Recruitment.

We will recruit participants who have a unilateral, transtibial amputation and have been using a LLP for at least six months. We will conduct data collections at the Veterans Affairs Puget Sound Health Care System in Seattle, WA and at the Minneapolis Veterans Affairs Health Care System in Minneapolis, MN. For participant recruitment, we will identify LLP users using electronic medical records or participant registries and will contact candidate participants to discuss their interest in participation. Additionally, we will provide recruitment materials (i.e., flyers) to local clinics to assist in recruitment efforts. LLP users who respond to these recruitment materials will be provided with information on the study and will be screened for participation eligibility.

We began initial recruitment and data collection on March 1, 2024. We estimate that we will complete participant recruitment by June 1, 2026 and complete data collection by July 31, 2026. Results will be disseminated after all data have been analyzed.

Inclusion/Exclusion criteria.

We will include participants that 1) are 18–89 years old, 2) have a unilateral, transtibial amputation, 3) are currently using a prosthesis and have used one for at least six months, 4) have a comfortably fitting prosthetic socket, 5) can walk continuously for at least two minutes at a time, and 6) are able to read and communicate in English. We will exclude participants that 1) are pregnant, 2) have an arm amputation or major amputation on the contralateral leg, and 3) weigh more than 263lbs (113 kg), due to the manufacturer weight limit of the study equipment (Humotech Caplex™ System).

Prosthetic feet.

We selected five prosthetic foot models to represent a clinically appropriate and commonly prescribed group of commercially-available (‘actual’) prosthetic feet with a range of sagittal and coronal plane stiffness properties. These foot models are from a variety of manufacturers, have different material properties (e.g., fiberglass versus carbon fiber), geometries (e.g., keel shape), and elastic components (e.g., fiber straps versus elastomeric bumpers). Consequently, these feet exhibit a range of behaviors in the sagittal and coronal planes. These feet will be used with each participant’s prescribed socket and interface/suspension.

In order to enhance the applicability of our findings, we will include participants with a range of mobility levels, and sort them into either a low mobility or a high mobility group (details below). Low mobility participants will walk on the following three prosthetic feet: A) a non-energy storage and return (non-ESR) flexible keel foot, B) a non-ESR multiaxial foot, and C) a split toe/heel ESR foot. High mobility participants will also walk on C) a split toe/heel ESR foot, as well as the following two prosthetic feet: D) an ESR multiaxial foot, and E) an extended keel ESR foot.

Multiaxial prosthetic foot emulator.

We will use a two degree-of-freedom prosthetic foot emulator (PFE)—the Humotech Caplex™ System (Humotech, Pittsburgh, PA) [25] configured with a multiaxial prosthetic foot end effector (PRO-003) which enables motion in the sagittal and coronal planes (Fig 3). The PFE has two off-board actuators (Fig 3A) that control the rotational stiffness of two ‘toes’ on the prosthetic foot end effector (Fig 3B). The PFE is adjusted to match the size, heel stiffness and weight of each actual prosthetic foot. The PFE end effector will be attached to each participant’s prescribed socket and interface/suspension and be aligned in the same manner as the actual feet (details below). In addition to walking in the laboratory with the actual feet detailed above, participants will walk with the PFE while it emulates the sagittal and coronal properties of each of the three feet (one at a time) assigned to their mobility group.

Download:

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

Fig 3. Prosthetic foot emulator.

(A) Two off-board actuators control the sagittal and coronal stiffness properties of the (B) prosthetic end effector by manipulating the torque produced by the two toes of the end effector.

https://doi.org/10.1371/journal.pone.0334497.g003

Baseline visit.

During visit 1, participants will provide written informed consent prior to any study-related activities. A certified prosthetist will then evaluate each participant, and we will collect pertinent amputation-related history and demographics information. Then we will collect self-reported measures (detailed below) and performance-based measures (detailed below), and determine participants’ comfortable, slow, and fast walking speeds while walking on a treadmill with their prescribed prosthetic foot. They will also be given the opportunity to walk with the PFE to become acclimated during visit 1.

Mobility group assignment.

During the first visit, participants will complete several performance-based measures (details below), which includes the two-minute walk test (2MWT) [26]. Participants that walk at least 375 feet (113m) will be categorized as high mobility, while those that walk less than 375 feet (113m) will be categorized as low mobility. Previous work has shown that a cutoff of 113m distinguishes between limited community ambulators (i.e., low mobility) and fully functional community ambulators (i.e., high mobility) within prosthesis users [27].

In-laboratory actual prosthetic foot ‘test-drive’ trials.

During either visit 2 or 3 (participants are randomized to the order of actual and emulated trials – see below), participants will complete walking activities with three actual prosthetic feet in a randomized order. Participants will complete all the walking activities with one foot at a time prior to switching foot models. A prosthetist will optimize the alignment of each prosthetic foot (as would be done in the clinical setting) prior to walking activities. These walking activities will include walking on a Sci-Fit AC5000 treadmill at three different speeds, at an incline, and on a cross slope (Fig 4), and walking on uneven terrain on a customized Woodway 4Front Pro motorized treadmill with custom built slats with a series of 69 rock climbing holds which equates to 29.3 rocks/m², placed in a quasi-random orientation. (Fig 5).

Download:

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

Fig 4. Cross-slope walking condition on Sci-Fit AC5000 treadmill.

Cross-slope walking activities will be completed on this treadmill with one side set higher than the other.

https://doi.org/10.1371/journal.pone.0334497.g004

Download:

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

Fig 5. Customized Woodway 4Front Pro motorized treadmill to simulate uneven terrain.

Walking activities at a comfortable walking speed will be completed on this treadmill.

https://doi.org/10.1371/journal.pone.0334497.g005

Participants can adjust their walking speeds as needed throughout the ‘test-drive’ visits, including selecting different comfortable walking speeds on the Sci-Fit and Woodway treadmills (all walking speeds will be finalized at the end of study visit 3 to ensure consistent within-participant walking speeds during biomechanical data collections). These in-laboratory activities are designed to simulate a range of environmental walking conditions often encountered in the community. Furthermore, we expect inclined walking to emphasize the mechanical behavior of the prosthetic feet in the sagittal plane, while cross-slopes walking will emphasize the coronal plane mechanical behavior of the prosthetic feet. We chose uneven ground walking to introduce terrain irregularities potentially affecting both the sagittal and coronal planes. Immediately following the walking activities, participants will rate their perceived stability, prosthetic foot preference, and perceived stiffness of each foot (Fig 1). Participants will rest as needed between each walking activity.

In-laboratory emulated prosthetic foot ‘test-drive’ trials.

Study methods will be the same as those described in the In-Laboratory Actual Prosthetic Foot ‘Test-Drive’ Trials, except for the following: 1) participants will walk with the PFE instead of the actual prosthetic feet and 2) participants will not walk on the uneven treadmill (because of geometric issues with the PFE foot pads interfacing with the rock features of the treadmill). Each participant will walk on three different emulated prosthetic feet (in a randomized order) corresponding to the three actual prosthetic feet.

Community trials.

At the end of visits 3, 4, and 5, a study prosthetist will fit participants with one of the three actual prosthetic feet in a randomized order. They will take each foot home and will be instructed to use it during regular daily activities for one week. After each week-long community trial, they will return to the laboratory for follow-up evaluation in the laboratory (visits 4, 5, and 6 respectively) at which time they will complete self-reported surveys, and performance-based measures (Fig 1). They will also complete the same walking activities as during the in-laboratory ‘test-drive’ trials while we collect biomechanical data via optical motion capture (detailed below). At the conclusion of the study (after completion of each participant’s sixth study visit activities), they will be re-fitted with their prescribed prosthetic foot.

Randomization and masking.

We will block randomize participants into different sequences of foot models that ensure each participant experiences a unique foot order across all study visits. Participants will also be randomized to either complete the actual ‘test-drive’ trials first (AE) or the emulated ‘test-drive’ trials first (EA). To mitigate participant expectation bias, we will mask participants to all foot conditions within each of the ‘test-drive’ trials and community trials. Feet will be referred to only as A, B, C, D, or E, and we will cover actual prosthetic feet with a sock to mask any identifiable markings. Data shared with our study biostatistician will be coded by foot condition (A, B, C).

Self-reported measures.

Fig 1 details at which study visits we will collect each of the following self-reported measures.

Perceived mobility will be assessed using a custom version of the Prosthetic Limb Users Survey of Mobility (PLUS-M). This custom short form was designed to improve precision in a narrow range of measurement [28] and enhance our ability to detect differences between prosthetic feet. PLUS-M items are individually calibrated, therefore scores on the custom short form are directly comparable to other PLUS-M short forms and related normative scores. PLUS-M shows evidence of reliability across the range of measurement, excellent test-retest reliability [29], and convergent construct validity in LLP users [30]. Higher scores indicate a greater level of perceived mobility.

Balance confidence will be evaluated by administering a revised version of the Activities-specific Balance Confidence (ABC) scale [31,32]. We will use the 5-point ordinal scale and scoring methods recommended by Sakakibara et al. [31] (i.e., items scored 0 (no confidence) to 4 (complete confidence), with summary scores created from average item scores). The ABC scale has demonstrated evidence of good internal consistency, test-retest reliability, and convergent construct validity in LLP users [33]. Higher ABC scores are indicative of greater balance confidence.

Perceived satisfaction and limitations to participating in social roles and activities will be assessed by administering two Patient-Reported Outcomes Measurement Information System (PROMIS) surveys, the Satisfaction with Social Roles and Activities (SSRA) Short Form 4a, and the Ability to Participate in Social Roles and Activities (APSRA) Short Form 6a [34,35], respectively. The PROMIS-APSRA and PROMIS-SSRA surveys consist of six and four questions, respectively. Both are answered on a five-point scale, with higher scores indicating fewer perceived limitations and greater satisfaction participating in social roles, respectively.

Prosthetic foot satisfaction will be assessed using a revised version of the Trinity Amputation and Prosthesis Experience Scales Satisfaction (TAPES-Sat) subscale [36,37]. This survey consists of five items that are scored on a 1–3 scale, where higher scores indicate greater satisfaction. This measure has high internal consistency in LLP users [36].

During visit 1, we will document the number of falls and near-falls participants recalled in the past year, using questions from the LLP User Fall Event Survey [38]. A fall will be defined as “a loss of balance or sudden loss of support where your body lands on the ground, floor, or another object” [39]. A near-fall will be defined as “a loss of balance where you caught yourself or recovered your balance without landing on the ground, floor, or another object” [39]. We will also assess whether participants experience any falls or near-falls during each community trial (e.g., at visit 4, we will ask whether they experienced any falls or near-falls during the week they wore the first actual foot home; Fig 1). We will also ask participants about their fear of falling, using four single-item questions [40–43].

Immediately following each walking activity, participants will rate their perceived stability (both actual and emulated) on an eleven-point scale (0–10); at the end of all walking activities with a given foot, participants will rate their overall preference for each prosthetic foot. These rating scales (see S1 File) are based on the previously published prosthetic foot preference rating scales used for this purpose [44]. Higher scores indicate greater levels of perceived stability and preference, respectively. Participants will also rate their perception of the stiffness of each prosthetic foot on a five-point ordinal rating scale (scored from 1 (extremely soft) to 5 (extremely stiff); see S1 File) [44]. Finally participants will rank their relative preference of the feet after trialing all feet (i.e., after each set of ‘test-drive’ trials (actual and emulated) and at the end of all three community trials; Fig 1).

Performance-based measures.

Fig 1 details at which study visits we will collect each of the following performance-based measures.

We will evaluate balance using the Berg Balance Scale (BBS), which is a 14-item performance test that includes static and dynamic balance activities. We will score each from 0 to 4 as determined by the participant’s ability to complete the activity. We will generate a summary score from the sum of all BBS item scores; a higher score indicates better balance. The BBS has demonstrated high inter-rater reliability in LLP users [45] and strong convergent validity with the ABC scale [5].

We will assess walking endurance using the 2MWT, a brief submaximal measure of aerobic capacity and walking performance [26]. Participants will walk at their fastest speed for two minutes on a level indoor walkway. Participants may use an assistive device or walker, if needed. We will record the total distance walked during the 2MWT; a greater distance covered indicates greater walking endurance. The 2MWT has evidence of excellent test-retest and inter-rater reliability in LLP users and is recommended for both high- and low-mobility individuals [46,47]. As detailed above, we will sort participants into mobility groups based on their baseline 2MWT score.

We will evaluate multi-directional dynamic balance using the Four Square Step Test (FSST), which assesses an individual’s ability to step forward, sideways, and backwards over low objects [48]. The layout of the FSST is a cross of PVC pipes that forms four “squares” on the ground. To perform the FSST, participants must complete a clockwise and then a counterclockwise stepping sequence through four squares formed by the PVC pipes, with both feet stepping into each square. Participants will complete two timed trials after being offered an opportunity to complete a practice trial. We will use the faster of the two times for scoring; shorter FSST times are indicative of better balance performance and lower fall risk. The FSST has good discriminant validity and reliability among people with a LLP [49] and has also shown evidence of high inter-rater and test–retest reliability among people with a LLP [50–52].

We will use the Four Stage Balance Test (4SBT), which is comprised of four standing balance tasks that increase in difficulty, to assess static balance [53]. These tasks include 1) standing with feet side-by-side, 2) standing with one foot placed in the instep of the other, 3) standing with tandem stance (i.e., one foot directly in front of the other), and 4) standing on one foot. We will time each task for up to ten seconds. If a participant is unable to stand in any of these positions for ten seconds, the test ends immediately, and we add their times on each task for their final score. An inability to complete at least three of the 4SBT tasks (i.e., total time > 30s) has been shown to be associated with an increased risk of falling [53]. The 4SBT has been used in LLP users as part of previous studies [54,55].

We will evaluate lateral balance using the Narrowing Beam Walking Test (NBWT) [56]. The NBWT is a performance-based test developed to challenge lateral balance by constraining step width and/or reducing the support surface [57]. To perform the NBWT participants will cross their arms over their trunk and walk along a narrowing beam that consists of four fixed width beam segments, each 1.83 meters in length and 3.8 cm in height. The beam segments become increasingly narrow in width, from 18.6 cm to 4.0 cm [57]. Consistent with developer guidelines, participants will attempt five trials, with each trial ending when the participant either uncrosses their arms, steps off the beam, or walks its full length, whichever comes first [57]. We will score NBWT performance based on the normalized distance walked (i.e., the mean distance walked across each trial relative to the overall beam length) during trials 3–5 [57]. Larger normalized distances are indicative of better balance performance. In LLP users, the NBWT exhibits evidence of construct validity with FSST and TUG and a moderate correlation with the ABC [56]. The NBWT also demonstrates excellent content validity, known groups construct validity, and diagnostic accuracy.

We will also use the L-Test, which is a modified version of the TUG, designed to assess many of facets of mobility [58], including transfers, walking and turning in LLP users. The L-Test setup consists of a chair and two cones forming an “L” shape: one cone placed three meters in front of the chair and the second cone seven meters to the right of the first. To complete the L-Test participants will stand from sitting in the chair and walk around the first cone, the second cone, and then the first cone again in the “L” shape” before returning to sit in the chair. Participants will each complete the L-Test three times, and the average of the three times will be their final score; a faster time indicates greater mobility. The L-Test has shown strong correlation with the TUG and 2MWT, and moderate correlation with the ABC. The L-Test has also shown evidence of high intra-rater and interrater reliability [58].

Biomechanical measures.

Throughout the walking activities during visits 4, 5, and 6 (i.e., after each one-week community trial with an actual prosthetic foot), we will collect biomechanical data. These data include kinematic data collected via motion capture using retroreflective markers on body segments and joints. Using the positional and kinematic data collected via motion capture, we will calculate the mediolateral (ML) and anteroposterior (AP) margins of stability (MoS) during each step [59,60]. We will also measure spatiotemporal gait outcomes (e.g., step width and time asymmetry). When available, we will also collect kinetic data via an in-ground instrumented split-belt AMTI treadmill, which we can use to explore secondary outcomes of lower-limb and pelvis motions, moments, and powers to better understand compensatory strategies used during different walking conditions (Fig 1).

Power analysis.

We performed power analyses for two of the primary study variables—NBWT and MoS—using pilot data collected on unilateral, transtibial prosthesis users walking with a multiaxial/split-toe prosthetic foot (n = 34) and with a solid keel foot (n = 11). Participants using the former foot walked a mean 3.4 ft farther than participants using the solid keel foot, with between and within participant standard deviation (SD) 4.1 and 2.7 ft, respectively. Using these estimates, we carried out a power analysis on the NBWT scores, we simulated 5000 datasets of NBWT with sample size equal to 40 and 50 participants, each wearing all 3 feet (within-participant design), and with 1 mean pair-wise difference in NBWT of 2 ft; smaller differences than found in the pilot data as that study estimated between-participant differences which tend to be larger. For each dataset, linear mixed effects regression was carried out as described in the methods section, and power was estimated based on a type 1 error of .01, to account for multiple comparisons. We estimate 81% and 91% power to detect differences in NBWT of 2 ft for sample sizes of 40 and 50 respectively. Using estimates obtained from Gates et al (2013) [59], we simulated MoS in participants with transtibial amputation walking on uneven surfaces, using 8.3 cm, 1.9 cm and 1.0 cm as the mean, between-participant and within-participant SDs respectively using a similar procedure as above. We estimate 98% power to estimate a pair-wise difference between prosthetic feet of 1 cm for 40 participants, and 87% power to estimate a difference of 0.7 cm for 50 participants. Accounting for potential attrition, we will aim to recruit 60 participants (30 at each site), so that even if 20% of the recruited participants drop out, we will still have enough participants for adequate statistical power.