Variability among Practitioners in Dynamic Observational Alignment of a Transfemoral Prosthesis
Mark D. Geil, PhD
ABSTRACT
The most innovative technology in prosthetic components cannot fulfill its objective without a proper patient interface and a proper alignment. Alignment, the orientation of components with respect to one another, is an area of prosthetic practice in which quantification and repeatability are not clearly defined. Researchers have sought an optimum standard and methods for automated alignment. Before automation is sought, variability among practitioners using current methods should be assessed. This investigation determined the outcomes of the alignment of five different prosthetic practitioners given the same subject and components using kinematic and kinetic gait analysis. Differences in static alignment were quantified through instrumented gait analysis; however, these differences were relatively small. Similar small differences were noted in gait velocity and ground reaction force, and bilateral joint angles during walking were very similar. This consistency among practitioners with varying levels of experience suggests that automated alignment is probably feasible but may not be necessary.
Keywords: Prostheses, alignment, lower-limb, gait analysis
Extensive research has been focused on the design of prosthetic components. New materials and geometries are investigated for use in prosthetic feet, and novel mechanical linkages are applied to prosthetic knees. However, even the most innovative component technology cannot fulfill its objective without a proper patient interface and a proper alignment. Although socket shape and fit have progressed through the years, little attention has been paid to alignment, the position and orientation of the prosthetic components, including the socket, joint(s), and foot, with respect to one another. Alignment remains an area of prosthetic practice in which quantification and repeatability are questionable.
Procedures do exist to document alignment, but they most often require special fixtures or measuring equipment. A method of measuring alignment based on a defined socket axis system was developed at the University of Strathclyde in 1975.1 Kerr et al.2 developed a means to measure angular alignment changes. A more automated alignment jig was developed and shown to be accurate and repeatable by Sin et al.3 Other researchers have developed means to quantify and document alignment outcomes. Hansen et al.4 have described the prosthetic foot roll-over shape as a means to measure the outcome of a particular alignment and a tool to optimize and eventually automate an a priori alignment. Zahedi et al.5 also stressed a need for automated alignment, or at least a means to help the prosthetist "visualize the alignment of a prosthesis and guide him [or her] to the true optimum alignment."
Quantification and automation of alignment assumes some optimal outcome is known. Hannah et al.6 suggested that increased bilateral symmetry is the factor to optimize. Zahedi et al.7,8 favored reduction in step-to-step variability. Hansen et al.9 suggest that the prosthetic foot roll-over shape should be matched as closely as possible to that of the physiological foot. Walking is complex, and a prosthesis contains a number of alignable degrees of freedom; in addition, patients present with prostheses having a variety of components and a range of technology and mechanical properties. Few would argue that the choice of cost factors in the optimization of alignment is difficult. It is also difficult to understand the cost factors chosen (sometimes subconsciously) by the prosthetist during a dynamic alignment. The prosthetist relies on observation of gait from multiple angles, and a few direct static measurements. It could be beneficial to assess the variability of alignment outcomes by multiple practitioners, given the same patient and the same components. If little variability exists, perhaps the need for automation in alignment is less compelling, and the prosthetists might be succeeding with their present methods. On the other hand, if significant variability exists, one must ascertain the source and understand whether there is no consensus on what is optimal or the simple observation of gait does not allow one to repeatedly come close to the known optimal outcome. This investigation assessed the variability of the outcomes of the dynamic alignments of multiple prosthetists given the same transfemoral amputee and the same modular components.
METHODS
The study measured the kinematics and kinetics of the gait of a single subject walking with a transfemoral prosthesis aligned by five different prosthetic practitioners. The subject was a 36-year-old active male unilateral left transfemoral amputee 10 years post-trauma. The prosthesis included an ischial containment socket, Mauch knee, and Otto Bock 1C40 foot. The prosthesis had been used previously by the subject but was not the subject's current prosthesis. Research was approved by the Georgia Tech Institutional Review Board for research involving human subjects and informed consent was obtained.
Practitioners were recruited from the Atlanta area. Practitioners ranged widely in certification and years of experience in prosthetic practice (Table 1
), but all were active in prosthetics practice (or residency) at the time of the investigation. The subject in the study was himself a certified prosthetist, and he was included among the five participating practitioners. Data are presented with the practitioners numbered in descending order of experience, save for the subject, who is listed as P5 given the unique circumstance of self-alignment.
For each analysis, the prosthesis was completely disassembled distal to the socket, and each practitioner assembled and aligned the modular components based on whatever measurement and gait observation he or she deemed necessary. Alignment devices were available for alignment of joint rotation and translation as well as pylon length. Once each practitioner was satisfied with the alignment, the subject was allowed time to walk and become accustomed to the alignment, and an instrumented gait analysis was conducted. After the analysis, the components were disassembled and the next practitioner repeated the process.
Kinematics were measured using a six-camera Motus real-time motion analysis system (Peak Performance, Englewood, CO). After the measurement of anthropometrics, 17 spherical retroreflective markers were attached to bony and prosthesis landmarks using a modified Helen Hayes marker set as described by Geil et al.9 The subject walked along the 20-m level walkway at a comfortable, self-selected speed. Kinetics were measured with a Kistler three-dimensional force platform mounted flush with the floor at the center of the walkway (Kistler, Amherst, NY). The subject was instructed to look straight ahead while walking to avoid targeting of the force platform. The subject's starting point was varied to enable a clean footfall of one and only one foot on the force platform. Sufficient walking trials were collected to enable two clean footfalls of the contralateral limb and three clean footfalls of the prosthesis for each alignment. An additional static trial was recorded with the subject standing still on the force platform with feet pointed straight ahead and shoulder width apart. Wherever possible, markers remained on the prosthetic components upon disassembly; care was taken to return any removed markers to the same location for each analysis.
Data were filtered using a fourth-order, zero-lag Butterworth filter with an optimized cutoff frequency.10 Data were reduced and analyzed using Motus 4.1. Maximum three-dimensional ground reaction forces and mean gait velocities were averaged for each trial for each alignment. The height off the ground in meters for each pelvic marker (right and left anterior superior iliac spine and sacrum) was calculated from static trials for each alignment. Intersegmental angles were calculated as the resultant three-dimensional angle between the segments proximal and distal to the knee and ankle joints. Angles were calculated for both static trials and walking trials.
RESULTS
Instrumented gait analysis revealed quantifiable differences in the alignments assembled by different practitioners. When the markers used in the gait analysis are displayed spatially and connected by lines, differences are apparent in the frontal and sagittal planes (Figure 1
) between the alignments of two practitioners with identical placement of markers. These differences are present in a static trial and are readily quantifiable with instrumented gait analysis.
Instrumented gait analysis is demonstrated as a means to quantify changes to alignment. When static trial intersegmental angles are compared among all practitioners, P5 revealed a prosthetic knee angle similar to that of the other practitioners, but a different contralateral knee angle (Figure 2
).
The same data show that the subject was able to achieve very similar foot progression angles for static trials with each practitioner, suggesting that P5's contralateral knee angle difference is probably due to prosthetic pylon length. A comparison of pelvic marker heights in the static trials
(Figure 3
) corroborates this conclusion. P5's alignment is the only one in which the pelvis is higher on the prosthesis side than the contralateral side. Implications on function must be drawn from analysis of motion.
Mean self-selected velocity of gait showed only small differences between practitioners' alignments
(Table 2
). The alignment producing the highest mean velocity (P4) showed a gain of 0.116 m/s over the smallest (P1). Typically a faster gait will produce larger vertical ground reaction forces. The trend was apparent here, with the exception that the alignment of P4 did not produce the largest ground reaction force on either side; instead, the alignment of P5, which showed the second highest mean velocity, produced a vertical ground reaction force 19.55 N higher than P4 on the prosthesis side. This outcome may be indicative of the previously quantified increased pylon length in the alignment of P5.
Given the differences in static alignment, the investigators were surprised to note that differences in intersegmental angles during gait were small (Figure 4
), suggesting that the alignment of the different practitioners produced a consistent gait pattern. Angles for the prosthetic knee were most consistent, whereas the contralateral knee showed only slight differences in the alignment of P5 in late swing/early stance phase. Ankle angles in the varying alignments were also very similar. Alignment changes to the ankle angle are apparent in differences in the swing phase portion of the prosthetic ankle angle curves, because the ankle is non-weight-bearing and is not affected by remote joints.
DISCUSSION
Researchers have long considered the advantages of the inclusion of objective numerical and measurement-based techniques into the alignment of prostheses. Techniques that make alignment more repeatable, more routinely optimized, and faster are naturally considered advantageous. The present study is a precursor to automated alignment and seeks to determine whether substantial differences are present in the outcome of the alignment of a set of different practitioners. The study is limited in that the source of any differences is not readily identifiable. The practitioners in the study varied in terms of education, background, and experience, so differences might be attributable to some factor of experience or some differently learned alignment techniques. More generally, alignments might differ because the prosthetists use different cost factors and a different optimal alignment outcome. Alternatively, the variability and subjectivity inherent in observational gait analysis could produce different alignments even when skill level and the optimization model are equal. This study is a necessary first step to identify and quantify differences in alignment outcomes; consequent implications for automated alignment and for the source of potential differences would require additional investigation.
Quantification of alignment is generally desirable but lacking in prosthetic practice. Zahedi et al.5 used a method to quantify alignments and measured repeatability and ranges of tolerable alignments. Although the study measured the alignments themselves and not the outcomes, it concluded that one patient could be satisfied with several alignments, that a prosthetist could not generally repeat a given alignment at will, and that different prosthetists produced different alignment ranges on any one patient. The present study would tend to agree with these conclusions, given that the alignments of the different practitioners were quantifiably different (although with a different quantification means than that of Zahedi et al.5), but the gait outcomes were very similar.
Several researchers have spoken of the goal of "optimum alignment." The data in the present study illustrate the difficulty of defining this optimum. For example, P3 aligned the subject such that in static standing, the bilateral knee angles were nearly identical, whereas the other practitioners aligned the prosthesis such that the contralateral limb was less flexed. The alignment of P3 also produced the greatest bilateral difference in static ankle angles. Despite these static differences, the joint angles measured in walking with the P3 alignment were very similar to the other practitioners. The question of optimum alignment persists. Did P3's alignment produce an ankle difference that allowed the knee angle similarity, or a knee similarity that produced the ankle difference? Which is preferable or more important? If the outcome is so similar, does it matter?
The practitioners in this study produced relatively similar static standing with their alignments, and even more similar walking, despite their varying levels of education and experience. When static differences were quantified, their impact on gait was either indirect or somehow masked by the amputee. It is apparent that there is a range of alignments that would probably be deemed acceptable by both amputee and prosthetist, although this investigation has not defined that range. This result suggests that automated alignment is feasible. In light of the literature, though, the present research suggests that until optimum alignment can be more clearly defined, automated or computerized alignment may not be necessary.
CONCLUSIONS
When five prosthetists conducted dynamic alignment of the same prosthesis for the same transfemoral amputee, kinematic, kinetic, and temporal/spatial gait outcomes were very similar across alignments. This similarity occurred despite the small but quantifiable differences in static weight-bearing alignments in the frontal and sagittal planes. The similar gait patterns suggest that either the alignment differences were too small to produce observable differences in gait patterns or, more likely, the amputee revealed an ability to mask the alignment differences and maintain a self-determined optimized gait strategy.
MARK D. GEIL, PhD, is affiliated with the School of Applied Physiology, Georgia Institute of Technology, Atlanta, GA. Correspondence to: Mark D. Geil, PhD, School of Applied Physiology, Georgia Institute of Technology, Atlanta, GA 30332-0356; E-mail: mark.geil@ap.gatech.edu
References:
- Berme N, Purdie CR, Solomonidis SE. Measurement of prosthetic alignment. Prosthet Orthot Int. 1978;2:73-76.
- Kerr G, Saleh M, Jarrett MO. An angular alignment protractor for use in the alignment of below-knee prostheses. Prosthet Orthot Int. 1984;8:56-57.
- Sin SW, Chow DH, Cheng JC. A new alignment jig for quantification and prescription of three-dimensional alignment for the patellar-tendon-bearing trans-tibial prosthesis. Prosthet Orthot Int. 1999;23:225-230.
- Hansen AH, Childress DS, Knox EH. Prosthetic foot roll-over shapes with implications for alignment of trans-tibial prostheses. Prosthet Orthot Int. 2000;24:205-215.
- Zahedi MS, Spence WD, Solomonidis SE, Paul JP. Alignment of lower-limb prostheses. J Rehabil Res Dev. 1986;23:2-19.
- Hannah RE, Morrison JB, Chapman AE. Prostheses alignment: effect on gait of persons with below-knee amputations. Arch Phys Med Rehabil. 1984;65:159-162.
- Zahedi MS, Spence WD, Solomonidis SE, Paul JP. Repeatability of kinetic and kinematic measurements in gait studies of the lower limb amputee. Prosthet Orthot Int. 1987;11:55-64.
- Zahedi MS, Spence WD, Solomonidis SE. The influence of alignment on prosthetic gait. In: Murdoch G, Donovan RG, eds. Amputation Surgery and Lower Limb Prosthetics. Oxford: Blackwell Scientific Publications; 1998:367-378.
- Geil MD, Parnianpour M, Berme N. Significance of nonsagittal power terms in analysis of a dynamic elastic response prosthetic foot. Trans ASME: J Biomech Eng. 1999;121:521-524.
- Winter DA, ed. Biomechanics and Motor Control of Human Movement. New York: John Wiley & Sons, Inc.; 1990
|
|
