Cost and outcomes of Hospital-based Usual cAre versus Tele-monitor self-directed Rehabilitation (HUATR) in patients with total knee arthroplasty: A randomized, controlled, non-inferiority trial.

OBJECTIVE: To evaluate the clinical efficacy and cost-effectiveness of telemonitored self-directed rehabilitation (TR) compared with hospital-based rehabilitation (HBR) for patients with total knee arthroplasty (TKA). DESIGN: In this randomized, non-inferiority clinical trial, 114 patients with primary TKA who were able to walk independently preoperatively were randomized to receive HBR (n = 58) or TR (n = 56). HBR comprised at least five physical therapy sessions over 10 weeks. TR comprised a therapist-led onboarding session, followed by a 10-week unsupervised home-based exercise program, with asynchronous monitoring of rehabilitation outcomes using a telemonitoring system. The primary outcome was fast-paced gait speed at 12 weeks, with a non-inferiority margin of 0.10 m/s. For economic analysis, quality-adjusted-life-years (QALY) was the primary economic outcome (non-inferiority margin, 0.027 points). RESULTS: In Bayesian analyses, TR had >95% posterior probability of being non-inferior to HBR in gait speed (week-12 adjusted TR-HBR difference, 0.02 m/s; 95%CrI, -0.05 to 0.10 m/s; week-24 difference, 0.01 m/s; 95%CrI, -0.07 to 0.10 m/s) and QALY (0.006 points; 95%CrI, -0.006 to 0.018 points). When evaluated from a societal perspective, TR was associated with lower mean intervention cost (adjusted TR-HBR difference, -S$227; 95%CrI, -112 to -330) after 24 weeks, with 82% probability of being cost-effective compared with HBR at a willingness to pay of S$0/unit of effect for the QALYs. CONCLUSIONS: In patients with uncomplicated TKAs and relatively good preoperative physical function, home-based, self-directed TR was non-inferior to and more cost-effective than HBR over a 24-week follow-up period. TR should be considered for this patient subgroup.

Using item response theory to estimate interpretation threshold values for the Frailty Index in community dwelling older adults.

BACKGROUND: Although the frailty index (FI) is designed as a continuous measure of frailty, thresholds are often needed to guide its interpretation. This study aimed to introduce and demonstrate the utility of an item response theory (IRT) method in estimating FI interpretation thresholds in community-dwelling adults and to compare them with cutoffs estimated using the receiver operating characteristics (ROC) method. METHODS: A sample of 1,149 community-dwelling adults (mean[SD], 68[7] years) participated in this cross-sectional study. Participants completed a multi-domain geriatric screen from which the 40-item FI and 3 clinical anchors were computed - namely, (i)self-reported mobility limitations (SRML), (ii)"fair" or "poor" self-rated health (SRH), and (iii) restricted life-space mobility (RLSM). Participants were classified as having SRML-1 if they responded "Yes" to either of the 2 questions regarding walking and stair climbing difficulty and SRML-2 if they reported having walking and stair climbing difficulty. Participants with a Life Space Assessment score <60 points were classified as having RLSM. Threshold values for all anchor questions were estimated using the IRT method and ROC analysis with Youden criterion. RESULTS: The proportions of participants with SRML-1, SRML-2, Fair/Poor SRH, and RLSM were 21 %, 8 %, 22 %, and 9 %, respectively. The IRT-based thresholds for SRML-2 (0.26), fair/poor SRH (0.29), and RLSM (0.32) were significantly higher than those for SRML-1 (0.18). ROC-based FI cutoffs were significantly lower than IRT-based values for SRML-2, SRH, and RLSM (0.12 to 0.17), and they varied minimally and non-systematically across the anchors. CONCLUSIONS: The IRT method identifies biologically plausible FI thresholds that could meaningfully complement and contextualize existing thresholds for defining frailty.

Estimating cutpoints of gait speed and sit-to-stand test values for self-reported mobility limitations in a cohort of community-dwelling older adults from Singapore: comparing receiver operating characteristic (ROC) analysis with adjusted predictive modelling.

OBJECTIVES: Clinical interpretability of the gait speed and 5-times sit-to-stand (5-STS) tests is commonly established by comparing older adults with and without self-reported mobility limitations (SRML) on gait speed and 5-STS performance, and estimating clinical cutpoints for SRML using the receiver operating characteristics (ROC) method. Accumulating evidence, however, suggests that the adjusted predictive modeling (APM) method may be more appropriate to estimate these interpretational cutpoints. Thus, we aimed to compare, in community-dwelling older adults, gait speed and 5-STS cutpoints estimated using the ROC and APM methods. DESIGN: Cross-sectional study. SETTING AND PARTICIPANTS: This study analyzed data from 955 community-dwelling independently walking older adults (73%women) aged ≥60 years (mean, 68; range, 60-88). METHODS: Participants completed the 10-metre gait speed and 5-STS tests. Participants were classified as having SRML if they responded "Yes" to either of the 2 questions regarding walking and stair climbing difficulty. Cutpoints for SRML and its component questions were estimated using ROC analysis with Youden criterion and the APM method. RESULTS: The proportions of participants with self-reported walking difficulty, self-reported stair climbing difficulty, and SRML were 10%, 19%, and 22%, respectively. Gait speed and 5-STS time were moderately correlated with each other (r=-0.56) and with the self-reported measures (absolute r-values, 0.39-0.44). ROC-based gait speed cutpoints were 0.14 to 0.16 m/s greater than APM-based cutpoints (P < 0.05) whilst ROC-based 5-STS time cutpoints were 0.8 to 3.3 s lower than APM-based cutpoints (P < 0.05 for walking difficulty). Compared with ROC-based cutpoints, APM-based cutptoints were more precise and they varied monotonically with self-reported walking difficulty, self-reported stair climbing difficulty, and SRML. CONCLUSIONS AND IMPLICATIONS: In a sample of 955 older adults, our findings of precise and biologically plausible gait speed and 5-STS cutpoints for SRML estimated using the APM method indicate that this promising method could potentially complement or even replace traditional ROC methods.

Associations of height, weight, and body mass index with handgrip strength: A Bayesian comparison in older adults.

BACKGROUND & AIMS: Handgrip strength is commonly normalized or stratified by body size to define subgroup-specific cut-points and reference limits values. However, it remains unclear which anthropometric variable is most strongly associated with handgrip strength. We aimed to, in older adults with no self-reported mobility limitations, determine whether height, weight, and body mass index (BMI) were meaningfully associated with handgrip strength. METHODS: This cross-sectional study included community-dwelling ambulant participants, and we identified 775 older adults who reported no difficulty walking 100 m, climbing stairs, and rising from the chair. Handgrip strength was measured with a digital dynamometer. Bayesian linear regression was used to estimate the probabilities that the positive associations of height, weight, and BMI with handgrip strength exceeded 0 kg (the null value) and 2.5 kg (the clinically meaningful threshold value). RESULTS: Mean handgrip strength was 22.1 kg (SD, 4) for women and 32.9 kg (SD, 6) for men. Body height, weight, and BMI had >99.9% probabilities of a positive association with handgrip strength; however, the associations of per interquartile increase in body weight and BMI with handgrip strength had low probabilities (<5%) of exceeding the clinically meaningful threshold of 2.5 kg. In contrast, body height had the highest probability (99.6%) of a clinically meaningful association with handgrip strength: adjusting for age and gender, handgrip strength was 3.2 kg (95% CrI, 2.7 to 3.8) greater in older adults 1.61 m tall than in older adults 1.51 m tall. CONCLUSIONS: In a large sample of mobile-intact older adults, handgrip strength differed meaningfully by body height. Although requiring validation, our findings suggest that future efforts should be directed at normalizing handgrip strength by body height to better define subgroup-specific handgrip weakness. A web-based application (https://sghpt.shinyapps.io/ippts/) was created to allow interactive exploration of predicted values and reference limits of age-, gender-, and height-subgroups.

Development and validation of a physical frailty phenotype index-based model to estimate the frailty index.

BACKGROUND: The conventional count-based physical frailty phenotype (PFP) dichotomizes its criterion predictors-an approach that creates information loss and depends on the availability of population-derived cut-points. This study proposes an alternative approach to computing the PFP by developing and validating a model that uses PFP components to predict the frailty index (FI) in community-dwelling older adults, without the need for predictor dichotomization. METHODS: A sample of 998 community-dwelling older adults (mean [SD], 68 [7] years) participated in this prospective cohort study. Participants completed a multi-domain geriatric screen and a physical fitness assessment from which the count-based PFP and the 36-item FI were computed. One-year prospective falls and hospitalization rates were also measured. Bayesian beta regression analysis, allowing for nonlinear effects of the non-dichotomized PFP criterion predictors, was used to develop a model for FI ("model-based PFP"). Approximate leave-one-out (LOO) cross-validation was used to examine model overfitting. RESULTS: The model-based PFP showed good calibration with the FI, and it had better out-of-sample predictive performance than the count-based PFP (LOO-R2, 0.35 vs 0.22). In clinical terms, the improvement in prediction (i) translated to improved classification agreement with the FI (Cohen's kw, 0.47 vs 0.36) and (ii) resulted primarily in a 23% (95%CI, 18-28%) net increase in FI-defined "prefrail/frail" participants correctly classified. The model-based PFP showed stronger prognostic performance for predicting falls and hospitalization than did the count-based PFP. CONCLUSION: The developed model-based PFP predicted FI and clinical outcomes more strongly than did the count-based PFP in community-dwelling older adults. By not requiring predictor cut-points, the model-based PFP potentially facilitates usage and feasibility. Future validation studies should aim to obtain clear evidence on the benefits of this approach.

Clinical Interpretability of Quadriceps Strength and Gait Speed Performance in Total Knee Arthroplasty: A Longitudinal Study.

OBJECTIVE: After a total knee arthroplasty, physical assessments of quadriceps strength and gait speed performance are often undertaken during rehabilitation. Our study aimed to improve their clinical interpretability by examining trajectory curves across levels of self-reported walking and stair climbing function. DESIGN: A sample of 2624 patients with primary total knee arthroplasty participated in this retrospective longitudinal study. Monthly, for 4 mos after surgery, quadriceps strength and gait speed were quantified. At the month-6 time point, self-reported walking and stair climbing function was measured. RESULTS: All physical measures improved nonlinearly over time. In mixed-effects models, greater quadriceps strength and gait speed over time were associated with higher month-6 self-reported walking and stair climbing function ( P < 0.001). Steeper gains in quadriceps strength and gait speed were associated with higher levels of walking and stair-climbing function (interaction P < 0.001). Among female patients who had great difficulty with stair ascent and ambulation, quadriceps strength trajectory curves plateaued after 8 wks after total knee arthroplasty. CONCLUSIONS: By stratifying trajectory curves across clinically interpretable functional levels, our findings potentially provide patients and clinicians a means to better interpret the continuous-scaled quadriceps strength and gait speed values. This information may be valuable when engaging patients in shared decision making and expectation setting. TO CLAIM CME CREDITS: Complete the self-assessment activity and evaluation online at http://www.physiatry.org/JournalCME. CME OBJECTIVES: Upon completion of this article, the reader should be able to: (1) Understand how self-reported walking and stair climbing abilities improved from baseline before total knee arthroplasty (total knee arthroplasty) to 6 mos postoperatively; (2) Describe the time course of the 2 performance-based measures of quadriceps strength and walking speed after a total knee arthroplasty; and (3) Relate the trajectories of post-total knee arthroplasty quadriceps strength and walking speed measurements across distinct levels of self-reported walking and stair climbing function. LEVEL: Advanced. ACCREDITATION: The Association of Academic Physiatrists is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians.The Association of Academic Physiatrists designates this Journal-based CME activity for a maximum of 1.0 AMA PRA Category 1 Credit(s) ™. Physicians should only claim credit commensurate with the extent of their participation in the activity.

Sit-to-Stand Weight-Bearing Symmetry Performance in Total Knee Arthroplasty: Recovery Curves, Correlates, and Predictive Validity With Gait Speed.

OBJECTIVE: After total knee arthroplasty, the ability to weight bear symmetrically during the sit-to-stand task provides important information regarding altered movement patterns. Despite this, comprehensive recovery curves and validity data for sit-to-stand weight-bearing symmetry are lacking in the total knee arthroplasty population. Our study aimed to (1) develop recovery curves with reference ranges, (2) identify the correlates of standard and constrained sit-to-stand weight-bearing symmetry, and (3) evaluate their predictive validity with gait speed. DESIGN: We performed a retrospective longitudinal study of 706 patients with primary unilateral total knee arthroplasty. Monthly, for 4 mos after surgery, sit-to-stand weight-bearing symmetry, knee pain, knee range of motion, quadriceps strength, and gait speed were quantified. RESULTS: Standard and constrained sit-to-stand weight-bearing symmetry measures improved nonlinearly over time. Standard sit-to-stand weight-bearing symmetry was most strongly associated with bilateral quadriceps strength, whereas constrained sit-to-stand weight-bearing symmetry was most strongly associated with ipsilateral quadriceps strength. Knee range of motion and contralateral knee pain were additional correlates. Both standard sit-to-stand and constrained sit-to-stand weight-bearing symmetry were independently and nonlinearly associated with gait speed in multivariable models. CONCLUSIONS: Our study provided recovery curves and validity data to support routine clinical measurement of sit-to-stand weight-bearing symmetry in total knee arthroplasty. Our results also indicate that constrained sit-to-stand may promote greater use of the operated limb than standard sit-to-stand.

Screening accuracy of percentage predicted gait speed for prefrailty/frailty in community-dwelling older adults.

AIM: In order to account for the variability in gait speed due to demographic factors, an observed gait speed value can be compared with its predicted value based on age, sex, and body height (observed gait speed divided by predicted gait speed, termed "GS%predicted" henceforth). This study aimed to examine the screening accuracy of an optimal GS%predicted threshold for prefrailty/frailty. METHODS: This cross-sectional study included 998 community-dwelling ambulant participants aged >50 years (mean age = 68 years). Participants completed a multi-domain geriatric screen and a physical fitness assessment, from which the 10-m habitual gait speed, GS%predicted, Physical Frailty Phenotype (PFP) index, and 36-item Frailty Index (FI) were computed. RESULTS: Based on the FI, ~49% of participants had pre-frailty or frailty. The optimal threshold of GS%predicted (0.93) had greater screening accuracy than the 1.0 m/s fixed threshold for gait speed (AUC, 0.65 vs. 0.60; DeLong's P < 0.001). Replacing gait speed with GS%predicted in the PFP improved its overall discrimination (AUC, 0.70 vs. 0.67 of original PFP; DeLong's P < 0.001). CONCLUSIONS: Defining a "slow" gait speed by a GS%predicted value of <0.93 provided greater screening accuracy than the traditional 1.0 m/s threshold for gait speed. Our results also support the use of GS%predicted-derived PFP to identify older adults at risk of prefrailty/frailty. Geriatr Gerontol Int 2022; 22: 575-580.

Reference Ranges for Gait Speed and Sit-to-Stand Performance in a Cohort of Mobility-Intact Community-Dwelling Older Adults From Singapore.

OBJECTIVES: Slow gait speed and sit-to-stand performance are associated with adverse clinical outcomes in older adults. Identifying older adults with functional performance "below norms" is the first step toward prevention. We aimed to (1) examine the associations of age, body height, and gender with gait speed and sit-to-stand performance and (2) develop subgroup-specific reference ranges in older adults with no self-reported mobility limitations. DESIGN: Cross-sectional study. SETTING AND PARTICIPANTS: This study analyzed data from 775 community-dwelling older adults who reported no difficulty walking 100 m, climbing stairs, and rising from the chair. METHODS: Gait speed and sit-to-stand performance were measured by the 10-m gait speed test and 5-times sit-to-stand test, respectively. Bayesian linear regression was used to derive 95% reference ranges for gait speed and sit-to-stand performance, defined by different levels of age, body height, and gender. RESULTS: Overall, 95% reference range was 0.89-1.79 m/s for habitual gait speed and 7.4-27.9 stands/30 s for sit-to-stand pace. Age had the highest posterior probability (>99%) of a meaningful association with both functional outcomes. Additionally, height was strongly associated with gait speed: a 10-cm increase in height was associated with 0.07 m/s (95% credible interval, 0.05-0.10) faster gait speed. For sit-to-stand test, the lower 95% reference range limits tended to be similar across gender and gender-specific height subgroups, owing to the associations of faster sit-to-stand pace with shorter height and male gender. Because extensive tables of reference ranges are impractical, a web-based application (https://sghpt.shinyapps.io/ippts/) is created to provide subgroup-specific reference ranges. CONCLUSIONS AND IMPLICATIONS: In a large sample of mobile-intact older adults, reference ranges for gait speed and sit-to-stand performance differed meaningfully by age. Furthermore, gait speed was stature dependent. Although requiring validation, our findings may be used to define subgroup-specific "below-range" values and to complement existing universal clinical cut points for gait speed and sit-to-stand performance.

Associations of the modified STarT back tool and Hospital Anxiety and Depression Scale (HADS) with gait speed and knee pain in knee osteoarthritis: a retrospective cohort study.

OBJECTIVE: The association of the modified STarT Back Tool (mSBT) psychosocial measure with gait speed and knee pain in knee osteoarthritis is not well defined. This study aimed to, in patients with knee osteoarthritis, (i) examine the convergent validity of mSBT with the Hospital Anxiety and Depression Scale (HADS) and (ii) compare the predictive validity of mSBT and HADS with gait speed and knee pain. METHODS: We performed a retrospective cohort analysis of mSBT, HADS, gait speed, and knee pain outcomes data collected from 119 patients who received outpatient physical therapy. Of these patients who were evaluated at their first (baseline) physical therapy visit, 55 had available data at the Week-16 follow-up visit. RESULTS: mSBT and HADS showed moderately strong pairwise correlations (Spearman correlation > 0.57; p < 0.001). After adjusting for age, sex, body weight, and knee impairment variables in multivariable linear mixed-effects analyses, mSBT was associated with gait speed (p < 0.001) and knee pain intensity (p < 0.001) and it had comparable strength of association as HADS. In within-patient regression analyses, change in mSBT was associated with changes in gait speed (p = 0.04) and knee pain (p = 0.01) over 16 weeks. CONCLUSION: The mSBT had convergent validity with HADS and it showed predictive validity with gait speed and knee pain in knee osteoarthritis. Although broader validation is required, the 5-item mSBT psychosocial measure may be applied as part of routine clinical care to assess psychological distress in patients with knee osteoarthritis.IMPLICATIONS FOR REHABILITATIONThe 5-item psychosocial subscale of the modified STarT Back tool (mSBT) showed good convergent validity with the 14-item Hospital Anxiety and Depression Scale in patients with knee osteoarthritis.The mSBT psychosocial subscale showed predictive validity, at both cross-sectional and longitudinal levels, with gait speed and knee pain in patients with knee osteoarthritis.The mSBT can potentially be used in the busy clinical setting to assess psychological distress in patients with knee osteoarthritis.

Rest Intervals during Virtual Reality Gaming Augments Standing Postural Sway Disturbance.

Immersive virtual reality (VR) can cause acute sickness, visual disturbance, and balance impairment. Some manufacturers recommend intermittent breaks to overcome these issues; however, limited evidence examining whether this is beneficial exists. The aim of this study was to examine whether taking breaks during VR gaming reduced its effect on postural sway during standing balance assessments. Twenty-five people participated in this crossover design study, performing 50 min of VR gaming either continuously or with intermittent 10 min exposure/rest intervals. Standing eyes open, two-legged balance assessments were performed immediately pre-, immediately post- and 40 min post-exposure. The primary outcome measure was total path length; secondary measures included independent axis path velocity, amplitude, standard deviation, discrete and continuous wavelet transform-derived variables, and detrended fluctuation analysis. Total path length was significantly (p < 0.05) reduced immediately post-VR gaming exposure in the intermittent rest break group both in comparison to within-condition baseline values and between-condition timepoint results. Conversely, it remained consistent across timepoints in the continuous exposure group. These changes consisted of a more clustered movement speed pattern about a lower central frequency, evidenced by signal frequency content. These findings indicate that caution is required before recommending rest breaks during VR exposure until we know more about how balance and falls risk are affected.

Knee performance and self-efficacy trajectory curves after ACL reconstruction: A longitudinal study.

OBJECTIVES: To examine knee flexion range-of-motion, quadriceps strength, and knee self-efficacy trajectory curves over 6 months after anterior cruciate ligament reconstruction (ACLR), stratified by patients' Month-6 sports activity level. DESIGN: Prospective longitudinal study. SETTING: Hospital outpatient physiotherapy department. PARTICIPANTS: 595 individuals after unilateral ACLR (mean age, 27 years). MAIN OUTCOME MEASURES: At 2-, 3-, and 6-months post-surgery, knee flexion range-of-motion, quadriceps strength, and self-efficacy were quantified. Flexion range-of-motion was additionally measured at 2- and 4-weeks post-surgery. Sports activity levels were assessed using the Tegner Activity Score at 6-months post ACLR. RESULTS: The various measures improved nonlinearly over time, with substantial improvements observed in the first 2-4 months post-surgery. In multivariable generalized least squares models, greater knee flexion range-of-motion, quadriceps strength, and self-efficacy over time were significantly associated with higher Month-6 Tegner levels (all P values < 0.01). Additionally, receiving a bone-patellar-tendon-bone graft or meniscal repair was associated with lower quadriceps strength trajectories (P-values<0.001) while female sex was associated with lower knee self-efficacy trajectories (P = 0.02). CONCLUSIONS: Greater knee flexion range-of-motion, quadriceps strength, and self-efficacy were associated with higher Month-6 Tegner levels. The derived trajectory curves may be useful for effective management decision making and adequate results interpretation during the rehabilitation process.

Immediate effects of valgus bracing on knee joint moments during walking in knee-healthy individuals: Potential modifying effects of body height.

BACKGROUND: The goal of valgus knee brace treatment is to reduce medial knee joint loading during walking, often indicated by external knee adduction moment (KAM) measures. However, existing healthy-subjects studies have been equivocal in demonstrating KAM reduction with valgus knee bracing. RESEARCH QUESTION: What are the immediate effects of valgus bracing at different tension levels on KAM during walking at a controlled speed and does body height modify the brace-KAM associations? METHODS: Data from 32 knee-healthy participants were analysed in this randomized crossover trial. Participants performed walking trials at controlled speed (1.3 ± 0.065 m/s) both with and without an Ossür Unloader One® brace. During the bracing condition, valgus tension was incrementally increased, from zero tension to normal tension and to maximum tolerable tension. RESULTS: Valgus bracing minimally increased knee flexion at heel-strike (P < 0.001) in a dose-dependent manner and minimally reduced gait velocity (∼0.015m/s) across all tension levels. Valgus bracing, overall, did not significantly reduce the various KAM measures. However, brace use at maximal tension was associated with a 0.04Nm/kg (9.2 %) increase in first peak KAM amongst participants with a body height of 1.75 m and a 0.03Nm/kg (7.6 %) decrease in first peak KAM amongst participants with a body height of 1.55 m. SIGNIFICANCE: Valgus bracing did not reduce the various KAM measures during walking; however, body height may play a moderating role. Given knee brace sizes vary more in circumference than length, this result may be due to the ratio between effective moment arm length relative to limb length. A deeper understanding of the potential neuro-biomechanical effects of valgus knee bracing and how these effects are potentially modified by body height may be critical to the design of effective knee braces.

Test-retest reliability and variability of knee adduction moment peak, impulse and loading rate during walking.

BACKGROUND: Previous studies have reported good test-retest reliability for peak knee adduction moment (KAM) during walking. However, reliability of other KAM measurements has not been established. RESEARCH QUESTION: What is the test-retest reliability of peak KAM, KAM impulse, and KAM loading rate measurements during walking in knee-healthy individuals? METHODS: Data from 32 knee-healthy participants were analysed in this test-retest reliability study. Various KAM measurements were reported for two sessions with kinematic and kinetic data obtained from a motion capture system synchronised with force plates, with a median of 1 week between sessions. RESULTS: For all KAM measures, intra-class correlation coefficients were above 0.90 and their lower bound 95 % confidence limits exceeded 0.81. However, absolute measurement variability differed across measures, with normalized SEM (8 %-15 %), normalized MDC95 (20 %-40 %), intra-session MAD (10 %-18 %), and inter-session MAD (12 %-22 %) varying over a 2-fold range. Overall and first peak KAM, KAM impulse over 50 % stance, and KAM loading rate (15 frame window) showed ≤10 % and ≤15 % intra- and inter-session MAD, respectively. SIGNIFICANCE: This study provided previously undefined test-retest reliability estimates for various KAM measures during walking. Researchers and clinicians should not assume that the various aspects of the KAM curve share similar reliability.

Machine learning methods are comparable to logistic regression techniques in predicting severe walking limitation following total knee arthroplasty.

PURPOSE: Machine-learning methods are flexible prediction algorithms with potential advantages over conventional regression. This study aimed to use machine learning methods to predict post-total knee arthroplasty (TKA) walking limitation, and to compare their performance with that of logistic regression. METHODS: From the department's clinical registry, a cohort of 4026 patients who underwent elective, primary TKA between July 2013 and July 2017 was identified. Candidate predictors included demographics and preoperative clinical, psychosocial, and outcome measures. The primary outcome was severe walking limitation at 6 months post-TKA, defined as a maximum walk time ≤ 15 min. Eight common regression (logistic, penalized logistic, and ordinal logistic with natural splines) and ensemble machine learning (random forest, extreme gradient boosting, and SuperLearner) methods were implemented to predict the probability of severe walking limitation. Models were compared on discrimination and calibration metrics. RESULTS: At 6 months post-TKA, 13% of patients had severe walking limitation. Machine learning and logistic regression models performed moderately [mean area under the ROC curves (AUC) 0.73-0.75]. Overall, the ordinal logistic regression model performed best while the SuperLearner performed best among machine learning methods, with negligible differences between them (Brier score difference, < 0.001; 95% CI [- 0.0025, 0.002]). CONCLUSIONS: When predicting post-TKA physical function, several machine learning methods did not outperform logistic regression-in particular, ordinal logistic regression that does not assume linearity in its predictors. LEVEL OF EVIDENCE: Prognostic level II.

Predicting Mobility Limitations in Patients With Total Knee Arthroplasty in the Inpatient Setting.

OBJECTIVE: To develop a prediction model for postoperative day 3 mobility limitations in patients undergoing total knee arthroplasty (TKA). DESIGN: Prospective cohort study. SETTING: Inpatients in a tertiary care hospital. PARTICIPANTS: A sample of patients (N=2300) who underwent primary TKA in 2016-2017. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURE: Candidate predictors included demographic variables and preoperative clinical and psychosocial measures. The outcome of interest was mobility limitations on post-TKA day 3, and this was determined a priori by an ordinal mobility outcome hierarchy based on the type of the gait aids prescribed and the level of physiotherapist assistance provided. To develop the model, we fitted a multivariable proportional odds regression model with bootstrap internal validation. We used a model approximation approach to create a simplified model that approximated predictions from the full model with 95% accuracy. RESULTS: On post-TKA day 3, 11% of patients required both walkers and therapist assistance to ambulate safely. Our prediction model had a concordance index of 0.72 (95% confidence interval, 0.68-0.75) when evaluating these patients. In the simplified model, predictors of greater mobility limitations included older age, greater walking aid support required preoperatively, less preoperative knee flexion range of movement, low-volume surgeon, contralateral knee pain, higher body mass index, non-Chinese race, and greater self-reported walking limitations preoperatively. CONCLUSION: We have developed a prediction model to identify patients who are at risk for mobility limitations in the inpatient setting. When used preoperatively as part of a shared-decision making process, it can potentially influence rehabilitation strategies and facilitate discharge planning.

Predicting individual knee range of motion, knee pain, and walking limitation outcomes following total knee arthroplasty.

Background and purpose - Up to 20% of patients are dissatisfied after total knee arthroplasty (TKA), mainly because of pain and restricted physical function. We developed a prediction model for 6-month knee range of motion, knee pain, and walking limitations in patients undergoing TKA surgery. Patients and methods - We performed a prospective cohort study of 4,026 patients who underwent elective, primary TKA between July 2013 and July 2017. Candidate predictors included demographic, clinical, psychosocial, and preoperative outcome measures. The outcomes of interest were (i) knee extension and flexion range of motion, (ii) knee pain rated on a 5-point ordinal scale, and (iii) self-reported maximum walk time at 6 months post TKA. For each outcome, we fitted a multivariable proportional odds regression model with bootstrap internal validation. Results - At 6 months post TKA, around 5% to 20% of patients had a flexion contracture ³ 10°, range of motion < 90°, moderate to severe knee pain, or a maximum walk time £â€¯15 minutes. The model c-indices (the probabilities to correctly discriminate between 2 patients with different levels of follow-up TKA outcomes) when evaluating these patients were 0.71, 0.79, 0.65, and 0.76, respectively. Each postoperative outcome was strongly influenced by the same outcome measure obtained preoperatively (all p-values < 0.001). Additional statistically significant predictors were age, sex, race, education level, diabetes mellitus, preoperative use of gait aids, contralateral knee pain, and psychological distress (all p-values < 0.001). Interpretation - We have developed models to predict, for individual patients, their likely post-TKA levels of knee extension and flexion range of motion, knee pain, and walking limitations. After external validation, they can potentially be used preoperatively to identify at-risk patients and to help patients set more realistic expectations about surgical outcomes.

Falls efficacy, postural balance, and risk for falls in older adults with falls-related emergency department visits: prospective cohort study.

BACKGROUND: Risk for falls in older adults has been associated with falls efficacy (self-perceived confidence in performing daily physical activities) and postural balance, but available evidence is limited and mixed. We examined the interaction between falls efficacy and postural balance and its association with future falls. We also investigated the association between falls efficacy and gait decline. METHODS: Falls efficacy, measured by the Modified Falls Efficacy Scale (MFES), and standing postural balance, measured using computerized posturography on a balance board, were obtained from 247 older adults with a falls-related emergency department visit. Six-month prospective fall rate and habitual gait speed at 6 months post baseline assessment were also measured. RESULTS: In multivariable proportional odds analyses adjusted for potential confounders, falls efficacy modified the association between postural balance and fall risk (interaction P = 0.014): increasing falls efficacy accentuated the increased fall risk related to poor postural balance. Low baseline falls efficacy was strongly predictive of worse gait speed (0.11 m/s [0.06 to 0.16] slower gait speed per IQR decrease in MFES; P < 0.001). CONCLUSION: Older adults with high falls efficacy but poor postural balance were at greater risk for falls than those with low falls efficacy; however, low baseline falls efficacy was strongly associated with worse gait function at follow-up. Further research into these subgroups of older adults is warranted. TRIAL REGISTRATION: ClinicalTrials.gov identifier: NCT01713543 .

Factors associated with gait speed recovery after total knee arthroplasty: A longitudinal study.

OBJECTIVES: Gait speed limitations can remain significant issues after a total knee arthroplasty (TKA) but their associated factors are not well understood. This study aimed to identify the factors associated with acute gait speed recovery post-TKA. METHODS: We performed a prospective longitudinal study of 1765 patients who underwent primary TKA between July 2013 and July 2015. At 4, 8, 12, and 16 weeks postsurgery, fast gait speed was measured. The factors associated with gait speed over time since TKA were identified using multivariable generalized least squares modeling. RESULTS: Lower postoperative quadriceps strength and knee flexion range of motion were closely associated with lower gait speed over time (0.084m/s, 0.064m/s, and 0.055m/s change in gait speed per interquartile range change in ipsilateral quadriceps strength, contralateral quadriceps strength, and knee flexion range of motion, respectively). Additional strong predictors of lower gait speed included older age (0.11m/s), lower levels of preoperative Short Form 36 physical function (0.066m/s), greater body mass (0.046m/s), and the preoperative use of a walking aid (overall P < 0.001). Patients who reported that they limited their daily activities due to a fear of falling also had poorer gait speed (0.033m/s and 0.054m/s slower gait speed for "Occasional" and "Often" categories, respectively, vs. "None"). CONCLUSIONS: Gait speed recovery post-TKA is driven by both physical and psychological factors, suggesting that identifying and treating the underlying physical and cognitive causes of gait speed limitations may be crucial to optimize functional recovery.

Development of a Prediction Model to Estimate the Risk of Walking Limitations in Patients with Total Knee Arthroplasty.

OBJECTIVE: Early and accurate risk prediction of walking limitations after total knee arthroplasty (TKA) is important for clinical and economic reasons. However, to our knowledge, no studies have systematically integrated multiple predictors into a single, clinically practical model. Our study aimed to develop a prediction model to estimate the risk of post-TKA walking limitations. METHODS: We performed a prospective cohort study of 1096 patients who underwent elective, primary TKA between July 2013 and September 2014. Candidate predictors included patient demographics, surgical factors, and pre- and early (1-mo) post-TKA functional measures. The outcome of interest was self-reported walking limitations at 6 months of post-TKA. We used multivariable proportional odds regression with bootstrap internal validation to develop the model. RESULTS: In all, 12% of patients reported walking limitations (maximum walk time ≤ 15 min) at 6 months postsurgery. The main predictors of increasing levels of walking limitations were preoperative walking limitations (overall p < 0.001), higher levels of body mass index [interquartile range (IQR)-OR 1.3, 95% CI 1.2-1.5], lower values of 1-month post-TKA gait speed (IQR-OR 1.9, 95% CI 1.3-2.6), the presence of contralateral knee pain (OR 1.9, 95% CI 1.2-3.0), and the use of a quadstick preoperatively (OR 3.5, 95% CI 1.7-7.3). The prediction model had an optimism-corrected concordance index of 0.71. CONCLUSION: A small but sizable proportion of patients with TKA had persistent mobility limitations. Our prediction model may help to risk-stratify patients, and external validation is required before the model can be used in clinical practice.