Skip to main content
NCBI home page
Osteoarthritis and Cartilage Open logoLink to Osteoarthritis and Cartilage Open
. 2021 Dec 25;4(1):100230. doi: 10.1016/j.ocarto.2021.100230

Feasibility of personalised hip load modification using real-time biofeedback in hip osteoarthritis: A pilot study

Laura E Diamond a,b,∗, Daniel Devaprakash a,b, Bradley Cornish b, Melanie L Plinsinga c, Andrea Hams a,b, Michelle Hall d, Rana S Hinman d, Claudio Pizzolato a,b, David J Saxby a,b
PMCID: PMC9718151  PMID: 36474469

Abstract

Objective

(i) Compare the feasibility of three load modification strategies to immediately increase hip contact force in people with hip osteoarthritis (OA) using real-time visual biofeedback during walking, and (ii) prospectively evaluate changes in pain and physical function following 6-weeks of walking using a prescribed personalised load modification strategy.

Design

Twenty participants with symptomatic mild-to-moderate hip OA walked on an instrumented treadmill while motion capture and electromyographic data were recorded (normal walk), then under three conditions: (i)neutral trunk lean; (ii)neutral pelvic obliquity; (iii)increased step length. The biomechanical parameter of interest and corresponding target value were displayed in real-time. Hip contact forces were subsequently computed using a calibrated electromyography-informed neuromusculoskeletal model. A decision tree was used to prescribe a personalised load modification strategy to each participant for integration into walking over 6-weeks.

Results

Only the step length modification significantly increased peak hip contact force compared to normal walking when performed by all participants (11.34 [95%CI 4.54,18.13]%, P ​< ​0.01). After participants were prescribed a personalised load modification strategy, both neutral pelvis (n ​= ​5, 11.88[95%CI -0.49,24.24]%) and step length (n ​= ​10, 12.79[95%CI 0.49,25.09]%) subgroups increased peak hip contact force >10%. After 6-weeks, 77% and 46% of participants reported a clinically important improvement in hip pain during walking and physical function, respectively.

Conclusion

Most participants with hip OA could immediately increase hip contact force through personalised movement retraining by a magnitude estimated to promote cartilage heath and reported an improvement in symptoms after 6-weeks. Findings provide preliminary support for a personalised load modification-based intervention for hip OA.

Keywords: Neuromusculoskeletal modelling, Gait retraining, Contact force, Biomechanics, Movement strategy, Electromyography

1. Introduction

Hip osteoarthritis (OA) is a chronic musculoskeletal disease predicted to affect one in four people during their lifetime [1]. Conservative non-drug interventions are recommended for clinical management of hip OA [2,3], though many patients report only small-to-modest improvements in pain, function, and quality of life [2,4]. These poor outcomes may exist because treatment targets are not well defined, not personalised, and are largely extrapolated from knee OA literature [4]. Further, current treatment approaches rarely lend themselves to self-management, which is central to hip OA clinical practice guidelines [2] and highly valued by patients [5]. Indeed, abnormal hip loading plays a fundamental role in OA worsening [6] and is potentially modifiable with movement retraining [7]. Personalised movement retraining designed to target a known mechanism of disease remains an entirely unexplored therapeutic strategy for people with hip OA.

Contrary to knee OA, under-rather than over-loading has been implicated in the pathogenesis of hip OA. Compared to healthy controls, people with symptomatic and radiographic mild-to-moderate [8] and end-stage [9,10] hip OA walk with lower magnitude hip contact forces. These observations contrast with the higher knee loads extensively proven to drive knee OA progression [11] and may explain why management strategies based on evidence from knee OA literature, applied to hip OA, have been largely ineffective [4]. Hip contact forces interact with cartilage mechanobiology (i.e. tissue mechanics) to regulate cartilage structure [12]. Collagen networks in the extracellular matrix of cartilage undergo strains during habitual activities (e.g. walking) and require adequate stimulus (i.e. joint loading) to avoid degeneration associated with disuse [12]. Ex vivo data suggest increasing joint load by even 5% may stimulate cartilage health [13]. A management strategy designed to target hip contact force and adequately stimulate hip cartilage during walking may have therapeutic benefits for people with hip OA.

Several biomechanical parameters are thought to be responsible for abnormal hip loading in hip OA [14]. People with mild-to-moderate hip OA descend stairs with greater frontal plane trunk lean [15] and people with severe hip OA walk with greater frontal plane trunk lean [16] than healthy controls. Trunk kinematics influence the position of the centre of mass relative to the hip joint centre [17] and thus have direct implications for hip loading. Further, musculoskeletal modelling simulations have shown that contralateral pelvic drop will directly affect loading at the hip during walking [10]. Spatiotemporal features of walking, including speed and step length, also contribute to loads sustained by the hip [10]. Lower walking speed [18], driven by a 10% shorter step length [19], is characteristic of people with hip OA across the spectrum of disease severity compared to healthy controls. Biomechanical parameters relating to trunk and pelvis position, and step length are mechanically linked to hip loading and are potentially modifiable with movement retraining [20,21].

Biofeedback is a critical adjunct to the learning of motor skills [22]. Visual biofeedback has been effective for altering surrogate measures of knee load in people with knee OA [7,20]. However, movement retraining using real-time biofeedback customised to each individual's walking biomechanics and designed to target the physiological load acting directly on the joint's tissue (i.e. hip contact force) [23] remains entirely unexplored in those with hip OA. Given the complex relationship between joint loading and symptoms [24], it is imperative to concurrently evaluate changes in symptoms when modifying joint load. This pilot study included individuals with symptomatic mild-to-moderate hip OA, who were prime candidates for targeted disease modifying interventions, and aimed to (i) compare the feasibility of three load modification strategies to immediately increase hip contact force when participants were provided with real-time visual biofeedback during walking, and (ii) prospectively evaluate changes in pain and physical function following 6-weeks of walking using a prescribed personalised load modification strategy.

2. Methods

This study used a within-participant design and reporting followed items of the CONSORT guidelines [25] applicable to non-randomised pilot and feasibility studies. Fig. 1 illustrates an overview of the study, assessment time points, and outcomes. Ethical approval was obtained from the institutional Human Research Ethics Committee (GU#2019/103) and participants provided their written informed consent prior to testing.

Fig. 1.

Fig. 1

Open in a new tab

Study overview including assessment time points and outcomes.

2.1. Participants

Twenty participants were recruited from the South East Queensland community via advertisements on social media between January 2020 and March 2021. Hip OA was classified according to the American College of Rheumatology clinical criteria for hip OA [26]. Participants were included if they: (i) were aged 50 years or older; (ii) reported hip pain on most days of the past month for >3 months; (iii) reported hip pain over the past week while walking of ≥4 on an 11-point numerical rating scale (NRS, 0 ​= ​‘no pain’ and 10 ​= ​‘worst pain possible’); (iv) could walk independently for at least 20 ​min 3-times a week; and (v) had a body mass index (BMI) ​< ​34 ​kg/m2. Exclusion criteria were: (i) previous hip replacement in the affected hip; (ii) any hip surgery in the past 6-months; (iii) current or past (<3 months) use of oral or intra-articular corticosteroid; (iv) hip pain symptoms associated with extra-articular or lumbar pathology; (v) any other joint or muscle pain, such as back, other hip, knees, ankles or feet, which was worse than hip pain on the (most) affected side; (vi) any neurological or cardiovascular condition affecting the ability to perform testing; and (vii) work restrictions or other commitments preventing engagement in study protocol over the 6-week intervention period.

2.2. Procedures

Volunteers were screened via an online survey followed by telephone screening to confirm eligibility (Supplementary Figure S1). Potentially eligible participants underwent a clinical examination [26] to confirm hip OA by a registered physiotherapist (AH) prior to data collection. For participants with bilateral hip pain, the most symptomatic hip was considered the study hip. Biomechanics data were collected at Griffith University (Time 1) by the same researchers (DD, BC, MLP). Baseline (Time 1) and 6-week follow-up (Time 2) participant-reported data were collected via paper-based questionnaires and telephone (Time 2 only). Participants recorded weekly data, including estimated total minutes walking using their prescribed load modification strategy and adverse events, in a paper-based logbook.

2.2.1. Biomechanics assessment

A full body marker set, consisting of 41 retro-reflective markers (Supplementary Figure S2) were fitted to each participant [23]. Surface electromyography (EMG) signals were synchronously recorded using a 16-channel wireless EMG system (2000 ​Hz; Cometa, Italy) from 12 lower-limb muscles of the study limb: gluteus maximus, gluteus medius, semitendinosus, biceps femoris, medial gastrocnemius, lateral gastrocnemius, soleus, vastus medialis, vastus lateralis, rectus femoris, tensor fascia latae, and tibialis anterior using standardised electrode placement [27]. Participants wore their regular walking shoes and were fitted with a safely harness that included shoulder, leg, and chest straps (ERGO Full Body Fall Arrest Harness, Spanset, Switzerland). The safety harness was secured such that participants could walk without impediment or discomfort (Fig. 2; Supplementary Figure S2).

Fig. 2.

Fig. 2

Open in a new tab

The biomechanical parameter of interest and a corresponding personalised target value (derived from the normal walk/control condition) were displayed in front of the treadmill. The target value was updated in real-time to provide immediate visual biofeedback to participants.

Participants walked on an instrumented split-belt treadmill (Bertec Corporation, USA), sampling at 1000 ​Hz, for 5-min at a self-selected speed (normal walk - control condition), then for 5-min under each of the following experimental conditions: (i) neutral trunk lean (medial/lateral); (ii) neutral pelvic obliquity (rise/drop); and (iii) increased step length. Participants maintained the same walking speed for each condition. The biomechanical parameter of interest and a corresponding personalised target value were displayed in front of the treadmill in real-time (Fig. 2). The personalised targets were calculated using the normal walk condition and were set to decrease trunk lean towards the affected side, decrease pelvic drop on the contralateral side, and increase step length by 5% (an amount estimated to increase hip contact force by the desired 5%) [13]. Participants were provided instructions and a demonstration prior to walking under each condition. Specifically, participants were instructed to “keep your body upright and avoid moving the upper body from side to side” (neutral trunk), to “keep both sides of your hips level during walking/do not drop your hip” (neutral pelvis), and to “take longer steps” (increased step length). Motion data were captured using a 10-camera motion capture system (200 ​Hz; Vicon MX, UK) operating with Vicon Nexus (version 2.9.1); the biomechanical parameter of interest was computed and then continuously streamed to the visual display using the drawnow function in MATLAB R2018a (MathWorks, USA). Following each walking condition, participants were asked to rate their hip, knee, foot/ankle, and back pain, perceived difficulty, and confidence on a numeric rating scale (NRS) with terminal descriptors ‘no pain’, ‘not hard at all’, ‘not confident at all’ (score 0) and ‘worst pain possible’, ‘extremely hard’, ‘completely confident’ (score 10), respectively, and their perceived exertion using the 20-point Borg scale [28].

2.2.2. Personalised load modification

A decision tree (Fig. 3) based on pain, confidence, and competence during each of the walking conditions was used to prescribe a personalised load modification strategy to each participant, from the pool of available movements strategies, for integration into their walking routine over 6-weeks. Immediately following 5-min of walking using each load modification strategy, participants were asked to rate their hip pain and confidence on a NRS. Competence was assessed visually by the researcher and used in combination with an objective measure (computed as the percentage of strides where the participant successfully achieved the personalised target). Prior to leaving the laboratory, participants walked for an additional 5-min using their prescribed personalised load modification strategy with assistance from visual biofeedback. Participants were expected to walk a minimum of 20 ​min 3-times per week (per study inclusion criteria) using their prescribed personalised load modification strategy. Participants wore pedometers (Alvita Ultimate, Omron, Japan) during the 7-days before their baseline (laboratory) and 6-week follow-up (phone) assessments. Number of steps per day were tracked to confirm walking targets were met during these periods.

Fig. 3.

Fig. 3

Open in a new tab

Decision tree used to prescribe a personalised load modification strategy to each participant for integration into their walking routine over 6-weeks.

2.2.3. Participant-reported outcomes

2.2.3.1. Pain (Time 1, Time 2)

Participants were asked to rate their hip pain intensity over the last week and during walking in both their (most) painful hip and their contralateral hip on NRS with terminal descriptors ‘no pain’ (score 0) and ‘worst pain possible’ (score 10). The NRS pain outcome has demonstrated reliability in OA [29]. A minimum clinically important change was defined as a change in pain of 1.8 units [30]. Participants were also asked to rate their change in hip pain since the beginning of the study at the 6-week follow up (Time 2) on a 7-item global rating of change scale [31].

2.2.3.2. Physical function (Time 1, Time 2)

Participants completed the physical function subscale of the Western Ontario and McMaster Universities (WOMAC) Osteoarthritis Index. The subscale contains 17 items, each answered on a Likert scale (no dysfunction ​= ​0; extreme dysfunction ​= ​4) and scores range from 0 to 68 where higher scores indicate poorer function. The WOMAC is a disease-specific instrument that has established validity, reliability, and responsiveness in OA [32]. A minimum clinically important change was defined as a change in physical function of 6 units [33].

2.2.3.3. Acceptability of intervention (Time 2)

Participants were asked about how confident they were walking with their personalised load modification strategy in regular daily life over the 6-week follow up period using an 11-point NRS ranging from ‘not confident at all’ (score 0) to ‘completely confident’ (score 10).

2.3. Data processing

Body marker motion and ground reaction forces from 7 trials for each walking condition (28 in total) were processed using MOtoNMS software (version 2.2) [34] in MATLAB R2019b (MathWorks, USA) for subsequent use in OpenSim (version 3.3) [35]. Body marker motion and ground reaction forces were low-pass filtered with a 4th order, dual-pass, zero-lag Butterworth filter with a nominal cut-off frequency of 6 ​Hz. Electromyograms from each muscle were band-pass filtered (30–300 ​Hz), full-wave rectified, and low-pass filtered with 4th order, dual-pass, zero-lag Butterworth filter (6 ​Hz). The EMG were subsequently amplitude-normalized to the maximum processed value recorded across all walking trials [36]. A generic full-body model [35] with 23° of freedom was linearly scaled to match individual anthropometry based on marker positions placed atop bony landmarks and estimated joint centers [37]. Optimal fibre and tendon slack lengths were morphometrically scaled to preserve dimensionless muscle and tendon operating curves [38]. Joint angles, moments, and musculotendon kinematics were calculated for each walking trial using inverse kinematics, inverse dynamics, and muscle analysis in OpenSim, respectively. These data were input, along with processed EMG, into the Calibrated EMG-informed Neuromusculoskeletal Modelling Toolbox (CEINMS) [39]. Using CEINMS, neuromusculoskeletal parameters were first calibrated to each participant using data from one randomly selected trial from each walking condition (4 in total). Once calibrated, EMG assisted neural mode was used within CEINMS [8] to calculate hip muscle forces (N) and hip contact forces (N) for the remaining 24 walking trials for each participant [40,41]. Within each walking trial, gait events were identified based on a vertical ground reaction force threshold of 30 ​N and biomechanical data were spline interpolated to 150 time-points for analysis of the stance phase. Data from two participants were excluded from biomechanical modelling as their foot strikes neared/crossed the split-belt treadmill's midline and could not be isolated to individual force plates. Walking speed (m/s) was calculated from the velocity of the calcaneus marker of the stance leg (Euclidean distance/time between frames). Treadmill speed could only be increased by increments of 0.05 ​m/s, and therefore participant walking speeds were rounded to the nearest 0.05, as required. Step length (m) was calculated as the Euclidean distance between left and right calcaneus markers during double stance. Hip contact force was normalized to body weight (N/BW) and the peak was extracted from each trial. Average peak hip contact force was calculated for each condition for each participant. Peak hip contact force change (%, relative to normal walk) was calculated for each load modification condition for each participant.

2.4. Statistical analysis

Peak hip contact force was compared between walking conditions for all participants using a repeated measures analysis of variance and post-hoc pairwise comparisons were made using paired sample t-tests (P ​< ​0.05). Peak hip contact force change (load modification condition minus normal walk) with 95% confidence intervals (CI) was calculated for participant subgroups based on their prescribed personalised load modification strategy. The number of participants (%) who reached the minimum clinically importance change in self-reported pain and physical function at 6-week follow-up was calculated for the entire cohort and for participant subgroups. Statistical analyses were performed using Statistical Package for the Social Sciences (SPSS, version 27, IBM Corporations).

3. Results

Of the 145 individuals who completed the online expression of interest, 118 (81%) were excluded at telephone screening, 7 (5%) were excluded at physical screening, and 20 (14%) were enrolled in the study (Supplementary Figure S1). Five (28%) participants did not return all 6-week follow up (Time 2) questionnaires or pedometer assessments. On average, participants walked 6460 ​± ​2644 steps/day and 7067 ​± ​1987 steps/day (mean difference −637 [95%CI -694, 1968] steps/day) during the 7-days before their baseline and 6-week follow-up assessments, respectively. Participant characteristics are reported in Table 1.

Table 1.

Participant characteristics at baseline (mean ​± ​standard deviation unless otherwise stated).

n ​= ​18
Age (years) 60.8 ​± ​6.3
Sex, female n (%) 13 (72%)
Height (m) 1.65 ​± ​0.10
Body mass (kg) 82.7 ​± ​15.2
Body mass index (kg/m2) 30.0 ​± ​4.1
Bilateral symptoms, n (%) 3 (17%)
Duration of symptoms, n (%)
 <1 year 1 (6%)
 1–2 years 4 (22%)
 2–5 years 10 (56%)
 >5 years 2 (11%)
Open in a new tab

Participants walked on the treadmill with an average speed of 0.93 ​± ​0.2 ​m/s. Peak hip contact force increased significantly when all participants walked using the step length modification compared to normal walking (mean difference 11.34 [95%CI 4.54, 18.13] %, P ​< ​0.01), though not when using the other modifications (neutral trunk: mean difference 1.27 [95%CI -3.26, 5.82] %; neutral pelvis: mean difference −0.13 [95%CI -5.53, 5.28] %; Fig. 4). Changes in peak hip contact force were driven by changes in the biomechanical parameter specific to each walking condition (trunk lean – Supplementary Figure S4; pelvic obliquity – Supplementary Figure S4; step length – Supplementary Figure S5).

Fig. 4.

Fig. 4

Open in a new tab

Ensemble average (±1 standard error, n ​= ​18) hip contact force across the stance phase of normal walking (red - dashed) and three load modifying conditions (neutral trunk lean (khaki), neutral pelvic obliquity (blue), increased step length (green)) (top left) with assistance from real-time visual biofeedback. Neutral trunk lean (top right), neutral pelvic obliquity (bottom left), and increased step length (bottom right) are also shown individually with normal walking. BW: bodyweight.

Following application of the decision tree, load modification subgroups included: neutral trunk n ​= ​3; neutral pelvis n ​= ​5; increased step length n ​= ​10. Thirteen (72%) participants increased peak hip contact force >5% when walking using their prescribed personalised load modification strategy (Fig. 5). These participants included 5 (100%) in the neutral pelvis subgroup and 8 (80%) in the step length subgroup. No participants in the trunk lean subgroup achieved a 5% increase in peak hip contact force, and on average, neutralised peak trunk lean by only ∼1° (mean difference 0.97 [95%CI -0.71, 2.65] degrees) compared to their normal walk condition (Supplementary Figure S4). On average, the neutral pelvis (11.88 [95%CI -0.49, 24.24] %) and step length (mean difference 12.79 [95%CI 0.49, 25.09] %) subgroups increased peak hip contact force by >10% (Fig. 5).

Fig. 5.

Fig. 5

Open in a new tab

Immediate percent change (±1 standard error) in peak hip contact force (left bars/y-axis; %, relative to normal walk) and 6-week change (±1 standard error) in hip pain during walking (right bars/y-axis; numeric rating scale (Time 2 minus Time 1 where a lower score indicates a greater decrease in hip pain)) when participants with hip OA walked using their prescribed load modification strategy (neutral trunk lean n ​= ​3; neutral pelvic obliquity n ​= ​5; increased step length n ​= ​10) with assistance from real-time visual biofeedback.

Participant-reported outcomes are summarised in Table 2. All but three participants (77%) reported a clinically important improvement in hip pain during walking (>1.8 NRS units) following the 6-week intervention (neutral trunk n ​= ​2 (67%); neutral pelvis n ​= ​4 (80%); increased step length n ​= ​4 (80%); Fig. 5). Six participants (46%) reported a clinically important change in physical function (>6 units on WOMAC) following the 6-week intervention (neutral trunk n ​= ​2 (67%); neutral pelvis n ​= ​3 (60%); increased step length n ​= ​1 (20%)). All but one participant (93%) rated their change in hip pain since the beginning of the study as ‘a little better’ (n ​= ​7 (50%)), ‘better’ (n ​= ​4 (28%)), or ‘much better’ (n ​= ​2 (14%)). Ten participants (71%) rated their confidence using their prescribed walking modification ≥ 7/10. No adverse events were reported and 67% (4/6) of those who reported pain in the contralateral hip during walking at Time 1 reported a decrease in pain in the contralateral hip at 6-week follow-up (Time 2).

Table 2.

Participant-reported outcomes.

Time 1 mean (SD) Time 2 mean (SD) Mean difference (95% CI)†
Overall hip pain during previous week, (0–10 NRS) 4.9 ​± ​1.9 2.9 ​± ​2.3 −1.7 (−2.5, −0.9)
Hip pain during walking during previous week, (0–10 NRS) 4.4 ​± ​2.2 2.3 ​± ​2.1 −1.9 (−2.8, −0.9)
Overall hip pain during previous week (contralateral hip), (0–10 NRS) 0.9 ​± ​1.6 0.7 ​± ​1.3 −0.4 (−1.4, 0.7)
Hip pain during walking during previous week (contralateral hip), (0–10 NRS) 1.1 ​± ​1.9 0.6 ​± ​1.2 −0.6 (−1.8, 0.6)
WOMAC physical function (0–68) 22.0 ​± ​10.1 14.7 ​± ​8.1 −6.0 (−14.0, 2.0)
Confidence walking using prescribed loan modification over 6-intervention period, (0–10 NRS)¢ – 7.6 ​± ​3.1 –
Open in a new tab

NRS: numeric rating scale (higher score indicates worse pain); ¢higher score indicates greater confidence; WOMAC: Western Ontario and McMaster Universities Osteoarthritis Index (higher score indicates worse function); †calculated as Time 2 (6-week follow-up) minus Time 1; SD: standard deviation; CI: confidence interval.

4. Discussion

This study compared the feasibility of three load modification strategies to immediately increase hip contact force when participants were provided with real-time visual biofeedback during walking, and prospectively evaluated changes in pain and physical function following 6-weeks of walking using a prescribed personalised load modification strategy. The primary findings were that, on average, only the step length load modification significantly increased peak hip contact force compared to normal walking when performed by all participants. However, after participants were prescribed a personalised load modification strategy, both the neutral pelvis and step length load modification subgroups increased peak hip contact force by >10% compared to normal walking, an amount more than double the 5% estimated to promote cartilage health [13]. Regardless of the personalised load modification strategy prescribed based on decision tree results, most participants reported an improvement in hip pain during walking after 6-weeks of walking using their prescribed strategy. These findings provide preliminary support for a personalised load modification-based intervention in people with hip OA.

On average, increasing step length was the only load modification to significantly increase peak hip contact force compared to normal walking when performed by all participants. After participants were prescribed a personalised load modification strategy based on decision tree results, both the neutral pelvis and step length load modification subgroups increased peak hip contact force by >10% compared to normal walking. Increasing step length requires greater muscle activation and elicits higher force [42] in hip extensor muscles (e.g. gluteus maximus) during stance, which may have contributed to the higher hip contact forces observed during this walking condition. Reducing contralateral pelvic drop during walking has a direct effect on the moment arm of the centre of mass relative to the hip joint centre [10], making the observed consequences for hip contact force unsurprising. Participants in the trunk lean load modification subgroup only neutralised peak trunk lean by ∼1°, which may explain why no participant achieved a 5% increase in peak hip contact force. The efficacy of the neutral trunk modification relied on participants using an excessive trunk lean strategy during their normal walking, though participant walking strategy was not factored into the decision tree used for subgrouping. Further, trunk and pelvis motion are often interconnected, particularly in populations with hip pathology [43], and may be difficult to target in isolation. Nevertheless, our results highlight that the hip's load response to each modification was unique across participants and emphasise the need for personalised interventions for management of hip OA [2].

Most participants with hip OA could immediately modulate their hip contact force during walking with assistance from real-time visual biofeedback of a target biomechanical parameter. Although mechanical joint loading is considered a modifiable mechanism of disease [6], internal loads acting directly on the hip have never before been targeted or assessed in conservative management programs for people with hip OA. Instead, gait modification interventions have largely focused on surrogate measures of joint load [7] and been primarily limited to people with knee OA [21,44]. Lower hip contact forces are characteristics of people with hip OA across the spectrum of disease severity compared to healthy controls [[8], [9], [10]] and increasing these internal loads to stimulate cartilage remodelling could have therapeutic benefit [45]. After participants were prescribed a personalised load modification strategy, most could immediately increase hip contact force up to 12% by neutralising frontal plane pelvic obliquity or increasing step length during walking. Although ex vivo and animal studies suggest a ∼1–4% increase in compressive force could be enough to stimulate aggrecan and protein synthesis in cartilage [13], robust in vivo studies in humans are lacking. It remains unclear what magnitude increase in hip contact force, if any, will restore anabolic activities without inducing catabolic cartilage degeneration [12] commonly observed in overloaded knees of people with knee OA [11]. Future research is essential to establish the relationship between cartilage strains during walking and cartilage health in people with hip OA.

After 6-weeks of walking using a prescribed personalised load modification strategy, 77% of participants reached or exceeded the minimum clinically important change for hip pain during walking. Almost half of these participants (46%) also achieved a clinically important change in physical function. Interestingly, the percentage of participants who reached a clinically important improvement in hip pain during walking was comparable between subgroups (neutral trunk ​= ​67%; neutral pelvis ​= ​75%; increased step length ​= ​80%). Given that biomechanical data were not acquired at 6-week follow-up, it is impossible to confirm whether observed improvements in symptoms corresponded with increases in hip contact force. The complex relationship between joint loading and symptoms is widely recognised [24] and future well-powered studies should aim to elucidate the extent to which hip contact force must be increased to be associated with a clinically meaningful improvement in symptoms. Many participants (71%) reported they were confident (≥ 7/10) walking using their prescribed personalised load modification strategy during the 6-week follow-up period. This study used laboratory-based visual biofeedback to assist with motor learning [22], though at least in some people with hip OA, wearable technology equipped with haptic biofeedback may be required to reinforce motor learning during walking in the real world [46].

Given the absence of a control group in this study, we caution extrapolation of results beyond those presented. Previous research has shown the placebo effect can be powerful in people with OA, leading in some cases to reductions in pain of up to 75% [47]. Further, biomechanics data were not acquired at 6-week follow-up, and thus we cannot confirm whether observed improvements in symptoms corresponded with increases in hip contact force. Future randomised controlled trials are required to establish causative links between reductions in pain and increases in hip contact force. Incorporating measures of joint structure will be essential to confirm whether increases in hip contact force are favourable, and not detrimental, to cartilage health. Our results may be used to inform sample size calculations for intervention studies that aim to investigate hip load and symptom altering effects of personalised load modification in people with hip OA. We used an EMG-informed modelling approach to estimate hip contact force, which has demonstrated excellent agreement with experimental data of people with hip OA and measurements from instrumented hip implants [48]. We calculated hip contact force via a neuromusculoskeletal model with some degree of personalisation (inclusion of EMG measurements, linear scaling, calibration of muscle parameters). However, higher levels of model personalisation, including accounting for personalised bone geometry [49], may further improve hip contact force estimates. Including estimates of regional hip contact loading (e.g. force orientation and distribution), loading frequency, and cartilage strain may provide further insight into the effects of a personalised load modification-based intervention. Participants in our study walked on a treadmill which was required to implement our real-time biofeedback technology, therefore, participants walked at slightly slower speeds than previous reports of people with hip OA [14]. Further, walking conditions were not randomised between participants in this pilot study. We included participants who had symptoms consistent with hip OA [26], though radiographs were not acquired. Some participants had bilateral hip OA, and notably, 4/6 of these participants reported an improvement in contralateral hip pain after 6-weeks. Our small sample size did not allow for unilateral/bilateral subgrouping, though this may be worthwhile in future well-powered studies. We assessed only immediate effects of the load modification interventions (one laboratory retraining session and a 6-week home-based follow-up period). Future studies should aim to establish the training period required to optimise outcomes of hip load modification-based interventions in hip OA.

Many people with hip OA continue to report unsatisfactory symptom relief [50]. There is an escalating need to improve the efficacy of conservative management strategies [2], while maintaining a persistent focus on patients’ desire to self-manage their condition [5]. Our results provide preliminary support for a personalised load modification-based intervention for the self-management of hip OA. These findings should be confirmed in future studies using an adequately powered robust randomised controlled trial design.

Author contributions

LED, MH, RSH, CP, DJS conceived the idea and designed the study. DD, BC, MLP, AH contributed to data acquisition. DD, BC, MLP, LED contributed to data analysis and interpretation. LED wrote the first draft of the article. All authors revised the paper and provided scientific input. All authors approved the final version of the manuscript. LED, on behalf of all authors, takes responsibility for the integrity of the work, from inception to manuscript.

Role of the funding source

Funding for this project was provided by an Arthritis Australia Project Grant and a Griffith University Health Group Seed Grant. Michelle Hall is supported by an Australian National Health and Medical Research Investigator Grant (#1172928). Rana S Hinman is supported by an Australian National Health and Medical Research Council Fellowship (#1154217). Study sponsors did not play any role in the study design, collection, analysis or interpretation of data, writing of the manuscript, or decision to submit the manuscript for publication.

Authorship

All authors should have made substantial contributions to all of the following: (1) the conception and design of the study, or acquisition of data, or analysis and interpretation of data, (2) drafting the article or revising it critically for important intellectual content, (3) final approval of the version to be submitted. By signing below each author also verifies that he (she) confirms that neither this manuscript, nor one with substantially similar content, has been submitted, accepted or published elsewhere (except as an abstract). Each manuscript must be accompanied by a declaration of contributions relating to sections (1), (2) and (3) above. This declaration should also name one or more authors who take responsibility for the integrity of the work as a whole, from inception to finished article. These declarations will be included in the published manuscript.

Acknowledgement of other contributors

All contributors who do not meet the criteria for authorship as defined above should be listed in an acknowledgements section. Examples of those who might be acknowledged include a person who provided purely technical help, writing assistance, or a department chair who provided only general support. Such contributors must give their consent to being named. Authors should disclose whether they had any writing assistance and identify the entity that paid for this assistance.

Conflict of interest

At the end of the text, under a subheading “Conflict of interest statement” all authors must disclose any financial and personal relationships with other people or organisations that could inappropriately influence (bias) their work. Examples of potential conflicts of interest include employment, consultancies, stock ownership, honoraria, paid expert testimony, patent applications/registrations, and research grants or other funding.

Declaration of funding

All sources of funding should be declared as an acknowledgement at the end of the text.

Role of the funding source

Authors should declare the role of study sponsors, if any, in the study design, in the collection, analysis and interpretation of data; in the writing of the manuscript; and in the decision to submit the manuscript for publication. If the study sponsors had no such involvement, the authors should state this.

Studies involving humans or animals

Clinical trials or other experimentation on humans must be in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2000. Randomized controlled trials should follow the Consolidated Standards of Reporting Trials (CONSORT) guidelines and be registered in a public trials registry.

Studies involving experiments with animals were in accordance with institution guidelines.

Please sign below to certify your manuscript complies with the above requirements and then upload this form at https://www.editorialmanager.com/oac/

Declaration of competing interest

The authors have no competing interests to declare.

Footnotes

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.ocarto.2021.100230.

Contributor Information

Laura E. Diamond, Email: l.diamond@griffith.edu.au.

Daniel Devaprakash, Email: d.devaprakash@griffith.edu.au.

Bradley Cornish, Email: bradley.cornish@griffith.edu.au.

Melanie L. Plinsinga, Email: m.plinsinga@griffith.edu.au.

Andrea Hams, Email: a.hams@griffith.edu.au.

Michelle Hall, Email: halm@unimelb.edu.au.

Rana S. Hinman, Email: ranash@unimelb.edu.au.

Claudio Pizzolato, Email: c.pizzolato@griffith.edu.au.

David J. Saxby, Email: d.saxby@griffith.edu.au.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Multimedia component 1
mmc1.docx (6.4MB, docx)

References

  • 1.Murphy L.B., Helmick C.G., Schwartz T.A., Renner J.B., Tudor G., Koch G.G., et al. One in four people may develop symptomatic hip osteoarthritis in his or her lifetime. Osteoarthritis Cartilage. 2010;18:1372–1379. doi: 10.1016/j.joca.2010.08.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Fernandes L., Hagen K.B., Bijlsma J.W., Andreassen O., Christensen P., Conaghan P.G., et al. EULAR recommendations for the non-pharmacological core management of hip and knee osteoarthritis. Ann. Rheum. Dis. 2013;72:1125–1135. doi: 10.1136/annrheumdis-2012-202745. [DOI] [PubMed] [Google Scholar]
  • 3.Egerton T., Diamond L.E., Buchbinder R., Bennell K.L., Slade S.C. A systematic review and evidence synthesis of qualitative studies to identify primary care clinicians' barriers and enablers to the management of osteoarthritis. Osteoarthritis Cartilage. 2017;25:625–638. doi: 10.1016/j.joca.2016.12.002. [DOI] [PubMed] [Google Scholar]
  • 4.Bannuru R.R., Osani M.C., Vaysbrot E.E., Arden N.K., Bennell K., Bierma-Zeinstra S.M.A., et al. OARSI guidelines for the non-surgical management of knee, hip, and polyarticular osteoarthritis. Osteoarthritis Cartilage. 2019;27:1578–1589. doi: 10.1016/j.joca.2019.06.011. [DOI] [PubMed] [Google Scholar]
  • 5.Claassen A., Kremers-van de K., van den Hoogen F.H.J., van der Laan W.H., Rijnen W.H.C., Koeter S., et al. The most important frequently asked questions of patients with hip or knee osteoarthritis: a best-worst scaling exercise. Arthritis Care Res. 2019 Jul;71(7):885–892. doi: 10.1002/acr.23719. [DOI] [PubMed] [Google Scholar]
  • 6.Felson D.T. Osteoarthritis as a disease of mechanics. Osteoarthritis Cartilage. 2013;21:10–15. doi: 10.1016/j.joca.2012.09.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Hunt M.A., Charlton J.M., Krowchuk N.M., Tse C.T.F., Hatfield G.L. Clinical and biomechanical changes following a 4-month toe-out gait modification program for people with medial knee osteoarthritis: a randomized controlled trial. Osteoarthritis Cartilage. 2018;26:903–911. doi: 10.1016/j.joca.2018.04.010. [DOI] [PubMed] [Google Scholar]
  • 8.Diamond L.E., Hoang H.X., Barrett R.S., Loureiro A., Constantinou M., Lloyd D.G., et al. Individuals with mild-to-moderate hip osteoarthritis walk with lower hip joint contact forces despite higher levels of muscle co-contraction compared to healthy individuals. Osteoarthritis Cartilage. 2020;28:924–931. doi: 10.1016/j.joca.2020.04.008. [DOI] [PubMed] [Google Scholar]
  • 9.Meyer C.A.G., Wesseling M., Corten K., Nieuwenhuys A., Monari D., Simon J.P., et al. Hip movement pathomechanics of patients with hip osteoarthritis aim at reducing hip joint loading on the osteoarthritic side. Gait Posture. 2018;59:11–17. doi: 10.1016/j.gaitpost.2017.09.020. [DOI] [PubMed] [Google Scholar]
  • 10.Wesseling M., de Groote F., Meyer C., Corten K., Simon J.P., Desloovere K., et al. Gait alterations to effectively reduce hip contact forces. J. Orthop. Res. 2015;33:1094–1102. doi: 10.1002/jor.22852. [DOI] [PubMed] [Google Scholar]
  • 11.Saxby D.J., Lloyd D.G. Osteoarthritis year in review 2016: mechanics. Osteoarthritis Cartilage. 2017;25:190–198. doi: 10.1016/j.joca.2016.09.023. [DOI] [PubMed] [Google Scholar]
  • 12.Pizzolato C., Lloyd D.G., Barrett R.S., Cook J.L., Zheng M.H., Besier T.F., et al. Bioinspired technologies to connect musculoskeletal mechanobiology to the person for training and rehabilitation. Front. Comput. Neurosci. 2017;11:96. doi: 10.3389/fncom.2017.00096. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Grodzinsky A.J., Levenston M.E., Jin M., Frank E.H. Cartilage tissue remodeling in response to mechanical forces. Annu. Rev. Biomed. Eng. 2000;2:691–713. doi: 10.1146/annurev.bioeng.2.1.691. [DOI] [PubMed] [Google Scholar]
  • 14.Diamond L.E., Allison K., Dobson F., Hall M. Hip joint moments during walking in people with hip osteoarthritis: a systematic review and meta-analysis. Osteoarthritis Cartilage. 2018;26:1415–1424. doi: 10.1016/j.joca.2018.03.011. [DOI] [PubMed] [Google Scholar]
  • 15.Hall M., Wrigley T.V., Kean C.O., Metcalf B.R., Bennell K.L. Hip biomechanics during stair ascent and descent in people with and without hip osteoarthritis. J. Orthop. Res. 2017;35:1505–1514. doi: 10.1002/jor.23407. [DOI] [PubMed] [Google Scholar]
  • 16.Moreside J., Wong I., Rutherford D. Altered erector spinae activity and trunk motion occurs with moderate and severe unilateral hip OA. J. Orthop. Res. 2018;36:1826–1832. doi: 10.1002/jor.23841. [DOI] [PubMed] [Google Scholar]
  • 17.Eng J.J., Winter D.A. Kinetic analysis of the lower limbs during walking: what information can be gained from a three-dimensional model? J. Biomech. 1995;28:753–758. doi: 10.1016/0021-9290(94)00124-m. [DOI] [PubMed] [Google Scholar]
  • 18.Constantinou M., Barrett R., Brown M., Mills P. Spatial-temporal gait characteristics in individuals with hip osteoarthritis: a systematic literature review and meta-analysis. J. Orthop. Sports Phys. Ther. 2014;44:291–297. doi: 10.2519/jospt.2014.4634. [DOI] [PubMed] [Google Scholar]
  • 19.Constantinou M., Loureiro A., Carty C., Mills P., Barrett R. Hip joint mechanics during walking in individuals with mild-to-moderate hip osteoarthritis. Gait Posture. 2017;53:162–167. doi: 10.1016/j.gaitpost.2017.01.017. [DOI] [PubMed] [Google Scholar]
  • 20.Simic M., Hunt M.A., Bennell K.L., Hinman R.S., Wrigley T.V. Trunk lean gait modification and knee joint load in people with medial knee osteoarthritis: the effect of varying trunk lean angles. Arthritis Care Res. 2012;64:1545–1553. doi: 10.1002/acr.21724. [DOI] [PubMed] [Google Scholar]
  • 21.Solomonow-Avnon D., Herman A., Levin D., Rozen N., Peled E., Wolf A. Positive outcomes following gait therapy intervention for hip osteoarthritis: a longitudinal study. J. Orthop. Res. 2017;35:2222–2232. doi: 10.1002/jor.23511. [DOI] [PubMed] [Google Scholar]
  • 22.Hodges N.J., Williams A.M., Hayes S.J., Breslin G. What is modelled during observational learning? J. Sports Sci. 2007;25:531–545. doi: 10.1080/02640410600946860. [DOI] [PubMed] [Google Scholar]
  • 23.Pizzolato C., Reggiani M., Saxby D.J., Ceseracciu E., Modenese L., Lloyd D.G. Biofeedback for gait retraining based on real-time estimation of tibiofemoral joint contact forces. IEEE Trans. Neural Syst. Rehabil. Eng. 2017;25:1612–1621. doi: 10.1109/TNSRE.2017.2683488. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Hall M., Chabra S., Shakoor N., Leurgans S.E., Demirtas H., Foucher K.C. Hip joint moments in symptomatic vs. asymptomatic people with mild radiographic hip osteoarthritis. J. Biomech. 2019;96:109347. doi: 10.1016/j.jbiomech.2019.109347. [DOI] [PubMed] [Google Scholar]
  • 25.Eldridge S.M., Chan C.L., Campbell M.J., Bond C.M., Hopewell S., Thabane L., et al. CONSORT 2010 statement: extension to randomised pilot and feasibility trials. BMJ. 2016;355:i5239. doi: 10.1136/bmj.i5239. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Altman R., Alarcon G., Appelrouth D., Bloch D., Borenstein D., Brandt K., et al. The American College of Rheumatology criteria for the classification and reporting of osteoarthritis of the hip. Arthritis Rheum. 1991;34:505–514. doi: 10.1002/art.1780340502. [DOI] [PubMed] [Google Scholar]
  • 27.Hermens H.J., Freriks B., Disselhorst-Klug C., Rau G. Development of recommendations for SEMG sensors and sensor placement procedures. J. Electromyogr. Kinesiol. 2000;10:361–374. doi: 10.1016/s1050-6411(00)00027-4. [DOI] [PubMed] [Google Scholar]
  • 28.Borg G.A. Psychophysical bases of perceived exertion. Med. Sci. Sports Exerc. 1982;14:377–381. [PubMed] [Google Scholar]
  • 29.Bellamy N. Osteoarthritis clinical trials: candidate variables and clinimetric properties. J. Rheumatol. 1997;24:768–778. [PubMed] [Google Scholar]
  • 30.Bellamy N., Carette S., Ford P.M., Kean W.F., le Riche N.G., Lussier A., et al. Osteoarthritis antirheumatic drug trials. III. Setting the delta for clinical trials-results of a consensus development (Delphi) exercise. J. Rheumatol. 1992;19:451–457. [PubMed] [Google Scholar]
  • 31.ten Klooster P.M., Drossaers-Bakker K.W., Taal E., van de Laar M.A. Patient-perceived satisfactory improvement (PPSI): interpreting meaningful change in pain from the patient's perspective. Pain. 2006;121:151–157. doi: 10.1016/j.pain.2005.12.021. [DOI] [PubMed] [Google Scholar]
  • 32.McConnell S., Kolopack P., Davis A.M. The Western Ontario and McMaster Universities osteoarthritis index (WOMAC): a review of its utility and measurement properties. Arthritis Rheum. 2001;45:453–461. doi: 10.1002/1529-0131(200110)45:5<453::aid-art365>3.0.co;2-w. [DOI] [PubMed] [Google Scholar]
  • 33.Tubach F., Ravaud P., Martin-Mola E., Awada H., Bellamy N., Bombardier C., et al. Minimum clinically important improvement and patient acceptable symptom state in pain and function in rheumatoid arthritis, ankylosing spondylitis, chronic back pain, hand osteoarthritis, and hip and knee osteoarthritis: results from a prospective multinational study. Arthritis Care Res. 2012;64:1699–1707. doi: 10.1002/acr.21747. [DOI] [PubMed] [Google Scholar]
  • 34.Mantoan A., Pizzolato C., Sartori M., Sawacha Z., Cobelli C., Reggiani M. MOtoNMS: a MATLAB toolbox to process motion data for neuromusculoskeletal modeling and simulation. Source Code Biol. Med. 2015;10:1–14. doi: 10.1186/s13029-015-0044-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Delp S.L., Anderson F.C., Arnold A.S., Loan P., Habib A., John C.T., et al. OpenSim: open-source software to create and analyze dynamic simulations of movement. IEEE Trans. Biomed. Eng. 2007;54:1940–1950. doi: 10.1109/TBME.2007.901024. [DOI] [PubMed] [Google Scholar]
  • 36.Devaprakash D., Weir G.J., Dunne J.J., Alderson J.A., Donnelly C.J. The influence of digital filter type, amplitude normalisation method, and co-contraction algorithm on clinically relevant surface electromyography data during clinical movement assessments. J. Electromyogr. Kinesiol. 2016;31:126–135. doi: 10.1016/j.jelekin.2016.10.001. [DOI] [PubMed] [Google Scholar]
  • 37.Kainz H., Hoang H.X., Stockton C., Boyd R.R., Lloyd D.G., Carty C.P. Accuracy and reliability of marker-based approaches to scale the pelvis, thigh, and shank segments in musculoskeletal models. J. Appl. Biomech. 2017;33:354–360. doi: 10.1123/jab.2016-0282. [DOI] [PubMed] [Google Scholar]
  • 38.Modenese L., Ceseracciu E., Reggiani M., Lloyd D.G. Estimation of musculotendon parameters for scaled and subject specific musculoskeletal models using an optimization technique. J. Biomech. 2016;49:141–148. doi: 10.1016/j.jbiomech.2015.11.006. [DOI] [PubMed] [Google Scholar]
  • 39.Pizzolato C., Lloyd D.G., Sartori M., Ceseracciu E., Besier T.F., Fregly B.J., et al. CEINMS: a toolbox to investigate the influence of different neural control solutions on the prediction of muscle excitation and joint moments during dynamic motor tasks. J. Biomech. 2015;48:3929–3936. doi: 10.1016/j.jbiomech.2015.09.021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Sartori M., Farina D., Lloyd D.G. Hybrid neuromusculoskeletal modeling to best track joint moments using a balance between muscle excitations derived from electromyograms and optimization. J. Biomech. 2014;47:3613–3621. doi: 10.1016/j.jbiomech.2014.10.009. [DOI] [PubMed] [Google Scholar]
  • 41.Hoang H.X., Pizzolato C., Diamond L.E., Lloyd D.G. Subject-specific calibration of neuromuscular parameters enables neuromusculoskeletal models to estimate physiologically plausible hip joint contact forces in healthy adults. J. Biomech. 2018;80:111–120. doi: 10.1016/j.jbiomech.2018.08.023. [DOI] [PubMed] [Google Scholar]
  • 42.Lim Y.P., Lin Y.C., Pandy M.G. Effects of step length and step frequency on lower-limb muscle function in human gait. J. Biomech. 2017;57:1–7. doi: 10.1016/j.jbiomech.2017.03.004. [DOI] [PubMed] [Google Scholar]
  • 43.Allison K., Wrigley T.V., Vicenzino B., Bennell K.L., Grimaldi A., Hodges P.W. Kinematics and kinetics during walking in individuals with gluteal tendinopathy. Clin. Biomech. 2016;32:56–63. doi: 10.1016/j.clinbiomech.2016.01.003. [DOI] [PubMed] [Google Scholar]
  • 44.Simic M., Hinman R.S., Wrigley T.V., Bennell K.L., Hunt M.A. Gait modification strategies for altering medial knee joint load: a systematic review. Arthritis Care Res. 2011;63:405–426. doi: 10.1002/acr.20380. [DOI] [PubMed] [Google Scholar]
  • 45.Bader D.L., Salter D.M., Chowdhury T.T. Biomechanical influence of cartilage homeostasis in health and disease. Arthritis. 2011;2011:979032. doi: 10.1155/2011/979032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Pizzolato C., Lloyd D.G., Zheng M.H., Besier T.F., Shim V.B., Obst S.J., et al. Finding the sweet spot via personalised Achilles tendon training: the future is within reach. Br. J. Sports Med. 2019;53:11–12. doi: 10.1136/bjsports-2018-099020. [DOI] [PubMed] [Google Scholar]
  • 47.Zhang W. The powerful placebo effect in osteoarthritis. Clin. Exp. Rheumatol. 2019;37(Suppl 120):118–123. [PubMed] [Google Scholar]
  • 48.Hoang H.X., Diamond L.E., Lloyd D.G., Pizzolato C. A calibrated EMG-informed neuromusculoskeletal model can appropriately account for muscle co-contraction in the estimation of hip joint contact forces in people with hip osteoarthritis. J. Biomech. 2019;83:134–142. doi: 10.1016/j.jbiomech.2018.11.042. [DOI] [PubMed] [Google Scholar]
  • 49.Kainz H., Killen B.A., Wesseling M., Perez-Boerema F., Pitto L., Garcia Aznar J.M., et al. A multi-scale modelling framework combining musculoskeletal rigid-body simulations with adaptive finite element analyses, to evaluate the impact of femoral geometry on hip joint contact forces and femoral bone growth. PLoS One. 2020;15 doi: 10.1371/journal.pone.0235966. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Basedow M., Hibbert P., Hooper T., Runciman W., Esterman A. Australians with osteoarthritis: satisfaction with health care providers and the perceived helpfulness of treatments and information sources. J. Multidiscip. Healthc. 2016;9:387–394. doi: 10.2147/JMDH.S110751. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Multimedia component 1
mmc1.docx (6.4MB, docx)

Articles from Osteoarthritis and Cartilage Open are provided here courtesy of Elsevier

RESOURCES