A Retrospective Study on the Feasibility of Using Low-burden Patient-reported Pain Scores to Track Recovery and Outcomes After Total Joint Replacement.

Background: Patients undergo total joint arthroplasty to improve function and resolve pain. Patient-reported outcome measures (PROMs) are often sought to determine the success of total joint arthroplasty but are time-consuming and patient response rates are often low. This study sought to determine whether pain numeric rating scores (NRSs) were associated with PROMs and objective mobility outcomes. Methods: This is a retrospective review of data in patients who utilized a smartphone-based care management application prior to and following total joint arthroplasty. NRS, Hip Disability and Osteoarthritis Outcome Score, Joint Replacement and Knee Injury and Osteoarthritis Outcome Score, Joint Replacement, and objective mobility data (step counts, gait speed, and gait asymmetry) were collected preoperatively and at 30 and 90 days postoperatively. Quantile regression was performed to evaluate the correlations between NRS and PROMs. Results: Total knee arthroplasty patients reported higher NRS than total hip arthroplasty patients postoperatively. NRS was significantly correlated with gait speed preoperatively and at 30 and 90 days postoperatively on quantile regression. Gait asymmetry was significantly associated with NRS at 30 days postoperatively. Regression results suggested significant correlations between NRS and PROMs scores; Hip Disability and Osteoarthritis Outcome Score, Joint Replacement, -0.46 (95% confidence interval: -0.48 to -0.44, P < .001) and Knee Injury and Osteoarthritis Outcome Score, Joint Replacement, -0.38 (95% confidence interval: -0.40 to -0.36, P < .001). Conclusions: NRS is correlated with both objective and subjective measures of function in patients undergoing arthroplasty. Simple pain ratings may be a valid measurement to help predict functional outcomes when collection of traditional PROMs is not feasible.

Early post-operative walking bouts are associated with improved gait speed and symmetry at 90 days.

INTRODUCTION: There is a paucity of literature on optimal patterns of daily walking following joint arthroplasty, which are now evaluated with consumer technologies like smartphones, and can enhance our understanding of post-operative mobility. When smartphone-recorded, daily walking patterns are captured, qualities of gait-recovery such as gait speed or symmetry can be analyzed in real-world environments. RESEARCH QUESTION: Are the daily distribution of walking bouts in the early post-operative period associated with 90-day gait quality measures following hip and knee arthroplasty? METHODS: Gait data was collected passively using a smartphone-based care management platform in patients undergoing hip and knee arthroplasty. As recorded via subjects' free-living smartphone-collected gait bouts, data were investigated as a function of the walking session length and were used to create a ratio to the total time logging bouts, representing the fraction of walking performed during a single session per day (aggregation score). Quantile regression was performed to evaluate the association between early walking session lengths or aggregation score at 30 days post-operatively and the gait-sampled speed and asymmetry of walking at 90 days. RESULTS: In total, 2255 patients provided evaluable data. The walking session length at 30 days was positively associated with 90-day mean gait speed across procedure types where quantile regression coefficients ranged from 0.11 to 0.17. In contrast, aggregation score was negatively associated with gait speed at 90 days, with coefficients ranging from -0.18 to -0.12. SIGNIFICANCE: The duration and frequency of walking bouts was associated with recovery of gait speed and symmetry following lower limb arthroplasty. The findings may help clinicians design walking protocols that are associated with improved gait metrics at 3 months.

Instrumented assessment of shoulder function: A study of inertial sensor based methods.

BACKGROUND: Inertial sensors have the potential to provide objective and practical methods to assess joint and limb function in the clinical setting. The aim of this study is to evaluate the psychometric properties of inertial sensor metrics in the assessment of patients with subacromial shoulder pain. METHODS: 25 patients with unilateral subacromial shoulder pain and 50 control subjects were recruited. Assessments were carried out on both shoulders for all participants during a short movement procedure. Patients had assessments repeated after receiving three months of physiotherapy. Inertial metrics evaluated included a smoothness measure and speed and power scores derived from the range of angular velocity and acceleration profiles. Individual shoulder scores and asymmetry scores were both evaluated in terms of reliability, known-group validity, convergent validity and responsiveness. FINDINGS: Regression analysis identified age to be a significant predictor for all scores, therefore an age matched sub-cohort of control subjects was used for comparative analyses. All scores demonstrated inter-rater reliability (ICC = 0.48-0.82), were able to differentiate pathological from healthy shoulders (AUC = 0.62-0.91) and displayed significant changes following treatment. Scores derived from the range of acceleration and velocity profiles demonstrated the largest effect sizes (Cohens d = 0.8-1.35), and displayed the highest correlation with the Oxford Shoulder Score (r = -0.40 - -0.58). INTERPRETATION: The scores investigated demonstrate good psychometric properties and have potential to complement existing methods of assessment in the clinical or research setting. Further work is required to fully understand their clinical relevance and optimise assessment methods and interpretation.

Adherence monitoring of rehabilitation exercise with inertial sensors: A clinical validation study.

BACKGROUND: Rehabilitation has an established role in the management of a wide range of musculoskeletal conditions. Much of this treatment relies on self-directed exercises at home, where adherence of execution is unknown. Demonstrating treatment fidelity is necessary to draw conclusions about the efficacy of rehabilitation interventions in both clinical and research settings. There is a lack of tools and methods to achieve this. RESEARCH QUESTION: This study aims to evaluate the feasibility of using a single inertial sensor to recognise and classify shoulder rehabilitation activity using supervised machine learning techniques. METHODS: Twenty patients with shoulder pain were monitored performing five rehabilitation exercises routinely prescribed for their condition. Accelerometer, gyroscope and magnetometer data were collected via a device mounted onto an arm sleeve. Non-specific motion data was included in the analysis. Time and frequency domain features were calculated from labelled data segments and ranked in terms of their predictive importance using the ReliefF algorithm. Selected features were used to train four supervised learning algorithms: decision tree, k-nearest neighbour, support vector machine and random forests. Performance of algorithms in accurately classifying exercise activity was evaluated with ten-fold cross-validation and leave-one-subject-out-validation methods. RESULTS: Optimal predictive accuracies for ten-fold cross-validation (97.2%) and leave-one-subject-out-validation (80.5%) were achieved by support vector machine and random forests algorithms, respectively. Time domain features derived from accelerometer, magnetometer and orientation data streams were shown to have the highest predictive value for classifying rehabilitation activity. SIGNIFICANCE: Classification models performed well in differentiating patient exercise activity from non-specific movement and identifying specific exercise type using inertial sensor data. A clinically useful account of home rehabilitation activity will help guide treatment strategies and facilitate methods to improve patient engagement. Future work should focus on evaluating the performance of such systems in natural and unsupervised settings.

Modeling perceived stress via HRV and accelerometer sensor streams.

Discovering and modeling of stress patterns of human beings is a key step towards achieving automatic stress monitoring, stress management and healthy lifestyle. As various wearable sensors become popular, it becomes possible for individuals to acquire their own relevant sensory data and to automatically assess their stress level on the go. Previous studies for stress analysis were conducted in the controlled laboratory and clinic settings. These studies are not suitable for stress monitoring in one's daily life as various physical activities may affect the physiological signals. In this paper, we address such issue by integrating two modalities of sensors, i.e., HRV sensors and accelerometers, to monitor the perceived stress levels in daily life. We gathered both the heart and the motion data from 8 participants continuously for about 2 weeks. We then extracted features from both sensory data and compared the existing machine learning methods for learning personalized models to interpret the perceived stress levels. Experimental results showed that Bagging classifier with feature selection is able to achieve a prediction accuracy 85.7%, indicating our stress monitoring on daily basis is fairly practical.