Study design & participants

We used a single-group design with assessments occurring at three time points (before the program (T1)), at midpoint (13 weeks, T2) and post-program (25-weeks, T3). Rolling admission was used, where prospective participants could begin the study at any time; thus, not all participants participated in all 25 weeks of the program.

Inclusion criteria for pSCI included age 18 years or older, a self-reported diagnosis of SCI with an injury level of C5 (ASIA A-D) or below and/or at least 4/5 elbow flexion strength bilaterally, ability to ambulate and/or propel a power or manual wheelchair independently, ability to communicate and read in English, and ability to provide transport to and from the fitness facility. Exclusion criteria for pSCI included failure to obtain medical clearance to exercise at a high intensity from a physician. Inclusion and exclusion criteria for CPs were identical to those with SCI other than SCI diagnosis.

Participants were recruited through local healthcare providers, social media, and word of mouth from enrolled participants. The program was free for participants. They received a t-shirt and water bottle for participating but no other compensation was provided. As the primary aim of this study was assessing feasibility, no power analysis was conducted. This study received approval from the Arkansas Colleges of Health Education’s Institutional Review Board (PT-2021-024) and was prospectively registered on ClinicalTrials.gov (NCT05221723). All participants provided written informed consent prior to any data collection.

HIFT programCoaches

There were two primary coaches, both licensed physical therapists. One coach with a Level 1 CrossFit certification, designed and led all exercise sessions. The other primary coach was a Board-Certified Clinical Specialist in Neurologic Physical Therapy (NCS). Intermittently, additional support was provided by another physical therapist with NCS, first year physical therapist students and/or first year osteopathic medicine students.

Facility & Equipment

Exercise sessions were held at CrossFit Fort Smith, an 8000 square-foot facility with accessible restrooms, parking, and floorplan. The facility was heated, but the only forms of cooling were two large ceiling fans and two large garage doors.

The facility contained common CrossFit equipment including squat racks, air-bikes, row ergometers, medicine balls and free weights. To increase exercise options we were able to add grip aids, lap belts, ski ergometers, bike ergometers, various ropes, lighter free weights, indoor sleds, and boxing equipment to the facility.

Exercise sessions

With one exception, group exercise sessions were held twice weekly on Tuesdays and Thursdays and were 60–75 min in duration. There were 49 sessions across 25 weeks, with a 5-week break between sessions 25 and 26. This break was required in July for safety as facility temperatures exceeded 32 °C (90 °F). Sessions were designed to meet or exceed SCI-specific exercise guidelines [18].

In general, sessions began with a 10-min structured warmup consisting of 3–5 min of participant-chosen aerobic activity, followed by a series of exercises designed to improve shoulder stability (Supplementary Appendix 1). Following the warmup, 20–60 min were spent on the workout of the day (WOD), which emphasized multi-modal, functional movements performed at a high intensity. Occasionally, a general cooldown consisting of lighter intensity exercise was performed for 5 min or less at the end of sessions, though sometimes the cooldown was passive rest.

Traditional exercise programs for pSCI prescribe a certain number of sets and repetitions for resistance exercises or a specific time for aerobic exercises. Conversely, the current study predominantly used HIFT methods common in CrossFit. These include As Many Repetitions (or rounds) As Possible (AMRAP): participants perform a series of exercises as many times as they can within a specific time frame, Every Minute On The Minute (EMOM): participants perform a certain number of repetitions for 1, 1.5, 2..etc. minutes, and when they complete the repetitions, they rest for the remainder of the time, and Repetitions (or rounds) for Time (RFT): participants complete a certain number of repetitions as fast as able [12]. Other methods included partner or team WODs: multiple participants work together to complete the prescribed exercise(s), interval training: participants exercise and rest for defined periods, and traditional training: participants perform a specific number of sets and repetitions or perform aerobic activity for a certain time or distance. The content for all exercise sessions is provided in Supplementary Appendix 2, including various exercise versions to accommodate participants with paraplegia, tetraplegia and those who were ambulatory.

To acclimate participants, the first four weeks of the program were prescribed with lower volumes (e.g., shorter AMRAPs) and subsequently increased in later sessions. Coaches provided extra guidance for participants who joined the program after the first four weeks, including monitoring exertion, helping choose appropriate intensity of exercise to ensure safety, and encouraging participants to “start low and go slow” as they became acclimated to the program. As is customary with HIFT performed in the community setting, participants were responsible for self-regulating their performance and intensity during exercise. Coaches pragmatically progressed or regressed exercises to ensure appropriate safety and intensity (e.g., increasing a weight during the start of a WOD and decreasing it later). Coaches progressed exercises for participants when the reported Rate of Perceived Exertion (RPE) was <5 out of 10 or when the participant was able to perform an exercise without any difficulty (e.g., no rest required or no change in movement speed).

Adapting & tailoring

All CPs actively participated in the exercise program. They did not physically assist pSCI as the focus of the program was HIFT for both pSCI and CPs. Occasionally during partner or team WODs, pSCI and their CPs were paired but more commonly, participants were paired by coaches based on ability levels and personalities. Because CrossFit was designed for able-bodied adults and the majority of our participants utilized wheelchairs, we did not prescribe many of CrossFit’s foundational movements (e.g., squats, push press, push jerk) and had to adapt others (e.g., deadlift, shoulder press, medicine ball clean). However, we prescribed exercises that targeted CrossFit’s ten physical skills (cardiovascular endurance, strength, flexibility, power, speed, coordination, agility, balance, accuracy, stamina), though cardiovascular endurance and muscular strength were prioritized based on SCI-exercise guidelines [18] (Supplementary Appendix 2).

Exercise programming was first tailored to manual wheelchair users and then adapted to meet the needs of powerchair users or ambulators (including all CPs) (Supplementary Appendix 2). Adapting was frequently used to ensure all participants’ safety, efficacy, and inclusion. For example, if the WOD prescribed 5 hoists, 10 medicine ball wall balls (arms only), and 15 medicine ball rotations for a manual wheelchair user, ambulators could instead perform 5 hoists, 10 medicine ball wall balls (added squat prior to throwing ball) and 15 medicine ball rotations. In contrast, powerchair users could perform 5 hoist holds (rather than grasping and releasing to hoist the weight), 10 medicine ball front raises (instead of throwing and catching the ball), and 15 medicine ball rotations. WODs were often intentionally self-limiting (e.g., AMRAP in 20 min) so that individuals with higher levels of fitness could complete a greater volume of work but all participants started and ended at the same time. We rarely used WODs that involved RFT as this drew attention to participants with more significant impairments or lower fitness levels.

Educational and motivational components of the program

Before the first exercise session, the lead coach provided education on safety with exercise, the community aspect of CrossFit, and how to use the Rate of Perceived Exertion scale to rate exercise intensity. Participants were oriented to the facility and learned how to safely use aerobic exercise equipment. Prior to each session, coaches described and demonstrated all exercises, including adapted versions for those with different abilities, including individuals with paraplegia, tetraplegia and those who were ambulatory. Select sessions ended with non-exercise components including mindfulness (sessions 13 and 20) and general nutritional advice from a registered dietician (session 11).

Participants were provided monthly newsletters (6 total), which provided education on CrossFit and its adaptive divisions, safety with exercise, local opportunities to engage in exercise, SCI-specific exercise guidelines [18], how to make a workout (with examples), and hydration with exercise. The newsletters also provided information about the coaches and select participants (e.g., where they were from, hobbies, etc.). While additional exercise was encouraged, no specific exercise was prescribed by coaches outside of sessions.

Various behavioral change techniques were used to enhance motivation and enjoyment. Self-efficacy [19] was targeted through vicarious experiences (observing coaches and other participants perform exercise), mastery experiences (the exercise program started with simple exercises at a low volume to allow early participant success), and verbal persuasion (coaches and fellow participants provided verbal encouragement and “fist bumps” during and after sessions) [20]. Social support was targeted through education on exercise performance, encouragement from coaches and fellow participants, providing all necessary equipment to safely participate in exercise, and the involvement of CPs, team and partner WODs [21]. Additionally, each session incorporated music chosen by participants.

Demographic & feasibility measures

Demographic data was collected prior to starting the exercise program. Measures of feasibility included recruitment, retention, attendance, safety, and fidelity (exercise intensity). Recruitment rate was the number of participants who underwent T1 and/or T2 assessments and participated in at least one exercise session compared to the number of prospective participants contacted. Retention rate was the number of participants who completed T2 and T3 assessments compared to the number of participants who completed T1. Attendance rate was the number of total exercise sessions attended divided by the number of available exercise sessions for each participant (i.e., 49 if a participant began the study at T1). To mimic real-word community programs, there were no attendance requirements. Rather, participants were encouraged to attend as many sessions as possible. Safety was measured as the number of adverse events that occurred as a result of the program and impacted exercise participation.

Fidelity was measured through session Rating of Perceived Exertion (session-RPE) [22]. Session-RPE was pragmatically chosen because large group data can be captured efficiently and donning/doffing heart rate monitors would increase the pre-exercise burden for both participants and coaches. RPE has been found independent of exercise mode and the level of SCI [23]. At the end of each session participants were shown a 0 to 10 RPE scale, then asked to verbally answer “how was your workout?” using the scale. Prior to beginning the study we operationally defined high-intensity as 5 (“hard”) or greater, which has been shown to correspond to anaerobic thresholds for people with and without SCI [23]. Adverse events and Session-RPE were the only participant data captured during exercise sessions.

Physical & psychosocial outcome measures

At each assessment point, physical and psychosocial data were collected. Cardiovascular endurance was measured via the 6-Min Arm Test (pSCI) [24] or 6-Min Walk Test (CPs only) [25]. Muscular strength was measured via hand-held dynamometry of the upper extremities [26] (pSCI) and/or Five Times Sit To Stand Test [27]. Walking speed was measured via the 10-Meter Walk Test [28] for both self-selected and fast speeds. Average power over one minute and anaerobic peak power (measured in watts) was measured with a ski ergometer (SkiErg®). Participants were instructed to “pull as hard and as fast as you can on each pull over the course of one minute”. Test position (i.e., standing or seated in chair/wheelchair) and damper setting were standardized across assessments.

Self-efficacy was measured via Exercise Self-Efficacy Scale (pSCI) [29] and Self-Efficacy for Exercise Scale (CPs) [30]. Social support was measured via Social Support and Exercise Survey [31]. Perceived physical function was measured via short forms of the Spinal Cord Injury-Functional Index (pSCI) [32]. Health-related quality of life was measured via short forms of the Spinal Cord Injury—Quality of Life (pSCI) [33] or Rand 36-item Short-Form 36 (SF-36) (CPs) [34].

A global rating of change (GRC) scale had participants rate their perceived level of change in weekly physical activity levels, ability to walk, push, or move fast, and ability to walk, push, or move for a long period of time compared to when they started the exercise program.

All assessments were conducted by the same trained physical therapists involved in the exercise program (RH or KH).

Data analysis

Descriptive statistics were used to present demographic and feasibility data. For physical and psychosocial outcomes, pSCI and CP data were analyzed separately. The Wilcoxon signed-rank test was used to examine within-group (pSCI or CPs) differences for those who completed T1, T2, and T3 assessments. Effect size, r, was calculated [35, 36]. Effect sizes of 0.5, 0.3, and 0.1 were considered large, medium, and small, respectively [37]. Alpha was set at ≤0.05. IBM SPSS Statistics for Macintosh, Version 26.0 (IBM SPSS, Chicago, IL) was used for data analysis.