Head injuries in Rugby, Part 1: Principle & high-level view

On January 1^st 2017 World Rugby announced some law clarifications and changes related to the high tackle laws.  They are part of a zero-tolerance attitude to head contact during the sport, aimed at reducing the risk of concussion.  These changes were not ‘guesswork’, but rather the result of an 8-month research project, the largest yet undertaken in the sport, to understand exactly how head injuries happen in professional rugby.

They have, of course, caused a spike in the number of high tackle penalties, yellow cards and red cards in the game, something I reviewed recently here on this site.  The most recent, and perhaps highest profile of them all, is shown in the clip below.  This is the story of how those laws were changed, and why.

Last month, two research articles on this process were published in the British Journal of Sports Medicine:

1. Video analysis of head injuries in Professional Rugby
2. Risk factors for head injuries and proposed intervention strategies

Academic journals are often not accessible to many, and so I want to use this site to translate and explain the research a bit more engagingly.

Below is a brief explanation of the process and the findings that led to the changes.  This is Part 1 of what will be a three-part series.  Part 2 will describe the characteristics of the tackle that increase the risk of head injury and Part 3 wraps it up with the recommendations of the expert group based on these findings.

Enjoy

In the sports injury field, there is a well-known model called “The Sequence of Prevention model” (van Mechelen) which describes how scientific research should be applied to reduce the risk of injury.  The model works in other situations too – it’s been applied to reducing gun violence, for instance.  You could apply it to figuring out why you get the flu every winter, or why you develop anterior knee pain on the bike every March, if you felt that way inclined.

The process must begin by understanding how often the event in question happens.  This is called the incidence, and in our case, the event is concussion.  An example of incidence would be to say “Employees miss six days of work per year due to illness”, or “Concussion happens once every two matches”.  There’s always a number of the event, plus a ‘denominator’ or time, making it a unit of frequency.

We know from various injury surveillance research studies that concussion incidence has been on the rise, and is now the most common injury in rugby.

That establishes the need to intervene to reduce the incidence.  In the last five years, rugby has introduced a number of initiatives to help identify, treat and then manage players who experience concussions, and these fall under the umbrella “secondary prevention strategies” – they detect and treat, rather than prevent (“primary prevention).

These initiatives include the HIA protocol, which gave teams access to a temporary substitution so that a player suspected of being concussed could be assessed for 10min away from the field. The HIA Protocol also identified a number of criteria that would cause a player to be immediately and permanently removed from play, the result being that players who suffered serious head knocks would not play on.

This HIA protocol has been part of successfully reducing the number of times a concussed player stays on the field.  There was a time when more than half of the players who would later be diagnosed as concussed kept playing on.  That figure is now down in the single digits.  Obviously, it needs to be zero, but it’s come down enormously, and a big part of it is the tool created to allow doctors to evaluate players away from the noise and pressure of the field.

Then there is also the Graduated Return to Play protocol, adopted to ensure that a player is not rushed back to match play too soon after concussion.

These initiatives are themselves works in progress, and are constantly being researched to look for improvements and changes that can be made in order to further protect players after concussions.

However, what we were really interested with this project in is whether the risk can be reduced. That is, prevention, rather than cure or management.  And that’s where the mechanism of Head injury becomes really important.

This is the second step of that sequence of prevention model, because until you understand exactly how the undesired event happens, how can you target your response to prevent it? That was the fundamental question addressed in this research study.

You might think it overkill to study in detail how head injuries happen.  That they’re obviously the result of impact, and so the solution should be obvious.

At a very superficial level, that’s true, but if you want to be really effective in your action to reduce the risk of concussion, then you have to be very detailed in your understanding of the mechanism of the injury.

Why?  Because whatever action is taken, there is a set of

1. Desired outcomes (in this case, preventing head injuries);
2. “Costs”. Not financial, but costs in the sense that all actions have implications – they may change the game, positively we hope, but potentially negatively.  If these changes are met with too much resistance, they are not viable, however theoretically well-meaning they may be;
3. Unintended consequences. Sometimes a change is intended to produce a positive outcome X, and it might, but it might also achieve outcome Y, which is negative.  The only way to anticipate this is to fully understand all the risks, and to avoid basing decisions on ‘gut instinct’ or intuition.

This is why the mechanism matters.  World Rugby had the capacity to do this study because of the HIA protocols that have been in place since 2013, allied to video footage of every single HIA event logged over this period.

That gave us a library of 611 HIA events across seven professional competitions – the 2015 World Cup, Super Rugby, European Cup and Challenge Cup, Pro 12, Top 14, Aviva Premiership.  Each of the 611 had good quality video footage, and we used that to address the following questions:

1. What is the highest risk activity in rugby?
2. Who is likely to experience a head injury?
3. How, specifically, does the head injury happen during tackles?

World Rugby also employs a team of full-time video analysts, one of whom was devoted to this project. He was the man who did a huge amount of the work in actually “coding” those 611 events.

What coding involves is basically labelling the events according to a set of pre-defined criteria.  To begin with, he had to label:

• Which activity caused the injury – tackle, kick contest, ruck, maul, lineout, open play etc.
• Which player got injured – by position
• When the injury happened during the match

Once that was done, we moved on to the tackle specifically.  We did this because we knew from experience and previous research that tackles are the highest risk activity in rugby.  We would confirm this in Part 1 of this study too.

So we came up with a detailed ‘coding template’ for the tackle, using previous research as a guide, and by consulting with experts from the sport, including coaches, referees and players. That coding template looked at the following aspects of tackles:

• The type of tackle
• The number of players involved in the tackle
• The preceding event
• The direction of the players involved in the tackle
• The speed of the players in the tackle
• The acceleration of players in the tackle
• The body position of players in the tackle
• Where the contact between the head of the injured player and other players took place
• Whether or not the tackle was ruled to be foul play

Our analyst, Ben Hester, got through a mountain of work, because not only did he label every one of the 611 Head injury events this way, he also analysed tackles in 20 “control” matches, where there was no head injury.  That meant another 3,160 tackles that did not cause a head injury were also labelled in this way.

Why was this necessary? Because once we know how many head injuries happen in a certain way, we need to contextualize this by comparing it to how many tackles happen that same way.  It’s the denominator I spoke of earlier, and you can think of it as exposure – how many of a given situation is a player exposed to, and then how often does that thing cause an injury?

For example, imagine you found that half of all head injuries happen during high speed tackles, but that high speed tackles make up 80% of all tackles in rugby.

That would mean a completely different thing compared to if high speed tackles make up only 10% of all tackles, but still caused half of the head injuries.  We are seeking to identify whether certain situations are disproportionately more likely to cause a head injury, because that’s where we’d put our efforts into reducing the risk.

So it’s important to have this “denominator” or a measure of exposure to a risk factor, because only then can you work out the true risk of that factor.

Ultimately, what we were aiming to do was to work out which aspects of tackles are more likely to cause a head injury. In the field, this is called “propensity”, and it’s the likelihood or risk of head injury for a given event.

Our approach would be to identify the aspects of tackles that have the highest propensity or risk to cause a head injury, and then put that set of situations before a group of rugby experts – players, coaches, referees, and ask them to help reduce the risk by making suggestions ranging from law change to education.

We knew this when we started.  We knew that solving the problem was not the domain of scientists, even though most fancy themselves to do that.  Our role was to describe the problem, understand it, and then explain it to people who knew more than we did and who could suggest ways to solve it.

So, our destination was some point in the future, when a group of expert coaches, players and officials would be brought together, to discover exactly which aspects of the sport posed the greatest risk of head injury.  They would discuss what the best strategies might be to reduce that risk.

At this point, remember that a key principle of risk reduction is that you can reduce risk in one of two ways:

1. Eliminate the risky behaviour completely.
2. Substitute the risky behaviour for something that is less risky.

As the figure below shows, we are trying to move behaviour from right to left along this spectrum of risk. Our job as researchers was to create the spectrum with research. The expert group’s job would be to use tools like law change, education, training and equipment to make the leftward shift.

1. The tackle has the risk and the volume, and risk is greatest for the TACKLER

The first set of questions we asked produced some expected, and one very surprising answer.

In the figure below, you’ll see a breakdown of the 611 HIAs resulting from each match activity, and on the right is the propensity, which is the number of head injuries for every 1000 of that match event.

So for tackles, the propensity was 1.94 Head injuries per 1000 tackles.  If you want to think about this differently, it means that a head injury happens once every 516 tackles.

Kick contests also had a relatively high propensity – 1.57 head injuries per 1000 kick contests, or once every 637 kick contests.  Then comes quite a large drop-off to the next activities – mauls, scrums, rucks and lineouts.

What is important to take into account here is not only how risky a given activity is, but also how often it happens during a match? That’s because practically, the interventions we would consider must address both aspects, and it would be inefficient to put a lot of effort into changing something that happens very rarely.

Tackles, of course, are far more common than kick contests. In fact, our analysis showed that are 26 tackles for every kick contest, and so even though a tackle and a kick contest have a relatively similar propensity, the tackle is the area of greatest concern because it is so much more frequent.

The combination of high propensity and high frequency makes the tackle responsible for a very large majority of the head injuries – 464 out of 611 head injuries, or 76%.  It’s clearly the area of focus moving forward.

What’s really interesting, and is our first unusual finding with implications for our actions later, is that the tackler is the one who is most at risk – see on the left of that figure that 335 out of those 464 tackle-induced head injuries happen to the tackler?  That’s 72%, and it means that making a tackle is 2.6 times more likely to cause a head injury than being tackled.  Other research studies looking at all injuries find either that the ball carrier is more likely to be injured, or that the risk is similar.

But for head injuries, the player who is most at risk is the one MAKING THE TACKLE.  Think about the implication of this for law change, because it’s very important.  The law almost exclusively protects the ball carrier – high tackles, tip tackles, no arms tackle – all written with the ball carrier as the object.  Meanwhile, the player more at risk is the ‘perpetrator’ of the tackle (legal or not).  We will come back to this one, because it’s central to understanding actions later.

In any event, the findings you see in the figure above tell a pretty clear story – the tackle is the most frequent and the highest risk activity, and so it is fully deserving of the more detailed analysis we did next.

And that is where we will pick it up early next week, for Part 2, which will look at the 464 tackles, and describe which are more likely to cause a head injury.  So do check in then to understand why the high tackle laws were changed.

Ross

The Science of Sport - Scientific comment and analysis of sports and sporting performance