Machine learning for automating subjective clinical assessment of gait impairment in people with acquired brain injury - a comparison of an image extraction and classification system to expert scoring.

BACKGROUND: Walking impairment is a common disability post acquired brain injury (ABI), with visually evident arm movement abnormality identified as negatively impacting a multitude of psychological factors. The International Classification of Functioning, Disability and Health (ICF) qualifiers scale has been used to subjectively assess arm movement abnormality, showing strong intra-rater and test-retest reliability, however, only moderate inter-rater reliability. This impacts clinical utility, limiting its use as a measurement tool. To both automate the analysis and overcome these errors, the primary aim of this study was to evaluate the ability of a novel two-level machine learning model to assess arm movement abnormality during walking in people with ABI. METHODS: Frontal plane gait videos were used to train four networks with 50%, 75%, 90%, and 100% of participants (ABI: n = 42, healthy controls: n = 34) to automatically identify anatomical landmarks using DeepLabCut™ and calculate two-dimensional kinematic joint angles. Assessment scores from three experienced neurorehabilitation clinicians were used with these joint angles to train random forest networks with nested cross-validation to predict assessor scores for all videos. Agreement between unseen participant (i.e. test group participants that were not used to train the model) predictions and each individual assessor's scores were compared using quadratic weighted kappa. One sample t-tests (to determine over/underprediction against clinician ratings) and one-way ANOVA (to determine differences between networks) were applied to the four networks. RESULTS: The machine learning predictions have similar agreement to experienced human assessors, with no statistically significant (p < 0.05) difference for any match contingency. There was no statistically significant difference between the predictions from the four networks (F = 0.119; p = 0.949). The four networks did however under-predict scores with small effect sizes (p range = 0.007 to 0.040; Cohen's d range = 0.156 to 0.217). CONCLUSIONS: This study demonstrated that machine learning can perform similarly to experienced clinicians when subjectively assessing arm movement abnormality in people with ABI. The relatively small sample size may have resulted in under-prediction of some scores, albeit with small effect sizes. Studies with larger sample sizes that objectively and automatically assess dynamic movement in both local and telerehabilitation assessments, for example using smartphones and edge-based machine learning, to reduce measurement error and healthcare access inequality are needed.

Controlling modified Tardieu scale assessment speeds to match joint angular velocities during walking impacts spasticity assessment outcomes.

OBJECTIVE: To investigate whether tailoring the speed of the Modified Tardieu Scale to reflect an individual's joint angular velocity during walking influences spasticity assessment outcomes. DESIGN: Observational trial. SETTING: Inpatient and outpatient neurological hospital department. SUBJECTS: Ninety adults with lower-limb spasticity. INTERVENTIONS: N/A. MAIN MEASURES: The Modified Tardieu Scale was used to assess the gastrocnemius, soleus, hamstrings and quadriceps. The V1 (slow) and V3 (fast) movements were completed as per standardised testing. Two additional assessments were completed, reflecting joint angular velocities during walking based on (i) a healthy control database (controlled velocity) and (ii) the individual's real-time joint angular velocities during walking (matched velocity). The agreement was compared using Cohen's and Weighted Kappa statistics, sensitivity and specificity. RESULTS: There was poor agreement when rating trials as spastic or not spastic at the ankle joint (Cohen's Kappa = 0.01-0.17). Trials were classified as spastic during V3 and not spastic during the controlled conditions in 81.6-85.1% of trials when compared to stance phase dorsiflexion angular velocities and 48.0-56.4% when compared to swing phase dorsiflexion angular velocities. The severity of muscle reaction demonstrated poor agreement at the ankle (Weighted Kappa = 0.01-0.28). At the knee, there was a moderate-excellent agreement between the V3 and controlled conditions when rating a trial as spastic or not spastic (Cohen's Kappa = 0.66-0.84) and excellent agreement when comparing severity (Weighted Kappa = 0.73-0.94). CONCLUSION: The speed of assessment impacted spasticity outcomes. It is possible that the standardised protocol may overestimate the impact spasticity has on walking, especially at the ankle.

Discordance Between Self-Reported and Performance-Based Physical Function in Patients Who Have Knee Osteoarthritis: Associations With Pain Intensity and Negative Affect.

BACKGROUND: Although self-reported measures of physical disability are strong indication criterion for total knee arthroplasty (TKA) in painful knee osteoarthritis (OA), some patients may report greater-than-observed disability. Contributing factors to this discordance are relatively unexplored. We aimed to examine whether pain and negative affect, including anxiety and depression, were associated with the discordance of self-reported measures with performance-based measures (PPM) of physical function. METHODS: We used cross-sectional data (n = 212) from two randomized rehabilitation trials in knee OA. All patients were assessed for knee pain intensity and symptoms of anxiety and depression. Self-reported function was assessed by the Western Ontario and McMaster Universities Arthritis Index (WOMAC) physical-function subscale. Objective performance-based measures (PPMs) of physical function were assessed by timed gait and stair tests. Continuous discordance scores were quantified by the difference in percentiles between WOMAC and PPMs (WOMAC-PPM), where a positive discordance, WOMAC-PPM >0, implied greater perceived than observed disability. RESULTS: Around 1 in 4 patients had >20 percentile units in WOMAC-PPM discordance. In Bayesian regression analyses, knee pain intensity had >99% posterior probability of positive associations with WOMAC-PPM discordance. Among patients awaiting TKA, anxiety intensity had approximately 99% probability of positive associations with discordance, and these associations had >65% probability of exceeding 10 percentile units. In contrast, depression had low (79% to 88%) probability of any association with discordance. CONCLUSION: In patients who have knee OA, a sizable proportion reported substantially greater physical disability than actually observed. Pain and anxiety intensity, but not depression, were meaningful predictors of this discordance. If validated, our findings may help in refining patient selection criteria for TKA.

The Effectiveness of Aquatic Plyometric Training in Improving Strength, Jumping, and Sprinting: A Systematic Review.

CONTEXT: Aquatic plyometric training may provide benefits due to reduced joint loading compared with land plyometric training; however, the reduced loading may also limit performance gains. OBJECTIVE: To systematically review the effect of aquatic plyometric training on strength, performance outcomes, soreness, and adverse events in healthy individuals. EVIDENCE ACQUISITION: Five databases were searched from inception to June 2020. Quality assessment and data extraction were independently completed by 2 investigators. When similar outcome measures were used, standardized mean differences were calculated. EVIDENCE SYNTHESIS: A total of 19 randomized controlled trials with 633 participants (mean age, range 14-30 y) were included. Aquatic plyometric training was most commonly performed in waist to chest deep water (12/19 studies), 2 to 3 times per week for 6 to 12 weeks (18/19 studies), with final program foot contacts ranging from 120 to 550. Meta-analyses were not completed due to the clinical and statistical heterogeneity between studies. Compared with land plyometric training, aquatic plyometric training exercises and dosage were replicated (15/16 studies) and showed typically similar performance gains (3/4 knee extensor strength measures, 2/4 leg extensor strength measures, 3/4 knee flexor strength measures, 7/10 vertical jump measures, 3/3 sprint measures). In total, 2 of 3 studies monitoring muscle soreness reported significantly less soreness following training in water compared with on land. Compared with no active training (no exercise control group or passive stretching), most effect sizes demonstrated a mean improvement favoring aquatic plyometric training (23/32 measures). However, these were not significant for the majority of studies measuring isokinetic knee strength, vertical jump, and sprinting. The effect sizes for both studies assessing leg press strength indicated that aquatic plyometric training is significantly more effective than no training. CONCLUSION: Aquatic plyometric training appears similarly effective to land plyometric exercise for improving strength, jumping, and sprinting and may be indicated when joint impact loading needs to be minimized. However, the low quality of studies limits the strength of the conclusions.

The reproducibility and responsiveness of subjective assessment of upper limb associated reactions in people with acquired brain injury during walking.

OBJECTIVE: The aim of this study is to determine inter-rater, test-retest and intra-rater reproducibility and responsiveness of subjective assessment of upper limb associated reactions in people with acquired brain injury using (1) the 'Qualifiers Scale' of the International Classification of Functioning, Disability and Health Framework, and (2) visually estimated elbow flexion angle during walking. DESIGN: Observational study. SETTING: A brain injury rehabilitation centre, Melbourne, Australia. SUBJECTS: People with acquired brain injury and upper limb associated reactions and experienced neurological physiotherapists. MAIN MEASURES: The Qualifiers Scale applied to individual upper limb joints and global associated reaction on a 5-point scale (0-4), a summed upper limb severity score and visually estimated elbow flexion angle. RESULTS: A total of 42 people with acquired brain injury (mean age: 48.4 ± 16.5 years) were videoed walking at self-selected and fast speeds. A subset of 30 chronic brain injury participants (mean time post injury: 8.2 ± 9.3 years) were reassessed one week later for retest reproducibility. Three experienced neurological physiotherapists (mean experience: 22.7 ± 9.1 years) viewed these videos and subjectively rated the upper limb associated reactions. Strong-to-very strong test-retest, intra- and inter-rater reproducibility was found for elbow flexion angle (ICC > 0.86) and the Qualifiers Scale applied to global and individual upper limb joints (ICC > 0.60). Responsiveness of change from self-selected to fast walking speed (mean increase 0.46 m/s) was highest for elbow flexion angle (effect size = 0.83) and low-to-moderate for the Qualifiers Scale. CONCLUSION: Subjectively rated associated reactions during walking demonstrated strong reproducibility and moderate responsiveness to speed change. The Qualifiers Scale and elbow flexion angle can both subjectively quantify associated reactions during walking in a clinical setting.

Multidisciplinary home-based rehabilitation in inoperable lung cancer: a randomised controlled trial.

BACKGROUND: Lung cancer is associated with poor health-related quality of life (HRQoL) and high symptom burden. This trial aimed to assess the efficacy of home-based rehabilitation versus usual care in inoperable lung cancer. METHODS: A parallel-group, assessor-blinded, allocation-concealed, randomised controlled trial. Eligible participants were allocated (1:1) to usual care (UC) plus 8 weeks of aerobic and resistance exercise with behaviour change strategies and symptom support (intervention group (IG)) or UC alone. Assessments occurred at baseline, 9 weeks and 6 months. The primary outcome, change in between-group 6 min walk distance (6MWD), was analysed using intention-to-treat (ITT). Subsequent analyses involved modified ITT (mITT) and included participants with at least one follow-up outcome measure. Secondary outcomes included HRQoL and symptoms. RESULTS: Ninety-two participants were recruited. Characteristics of participants (UC=47, IG=45): mean (SD) age 64 (12) years; men 55%; disease stage n (%) III=35 (38) and IV=48 (52); radical treatment 46%. There were no significant between-group differences for the 6MWD (n=92) at 9 weeks (p=0.308) or 6 months (p=0.979). The mITT analyses of 6MWD between-group differences were again non-significant (mean difference (95% CI): 9 weeks: -25.4 m (-64.0 to 13.3), p=0.198 and 6 months: 41.3 m (-26.7 to 109.4), p=0.232). Significant 6-month differences, favouring the IG, were found for HRQoL (Functional Assessment of Cancer Therapy-Lung: 13.0 (3.9 to 22.1), p=0.005) and symptom severity (MD Anderson Symptom Inventory-Lung Cancer: -2.2 (-3.6 to -0.9), p=0.001). CONCLUSIONS: Home-based rehabilitation did not improve functional exercise capacity but there were improvements in patient-reported exploratory secondary outcomes measures observed at 6 months. TRIAL REGISTRATION: Australian New Zealand Clinical Trials Registry (ACTRN12614001268639).

Toward Accurate Clinical Spasticity Assessment: Validation of Movement Speed and Joint Angle Assessments Using Smartphones and Camera Tracking.

OBJECTIVE: To investigate whether a three-dimensional (3-D) camera (Microsoft Kinect) and a smartphone can be used to accurately quantify the joint angular velocity and range of motion (ROM) compared to a criterion-standard 3-D motion analysis system during a lower limb spasticity assessment. DESIGN: Observational, criterion-standard comparison study. SETTING: Large rehabilitation center. PARTICIPANTS: A convenience sample of 35 controls, 35 patients with a neurologic condition, and 34 rehabilitation professionals (physiotherapists and rehabilitation doctors) participated (N=104). INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: The Modified Tardieu Scale was used to assess spasticity of the quadriceps, hamstrings, soleus, and gastrocnemius. Data for each trial were collected concurrently using the criterion-standard Optitrack 3-D motion analysis (3DMA) system, Microsoft Kinect, and a smartphone. Each healthy control participant was assessed by 1 health professional and each patient with a neurological condition was assessed by 3 health professionals. Spearman correlation coefficient and intraclass correlation coefficient with 95% confidence intervals were used to report the strength of the relationships investigated. RESULTS: The smartphone and Microsoft Kinect demonstrated excellent concurrent validity with the 3DMA system. Overall, 74.8% of the relationships investigated demonstrated a very strong (≥0.80) correlation across all of the testing parameters. The Microsoft Kinect was superior to the smartphone for measuring joint start and end angle, the smartphone was superior for measuring joint angular velocity, and the 2 systems were comparable for measuring total joint ROM. CONCLUSIONS: These findings provide preliminary evidence that user-friendly, low-cost technologies can be used to facilitate accurate measurements of joint angular velocity and angles during a lower limb spasticity assessment in a clinical setting.

Do video game interventions improve motor outcomes in children with developmental coordination disorder? A systematic review using the ICF framework.

BACKGROUND: Children with developmental coordination disorder (DCD) experience a range of difficulties that can potentially limit their academic, social and physical ability. Recent research has developed interventions that aim to improve motor outcomes in a variety of paediatric cohorts using video gaming equipment. Therefore, we aimed to systematically review the literature on virtual reality or video game interventions that aim to improve motor outcomes in children with DCD. METHODS: Seven databases were searched for studies using the following criteria: a) virtual reality or video game based intervention; b) children with DCD; and c) motor outcomes relating to body structure and function, activity or participation. Data were extracted relating to study design, participant characteristics, details of the intervention, outcome measures, results, and feasibility/adherence. RESULTS: Fifteen articles were included for review, including eight randomised controlled trials. No studies used virtual reality equipment, with all interventions using video games (Nintendo Wii in 12/15 articles). Mixed effects of video game intervention on outcome were found, with conflicting evidence across studies. Studies that reported on feasibility found most children enjoyed and adhered to the video game interventions. CONCLUSIONS: This review found limited evidence for the effectiveness of video game interventions for children with DCD to improve motor outcomes due to limitations in the research including low sample sizes and low to moderate methodological quality. Further research is needed to determine the effect of video game or virtual reality interventions on motor outcomes in children with DCD. PROTOCOL REGISTRATION: The protocol for this systematic review can be found on PROSPERO ( CRD42017064427 ).

The nature and extent of upper limb associated reactions during walking in people with acquired brain injury.

BACKGROUND: Upper limb associated reactions (ARs) are common in people with acquired brain injury (ABI). Despite this, there is no gold-standard outcome measure and no kinematic description of this movement disorder. The aim of this study was to determine the upper limb kinematic variables most frequently affected by ARs in people with ABI compared with a healthy cohort at matched walking speed intention. METHODS: A convenience sample of 36 healthy control adults (HCs) and 42 people with ABI who had upper limb ARs during walking were recruited and underwent assessment of their self-selected walking speed using the criterion-reference three dimensional motion analysis (3DMA) at Epworth Hospital, Melbourne. Shoulder flexion, abduction and rotation, elbow flexion, forearm rotation and wrist flexion were assessed. The mean angle, standard deviation (SD), peak joint angles and total joint angle range of motion (ROM) were calculated for each axis across the gait cycle. On a group level, ANCOVA was used to assess the between-group differences for each upper limb kinematic outcome variable. To quantify abnormality prevalence on an individual participant level, the percentage of ABI participants that were outside of the 95% confidence interval of the HC sample for each variable were calculated. RESULTS: There were significant between-group differences for all elbow and shoulder abduction outcome variables (p < 0.01), most shoulder flexion variables (except for shoulder extension peak), forearm rotation SD and ROM and for wrist flexion ROM. Elbow flexion and shoulder abduction were the axes most frequently affected by ARs. Despite the elbow being the most prevalently affected (38/42, 90%), a large proportion of participants had abnormality, defined as ±1.96 SD of the HC mean, present at the shoulder (32/42, 76%), forearm (20/42, 48%) and wrist joints (10/42, 24%). CONCLUSION: This study provides valuable information on ARs, and highlights the need for clinical assessment of ARs to include all of the major joints of the upper limb. This may inform the development of a criterion-reference outcome measure or classification system specific to ARs.

Greater knee flexion excursion/moment in hopping is associated with better knee function following anterior cruciate ligament reconstruction.

PURPOSE: Individuals with impaired knee function after anterior cruciate ligament reconstruction (ACLR) may be at greater risk of developing knee osteoarthritis related to abnormal knee joint movement and loading. The aim of this study was to assess the association between knee biomechanics and knee laxity during hopping and clinically assessed knee function (i.e., patient-reported knee function and hop tests) following ACLR. METHODS: Sixty-six participants (23 women, mean age 28 ± 6 years, mean 18 ± 3 months following ACLR) completed a standardized single-leg hopping task. Three-dimensional movement analysis was used to assess knee flexion excursion and body weight/height normalized knee flexion moments during landing for the involved limb. Anterior-posterior knee laxity was assessed with a KT-1000 knee arthrometer. Participants then completed a patient-reported knee function questionnaire and three separate hop tests (% of uninvolved limb) and were divided into poor and satisfactory knee function groups (satisfactory: ≥85% patient-reported knee function and ≥ 85% hop test symmetry). Associations between knee function and hop biomechanics/knee laxity were assessed using logistic regression and interquartile range scaled odds ratios (ORIQR). RESULTS: Greater knee flexion excursion (ORIQR 2.9, 95%CI 1.1-7.8), greater knee flexion moment (ORIQR 4.9, 95%CI 1.6-14.3) and lesser knee laxity (ORIQR 4.7, 95%CI 1.5-14.9) were significantly associated with greater odds of having satisfactory knee function (≥ 85% patient-reported knee function and ≥ 85% hop test symmetry). CONCLUSION: Greater knee flexion excursion/moment during hop-landing and lesser knee laxity is associated with better patient-reported knee function and single-leg hop test performance following ACLR. Patients with lower levels of knee function following ACLR demonstrated hop-landing biomechanics previously associated with early patellofemoral osteoarthritis. Therefore, interventions aimed at improving hop landing biomechanics in people with poor knee function are likely required. LEVEL OF EVIDENCE: III, Cross-sectional study.

Calf muscle density is independently associated with physical function in overweight and obese older adults.

OBJECTIVES: To determine whether associations of calf muscle density with physical function are independent of other determinants of functional decline in overweight and obese older adults. METHODS: This was a secondary analysis of a cross-sectional study of 85 community-dwelling overweight and obese adults (mean±SD age 62.8±7.9 years; BMI 32.3±6.1 kg/m2; 58% women). Peripheral quantitative computed tomography assessed mid-calf muscle density (66% tibial length) and dual-energy X-ray absorptiometry determined visceral fat area. Fasting glucose, Homeostatic Model Assessment of Insulin Resistance (HOMA-IR) and C-reactive protein (CRP) were analysed. Physical function assessments included hand grip and knee extension strength, balance path length (computerised posturography), stair climb test, Short Physical Performance Battery (SPPB) and self-reported falls efficacy (Modified Falls Efficacy Scale; M-FES). RESULTS: Visceral fat area, not muscle density, was independently associated with CRP and fasting glucose (B=0.025; 95% CI 0.009-0.042 and B=0.009; 0.001-0.017, respectively). Nevertheless, higher muscle density was independently associated with lower path length and stair climb time, and higher SPPB and M-FES scores (all P⟨0.05). Visceral fat area, fasting glucose and CRP did not mediate these associations. CONCLUSIONS: Higher calf muscle density predicts better physical function in overweight and obese older adults independent of insulin resistance, visceral adiposity or inflammation.

Do Trials of Resistance Training to Improve Mobility After Stroke Adhere to the American College of Sports Medicine Guidelines? A Systematic Review.

OBJECTIVE: To determine whether resistance training to improve mobility outcomes after stroke adheres to the American College of Sports Medicine (ACSM) guidelines, and whether adherence was associated with better outcomes. DATA SOURCES: Online databases searched from 1975 to October 30, 2016. STUDY SELECTION: Randomized controlled trials examining the effectiveness of lower limb strength training on mobility outcomes in adult participants with stroke. DATA EXTRACTION: Two independent reviewers completed data extraction. Quality of trials was determined using the Cochrane Risk of Bias Tool. Trials were scored based on their protocol's adherence to 8 ACSM recommendations. To determine if a relation existed between total adherence score and effect size, Spearman ρ was calculated, and between individual recommendations and effect size, Mann-Whitney U or Kruskal-Wallis tests were used. DATA SYNTHESIS: Thirty-nine trials met the inclusion criteria, and 34 were scored on their adherence to the guidelines. Adherence was high for frequency of training (100% of studies), but few trials adhered to the guidelines for intensity (32%), specificity (24%), and training pattern (3%). Based on the small number of studies that could be included in pooled analysis (n=12), there was no relation between overall adherence and effect size (Spearman ρ=-.39, P=.21). CONCLUSIONS: Adherence to the ACSM guidelines for resistance training after stroke varied widely. Future trials should ensure strength training protocols adhere more closely to the guidelines, to ensure their effectiveness in stroke can be accurately determined.

Factors Associated With Post-Stroke Physical Activity: A Systematic Review and Meta-Analysis.

OBJECTIVE: To integrate the literature investigating factors associated with post-stroke physical activity. DATA SOURCES: A search was conducted from database inception to June 2016 across 9 databases: Cochrane, MEDLINE, ProQuest, Web of Science, PsycINFO, Scopus, Embase, CINAHL, and Allied and Complementary Medicine Database. The reference lists of included articles were screened for secondary literature. STUDY SELECTION: Cohort and cross-sectional studies were included if they recruited community-dwelling stroke survivors and measured factors associated with physical activity. DATA EXTRACTION: Risk of bias was evaluated using the Quality in Prognosis Studies checklist. A meta-analysis was conducted for correlates where there were at least 2 studies that reported a correlation value. Correlation values were used in an effect size measure and converted to a standardized unit with Fisher r to z transformation and conversion back to r method. Results were described qualitatively for studies that could not be pooled. DATA SYNTHESIS: There were 2161 studies screened and 26 studies included. Age (meta r=-.17; P≤.001) and sex (meta r=-.01; P=.02) were the nonmodifiable factors that were found to be associated with post-stroke physical activity. The modifiable factors were physical function (meta r=.68-.73; P<.001), cardiorespiratory fitness (meta r=.35; P≤.001), fatigue (meta r=-.22; P=.01), falls self-efficacy (meta r=-.33; P<.001), balance self-efficacy (meta r=.37; P<.001), depression (meta r=-.58 to .48; P<.001), and health-related quality of life (meta r=.38-.43; P<.001). The effect of side of infarct, neglect, and cognition on post-stroke physical activity was inconclusive. CONCLUSIONS: Age, sex, physical function, depression, fatigue, self-efficacy, and quality of life were factors associated with post-stroke physical activity. The cause and effect of these relations are unclear, and the possibility of reverse causality needs to be addressed.

Feasibility of Ballistic Strength Training in Subacute Stroke: A Randomized, Controlled, Assessor-Blinded Pilot Study.

OBJECTIVE: To establish the feasibility and effectiveness of a 6-week ballistic strength training protocol in people with stroke. DESIGN: Randomized, controlled, assessor-blinded study. SETTING: Subacute inpatient rehabilitation. PARTICIPANTS: Consecutively admitted inpatients with a primary diagnosis of first-ever stroke with lower limb weakness, functional ambulation category score of ≥3, and ability to walk ≥14 m were screened for eligibility to recruit 30 participants for randomization. INTERVENTIONS: Participants were randomized to standard therapy or ballistic strength training 3 times per week for 6 weeks. MAIN OUTCOME MEASURES: The primary aim was to evaluate feasibility and outcomes included recruitment rate, participant retention and attrition, feasibility of the exercise protocol, therapist burden, and participant safety. Secondary outcomes included measures of mobility, lower limb muscle strength, muscle power, and quality of life. RESULTS: A total of 30 participants (11% of those screened) with mean age of 50 years (SD 18) were randomized. The median number of sessions attended was 15 of 18 and 17 of 18 for the ballistic and control groups, respectively. Earlier than expected discharge to home (n=4) and illness (n=7) were the most common reasons for nonattendance. Participants performed the exercises safely, with no study-related adverse events. There were significant (P<.05) between-group changes favoring the ballistic group for comfortable gait velocity (mean difference [MD] 0.31m/s, 95% confidence interval [CI]: 0.08-0.52), muscle power, as measured by peak jump height (MD 8cm, 95% CI: 3-13), and peak propulsive velocity (MD 64cm/s, 95% CI: 17-112). CONCLUSIONS: Ballistic training was safe and feasible in select ambulant people with stroke. Similar rates of retention and attrition suggest that ballistic training was acceptable to patients. Secondary outcomes provide promising results that warrant further investigation in a larger trial.

Three-dimensional assessment of squats and drop jumps using the Microsoft Xbox One Kinect: Reliability and validity.

The Microsoft Xbox One Kinect™ (Kinect V2) contains a depth camera that can be used to manually identify anatomical landmark positions in three-dimensions independent of the standard skeletal tracking, and therefore has potential for low-cost, time-efficient three-dimensional movement analysis (3DMA). This study examined inter-session reliability and concurrent validity of the Kinect V2 for the assessment of coronal and sagittal plane kinematics for the trunk, hip and knee during single leg squats (SLS) and drop vertical jumps (DVJ). Thirty young, healthy participants (age = 23 ± 5yrs, male/female = 15/15) performed a SLS and DVJ protocol that was recorded concurrently by the Kinect V2 and 3DMA during two sessions, one week apart. The Kinect V2 demonstrated good to excellent reliability for all SLS and DVJ variables (ICC ≥ 0.73). Concurrent validity ranged from poor to excellent (ICC = 0.02 to 0.98) during the SLS task, although trunk, hip and knee flexion and two-dimensional measures of knee abduction and frontal plane projection angle all demonstrated good to excellent validity (ICC ≥ 0.80). Concurrent validity for the DVJ task was typically worse, with only two variables exceeding ICC = 0.75 (trunk and hip flexion). These findings indicate that the Kinect V2 may have potential for large-scale screening for ACL injury risk, however future prospective research is required.

Poor knee function after ACL reconstruction is associated with attenuated landing force and knee flexion moment during running.

PURPOSE: Poor knee function after anterior cruciate ligament reconstruction (ACLR) may increase the risk of future knee symptoms and knee osteoarthritis via abnormal knee joint loading patterns, particularly during high-impact activity. This study aimed to assess the relationship between poor self-reported or clinically measured knee function and knee moments/vertical ground reaction force (vGRF) in individuals following ACLR. METHODS: 61 participants (mean 16.5 ± 3 months following ACLR, 23 women) completed a patient-reported knee function questionnaire and three hop tests (% of uninvolved limb). Participants were divided into satisfactory and poor knee function groups (poor < 85% patient-reported knee function and/or < 85% hop test symmetry). The knee biomechanics of both groups were assessed with three-dimensional motion analysis during the stance phase of overland running at self-selected speeds, and the association between knee function and knee moments was assessed using analysis of covariance with running speed as a covariate. RESULTS: Participants with poor knee function (n = 30) ran with significantly smaller peak knee flexion moments (moderate effect size 0.7, p = 0.03) and significantly smaller peak vGRFs (large effect size 1.0, p = 0.002) compared to those with satisfactory knee function (n = 31). No significant differences were observed for knee adduction and knee external rotation moments or knee kinematics. CONCLUSION: Individuals following ACLR with poor self-reported knee function and/or hop test performance demonstrate knee moments during running that may be associated with lower knee joint contact forces. These findings provide greater understanding of the relationship between knee biomechanics during running and clinical assessments of knee function. LEVEL OF EVIDENCE: III. Cross-sectional study.

Open-source 3D printed sensors for hand strength assessment: Validation of low-cost load cell and fabric sensor-based systems.

BACKGROUND/AIM: Pinch and grip strength assessment is commonly performed in occupational therapy practice. However, the typically utilised methods are limited to pinch or grip dynamometers and do not readily translate to handling everyday objects. With the advent of consumer-grade 3D printing and low-cost sensor systems, the possibilities for creating customised assessment devices are expanding. As a first step in determining the validity of low-cost sensor systems, their data must be compared to a gold standard. Consequently, this study examined the criterion validity of two such systems for measuring pinch strength, specifically a small load cell and pressure-sensing fabric, with a mechanical pinch gauge. METHODS: A total of 33 participants performed strength tests using a mechanical pinch grip device, which had a plastic 3D printed cover with a pressure-sensing fabric overlaid on it to allow for simultaneous criterion validation, and a small load cell with a plastic 3D printed casing designed for comfortable pinch grip assessment. RESULTS: The simultaneously assessed fabric sensor and mechanical pinch grip device showed excellent absolute (ICC2,k  = 0.94) and relative (Pearson's R = 0.90) agreement. Both devices showed similar excellent relative (R > 0.75) agreement with the load cell despite non-simultaneous assessment. These findings indicate that 3D printed sensors incorporating a load cell and a pressure-sensing fabric can be used to replicate a pinch grip assessment performed with a mechanical pinch gauge. CONCLUSIONS: This study lays the foundation for these sensor systems to be modified for use as assessment tools during the performance of functional tasks using everyday objects. Additionally, because both systems generate real-time force data they could be used for biofeedback as part of rehabilitation and strengthening programs. To aid uptake and future research using these systems, the 3D print models, step-by-step hardware design and software programs used are provided in an open-source format at www.rehabtools.org/otsensors.html.

Benefits of home-based multidisciplinary exercise and supportive care in inoperable non-small cell lung cancer - protocol for a phase II randomised controlled trial.

BACKGROUND: Lung cancer is one of the most commonly diagnosed cancers, and is a leading cause of cancer mortality world-wide. Due to lack of early specific symptoms, the majority of patients present with advanced, inoperable disease and five-year relative survival across all stages of non-small cell lung cancer (NSCLC) is 14%. People with lung cancer also report higher levels of symptom distress than those with other forms of cancer. Several benefits for survival and patient reported outcomes are reported from physical activity and exercise in other tumour groups. We report the protocol for a study investigating the benefits of exercise, behaviour change and symptom self-management for patients with recently diagnosed, inoperable, NSCLC. METHODS: This multi-site, parallel-group, assessor-blinded randomised controlled trial, powered for superiority, aims to assess functional and patient-reported outcomes of a multi-disciplinary, home-based exercise and supportive care program for people commencing treatment. Ninety-two participants are being recruited from three tertiary-care hospitals in Melbourne, Australia. Following baseline testing, participants are randomised using concealed allocation, to receive either: a) 8 weeks of home-based exercise (comprising an individualised endurance and resistance exercise program and behaviour change coaching) and nurse-delivered symptom self-management intervention or b) usual care. The primary outcome is the between-group difference in the change in functional exercise capacity (six-minute walk distance) from baseline to post-program assessment. Secondary outcomes include: objective and self-reported physical activity levels, physical activity self-efficacy, behavioural regulation of motivation to exercise and resilience, muscle strength (quadriceps and grip), health-related quality of life, anxiety and depression and symptom interference. DISCUSSION: There is a lack of evidence regarding the benefit of exercise intervention for people with NSCLC, particularly in those with inoperable disease receiving treatment. This trial will contribute to evidence currently being generated in national and international trials by implementing and evaluating a home-based program including three components not yet combined in previous research, for people with inoperable NSCLC receiving active treatment and involving longer-term follow-up of outcomes. This trial is ongoing and currently recruiting. TRIAL REGISTRATION: This trial was prospectively registered on the Australian New Zealand Clinical Trials Registry ( ACTRN12614001268639 : (4/12/14).

A Longitudinal Examination of Postural Impairments in Children With Mild Traumatic Brain Injury: Implications for Acute Testing.

OBJECTIVE: To examine how postural control changes following pediatric mild traumatic brain injury. SETTING: Urban pediatric emergency department. PARTICIPANTS: Children 11 to 16 years old who presented within 6 hours of sustaining mild traumatic brain injury. DESIGN: Prospective observational cohort followed for 1 month. MAIN MEASURES: Total center of pressure path velocity and path velocity within distinct frequency bands, ranging from moderate to ultralow, were recorded by the Nintendo Wii Balance Board during a 2-legged stance. Measurements were recorded in 2 separate tests with eyes open and closed. The scores of the 2 tests were compared, and a Romberg quotient was computed. RESULTS: Eleven children were followed for 1 month postinjury. The ultralow frequency, which reflects slow postural movements associated with exploring stability boundaries, was lower (p = .02) during the eyes closed stance acutely following injury. The Romberg quotient for this frequency was also significantly lower acutely following injury (p = .007) than at 1 month. CONCLUSION: Following mild traumatic brain injury, children acutely demonstrate significantly more rigid sway patterns with eyes closed than with eyes open, which were highlighted by the Romberg quotient. The Romberg quotient could allow for accurate identification and tracking of postural instability without requiring knowledge of preinjury balance ability.

Balance and Falls in Acute Exacerbation of Chronic Obstructive Pulmonary Disease: A Prospective Study.

Individuals with chronic obstructive pulmonary disease (COPD) have demonstrated balance impairment and a higher fall incidence. However, these have not been investigated in acute exacerbations of the disease (ECOPD). This study evaluates balance in patients during an ECOPD compared to stable COPD and healthy controls, and examines the fall incidence rate after hospitalisation due to ECOPD compared to individuals with stable COPD. Balance performance of 26 hospitalised patients with ECOPD was compared to 26 community-dwelling participants with stable COPD and 25 matched healthy controls. Balance was evaluated using computerised posturography and the Berg Balance Scale (BBS). Prospective falls were monitored by monthly calendars for 12 months in both COPD groups. Compared to controls, greater balance impairment was observed during ECOPD for most posturography variables across standing conditions (p ≤ 0.05). Both COPD groups had worse BBS scores (p ≤ 0.05) compared to controls. Increased dyspnoea and reduced quadriceps' strength were associated with impaired balance performance. A higher fall incidence (1.76 falls/person/year) was observed following hospitalisation in patients with ECOPD compared to stable COPD (0.53 falls/person/year) at 12 months. Patients with ECOPD demonstrate balance impairments which are associated with increased dyspnoea and reduced muscle strength. Balance impairment during ECOPD may contribute to a high incidence of falls following hospitalisation.