Effect of investigator observation on gait parameters in individuals with stroke

Machine learning model identifies patient gait speed throughout the episode of care, generating notifications for clinician evaluation

INTRODUCTION

The advent of digital and mobile health innovations, especially use of wearables for passive data collection, allows remote monitoring and creates an abundance of data. For this information to be interpretable, machine learning (ML) processes are necessary.

RESEARCH QUESTION

Can a machine learning model successfully identify patients expected to have low gait speed in the early recovery period following joint replacement surgery?

METHODS

A commercial database from a smartphone-based care management platform passively collecting mobility data pre- and post-lower limb arthroplasty was used. We sought to create a ML model to predict gait speed recovery curves and identify patients at risk of poor gait speed outcome, a measure associated with range of motion and patient-reported outcomes. Model performance including sensitivity, specificity, precision, and accuracy were determined. Receiver operator curve (ROC) analysis was used to compare true and false positive rates. To benchmark our model, we compared threshold-based notifications based on the patient's current gait speed.

RESULTS

The performance of the predictive model was significantly improved compared to baseline of threshold-based exceptions using current gait speed. The ML model currently provides 53 % precision, 88 % accuracy, 36 % sensitivity, and 95 % specificity on the held-out test set. The ROC analysis suggests good clinical performance (AUC=0.81).

SIGNIFICANCE

Utilization of ML to predict gait recovery following total joint replacement is feasible and provides results with excellent specificity. This model will allow inclusion of additional data for retraining as patient populations evolve. Clinician feedback regarding notifications, including resulting actions and outcomes, can be used to further inform the model and improve clinical utility.

Hang-Time HAR: A Benchmark Dataset for Basketball Activity Recognition Using Wrist-Worn Inertial Sensors

We present a benchmark dataset for evaluating physical human activity recognition methods from wrist-worn sensors, for the specific setting of basketball training, drills, and games. Basketball activities lend themselves well for measurement by wrist-worn inertial sensors, and systems that are able to detect such sport-relevant activities could be used in applications of game analysis, guided training, and personal physical activity tracking. The dataset was recorded from two teams in separate countries (USA and Germany) with a total of 24 players who wore an inertial sensor on their wrist, during both a repetitive basketball training session and a game. Particular features of this dataset include an inherent variance through cultural differences in game rules and styles as the data was recorded in two countries, as well as different sport skill levels since the participants were heterogeneous in terms of prior basketball experience. We illustrate the dataset's features in several time-series analyses and report on a baseline classification performance study with two state-of-the-art deep learning architectures.

Influence of the Hawthorne effect on spatiotemporal parameters, kinematics, ground reaction force, and the symmetry of the dominant and nondominant lower limbs during gait

The Hawthorne effect is a change in behavior resulting from awareness of being observed or evaluated. This study aimed to determine whether awareness of being evaluated or presence of an observer influence gait. Twenty-one young women were asked to walk in three conditions. In the first condition (unawareness of evaluation; UE), participants were aware that it was a practice trial, and there was no observer. In the second condition (awareness of evaluation; AE), participants were aware that their gait was being evaluated. The third condition (AE + researcher observation; RO) was similar to the second condition except that an additional researcher observed the participant' gait. The spatiotemporal, kinematic, ground reaction forces, and ratio index (symmetry of both lower limbs) were compared among the three conditions. A higher ratio index indicated a relative increase in the value on left versus right. Gait speed (P = 0.012) and stride length (right and left; P = 0.006 and 0.007, respectively) were significantly increased in the AE + RO than in UE. Range of motion of the right hip and left ankle was significantly greater in AE than in UE (P = 0.039 and 0.012, respectively). The ratio index of ground reaction force during push-off was significantly higher in AE and AE + RO conditions than in UE (P < 0.001 and P = 0.004, respectively). The Hawthorne effect (awareness of being evaluated or presence of an observer) potentially influences gait. Thus, factors that influence gait analysis should be considered when evaluating normal gait.

A scoping review on recent trends in wearable sensors to analyze gait in people with stroke: From sensor placement to validation against gold-standard equipment

The purpose of the review is to evaluate wearable sensor placement, their impact and validation of wearable sensors on analyzing gait, primarily the postural instability in people with stroke. Databases, namely PubMed, Cochrane, SpringerLink, and IEEE Xplore were searched to identify related articles published since January 2005. The authors have selected the articles by considering patient characteristics, intervention details, and outcome measurements by following the priorly set inclusion and exclusion criteria. From a total of 1077 articles, 142 were included in this study and classified into functional fields, namely postural stability (PS) assessments, physical activity monitoring (PA), gait pattern classification (GPC), and foot drop correction (FDC). The review covers the types of wearable sensors, their placement, and their performance in terms of reliability and validity. When employing a single wearable sensor, the pelvis and foot were the most used locations for detecting gait asymmetry and kinetic parameters, respectively. Multiple Inertial Measurement Units placed at different body parts were effectively used to estimate postural stability and gait pattern. This review article has compared results of placement of sensors at different locations helping researchers and clinicians to identify the best possible placement for sensors to measure specific kinematic and kinetic parameters in persons with stroke.

Defining and evaluating the Hawthorne effect in primary care, a systematic review and meta-analysis

In 2015, we conducted a randomized controlled trial (RCT) in primary care to evaluate if posters and pamphlets dispensed in general practice waiting rooms enhanced vaccination uptake for seasonal influenza. Unexpectedly, vaccination uptake rose in both arms of the RCT whereas public health data indicated a decrease. We wondered if the design of the trial had led to a Hawthorne effect (HE). Searching the literature, we noticed that the definition of the HE was unclear if stated. Our objectives were to refine a definition of the HE for primary care, to evaluate its size, and to draw consequences for primary care research. We designed a Preferred Reporting Items for Systematic reviews and Meta-Analyses review and meta-analysis between January 2012 and March 2022. We included original reports defining the HE and reports measuring it without setting limitations. Definitions of the HE were collected and summarized. Main published outcomes were extracted and measures were analyzed to evaluate odds ratios (ORs) in primary care. The search led to 180 records, reduced on review to 74 for definition and 15 for quantification. Our definition of HE is "an aware or unconscious complex behavior change in a study environment, related to the complex interaction of four biases affecting the study subjects and investigators: selection bias, commitment and congruence bias, conformity and social desirability bias and observation and measurement bias." Its size varies in time and depends on the education and professional position of the investigators and subjects, the study environment, and the outcome. There are overlap areas between the HE, placebo effect, and regression to the mean. In binary outcomes, the overall OR of the HE computed in primary care was 1.41 (95% CI: [1.13; 1.75]; I 2 = 97%), but the significance of the HE disappears in well-designed studies. We conclude that the HE results from a complex system of interacting phenomena and appears to some degree in all experimental research, but its size can considerably be reduced by refining study designs.

Hawthorne Effect in Gait Analysis of Children with In-Toeing Caused by Increased Femoral Anteversion

Background

In-toeing is one of the main reasons children are applying to the orthopedics clinic. In the clinical settings, during in-toeing gait assessment parents often define that their child does not walk same at the clinic as at home, linked possibly to Hawthorne effect.

Research Question

How does the in-toeing angle differ when children are aware, versus when they are not aware of their gait inspection?

Methods

This single center, clinical, cross-sectional, observational study looked into the variation in gait pattern of twelve children with in-toeing, with and without their awareness. Two videos for each child was recorded with a smart phone, once at the clinic while aware and once by the family without awareness, and uploaded into Kinovea software for gait analysis. The angle of foot in-toeing was measured and analyzed using SPSS comparison of means and correlations.

Results

The gait pattern evaluated with the angle of in-toeing showed a significant difference between the two videos of the same child. The angle difference returned a p value of 0.000 using paired sample t test and a Cohen's d value of 1.4, representing the large significance between clinic and family recorded videos. The initial foot in-toeing angle showed a moderate positive Pearson's correlation of 0.031 when compared with the angle difference in both settings.

Significance

These results highlight the importance of including family recorded videos in gait pattern analysis without the patient's awareness. The study not only shows the significant difference found which can be explained by Hawthorne Effect but also suggests a clearer path of understanding the child's condition with the family.

Analysing Gait Patterns in Degenerative Lumbar Spine Disease Using Inertial Wearable Sensors: An Observational Study

OBJECTIVE

Using a chest-based inertial wearable sensor, we examined the quantitative gait patterns associated with lumbar disc herniation (LDH), lumbar spinal stenosis (LSS), and chronic mechanical low back pain (CMLBP). 'Pathological gait signatures' were reported as statistically significant group difference (%) from the 'normative' gait values of an age-matched control population.

METHODS

A sample of patients presenting to the Prince of Wales Private Hospital (Sydney, Australia) with primary diagnoses of LDH, LSS, or CMLBP were recruited. Spatial, temporal, asymmetry, and variability metrics were compared with age-matched (±2 years) control participants recruited from the community. Participants were fitted at the sternal angle with an inertial measurement unit, MetaMotionC, and walked unobserved (at a self-selected pace) for 120 m along an obstacle-free, carpeted hospital corridor.

RESULTS

LDH, CMLBP, and LSS groups had unique pathological signatures of gait impairment. The LDH group (n = 33) had marked asymmetry in terms of step length, step time, stance, and single-support asymmetry. The LDH group also involved gait variability with increased step length variation. However, distinguishing the CMLBP group (n = 33) was gait variability in terms increased single-support time variation. The gait of participants with LSS (n = 22) was both asymmetric and variable in step length.

CONCLUSIONS

Wearable sensor-based accelerometry was found to be capable of detecting the gait abnormalities present in patients with LDH, LSS, and CMLBP, when compared to age-matched controls. Objective and quantitative patterns of gait deterioration uniquely varied between these subtypes of lumbar spine disease. With further testing and validation, gait signatures may aid clinical identification of gait-altering pathologies.

Measuring Gait Parameters from Structural Vibrations

Measuring gait parameters (e.g. speed, cadence, step duration) accurately is invaluable for evaluation during treatment of older adults who struggle with disability onset, disease progression, balance, and injurious falls. Traditionally stopwatches or timing gates are used to measure gait speed in clinical settings, and these are limited to measuring gait speed. Other wearable and non-wearable technologies offer the ability to measure additional gait parameters though patients are known to walk differently with the devices and even tend to slow down before engaging with a non-wearable such as a floor mat. Floor vibrations are a promising option to measuring gait parameters while not being intrusive and not requiring line-of-sight to the patient for measurements. This paper presents methodology for extracting gait parameters using vibrations with comparisons to APDM Wearable Technologies Mobility Lab sensors and stopwatch measurements. Performance is examined across 97 participants for self-selected speed forward, full speed forward, and backwards walks at three different testing sites for a total of 1039 walks. Gait speed vibrations measurements demonstrated excellent reliability with APDM Mobility Lab (ICC: 0.98; 99% CI: 0.01±0.01 m/s) and stopwatch (ICC: 0.97; 99% CI: -0.01±0.01 m/s) measurements. Similar excellent results are reported for cadence, gait cycle duration, step duration, and stride length parameters.

Field Study of Postural Characteristics of Standing and Seated Smartphone Use

Prior investigations have been primarily conducted in a laboratory to examine the effects of the smartphone use on the neck and head positions, whether these results are applicable to actual conditions is still unknown. This field survey thus analyzed the neck flexion (NF), head flexion (HF), gaze angle (GA), and viewing distance (VD) of smartphone users in public areas in Taipei, Taiwan. Six hundred smartphone users (300 men and 300 women) were photographed sagittally in standing, supported sitting, or unsupported sitting postures while using a smartphone. Results showed that women had significantly less NF and HF and shorter VDs than male users. Regardless of gender, higher NF was observed for standing than for sitting. Women had similar NF and HF while sitting supported and unsupported, but both were significantly lower than those while standing. By contrast, male users had higher NF and HF during unsupported sitting than during supported sitting. The NF (45°-50°) was much greater than the recommended maximum safe NF of 15°. Women may be at higher risk of visual strain because of shorter VD.

Does Time of Day influence postural control and gait? A review of the literature

BACKGROUND

Like many physiologic processes, Time of Day may influence postural control and gait. A better understanding of diurnal variations in postural control and gait may help to improve diagnoses, reduce falls, and optimize rehabilitation and training routines. This review summarizes the current literature that addresses these questions.

RESEARCH QUESTION

Does time of day affect postural control and gait?

METHODS

We searched PubMed, Google Scholar, and IEEE using a combination of keyword and MeSH terms. We included papers that studied human subjects and assessed gait or postural control as a function of time of day. We evaluated the quality of the identified papers based on nine assessment criteria and analyzed them considering the topic (postural control or gait), age, and characteristics of the conducted assessments. We then quantitatively synthesized the results across studies using a meta-analytical approach (i.e., Hedges' g model).

RESULTS

Twenty-two papers considered the relationship between time of day and postural control, and eleven considered the relationship between time of day and gait. Six studies found that postural control was best in the morning, four described postural control being best in the afternoon, four described optimal postural control in the evening, and eight reported no time of day effect. Two studies found gait best in the morning, five described gait best in the afternoon, two described optimal gait in the evening, and two reported no time of day effect. The results of the quantitative analysis suggest that both postural control and gait were best in the evening.

SIGNIFICANCE

While there is no clear consensus on whether there is a time of day effect for postural control and gait, the findings of this review provide initial evidence suggesting that a small but statistically significant effect exists in favor of the evening. Standardized testing, including repeated and continuous evaluations, may help provide more definitive information on time of day influences on postural control and gait.

Verbal feedback enhances motor learning during post-stroke gait retraining

BACKGROUND

Fast treadmill walking combined with functional electrical stimulation to ankle muscles (FastFES) is a well-studied gait intervention that improves post-stroke walking function. Although individualized verbal feedback is commonly incorporated during clinical gait training, and a variable practice structure is posited to enhance learning, the influence of these two factors on motor learning during locomotor interventions such as FastFES is poorly understood.

OBJECTIVES

To determine if the addition of individualized verbal feedback or variable practice to a FastFES training session enhances motor learning of targeted gait patterns.

METHODS

Nine individuals with post-stroke hemiparesis completed a crossover study comprising exposure to 3 dose-matched types of gait training: (1) FastFES (FF), comprising five 6-minute bouts of training with intermittent FES, (2) FF with addition of individualized verbal instructions and faded feedback delivered by a physical therapist (FF+PT), (3) FF with variable gait speed and FES timing (FF+Var). Gait biomechanics data were collected before (Pre), immediately after (Post), and 24-h following (Retention) each training type. Within-session and retention change scores of 3 targeted gait variables were calculated to assess locomotor learning.

RESULTS

FF+PT resulted in larger improvements within-session and at retention in trailing limb angle, and a trend for larger improvements in paretic pushoff compared to FF. FF+Var failed to show greater learning of biomechanical variables compared to FF.

CONCLUSIONS

Addition of individualized verbal feedback (FF+PT) to a single session of gait training may enhance within- and across-session learning of targeted gait variables in people post-stroke, and merits more investigation.