Within-Session Reliability and Minimum Detectable Differences for Discrete Lower-Extremity Angles and Moments During Walking

in Journal of Applied Biomechanics
View More View Less
  • 1 Drexel University

Differences in walking biomechanics between groups or conditions should be greater than the measurement error to be considered meaningful. Reliability and minimum detectable differences (MDDs) have not been determined for lower-extremity angles and moments during walking within a session, as needed for interpreting differences in cross-sectional studies. Thus, the purpose of this study was to determine within-session reliability and MDDs for peak ankle, knee, and hip angles and moments during walking. Three-dimensional gait analysis was used to record walking at 1.25 m/s (±5%) in 18 men, 18–50 years of age. Peak angles and moments were calculated for 2 sets of 3 trials. Intraclass correlation coefficients (3, 3) were used to determine within-session reliability. In addition, MDDs were calculated. Within-session reliability was good to excellent for all variables. The MDDs ranged from 0.9° to 3.6° for joint angles and 0.06 to 0.15 N·m/kg for joint moments. Within-session reliability for peak ankle, knee, and hip angles and moments was better than the between-session reliability reported previously. Overall, our MDDs were similar or smaller than those previously reported for between-session reliability. The authors recommend using these MDDs to aid in the interpretation of cross-sectional comparisons of lower-extremity biomechanics during walking in healthy men.

Differences in discrete walking biomechanical variables between groups or conditions should be greater than the measurement error to be considered meaningful. Reliability and the minimum detectable difference (MDD) are used to determine what is considered a meaningful difference. Reliability has been determined for several lower-extremity biomechanical variables, but only for between-session comparisons.1,2 Within- and between-session reliability are influenced by differences in participants’ performance between trials; in addition, between-session reliability is influenced by marker placement and day-to-day differences in participant performance. Within-session reliability, where markers are not removed between sessions, is not subject to either of these factors and, therefore, may have higher reliability. Reliability calculations can be used to determine the MDD. The MDD indicates how large these differences must be to have a 95% likelihood of exceeding the measurement error. Similar to reliability, MDDs have been determined for lower-extremity biomechanical variables based on between-session comparisons.1,2 Thus, within-session reliability and MDDs are needed to aid the interpretation of differences in walking biomechanics between groups or conditions in cross-sectional study designs.

Many biomechanical variables have been compared during walking between conditions or groups of adults in cross-sectional studies with a single study visit. For example, the walking biomechanics of healthy adults have been compared before and after fatigue protocols, walking overground compared with on a treadmill, wearing different footwear, and in adults with different knee alignment.36 The variables of interest for these studies often include lower-extremity joint biomechanics measured during weight acceptance. Common variables of interest at the ankle include peak plantarflexion and eversion angles, as well as peak dorsiflexor moment.3,4,6,7 At the knee and hip, common variables include peak flexion and adduction angles and peak extensor and abductor moments.37 These biomechanical variables are typically compared between different conditions or groups using statistical tests. However, it is important to take the measurement error into account when interpreting differences in groups in cross-sectional studies. Within-session reliability and MDDs provide insight into the magnitude of measurement error for these variables of interest. Therefore, the purpose of this study was to determine within-session reliability and MDDs for peak ankle, knee, and hip angles and moments during the stance phase of walking in healthy men. We hypothesized that within-session reliability would be good to excellent for all variables.

Methods

As part of a larger study, healthy men between 18 and 50 years of age were recruited. All participants provided written informed consent, and all study procedures were approved by the Drexel University Institutional Review Board. Participants were excluded if they had any previous major lower-extremity surgery or an injury in the previous 6 months.

Eighteen men (age = 28 [7] y; height = 1.82 [0.07] m; body mass = 73.3 [8.1] kg) reported to the gait laboratory. A power analysis determined the minimum sample size for determining reliability using an intraclass correlation coefficient.8 With α = .05, β = 0.2, minimum acceptable reliability of 0.7 and anticipated reliability of 0.9, a minimum sample size of 18 was indicated. For the session, the participants wore laboratory-provided shoes with cutouts, socks with cutouts, shorts, and a tank top, so that markers could be placed directly on the skin. The participants stood in a standardized position9 while markers were placed on them. All markers were placed by the same biomechanist, with 5 years of experience in gait analysis. Reflective markers were placed on anatomical landmarks at the acromion processes, iliac crests, greater trochanters, medial and lateral epicondyles, medial and lateral malleoli, first metatarsals, fifth metatarsals, and inferior calcaneus using skin safe tape. These anatomical markers were used to define the joint coordinate system and define segments. Pelvis and lower-extremity segments were tracked using thermoplastic shells with 4 noncollinear reflective markers attached.10 The posterior pelvis shell was attached directly on the skin. The shells for the lower extremity were attached to neoprene wrapped around the proximal thighs and distal shanks and secured with hook and loop tape.10 In addition, 3 markers were placed directly on the skin of the superior, lateral, and medial calcaneus to track the foot. Following marker placement, a standing calibration trial was recorded in the standardized position. Then, anatomical markers were removed, and the participants walked over ground at 1.25 m/s (±5%) for 6 good trials. A trial was good if the participants walked at the appropriate speed, each foot landed on a force plate, and all markers were present and visible to the motion capture system.

Data Analyses

Three-dimensional marker positions were recorded with an 8-camera motion capture system (Vicon, Oxford, United Kingdom) sampling at 200 Hz. This was synchronized to ground reaction force data, which were collected by force plates sampling at 1000 Hz (AMTI Inc, Watertown, MA). The raw data were filtered using a fourth-order low-pass Butterworth filter at 6 Hz. The cutoff frequency was determined from a previous residual analysis performed on data collected in our laboratory. Next, the data were processed using rigid body analysis in Visual 3D software (C-Motion Inc, Germantown, MD). The stance phase was determined using a threshold of 20 N for ground reaction force to determine foot strike. Joint angles were determined using the joint coordinate system.11 Following previously established procedures,11 a medial-lateral (X), anterior–posterior (Y), and proximal–distal (Z) order of rotation was used. Internal joint moments were determined using inverse dynamics relative to the proximal segment and normalized to body mass in kilograms per International Society of Biomechanics standards.12 The positive force directions were proximal, medial, and anterior.12 Peak joint angles and moments were determined for each trial during weight acceptance, the first 50% of stance, using custom MATLAB software (MathWorks, Natick, MA). The variables of interest were extracted from the data set for the right side. The ankle variables were peak plantarflexion angle, peak eversion angle, and peak dorsiflexor moment. The knee variables were peak flexion and peak adduction angles and peak extensor and peak abductor moments. The hip variables were peak flexion and peak adduction angles and peak extensor and peak abductor moments.

The descriptive statistics were calculated for 2 sets of 3 trials of walking per participant. The first set was the average of the first 3 trials, and the second set was the average of the second 3 trials. Within-session reliability was calculated using a 2-way mixed model intraclass correlation coefficient (3, 3) to compare the averages of the 2 sets of 3 trials. Reliability was considered excellent if the r coefficient was at least .9 and good if it was at or above .75.13 F tests of P < .05 were used to confirm that the sets of data were different enough to be compared. In addition, standard error of measurement was calculated for each variable of interest with 95% confidence. The MDD with 95% CI was calculated using the following equation: MDD = standard error of measurement × 1.96 × √2.13

Results

Overall, within-session reliability was good to excellent for all ankle, knee, and hip variables (Table 1). At the ankle, the peak plantarflexion angle had good reliability. The peak eversion angle and peak dorsiflexor moment at the ankle demonstrated excellent reliability. In addition, all peak angles and moments at the knee and hip had excellent reliability. The MDDs for peak joint angles were between 0.9° and 3.6°. For peak joint moments, the MDDs were between 0.03 and 0.15 N·m/kg despite the large variation in the magnitude of peak moment among variables.

Table 1

Descriptive Statistics and Within-Session Reliability of Peak Ankle, Knee, and Hip Angles and Moments During Walking

JointVariableMean (SD)RangeICC3,3 (95% CI)SEMMDD
First set of trialsSecond set of trials
AnklePeak plantarflexion angle, °10.6 (2.1)11.1 (3.3)6.4 to 21.7.850 (.607–.944)1.02.9
Peak eversion angle, °5.1 (5.6)5.6 (5.0)−3.9 to 13.5.981 (.949–.993)0.72.0
Peak dorsiflexor moment, N·m/kg0.20 (0.04)0.21 (0.03)0.12 to 0.29.933 (.821–.975)0.010.03
KneePeak flexion angle, °15.7 (6.4)14.9 (6.3)3.1 to 26.7.959 (.894–.985)1.33.6
Peak adduction angle, °1.9 (3.1)2.1 (3.2)−3.6 to 7.1.989 (.971–.996)0.30.9
Peak extensor moment, N·m/kg0.50 (0.25)0.48 (0.24)−0.01 to 0.88.951 (.871–.982)0.050.15
Peak abductor moment, N·m/kg0.50 (0.17)0.51 (0.17)0.28 to 0.97.976 (.938–.991)0.030.07
HipPeak flexion angle, °24.4 (5.7)23.9 (5.9)13.2 to 33.9.986 (.962–.995)0.71.9
Peak adduction angle, °8.7 (3.1)8.7 (2.8)4.6 to 16.3.985 (.961–.995)0.41.0
Peak extensor moment, N·m/kg0.56 (0.13)0.56 (0.14)0.36 to 0.85.973 (.928–.990)0.020.06
Peak abductor moment, N·m/kg0.92 (0.18)0.92 (0.15)0.56 to 1.25.973 (.929–.990)0.030.07

Abbreviations: ICC, intraclass correlation coefficient; MDD, minimum detectable difference; SEM, standard error of measurement; 95% CI, 95% confidence interval.

Discussion

The purpose of the study was to determine the within-session reliability and MDDs for peak ankle, knee, and hip angles and moments during walking. Our hypothesis that all variables of interest would have good to excellent reliability for within-session reliability was supported. Within-session reliability and MDDs for lower-extremity biomechanics during walking have not been reported previously in the literature for healthy adults.

We can evaluate the within-session reliability in the context of between-session values previously reported in the literature for some of our variables of interest. At the ankle, between session reliability for the peak eversion angle (0.88) and peak dorsiflexor moment (0.67 to 0.81)1,2 was moderate to good. However, we found excellent reliability for both within-session. Similarly at the knee, the between-session reliability for the peak abductor moment was good (0.88),2 whereas the within-session reliability was excellent. This trend continued at the hip, with moderate or good between-session reliability1,2 for the peak flexion angle (0.89), peak adduction angle (0.82), peak extensor moment (0.77–0.89), and peak abductor moment (0.61–0.88) compared with excellent within-session reliability. As anticipated, within-session reliability is consistently higher than previously reported between-session reliability.

There are several key differences in between- and within-session comparisons. The major difference is that markers are removed and replaced between sessions, but not between trials within a session. Removing and replacing markers introduces the potential for error due to differences in marker placement from session to session. For example, pelvis markers may be difficult to place in the identical location from one session to the next due to adipose tissue overlaying bony landmarks in this region. Thus, it may be more difficult to consistently identify the bony landmark and place the marker directly over it. Our observation that hip variables had notably higher reliability within-session compared to between-session reliability reported in the literature supports this assertion. In addition, participants may move differently from day to day. This difference is likely small in healthy young adults, but may become more pronounced in older adults or those with movement pathologies. These differences likely account for the higher reliability within a session compared with between sessions.

As expected, within-session MDDs were similar or smaller to those reported previously by Wilken et al.2 Most within-session MDDs were less than half of the equivalent between-session values. This was the case for peak ankle dorsiflexor moment, peak hip flexion and adduction angles, and peak hip extensor and abductor moments.2 The only variable with within-session MDD that was more than half of the between-session value was the peak knee abductor moment (within 0.07 N·m/kg and between 0.10 N·m/kg).2 Thus, smaller differences between groups or conditions in cross-sectional studies may be considered meaningful when within-session MDDs are used to interpret the findings. The ability to demonstrate that small differences between groups or conditions are likely greater than the measurement error may be critical for a meaningful interpretation of cross-sectional studies.

It should be noted that the study sample only included young men, which may limit the generalizability of the findings. However, we have no reason to expect that healthy young women would have different trial-to-trial consistency than healthy young men. Nevertheless, caution should be used when extending these findings to other populations, such as older adults or clinical populations. Findings for other populations have demonstrated less trial-to-trial consistency for some discrete walking biomechanical variables.14 Therefore, within-session reliability specific to these populations should be determined.

Conclusion

Within-session reliability of peak ankle, knee, and hip angles and moments during walking was good to excellent in healthy men. Furthermore, MDDs were less than 4° for peak angles and 0.15 N·m/kg or less for peak moments. We recommend using the MDDs reported here to aid the interpretation of cross-sectional comparisons of lower-extremity biomechanics during walking in healthy men.

Acknowledgments

The authors would like to thank Trey Brindle for collecting the data for this study. The authors have no conflicts of interest to report.

References

  • 1.

    Monaghan K, Delahunt E, Caulfield B. Increasing the number of gait trial recordings maximises intra-rater reliability of the CODA motion analysis system. Gait Posture. 2007;25(2):303315. PubMed ID: 16730177 doi:10.1016/j.gaitpost.2006.04.011

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 2.

    Wilken JM, Rodriguez KM, Brawner M, Darter BJ. Reliability and minimal detectible change values for gait kinematics and kinetics in healthy adults. Gait Posture. 2012;35(2):301307. PubMed ID: 22041096 doi:10.1016/j.gaitpost.2011.09.105

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 3.

    Hunt MA, Hatfield GL. Ankle and knee biomechanics during normal walking following ankle plantarflexor fatigue. J Electromyogr Kinesiol. 2017;35:2429. PubMed ID: 28587934 doi:10.1016/j.jelekin.2017.05.007

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 4.

    Lee SJ, Hidler J. Biomechanics of overground vs. treadmill walking in healthy individuals. J Appl Physiol. 2008;104(3):747755. PubMed ID: 18048582 doi:10.1152/japplphysiol.01380.2006

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 5.

    Longpre HS, Potvin JR, Maly MR. Biomechanical changes at the knee after lower limb fatigue in healthy young women. Clin Biomech. 2013;28(4):441447. doi:10.1016/j.clinbiomech.2013.02.010

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 6.

    Zhang X, Paquette MR, Zhang S. A comparison of gait biomechanics of flip-flops, sandals, barefoot and shoes. J Foot Ankle Res. 2013;6(1):18. doi:10.1186/1757-1146-6-45

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 7.

    Barrios JA, Davis IS, Higginson JS, Royer TD. Lower extremity walking mechanics of young individuals with asymptomatic varus knee alignment. J Orthop Res. 2009;27(11):14141419. PubMed ID: 19402149 doi:10.1002/jor.20904

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 8.

    Walter S, Eliasziw M, Donner A. Sample size and optimal designs for reliability studies. Stat Med. 1998;17(1):101110. PubMed ID: 9463853 doi:10.1002/(SICI)1097-0258(19980115)17:1%2C101::AID-SIM727%2E3.0.CO;2-E

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 9.

    McIlroy W, Maki B. Preferred placement of the feet during quiet stance: development of a standardized foot placement for balance testing. Clin Biomech. 1997;12(1):6670. doi:10.1016/S0268-0033(96)00040-X

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 10.

    Manal K, McClay I, Stanhope S, Richards J, Galinat B. Comparison of surface mounted markers and attachment methods in estimating tibial rotations during walking: an in vivo study. Gait Posture. 2000;11(1):3845. PubMed ID: 10664484 doi:10.1016/S0966-6362(99)00042-9

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 11.

    Cole GK, Nigg BM, Ronsky JL. Application of the joint coordinate system to three-dimensional joint attitude and movement representation: a standardization proposal. J Biomech Eng. 1993;115(4A):344349. PubMed ID: 8309227 doi:10.1115/1.2895496

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 12.

    Derrick TR, van den Bogert AJ, Cereatti A, Dumas R, Fantozzi S, Leardini A. ISB recommendations on the reporting of intersegmental forces and moments during human motion analysis. J Biomech. 2020;99:109533. PubMed ID: 31791632 doi:10.1016/j.jbiomech.2019.109533

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 13.

    Portney L, Watkins M. Foundations of clinical research. Application to Practice. 1993:457484.

  • 14.

    Klejman S, Andrysek J, Dupuis A, Wright V. Test-retest reliability of discrete gait parameters in children with cerebral palsy. Arch Phys Med Rehabil. 2010;91(5):781787. doi:10.1016/j.apmr.2010.01.016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation

The authors are with the Department of Physical Therapy and Rehabilitation Sciences, Drexel University, Philadelphia, PA, USA.

Hawkins (jh3266@dragons.drexel.edu) is corresponding author.
  • 1.

    Monaghan K, Delahunt E, Caulfield B. Increasing the number of gait trial recordings maximises intra-rater reliability of the CODA motion analysis system. Gait Posture. 2007;25(2):303315. PubMed ID: 16730177 doi:10.1016/j.gaitpost.2006.04.011

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 2.

    Wilken JM, Rodriguez KM, Brawner M, Darter BJ. Reliability and minimal detectible change values for gait kinematics and kinetics in healthy adults. Gait Posture. 2012;35(2):301307. PubMed ID: 22041096 doi:10.1016/j.gaitpost.2011.09.105

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 3.

    Hunt MA, Hatfield GL. Ankle and knee biomechanics during normal walking following ankle plantarflexor fatigue. J Electromyogr Kinesiol. 2017;35:2429. PubMed ID: 28587934 doi:10.1016/j.jelekin.2017.05.007

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 4.

    Lee SJ, Hidler J. Biomechanics of overground vs. treadmill walking in healthy individuals. J Appl Physiol. 2008;104(3):747755. PubMed ID: 18048582 doi:10.1152/japplphysiol.01380.2006

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 5.

    Longpre HS, Potvin JR, Maly MR. Biomechanical changes at the knee after lower limb fatigue in healthy young women. Clin Biomech. 2013;28(4):441447. doi:10.1016/j.clinbiomech.2013.02.010

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 6.

    Zhang X, Paquette MR, Zhang S. A comparison of gait biomechanics of flip-flops, sandals, barefoot and shoes. J Foot Ankle Res. 2013;6(1):18. doi:10.1186/1757-1146-6-45

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 7.

    Barrios JA, Davis IS, Higginson JS, Royer TD. Lower extremity walking mechanics of young individuals with asymptomatic varus knee alignment. J Orthop Res. 2009;27(11):14141419. PubMed ID: 19402149 doi:10.1002/jor.20904

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 8.

    Walter S, Eliasziw M, Donner A. Sample size and optimal designs for reliability studies. Stat Med. 1998;17(1):101110. PubMed ID: 9463853 doi:10.1002/(SICI)1097-0258(19980115)17:1%2C101::AID-SIM727%2E3.0.CO;2-E

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 9.

    McIlroy W, Maki B. Preferred placement of the feet during quiet stance: development of a standardized foot placement for balance testing. Clin Biomech. 1997;12(1):6670. doi:10.1016/S0268-0033(96)00040-X

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 10.

    Manal K, McClay I, Stanhope S, Richards J, Galinat B. Comparison of surface mounted markers and attachment methods in estimating tibial rotations during walking: an in vivo study. Gait Posture. 2000;11(1):3845. PubMed ID: 10664484 doi:10.1016/S0966-6362(99)00042-9

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 11.

    Cole GK, Nigg BM, Ronsky JL. Application of the joint coordinate system to three-dimensional joint attitude and movement representation: a standardization proposal. J Biomech Eng. 1993;115(4A):344349. PubMed ID: 8309227 doi:10.1115/1.2895496

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 12.

    Derrick TR, van den Bogert AJ, Cereatti A, Dumas R, Fantozzi S, Leardini A. ISB recommendations on the reporting of intersegmental forces and moments during human motion analysis. J Biomech. 2020;99:109533. PubMed ID: 31791632 doi:10.1016/j.jbiomech.2019.109533

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 13.

    Portney L, Watkins M. Foundations of clinical research. Application to Practice. 1993:457484.

  • 14.

    Klejman S, Andrysek J, Dupuis A, Wright V. Test-retest reliability of discrete gait parameters in children with cerebral palsy. Arch Phys Med Rehabil. 2010;91(5):781787. doi:10.1016/j.apmr.2010.01.016

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 165 165 165
PDF Downloads 94 94 94