Within a research setting, the efficacy of a particular prosthetic device is typically determined by comparing group mean biomechanical outcome variables (e.g., peak residual knee flexion during stance) from a cohort of participants with lower-limb amputations performing locomotor tasks when using one device versus using a different device, typically with differing design features.1–8 Findings from such research is used by clinicians to objectively inform judgments regarding the potential (dis)advantages of prescribing one prosthetic device compared with another. Interpreting the findings from research in this way is based upon the assumption that the differences in biomechanical outcome variables between prosthetic conditions are solely a result of the prosthetic components and that the observed effect would be present for any patient with a similar level of amputation, activity level, and health status as those reported in the research. However, this assumption may not necessarily be valid, and even if it is, the applicability of such research findings to a clinical setting, where decisions regarding prosthetic prescription are made on a patient by patient basis, is questionable. For example, in a cohort study comparing the efficacy of one prosthetic device versus another, it is entirely possible that group mean biomechanical outcome variables may indicate a statistically significant effect in one direction, even though some members of the group display minimal changes, or even changes in the opposing direction. On the basis of the findings from such studies, a clinician could make an evidence-based decision regarding prosthetic prescription that may have a negative outcome for a patient. Accordingly, before finalizing prosthetic prescription, clinicians will typically ask a patient to “try out” any recommended change in prosthetic device while the patient is at their clinic, with the necessary adjustments to alignment and the like being made with a view to optimizing the device's function during gait. Any differences in gait observed at the clinic when the patient switches to using the new device are expected to, and may indeed, be a result of the function of the new prosthetic device. However, it is possible that evaluation of a patient's gait made on any day can also be affected by patient-related factors (e.g., weight, physical condition, motivation) rather than solely device-related factors.
Therefore, to investigate the efficacy of carrying out such single-session evaluations, the current study undertook quantitative gait analyses of an individual with a transtibial amputation when using two different foot-ankle devices, on two separate occasions, 14 months apart. It was reasoned that if quantitative gait analysis showed that the differences observed when using one device compared with the other were in the same direction and of similar magnitude on each of the two occasions, this would indicate that device-related factors were more important than patient-related factors. This would highlight that a single-session qualitative comparison of a person with amputation's gait using two different foot-ankle devices, as typically occurs in a clinic, is a valid approach in finalizing decisions regarding prosthetic prescription and, furthermore, that prosthetic prescription decisions should not be made solely using evidence from research evaluating group mean response to using a new device.
One healthy adult man (age, 35.8 years; mass, 90.4 kg; height, 1.86 m at the time of the first data collection session) with a unilateral transtibial amputation and described as being K4 on the Medicare Scale by his prescribing clinician participated. Amputation of the right limb had been conducted, because of trauma, 9.2 years before the first data collection session. The habitual prosthetic foot device (at the time of both data collection sessions) was an Echelon VT (Chas. A Blatchford and Sons Ltd, Basingstoke, UK). The participant had used a full-contact, suction socket with silicon liner for 12 months before the first data collection session and was using the same at the time of the second data collection session. Data were recorded while the participant completed repeated walking trials using two different types of prosthetic feet. A second, identical data collection session was conducted 14 months later (at which time the participant's mass had increased by 6.0 kg). The study gained ethical approval from the University of Bradford bioethics committee, with written informed consent being obtained from the participant prior to participation.
In the present study, the protocol, trial order, laboratory setup, experimenters, and prosthetist were identical for both data collection sessions. Segmental kinematic and ground reaction force data were recorded at 200 Hz using a ten camera motion capture system (Vicon MX, Oxford, UK) and two floor-mounted force platforms (AMTI, Watertown, MA, USA), while the participant completed overground walking trials along a flat and level 8-m walkway (full details of the marker configuration used to determine segmental kinematics are reported in De Asha et al.9). The participant completed 12 walking trials at a self-selected walking speed, using each of two prosthetic foot devices (details below) with stance phase kinetic data being recorded for the intact and prosthetic limbs (six trials for each limb in each prosthetic condition; 24 trials in total).
To avoid habituation affecting the comparisons made between prosthetic feet, during both data collection sessions, the participant used foot devices (an Elan and Epirus; both Chas. A Blatchford and Sons Ltd), which were different from his currently prescribed foot. Both the Elan and Epirus feet incorporate a dynamic-response foot base with the same shape and design of heel and fore-foot keels and both have an ankle device that passively articulates during stance. In both feet, deflection of the heel and forefoot keels provides simulated “ankle” motion, with actual articulation occurring at the ankle device, which in the Elan is governed by a microprocessor -controlled hydraulic unit, while in the Epirus, it is governed by the elastic resistance offered by a rubber ball joint. The overall prosthetic length, socket, and suspension were unchanged between devices. However, as it was impossible to replicate exactly the foot alignment of the first data collection during the second, it was decided to use the same approach to obtain “optimal” alignment for each foot at each session. Thus, as is common practice clinically, foot alignment was decided upon by a mixture of feedback regarding perceived function and comfort from the participant and the expertise of the prosthetist. After each foot device had been fitted, the participant walked on it for a period of approximately 20 minutes before data collection to enable the participant to become accustomed to it.
Initial processing of marker trajectories was undertaken within Nexus software (Vicon). Marker trajectory and ground reaction force data were then exported in C3D format to Visual 3D software (C-Motion, Germantown, MD, USA), where it was then filtered using a fourth-order, zero-lag Butterworth filter with a 6 Hz cutoff, and all further processing was completed (for more details regarding data processing see De Asha et al.9). Walking speed and sagittal plane joint kinematics and kinetics for both limbs have been shown by previous research to be important outcome variables in gait of persons with unilateral amputation.10 Therefore, we assessed average walking speed; positive and negative peaks in sagittal plane joint moments and powers; peak flexion and extension at hips, knees, and ankles (dorsiflexion and plantarflexion); and the joint angle at specific gait events (e.g., initial contact, peak loading response, toe-off). Clinically meaningful differences in these biomechanical outcome variables were defined as those with an effect size difference (d) between prosthetic conditions of at least 0.4 (medium effect size).11 No inferential statistical tests were applied.
In total, there were eight biomechanical variables where clinically meaningful differences (d > 0.4) were evident between foot types during the first data collection session. These variables were walking speed, peak plantarflexion and dorsiflexion at the prosthetic ankle, residual knee loading response flexion, peak positive power during early stance at the prosthetic ankle, peak negative residual knee power during late stance, and peak stance-phase extension and flexion moments at the residual knee (direction and magnitude of differences are shown at Table 1 and Figures 1 and 2). The mean (SD) effect size difference between foot types was 0.99 (0.48). During the second data collection session, differences between foot types in the same eight variables were in the same direction and had much the same effect size, mean 0.89 (0.51), as those determined during the first session (Table 1).
The aim of the current study was to investigate the efficacy of making decisions regarding prosthetic foot prescription for an individual patient, based on a single-session comparison of his gait using two different foot-ankle devices. To fulfill this aim, we undertook quantitative gait analyses of an individual with a transtibial amputation using two different foot-ankle devices on two separate occasions, 14 months apart.
The results indicated that eight biomechanical variables had medium or large effect size differences between prosthetic foot types (d > 0.4) during the first data collection session. All eight of these variables had the same directional and comparable effect size differences between prosthetic conditions during the second data collection session, which was conducted 14 months after the initial session with the participant now older and heavier by 6 kg, indicating an increase in body mass index from 26.1 to 27.9 (implying his physical conditioning had altered). The average effect size difference between prosthetic foot types was slightly greater in the first session than the second (session 1, d = 0.99; session 2, d = 0.89), although average effect sizes were similarly “large” for both data collection sessions. Therefore, these results suggest that a single-session comparison of the gait of an individual with unilateral transtibial amputation using two different foot-ankle devices, as is typically undertaken in a clinical setting (albeit in a qualitative manner), is appropriate for identifying biomechanical differences between prosthetic devices. Hence, it is a valid type of evaluation to conduct when finalizing decisions about prosthetic prescription.
Despite an increase in participant mass of 6 kg between sessions, both feet had the same category of heel and forefoot keel stiffness on both occasions. With such an increase in body mass, absolute joint kinetic values would have increased between testing sessions, so to ensure that comparisons were valid, all joint kinetic parameters were normalized to body mass. These normalized values, which had comparable magnitudes between sessions, showed meaningful effect size differences between prosthetic foot types in both data collection sessions, suggesting the change in foot type was the main cause of such differences, not patient-related factors. These findings support our stated supposition.
Self-selected walking speed was higher during both sessions (“medium” effect sizes) when the participant used the Elan, compared with the Epirus foot device. Increased self-selected walking speed can be considered a global descriptor of improved gait function.12 Thus, speed increases when using a particular foot device provide an indication that any accompanying changes in other parameters when using the new device may also be beneficial. Thus, in the present study, the reduction in peak negative residual knee power during late stance when using the Elan compared with the Epirus foot device, for example, can be interpreted as a beneficial change.
In general, the magnitudes of effect size differences between prosthetic conditions appear to reflect relatively low intertrial variability rather than large changes to average values (see Figures 1 and 2). This is likely because of the participant, who was assessed as being K4 on the Medicare scale, having excellent gait function. Although data from such a participant may be limited in terms of its generalizability to the wider population group, such a participant would be better able to adjust to using different, nonhabitual feet than would a participant with a lower level of function. It is quite possible that the differences between prosthetic conditions observed in the present study may well be amplified in an individual with a lower Medicare classification due to them being more affected by the function and/or design features of a particular prosthetic device, compared with an individual with higher levels of function.
In conclusion, the findings of this study indicate that the direction and effect sizes of differences in biomechanical outcomes when using one prosthetic foot versus another remained more or less constant when such assessment was conducted 14 months later, even though certain participant specific parameters (e.g., body mass, age) were different between sessions. These findings support our stated (a priori) supposition and lends support to the current clinical practice of finalizing decisions regarding prosthetic prescription for an individual, based on a single-session comparison of their gait using two different devices (typically, new device compared with habitual device). Potentially, future studies could undertake a case series using the same approach as the present study to support, or refute, the findings of this study.
The authors thank Chas. A Blatchford & Sons, Ltd, for providing the prosthetic feet used during this study.
1. Zmitrewicz RJ, Neptune RR, Walden JG, et al. The effect of foot and ankle prosthetic components on braking and propulsive impulses during transtibial amputee gait. Arch Phys Med Rehabil
2. Sensinger JW, Intawachirarat N, Gard SA. Contribution of prosthetic knee and ankle mechanisms to swing-phase foot clearance. IEEE Trans Neural Syst Rehabil Eng
3. Klodd E, Hansen A, Fatone S, Edwards M. Effects of prosthetic foot forefoot flexibility on gait of unilateral transtibial prosthesis users. J Rehabil Res Dev
4. Gitter A, Czerniecki JM, DeGroot DM. Biomechanical analysis of the influence of prosthetic feet on below-knee amputee walking. Am J Phys Med Rehabil
5. Barth DG, Schumacher I, Sienko Thomas S. Gait analysis
and energy cost of below-knee amputees wearing six different prosthetic feet. J Prosthet Orthot Int
6. Underwood HA, Tokuno CD, Eng JJ. A comparison of two prosthetic feet on the multi-joint and multi-plane kinetic gait compensations in individuals with a unilateral trans-tibial amputation. Clin Biomech (Bristol, Avon)
7. Ventura JD, Klute GK, Neptune RR. The effects of prosthetic ankle dorsiflexion and energy return on below-knee amputee leg loading. Clin Biomech (Bristol, Avon)
8. De Asha AR, Munjal R, Kulkarni J, Buckley JG. Impact on the biomechanics of overground gait of using an 'Echelon' hydraulic ankle-foot device in unilateral trans-tibial and trans-femoral amputees. Clin Biomech (Bristol, Avon)
9. De Asha AR, Johnson L, Munjal R, et al. Attenuation of centre-of-pressure trajectory fluctuations under the prosthetic foot when using an articulating hydraulic ankle attachment compared to fixed attachment. Clin Biomech (Bristol, Avon)
10. Sagawa Y Jr, Turcot K, Armand S, et al. Biomechanics and physiological parameters during gait in lower-limb amputees: a systematic review. Gait Posture
11. Cohen J. Statistical Power Analysis for the Behavioral Sciences
. New York: Academic Press Inc; 1977.
12. Andriacchi TP, Ogle JA, Galante JO. Walking speed as a basis for normal and abnormal gait measurements. J Biomech