Stronger by Science

What I Learned About Injury Rates from Surveying 1,900 Powerlifters

Last year, I sent out a survey just to get a big picture overview of the state of powerlifting.  I meant to publish the results a long time ago, but never quite got around to it.  Luckily, my friend Andrew Patton has been going through the data – specifically the data on injuries (Andrew is a part of Cartolytics, a data science and spatial analysis consulting group; if you need help with data analysis, you should definitely get in touch with him).

This article is a cross-sectional analysis of the injury data that we collected.  Unfortunately, it probably wouldn’t be publishable in a journal because I wasn’t specific enough in the way I phrased the questions (I’m not an epidemiologist; live and learn, I suppose). Still, this article should still be interesting and useful, and it gives us a starting point for a more rigorous analysis.  Andrew and I are currently collaborating on a more thorough prospective study on the injury risks of powerlifting.  We’ll keep you guys updated on the results as we’re able.

First, some background about these data:

I asked if people had ever experienced an acute injury when training.  If they had, I asked what types of injuries they’d sustained.  The strength of this approach is that it allows you to get data on less-common types of injuries (for example, if you follow 100 people for one year, you may never see a neck injury or a triceps injury).  The drawbacks are that people may not remember old injuries (general recall issues you run into with any retrospective analysis), the frequency of each injury may be off (i.e. if someone had one acute pec injury and 3 acute SI joint injuries, we know they injured both their pec and their SI joint, but we only register one person with an SI joint injury versus 3 total SI joint injuries), and we can’t quantify injuries per 1,000 hours of training to compare the injury risk of powerlifting to other sports (we know what proportion of these lifters have been injured, but not how frequently injuries occur).

Here’s the big picture overview of the current analysis:

  1. Men were much more likely to have sustained an acute injury than women; roughly ⅔ of men had sustained at least one acute injury in their training career, whereas only about ½ of women had sustained at least one acute injury.  This held true even when controlling for training age and competitive success.
  2. Unsurprisingly, people who’d been training longer were more likely to have sustained an acute injury than newer lifters.
  3. Lifters who had sustained at least one acute injury were relatively stronger (assessed via allometric scaling), competed more frequently, and were more likely to have previously dealt with a chronic injury than lifters who had never sustained an acute injury.  All of those factors were also (unsurprisingly) associated with time spent training, but were still independently predictive of having experienced an acute injury, meaning stronger lifters and lifters who competed more often were more likely to have experienced an injury regardless of training experience.

The most surprising finding of this analysis was that no training variable meaningfully predicted injury risk, including weekly training volume, per-lift training frequency, or proportion of training with loads in excess of 85% of 1RM.





If you’re interested in the complete results, I’ll let Andrew take it from here.


Context: Alteration of training style and methodology based on skill level and gender may change the risk of acute injury for competitive raw powerlifters.

Objective: To determine training-based risk factors for acute injury in competitive raw powerlifters.

Design, Setting, and Patients: Cross-sectional analysis of 1,543 responses to an online survey directed at competitive strength athletes on Facebook and Reddit. The survey was administered from April 10, 2016, to April 15, 2016. Acute injury was defined by answering in the affirmative to “Have you ever sustained an acute (sudden injury/pain) injury directly resulting from your training?”

Main Outcome Measure: Odds of having an acute injury resulting from powerlifting training while controlling for training age, sex, and absolute load.

Results: 63.9 percent of the respondents reported having an acute powerlifting-related injury during their powerlifting careers. 66.9 percent of men sustained an acute injury while 49.2 percent of women sustained an acute injury (p < 0.01). Multivariate risk factors positively associated with the odds of an acute injury included increased training age (p < 0.05), having a chronic injury (p < 0.01), increased frequency of competition (p < 0.05), and increased level of lifting ability (p < 0.05). The risk factor negatively associated with the odds of having an acute injury was being female. No significance was found with any variable associated with training style or methodology.

Conclusions: While the cross-sectional nature of the data set removes the ability to analyze temporality of exposure and outcome, it is somewhat surprising that no training methodologies were associated with the odds of injury risk. All other positive predictors are generally associated with length of training and load lifted. The cause for significant decrease in odds of having an acute injury for women was not determined and warrants further inquiry, although possible explanations include generally larger natural range of motion and better decision making.


There is a substantial body of literature regarding the injury rates (injuries/hour played) in athletics at the recreational and professional level. However, these are generally “stick and ball” or Olympic sports. The few studies that have studied strength sports have found injury rates generally comparable to team sports. However, with the explosion in the popularity of recreational strength training in the United States in the recent years, characterizing the actual potential causes for acute injury is necessary. This analysis is far easier in strength sports, and particularly powerlifting, as the training is fully quantized with discrete numbers of sets, repetitions, and load. Anecdotally, there is evidence that extreme volumes and/or extreme intensities can produce injuries, but there is little established data. This study will attempt to characterize specific demographic and training variables that are plausibly related to acute injury risk in powerlifters.


In early April 2016, survey responses were solicited from Facebook and Reddit. Approximately 1,900 responses were collected in total. Questions covered a wide range of standard demographic information and a variety of training, injury, and lifestyle-related factors as well. The outcome, acute injury, was assessed by the dichotomous answers to “have you ever sustained an acute (sudden injury/pain) injury directly resulting from your training?” Respondents were excluded from the study population if they did not record injury status, age, sex, bodyweight, or a full three-lift total. In addition, respondents were excluded if they were ever users of anabolic steroids/PEDs or recorded exclusively non-raw totals. Following application of inclusion criteria, 1,543 responses were retained for analysis.

Given the dichotomous nature of the outcome, logistic regression was used to model the relationship between predictors and odds of acute injury. Certain model parameters including sex, strength levels, amount of training over 85% of 1RM, weekly volume, and weekly frequency were chosen a priori, based on biological evidence and experience in the field. Strength levels were categorized via adjusted strength score and standardized to results from 2016 Raw Nationals to develop rankings to make sexes comparable. As an example, men and women who had ASS scores in the same percentile (when compared within sexes) were placed in the same strength category.  Full population characteristics are presented in Table 1.

Table 1: Study Population Characteristics

Following univariate analysis, training age, strength category, competition frequency, chronic injury, weekly volume, and weekly frequency were found to have significant (p < 0.05) relationships, whereas the amount of heavy training over 85% of 1RM was not significant. During the course of the univariate analysis, a spline term for training age was created at greater than two years of training. With the full model, backward and forward stepwise regression was performed, returning the same results. The goodness of fit was assessed with Pearson’s Chi Squared (p > 0.05) and found to be significant with a mean variance inflation factor of 1.08. The odds ratios resulting from the regressions are presented in Table 2.

Table 2: Regression Results


The regression results indicated that being a female was negatively associated with the presence of acute injury, whereas increased training age, increased competition frequency, increased strength ability, and the presence of a chronic injury were found to be positively associated with the presence of acute injury. Neither volume, frequency, or percent of training above 85% of 1RM were present in the final model. When comparing standardized Pearson residuals against the probability of acute injury, male outliers trended negative and female outliers trended positive. The residuals are plotted in Figure 1.

Figure 1: Standardized Pearson Residuals by Probability of Acute Injury and Sex


While interesting and somewhat counterintuitive to conventional thinking, the results from this analysis should not be overinterpreted. The issue of temporality is a major concern. The exposure in this situation, training protocol, is likely to be adjusted either temporarily or permanently post-acute injury, and the wording of the questionnaire did not capture this nuance. Furthermore, the respondents gave fairly broad answers to their definition of an acute injury from a strained muscle to severe ligament and orthopedic injuries. Nonetheless, the primary outcome of interest from this analysis was the significant differential in predicted acute injury probability by sex across all training age levels, as seen in Figure 2.

Figure 2: Predicted Probability of Acute Injury by Training Age and Sex

It appears that to some degree, developing an acute injury is merely a function of time, in that 10 years of training creates more individual repetitions and chance for a random or fluke injury. Within that gradient though, the substantial probability differences based on sex is immediately apparent. This difference is hypothesized to be caused by one of three factors. First, women tend to have greater flexibility and range of motion, which could be protective to muscle and joint injuries associated with compromised movement patterns when performing exercises. Secondly, women handle lighter loads, which results in less absolute systemic load. Lastly, women tend to make fewer exceptionally poor choices regarding weight selection in training. However, further research is needed to validate these claims.

Acute Injury Study Part II

In an attempt to create a more robust and detailed analysis, a second study is under way. It’s a prospective study, requiring participants to fill a short survey every month for one year on their training, injury status, and lifestyle. 

We’ll keep you guys updated on the results as we’re able.  Complete results should be in by October, 2018.