Introduction

Running injuries are common in recreational athletes. A while ago we posted a research review that assessed the effectiveness of running adaptations for runners with patellofemoral pain. Besides patellofemoral pain, the whole lower limb and lower back can become injured while running. Most of the research evidence focuses on biomechanical injury risk factors. However, we know that injuries are multifactorial, and therefore, we should assess more than biomechanics alone. Since nearly every runner has a wearable GPS device, much data is available. Data obtained from these devices can give us valuable information on training factors, running mechanics, running performance, and history. In a study by Cloosterman et al. (2022), GPS-acquired data was found to be associated with running-related knee injuries and they found that this could be a valuable method for assessing runners in practice. Therefore, this paper wanted to investigate whether data from runners’ wearable devices could explain running injury risk besides knee injuries. This can be useful since it can help identify modifiable risk factors while allowing for individualized risk assessment.

Methods

MethodsThe current study was a prospective longitudinal study recruiting healthy runners. The primary aim of this study was to explore the feasibility and usability of GPS data in investigating the associations between training load and running-related knee injuries in recreational runners.

• To assess feasibility, the study set specific thresholds for recruitment, acceptance, adherence, and data collection. The recruitment period lasted for 47 days, and the acceptance rate was calculated as 133 out of 149 participants, which corresponds to 89%. Adherence was measured as 70%, indicating that 93 out of 133 participants completed the study requirements. Data collection was achieved for 86 out of the 93 participants, resulting in a data collection rate of 92%.

The secondary aim of the study was to explore whether the baseline data acquired from the wearable devices and questionnaires were prospectively associated with running injuries.

• Healthy recreational runners were included to assess running injury risk from runners’ wearable devices. They were between 18-45 years, they ran at least 3 times per week for a minimum of 60 minutes per week. They had to participate in running for at least the past year, to ensure they were not novice runners. They were free from pain and did not sustain a running injury in the past 6 months. Their running activities had to be their main exercise, so one criterion was that they did not participate in more than two additional forms of exercise per week besides running.

The participants were required to complete three patient-reported outcome measures (PROMs) related to their psychological health, sleep quality, and intrinsic motivation to run.

• The short Warwick-Edinburgh Mental Wellbeing Scale was filled to evaluate mental wellbeing in the past two weeks
• The brief version of the Pittsburgh Sleep Quality Index was used to measure sleep quality in the past month
• The Sport Motivation Scale-6 was used to evaluate self-determined motivation

Baseline anthropometric, biomechanical, metabolic, and training load data were extracted from their GPS wristwatch for analysis. This included:

• Weekly running frequency (days per week)
• Weekly distance (km)
• Critical power (W)
• Cadence (steps per minute)
• Ground contact time (ms)
• Stride length (m)

The acute load by distance (km) and effort (no unit) was calculated from seven days prior to enrollment and the chronic load from 28 days before enrollment. By dividing the acute load by the chronic load, the acute-to-chronic workload ratio (ACWR) was calculated. A high ACWR was defined when the value exceeded 1.5. For example, when someone ran 20 km in the past 7 days and they only ran 12.5 km in the past 28 days, this leads to an ACWR of 1.6 (since 20km/12.5km=1.6), which is then classified as high.

For the 12-week study period, participants were asked to complete a weekly injury status surveillance questionnaire. This allowed the researchers to monitor and track any running-related injuries that occurred during the study. A running-related injury was defined as an episode of pain that stopped or limited them for 3 consecutive runs or that persisted for seven days or led to the runner seeking medical advice.

Results

A total of 133 participants registered their training data, 93 completed the study and GPS data from 86 participants were obtained.

Of those participants who shared their training data, 21 participants (24%) sustained a running-related injury and 65 remained uninjured. Altogether, they covered 45231km.

No significant association was found between anthropometric data, self-determined motivation and weekly running volume or chronic load by effort and running injury risk. Neither for gender, inadequate sleep quality, high ACWR by distance or effort and subsequent running-related injury.

There was, however, a significant association between the acute load by calculated effort and subsequent running-related injury.

Questions and thoughts

There was no significant association between a high ACWR calculated by distance or effort and running injuries. However, the current study found that the acute load by calculated effort was prospectively associated with an increased running injury risk. We have to bear in mind that the primary aim of this study was to investigate the feasibility of the data collection. Yet, it seems logical when you consider other studies that elaborate on this topic such as Johnston et al. (2019). The only question to study remains whether we can use the data acquired from wearable GPS devices to analyze the association between training and running injury risk. In the meantime, it seems important to keep an eye out for sudden spikes in training load increases. Although not significant, the fact that a higher percentage of runners in the injured group had ACWR values above 1.5 compared to non-injured runners may mean something.

Running injuries were analyzed all together. There was no differentiation between acute sudden (for example a lateral ankle sprain) or acute injuries that developed gradually (such as a stress fracture). For the majority of injuries that developed gradually, I think that training history is a major determining factor. On the other hand, acute injuries often occur suddenly and may be due to surrounding factors such as traffic, visibility, terrain et cetera. Therefore, it would be interesting to follow up on this study and analyze these different types of injuries separately.

To calculate the acute load by calculated effort, the following formula can be used:

([Power]/[Critical Power]) for each second run in a session divided by 7 days. Critical power is obtained from the following equation: (([w3min] + [w9min]) / 2) * 0.90. Where w3min and w9min represent the maximum watts produced in a three- and nine-minute period during a workout, respectively.

As this is quite a calculation, I would stick to the ACWR. Even though this association was not significant, I think it can be a good help to track someone’s training over time. Keep in mind that the acute workload should not exceed the limits of the chronic workload. This was found in ultramarathon runners by Craddock et al. (2020) and marathon runners by Toresdahl et al. (2023). But importantly, an ACWR that is too low may also lead to a higher running injury risk Nakaoka et al. (2021). This study sheds light on possible associations in recreational running.

Talk nerdy to me

The injury rate in this study was calculated per 1000 kilometers, instead of 1000 hours. Therefore, you should bear this in mind when comparing to other studies on this topic that used another metric to define the incidence rate. The authors point to the possibility that this could lead to different outcomes when participants’ pace highly differed between the participants.

There was no division between the distances that someone ran throughout the study. Shorter and longer distances may lead to different types of injury.

The feasibility study was not powered to detect associations between the collected variables and running injury risk. Therefore, these results shed light on a topic that should be examined further in detail. In the meantime, these results are merely exploratory.

Workload can be defined as internal or external, based on the effort experienced by the participants and the distance covered, respectively. When someone is sick or tired, a 3-kilometer run may seem impossible (internal workload), while the external workload is really limited. Therefore, it is best to consider both when you guide a runner and refrain from using arbitrary cut-off thresholds for high workloads.

Baseline conditioning could have affected the results between the participants. Although, since the eligibility criteria required runners who ran 1 hour per week for at least 3 times per week over the last 12 months, I think that their baseline condition would be sufficient.

Take home messages

This study showed that acute load by calculated effort was associated with sustaining a subsequent running-related injury. It seems that spikes in running intensity or sudden increases in training may be related to running injury risk. However, the primary aim of the study was to investigate the feasibility of the data collection process. This leads us to remain cautious about the association between acute load and running injury risk. The authors should now conduct a prospective cohort study with an adequately powered design to fully examine this association and to investigate whether GPS data can be used. In the meantime, it seems reasonable to keep an eye out for sudden spikes in running training, since earlier studies have already warned us of these risk factors.

Reference

Neal BS, Bramah C, McCarthy-Ryan MF, Moore IS, Napier C, Paquette MR, Gruber AH. Using wearable technology data to explain recreational running injury: A prospective longitudinal feasibility study. Phys Ther Sport. 2024 Jan;65:130-136. doi: 10.1016/j.ptsp.2023.12.010. Epub 2023 Dec 30. PMID: 38181563.