2.1. Design

Constructivist grounded theory was deemed an appropriate methodological choice for the current study, as a theory addressing the factors affecting the adoption of injury focused running technologies in recreational runners is yet to be identified. Grounded theory (GT) consists of strategies for developing theories through the analysis of qualitative data [29, 30]. It allows for the investigation of how and why people, communities or organisations experience and respond to events, challenges and problematic situations [31], and elicits rich, narrative accounts of this experience in order to generate an inductive theory [32]. Constructivist grounded theory (CGT) is similar to GT in the sense that it involves constant comparative analysis and saturation; however, CGT assumes that rather than theories being discovered as in GT, we construct theories through past and present experiences and interactions with people, perspectives and practices [32]. Constructivist grounded theory is an iterative process that follows repeated cycles of data collection and analysis to allow for continuous improvement, expansion and clarity of the emerging theory [33]. There was a need to identify both the perceived facilitators and barriers to adoption as certain factors may act in a bi-directional manner, serving as both facilitators and barriers [34, 35]. Ethical approval was granted by the local university’s Ethics Committee. The Standards for Reporting Qualitative Research [36] (S1 Table) were adhered to. A semi-structured focus group schedule was developed by the researchers, and followed an iterative process throughout the pilot study phase (S2 Table).

2.2. Participants

A purposive sample of 27 adult recreational runners were recruited. Local running clubs were contacted via email and asked to distribute research information and contact details to potential participants. Those interested then contacted the researchers. Eligible participants had to be aged between 20 and 60 years and meet Mulvad and colleagues’ [37] definition of a recreational runner: someone running at least once per week for at least six months. A minimum/maximum running volume or the use of wearable technologies was not included in the inclusion criteria.

2.3. Pilot study

A pilot study was conducted to educate and train the primary author in efficient focus group moderation techniques, and in the use of an analytical framework for analysing qualitative data. The results of the pilot study were not included in the main study results. The focus group schedule was updated throughout this pilot phase to include more open-ended questions and additional probes to include all participants in the discussion (e.g., “What do you think M#, have you any thoughts on that?”). Four male and five female recreational runners were recruited as a convenience sample, aged 25.1 years ± 2.2 years. Three separate pilot study focus groups were facilitated by the primary author, each taking place via remote video conferencing software (Zoom, version 5.7.0) and lasted 39.1 minutes ± 5.4 minutes.

2.4. Main study procedures

Prior to taking part in a focus group, participants were required to provide informed written consent and complete a short pre-focus group questionnaire. The questionnaire was used to gather demographic information, as well as details on participants’ running habits, their usage of running technologies and their experience with RRI’s (S1 Appendix). A RRI was defined as any musculoskeletal pain in the lower back/lower limbs that causes a restriction to or stoppage of running for at least seven days or three consecutive scheduled sessions, or that causes a runner to consult a healthcare professional [38]. On completion of the questionnaire, participants were contacted via email to arrange a suitable focus group time. To encourage as much interaction as possible, the focus groups were stratified to include participants of similar age, with similar running backgrounds.

Nine separate focus groups took place with 27 recreational runners (range = 2–4, median = 3 participants per group). Focus groups were moderated by the primary author and lasted 45.1 minutes ± 11.4 minutes. Each focus group began with a brief introduction to the study and the aims of the focus group were outlined (S2 Table). Participants were encouraged to speak freely and given the opportunity to ask questions throughout. Group discussion began by each participant describing the types of running technologies they use. Following this, a discussion regarding the facilitators and barriers to technology use progressed, with a specific emphasis placed on injury focused running technologies. To aid discussion of injury focused technologies, it was suggested that a hypothetical smartphone application (app) could collect both sensed data (from a sensor) and data that users would be required to input manually (potentially before and/or after a run). Participants were probed to discuss this in relation to their experience with other running- and/or health-related apps, as well as their perceived use of a new injury focused technology (i.e., the hypothetical technology described). Conversation then moved to discuss participants’ perceived risk factors for RRIs, and the metrics they deemed important to monitor for RRI risk. On the closing of the focus groups, participants were given another opportunity to ask questions and to provide further comments or statements that they felt may be important. A reflective and iterative approach was taken with regard to focus group moderation and the content of the focus group schedule. Additional probes were included in the focus group schedule and adjustments to moderation techniques, (e.g., ensuring equal speaking opportunities for all participants) were made during this data collection phase.

2.5. Data analysis

Frequencies and descriptive statistics were generated from the questionnaire responses using SPSS [version 27.0; IBM Corporation]. Focus groups were audio and video recorded using built in software available in Zoom [version 5.7.0], and transcribed verbatim by the primary author. Participants were allocated an identification number during transcription to maintain anonymity and protect their confidentiality, with responses coded by participant gender (e.g., male = M; female = F). The transcribed focus groups were coded by the primary author using NVivo (QSR International) Constant comparative analysis was conducted, initiated after transcription of the first focus group, and continued throughout the data collection phase [39], and theoretical sampling continued until data saturation was reached [40]. A coding framework was developed and updated by the primary author throughout the data collection phase, and was used in the coding of the transcribed focus groups (S3 Table). Braun and Clarke’s [41] methodology for thematic analysis was utilised during data analysis, which involved six key features: familiarisation with the data, generating initial codes, searching for themes, reviewing themes, defining and naming themes, and producing the report [41]. From the identified codes, core categories were identified, with subsequent themes and sub-themes emerging.

2.6. Trustworthiness

Multiple methods of trustworthiness were undertaken to ensure the rigorous and accurate presentation of findings. A critical friend approach was used to enhance the analytical process [42], and to establish reliability and ensure rigour of results [43]. The goal of critical friends is not to reach consensus or agree on all aspects of the findings, but rather ‘encourage reflexivity by challenging each other’s construction of knowledge’ [43, 44]. The approach also gives the opportunity for researchers to explore multiple interpretations of the data, reducing the effect of researcher bias [42, 45]. After all transcripts had been coded by the primary author, a percentage of transcripts were coded by an external researcher with experience in qualitative research (SOK). Taking a critical friend approach, researchers (AL and SOK) met on multiple occasions to conduct a coding consistency check on the coded transcripts. Codes, sub-themes, themes and core categories were critically reviewed and discussed. A high level of agreement was reached, while any disagreements during the analysis were discussed, with varying interpretations presented. This stage of analysis led to the development of some additional codes, as well as the merging of existing codes.

Following this, trustworthiness was further enhanced through investigator triangulation, in which the primary author met with two other members of the research team (KM and EW). Similar approaches were taken to review and discuss the coded data, with any disagreements discussed and appropriate changes made.

Additionally, in the presentation of the representative and accurate findings, multiple examples and direct quotations from transcripts are provided (S4 Table), indicating a broad and diverse contribution from participants during focus groups, reducing the chance of individual bias [46]. Included quotations were agreed upon by researchers.

3.1. Metrics perceived as important for monitoring RRI risk

Three core categories of risk factors were identified as important for monitoring with injury focused running technologies: overtraining, training-related risk factors, and individual-related risk factors. Within each core category, various themes and sub-themes emerged (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Running-related injury risk factors perceived as important to monitor using wearable technology devices by recreational runners.

https://doi.org/10.1371/journal.pone.0265475.t002

3.2. Facilitators to the use of injury focused running technologies

Ease of use and receiving useful feedback were identified as core categories of facilitators to the use of injury focused running technologies (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Facilitators and barriers to the use of injury-focused wearable technologies.

https://doi.org/10.1371/journal.pone.0265475.t003

3.3. Barriers to the use of injury focused running technologies

Difficulty of use and ineffective feedback were identified as core categories of barriers to the use of injury focused running technologies (Table 3).

4.1. Metrics important for monitoring RRI risk

The broad range metrics perceived as important to monitor for RRI risk highlights participants’ awareness of the multifactorial aetiology associated with RRI’s, as shown by multiple systematic reviews [47–50]. Overtraining, consisting of excessive loading and inadequate recovery, was perceived as a leading risk factor for the development of RRI’s in the current study, in line with current knowledge about RRIs [51] and similar to the perceptions of recreational runners in previous studies [28, 52]. Also similar to the findings of Clermont and colleagues [8], the current participants identified longer distances and higher intensity sessions to be important metrics to monitor for excessive load, and subsequent injury risk. Additionally, inadequate recovery (which included the sub-themes of fatigue and poor sleep, insufficient rest days, and poor nutrition) was also perceived to contribute to overtraining. As in similar research, the importance of sleep and food intake for injury prevention have previously been reported by recreational runners [8]. Overtraining, as reported by participants of the current study, also maps to the biomechanical model of injury, whereby loading of tissues beyond their adaptive capability, combined with insufficient recovery, results in injury [53, 54].

Participants identified the importance of monitoring certain training-related metrics for risk of RRI’s. Terrain received significant attention as an important risk factor to monitor. While some perceived harder terrains to increase the risk of injury, there was a lack of consensus as to which type of terrain poses greater risks. Harder terrains with less deformation have been hypothesized to result in higher impact forces, increasing the risk of injury [55, 56]. However, while some individual studies have found harder surfaces to produce higher loading [55, 57, 58], other studies have not [56, 59]. Previous systematic reviews [50, 60] have not found terrain to be a significant risk factor for injury. Our participants perceived that participation in other sports (such as rugby, Gaelic football, golf and track events), increased a runner’s risk of RRIs. While it has been suggested that additional participation in other sports adds to the cumulative stress placed on the body, increasing the risk of injury [61], a prospective study found that increased weekly volume of other sport participation (i.e., concurrent training) reduced the risk of RRI’s [62]. With running technique, it was suggested that foot strike technique, cadence, and bilateral asymmetry were important to monitor, although participants did not describe how these factors influenced RRI risk. In a similar study, certain aspects of running technique (such as joint motion, ground contact time, and centre of mass motion) were the lowest ranked metrics by participants amongst a list of factors presented to them by the authors as potentially preventing RRI’s [8]. Systematic reviews and meta-analyses have been unable to identify strong justifications for the role of specific biomechanical risk factors in the onset of RRI’s [63, 64]. While foot strike technique has been suggested to be causative of RRI’s based on the increased load that some techniques produce (especially rear-foot strike [65, 66]), a systematic review concluded that there is very low evidence to suggest a relationship with RRI’s in general [67]. In relation to increased cadence, while a systematic review found that increasing cadence reduces the magnitude of key biomechanical factors (such as joint kinematics and kinetics, and whole body loading) associated with RRI’s [68], a recent systematic review and meta-analysis concluded that average cadence does not differ between injured and uninjured runners [69]. Bilateral asymmetry, which relates to differences between the left and right lower limbs, has been suggested as a risk factor for RRI’s based on the premise that because one leg is subjected to more loading, it is predisposed to injury [70, 71]. Again the literature is contrasting, with some studies finding significant limb asymmetries in injured runners both retrospectively [72] and prospectively [73] compared to uninjured runners, while some studies report no differences in asymmetry [70, 74]. No systematic review drawing an overall conclusion has been published to date. Footwear was the final sub-theme of training-related metrics identified, with perceptions that older shoe age increased injury risk. This perception may be associated with the theory that shoe cushioning decreases loading on the body [75, 76], and a decrease in cushioning capacity with extended use increases the risk of RRI’s [77, 78]. However, a recent systematic review concluded that no evidence-based recommendations could be made for shoe age and preventing RRI’s [79].

The final core category identified as important for monitoring RRI risk was individual-related factors. Ongoing ‘niggles’ were suggested as an important risk factor for RRI onset in the current study. Different from an injury, in which a runner is forced to reduce or stop training for a period of time [38], our participants’ perception of ‘niggles’ is similar to previous research where runners described ‘complaints’ as ‘small pains’ with which they can continue to run [28]. Interestingly, previous injury was only discussed in one third of focus groups, despite being found to be the strongest risk factor for further RRI’s in a recent systematic review [50]. Runners have also failed to acknowledge the importance of a previous injury as a risk factor for injury in an earlier qualitative study [52]. While this may reflect a sense of being ‘unable to change’ the occurrence of having a previous injury, it clearly should be taken into account (via an application) when monitoring for the purpose of preventing re-injury. Population characteristics, including age and BMI, were mentioned by some participants in the current study. It was perceived that older age and greater BMI increased the risk of RRI; however a recent systematic review found conflicting and inconsistent findings for both age and BMI as a risk factor for RRI in short and long-distance runners [50].

It is also important to note that some risk factors for RRI’s were not mentioned in the current study, despite being shown as potential risk factors in the literature. For example, sex was not mentioned but has received some attention in the literature. Although findings are mixed, systematic reviews have reported males [50, 80] and females [81, 82] to be at a greater risk for specific RRI’s. Additionally, monitoring ground reaction forces (peak and rate) as an indication of how hard someone strikes the ground was not mentioned by participants in the current study, but previous systematic reviews [83] and meta-analyses [63, 76] have investigated the relationship to RRI risk. While there are ‘conflicting’ [63] and ‘inconsistent’ [64] results for a relationship with RRI’s in general, high peak and rates of loading have been found to contribute to the development of specific RRI’s, such as bony and soft tissue injuries [75] and stress fractures [76].

These findings also raise the question about how runners form their opinions that a metric is a risk factor for RRIs, when the research evidence would suggest it is not a risk factor. These perceptions may be due to widely available information on popular running websites. There are many examples of low cadence [84], heel-striking [85], and harder terrains such as concrete [86] being described as risk factors for RRI’s on websites, despite a lack of supporting scientific evidence. Clearly there is a need for the scientific community to better educate runners.

These findings expand on the current evidence and report new findings in relation to the metrics deemed important by runners for monitoring RRI risk when using wearable technologies. Firstly, injury focused technologies should monitor risk factors that are deemed important by runners, where evidence-based research supports their relevance (e.g. excessive loading and inadequate recovery). The challenge for app developers is whether to include metrics that monitor risk factors that are: (i) not deemed important by runners, but research does support their relevance (e.g. previous injury), or/and (ii) that are deemed important by runners, but current research does not support their relevance (e.g. terrain and foot strike technique). In the case of the first point, the authors would strongly advocate for including factors supported by evidence-based research (e.g., previous injury), with efforts made by app developers to educate runners in potentially valuable metrics. This is important in order to improve the perceived usefulness of devices [19, 20, 21]. In the case of the second point, the inclusion of these metrics (e.g., terrain and foot strike technique) may be useful if they encourage technology adoption and uptake. This must be balanced against the additional time needed by the user to input this data, which itself may be a barrier to app adoption and continued utilization (discussed below). Also, a lack of research evidence (or mixed evidence) to support a relationship between a metric and an increased risk of a RRI does not necessarily indicate that there is no relationship, but may more reflect the inability of current research to effectively examine the relationship. For example, examining the relationship between running impact loading and injury has been predominantly limited to a one-off assessment, frequently in a laboratory environment [87]. Further research is required to further support the perceived usefulness of metrics that are not currently evidence based but were deemed important by runners of the current study. In addition, future research should include clinicians and running coaches, as their thoughts and opinions may yield further insight into the metrics deemed important for monitoring RRI risk. Development of an app which incorporates a wearable sensor (e.g. an accelerometer) to monitor impact loading and collect user input data on injury status would allow long-term and ongoing monitoring of runners in their natural environment. This would provide more precise and ecologically valid data to better explore whether a relationship does exist. The above findings are also important to coaches and clinicians in developing intervention strategies for injury prevention, where uptake and adherence by runners is improved when runner perception aligns with intervention design [21].

4.2. Difficulty/Ease of use

The first identified core category of both facilitators and barriers was in relation to the perceived difficulty and ease of use of injury focused technologies. A table of recommendations for the design focused smartphone app and wearable sensor is provided in the Supporting Information (S5 Table).

4.3. Receiving useful feedback

One final core category of facilitators identified was receiving useful feedback. Receiving relevant, useful and accurate data regarding RRI risk was identified as a facilitator, with participants describing their desire for feedback that could reduce their injury risk, monitor their rehabilitation from injury, and help them understand the mechanisms of injury. It is well understood that maintaining user interest [95, 99] and receiving useful and accurate data [90, 99] can facilitate the use of wearable technologies; while the collection and reporting of inaccurate data and irrelevant information have been suggested as barriers to use of physical activity tracking technologies [26, 88–91, 93].

In line with the TRIPP model (Stage 5: intervention context to inform implementation strategies), the successful implementation of injury prevention practices will be determined by, among other factors, the likelihood of its uptake [21]. In order to improve uptake, researchers (and those issuing injury prevention programmes) must understand why injury prevention practices may or may not be adopted by the target population and provide confidence that adoption of the intervention will reduce the likelihood of injury [21]. Additionally, the TAM [19] and UTUAT [20] models suggest that technologies are more likely to be adopted if they are perceived as both useful and easy to use. Receiving relevant feedback is one way of improving the perceived usefulness of a device, as suggested by participants of the current study.

Some participants also suggested that receiving enhanced data, specifically related to running performance, beyond what they are currently collecting would facilitate their use of injury focused technologies. In the interest of developing a useful injury focused device, these findings are particularly beneficial as they may help to improve perceived usefulness, and ultimately adoption and usage behaviour.