The Dutch CrossFit Study, or How to Correctly Cite Injury Data

Photo: Derek Kusmus, CrossFit Powerhouse 040,

The Orthopaedic Journal of Sports Medicine recently published a study entitled “Injury Incidence and Patterns Among Dutch CrossFit Athletes.” The study is the result of a questionnaire, distributed to CrossFit affiliates in the Netherlands and posted on Facebook. From the 449 qualifying responses received, the authors calculated that injury incidence for athletes participating in CrossFit was 56.1 percent.

So why would we need to write a post guiding people through reading and interpreting this study? Because as with similar studies we have covered in the past, the vast majority of people don’t understand the difference between injury incidence (sometimes called injury incidence rate) and injury rate. Unfortunately, neither do the journalists, critics, and competitors of CrossFit. Sadly, researchers and the journals that publish their work also fail to make this distinction at times. The result is a flurry of social media posts and poorly researched articles claiming that CrossFit is “dangerous” when in reality the study says nothing of the sort.

So, if you plan to cite “Injury Incidence and Patterns Among Dutch CrossFit Athletes,” you should first make sure you understand the distinction between injury incidence and injury rate.

A good definition of injury rate comes from Dr. Will Hopkins, Professor of Research Design and Statistics at Auckland University of Technology. He writes, “An injury rate is the number of injuries over a period of time divided by a measure of the total exposure to sport during this period. The measure of exposure is athlete-time at risk: the total number of athletes (injured and uninjured) multiplied by their average participation time, expressed as the number of practices and/or games or as actual time spent in these activities.”

This rate is commonly expressed as X number of injuries per 1,000 hours of training.
Injury incidence, on the other hand, is a measure of how many persons in a given population have experienced an injury within a specified period of time. In the study in question, that period was the 12 months prior to answering the questionnaire.

The distinction is important because injury prevalence can’t tell us anything about how risky a particular activity is. For example, if I state that out of 100 CrossFit participants questioned, all 100 of them have sustained exactly one injury at some point in the last year, it would be tempting to assume that CrossFit must be extremely dangerous. After all, it seems safe to infer that your chances of getting injured doing CrossFit are 100 percent!

But this would be a serious error in reasoning. Why? Because we know nothing about how much time these athletes have been exposed to the program. It is possible that each of the 100 people questioned has been performing 3 hours of training, every day, for 365 days. That would mean CrossFit produces roughly one injury per 1,000 hours of exposure. That would mean it is very safe. It’s also possible that the 100 people questioned have only performed 5 minutes of CrossFit, which would mean the program is exceptionally dangerous.

If you understand the distinction between injury incidence and injury rate, you quickly see that the 56.1 percent injury incidence for athletes participating in the original study tells us nothing about the risk of CrossFit in general. The uselessness of this data for assessing CrossFit’s risk is compounded by the fact that the injury questionnaire was solicited, which incentivizes people who have been injured to respond. The authors acknowledge this phenomenon, writing, “This study has several limitations. Because we distributed our survey electronically, this may have generated a selection bias, as certain athletes might have felt more inclined to answer the survey.”

Thankfully, the authors of this study seem to have a good grasp on the differences between injury rate, incidence, and prevalence. It is encouraging that the reference list at the end of their study includes our previous blog articles on this topic. This, however, apparently doesn’t stop those looking to cast CrossFit in a negative light from drawing inaccurate conclusions about the study’s data to promote their own programs.

So did the authors calculate an injury rate? No, they didn’t. But it’s worth noting that the studies that have calculated CrossFit’s injury rate have found our program to be as safe as (or safer) than other common fitness activities. The authors are careful to acknowledge this, writing “Because we did not calculate an injury rate (injuries per 1000 hours of training) and because definitions of injury vary by study and sport, we have to be cautious when comparing injury incidence rates from our study with the results of others.”

It’s too bad businesses like Spark Fitness don’t understand these differences, but at least now you can help educate them by directing them to this article.


  1. I think there is a way from the data in the study to guess at injury rate. The accuracy ranges from “this is a guess” to “not accurate”.

    449 athletes

    33% said they train multiple times
    67% said they train once per day

    Average Training days/week


    Going on a small limb and saying each training is 1 hour of time.
    Gonna guess multiple times training per day probably averages to 2, or 2 hours.

    449*67%*3.9 * 52 ( weeks in year) = 61,000 training hours.

    449*33%*3.9*2 (multiple times per day) * 52 (weeks in year) = 60,000 training hours

    total of 121K training hours. This is probably on the high side as injured athletes wouldn’t be constantly training 52 weeks in a year.

    Now…injury total:

    184 (1 injury in year) + 56 * 2 (2 injuries) + 12 * 3.5 ( 3 or more injures…3.5 is arbitrary inflation). = 338.

    338/121 (121 THOUSAND training hours) = 2.79 injuries per 1000 Thousand training hours.

    Now this 2.79, roughly calculated, should be higher because injured athletes wouldn’t be contributing any more training hours to the grand total calculated before at 121K. But, I’m gonna guess I made up for some of that by saying each athlete workouts out 52 weeks, 4 days a week 🙂

    Here is a meta study of running injury rates.

    2.5 (injury per 1000 thousand training hours) in experienced runners
    33.3 in novice runners.

    • I thought that might be possible as well, but didn’t take the time to do the math. That looks right, and your result of 2.79 injuries per 1000 hours of training would fall right in the middle of the other published injury rates for CrossFit.

Leave a Reply to Russell Berger Cancel reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: