlete’s mental health when they’re forced to take time away from the sport. “The cycle includes denial and anger before you can start recovering,” Dr Misia Gervis, a sports psychologist, told the Guardian in 2017. “Left to your own devices it can take a long time to accept what is happening. You think about where you are in your contract, where you are in your career, if a new manager might come in who doesn’t know you – will he buy someone to replace you? That’s the worst thing about it. There is so much time to think.”
Herrington believes that the injury crisis in football will “continue to have a domino effect” if artificial intelligence is not used to intervene. “What’s happening is extra load is being put on the players who aren’t injured, to cover for the ones that are,” Herrington said. “As a result those players get injured too from the demand. So generally once a club gets a few injuries it spirals into a whole lot more. Everyone could get injured and once that happens there’s no coming back from it.”
There is a limit to how much information the human brain can process at once. Herrington, Moore and Powell agree that AI can be useful in collecting and interpreting information based on thousands of factors. Then, human brains can ascribe that risk to the context of the season and make decisions accordingly.
“As humans there’s no way we could interpret all the factors like a computer can. Our individual heads would explode,” Herrington said. “But we do need some intervention to reduce this multiplier effect.”
If the football community wants to continue using the technology in the future, AI analysis must be used correctly, productively and actually show positive results – for now, the data bank is too small.
Considering that Premier League clubs hold the power (and the check book), Herrington believes they need to work together for AI to actually benefit the sport. In his eyes, there’s not enough data to make informed decisions unless the clubs share information among themselves.
“What we need is for Premier League clubs to share their injury data anonymously, for it to be put into computers and the outputs shared with everyone. I don’t know if there’s a willingness in professional sport to do that. But if they don’t, we’ll probably never get the best out of artificial intelligence,” Herrington said.
More holistic monitoring of players off the field for metrics such as sleep and physical activity, investing in the amateur game, and understanding the technology are other logical next steps.
“Premier League clubs have a real opportunity to showcase best practices and be a positive catalyst for change,” Herrington says. “We all need to understand, recognize and recalibrate how we view injuries, integrate the right tools, and begin distributing sufficient resources across the sport to remedy the situation.”
