Fall Risk Assessment Through Automatic Combination of Clinical Fall Risk Factors and Body-Worn Sensor Data

Performance of digital technologies in assessing fall risks among older adults with cognitive impairment: a systematic review

Older adults with cognitive impairment (CI) are twice as likely to fall compared to the general older adult population. Traditional fall risk assessments may not be suitable for older adults with CI due to their reliance on attention and recall. Hence, there is an interest in using objective technology-based fall risk assessment tools to assess falls within this population. This systematic review aims to evaluate the features and performance of technology-based fall risk assessment tools for older adults with CI. A systematic search was conducted across several databases such as PubMed and IEEE Xplore, resulting in the inclusion of 22 studies. Most studies focused on participants with dementia. The technologies included sensors, mobile applications, motion capture, and virtual reality. Fall risk assessments were conducted in the community, laboratory, and institutional settings; with studies incorporating continuous monitoring of older adults in everyday environments. Studies used a combination of technology-based inputs of gait parameters, socio-demographic indicators, and clinical assessments. However, many missed the opportunity to include cognitive performance inputs as predictors to fall risk. The findings of this review support the use of technology-based fall risk assessment tools for older adults with CI. Further advancements incorporating cognitive measures and additional longitudinal studies are needed to improve the effectiveness and clinical applications of these assessment tools. Additional work is also required to compare the performance of existing methods for fall risk assessment, technology-based fall risk assessments, and the combination of these approaches.

Systematic review of candidate prognostic factors for falling in older adults identified from motion analysis of challenging walking tasks

The objective of this systematic review is to identify motion analysis parameters measured during challenging walking tasks which can predict fall risk in the older population. Numerous studies have attempted to predict fall risk from the motion analysis of standing balance or steady walking. However, most falls do not occur during steady gait but occur due to challenging centre of mass displacements or environmental hazards resulting in slipping, tripping or falls on stairs. We conducted a systematic review of motion analysis parameters during stair climbing, perturbed walking and obstacle crossing, predictive of fall risk in healthy older adults. We searched the databases of Pubmed, Scopus and IEEEexplore.A total of 78 articles were included, of which 62 simply compared a group of younger to a group of older adults. Importantly, the differences found between younger and older adults did not match those found between older adults at higher and lower risk of falls. Two prospective and six retrospective fall history studies were included. The other eight studies compared two groups of older adults with higher or lower risk based on mental or physical performance, functional decline, unsteadiness complaints or task performance. A wide range of parameters were reported, including outcomes related to success, timing, foot and step, centre of mass, force plates, dynamic stability, joints and segments. Due to the large variety in parameter assessment methods, a meta-analysis was not possible. Despite the range of parameters assessed, only a few candidate prognostic factors could be identified: older adults with a retrospective fall history demonstrated a significant larger step length variability, larger step time variability, and prolonged anticipatory postural adjustments in obstacle crossing compared to older adults without a fall history. Older adults who fell during a tripping perturbation had a larger angular momentum than those who did not fall. Lastly, in an obstacle course, reduced gait flexibility (i.e., change in stepping pattern relative to unobstructed walking) was a prognostic factor for falling in daily life. We provided recommendations for future fall risk assessment in terms of study design.In conclusion, studies comparing older to younger adults cannot be used to explore relationships between fall risk and motion analysis parameters. Even when comparing two older adult populations, it is necessary to measure fall history to identify fall risk prognostic factors.

Clinical Sensor-Based Fall Risk Assessment at an Orthopedic Clinic: A Case Study of the Staff's Views on Utility and Effectiveness

In-hospital falls are a serious threat to patient security and fall risk assessment (FRA) is important to identify high-risk patients. Although sensor-based FRA (SFRA) can provide objective FRA, its clinical use is very limited and research to identify meaningful SFRA methods is required. This study aimed to investigate whether examples of SFRA methods might be relevant for FRA at an orthopedic clinic. Situations where SFRA might assist FRA were identified in a focus group interview with clinical staff. Thereafter, SFRA methods were identified in a literature review of SFRA methods developed for older adults. These were screened for potential relevance in the previously identified situations. Ten SFRA methods were considered potentially relevant in the identified FRA situations. The ten SFRA methods were presented to staff at the orthopedic clinic, and they provided their views on the SFRA methods by filling out a questionnaire. Clinical staff saw that several SFRA tasks could be clinically relevant and feasible, but also identified time constraints as a major barrier for clinical use of SFRA. The study indicates that SFRA methods developed for community-dwelling older adults may be relevant also for hospital inpatients and that effectiveness and efficiency are important for clinical use of SFRA.

Characterizing fall risk factors in Belgian older adults through machine learning: a data-driven approach

BACKGROUND

Falls are a major problem associated with ageing. Yet, fall-risk classification models identifying older adults at risk are lacking. Current screening tools show limited predictive validity to differentiate between a low- and high-risk of falling.

OBJECTIVE

This study aims at identifying risk factors associated with higher risk of falling by means of a quality-of-life questionnaire incorporating biological, behavioural, environmental and socio-economic factors. These insights can aid the development of a fall-risk classification algorithm identifying community-dwelling older adults at risk of falling.

METHODS

The questionnaire was developed by the Belgian Ageing Studies research group of the Vrije Universiteit Brussel and administered to 82,580 older adults for a detailed analysis of risk factors linked to the fall incidence data. Based on previously known risk factors, 139 questions were selected from the questionnaire to include in this study. Included questions were encoded, missing values were dropped, and multicollinearity was assessed. A random forest classifier that learns to predict falls was trained to investigate the importance of each individual feature.

RESULTS

Twenty-four questions were included in the classification-model. Based on the output of the model all factors were associated with the risk of falling of which two were biological risk factors, eight behavioural, 11 socioeconomic and three environmental risk factors. Each of these variables contributed between 4.5 and 6.5% to explaining the risk of falling.

CONCLUSION

The present study identified 24 fall risk factors using machine learning techniques to identify older adults at high risk of falling. Maintaining a mental, physical and socially active lifestyle, reducing vulnerability and feeling satisfied with the living situation contributes to reducing the risk of falling. Further research is warranted to establish an easy-to-use screening tool to be applied in daily practice.

A Systematic Review of Wearable Sensor-Based Technologies for Fall Risk Assessment in Older Adults

Falls have been recognized as the major cause of accidental death and injury in people aged 65 and above. The timely prediction of fall risks can help identify older adults prone to falls and implement preventive interventions. Recent advancements in wearable sensor-based technologies and big data analysis have spurred the development of accurate, affordable, and easy-to-use approaches to fall risk assessment. The objective of this study was to systematically assess the current state of wearable sensor-based technologies for fall risk assessment among community-dwelling older adults. Twenty-five of 614 identified research articles were included in this review. A comprehensive comparison was conducted to evaluate these approaches from several perspectives. In general, these approaches provide an accurate and effective surrogate for fall risk assessment. The accuracy of fall risk prediction can be influenced by various factors such as sensor location, sensor type, features utilized, and data processing and modeling techniques. Features constructed from the raw signals are essential for predictive model development. However, more investigations are needed to identify distinct, clinically interpretable features and develop a general framework for fall risk assessment based on the integration of sensor technologies and data modeling.

Identification of Individually Altered Gait Behavior Using an Unobtrusive IMU Sensor Setup

Gait behavior is considered an important indicator for the assessment of the general health status and provides a diagnostic observation for neuro-degenerative and musculo-skeletal diseases. Individual changes in gait behavior often reflect a deterioration of the current health status in a general sense and therefore provide significant information for clinicians and care-givers. In this work, we have used an unobtrusive sensor setup comprising three inertial measurement units (IMUs) located at the wrist, the chest and the thigh to obtain an objective measure of the human locomotion. We conducted a clinical trial in a movement laboratory environment to obtain a database of gait data at different walking speeds and conditions. The aging-simulation suit GERT was used to deteriorate the individual gait behavior during the experiments. Treadmill walking trials were used to train different classifiers to discriminate normal walking from GERT-affected walking patterns. Level-ground walking trials were used to validate the previously generated classifiers. A five-fold cross validation during the training process yielded overall F₁-scores between 0.965 and 0.986. The validation tests showed promising results with prediction accuracies of more than 80%. Clinical relevance- The clinical relevance of this contri-bution can be considered two-fold. First we demonstrate the possibility of an unobtrusive monitoring system to iden-tify individual deterioration of gait behavior. Second we also validate the use of aging-simulation suits to introduce individual changes of gait patterns in healthy subjects to create a database of simulated yet realistic gait impairments associated with aging.

Human Resource Planning and Configuration Based on Machine Learning

Human resources are the core resources of an enterprise, and the demand forecasting plays a vital role in the allocation and optimization of human resources. Starting from the basic concepts of human resource forecasting, this paper employs the backpropagation neural network (BPNN) and radial basis function neural network (RBFNN) to analyze human resource needs and determine the key elements of the company's human resource allocation through predictive models. With historical data as reference, the forecast value of current human resource demand is obtained through the two types of neural networks. Based on the prediction results, the company managers can carry out targeted human resource planning and allocation to improve the efficiency of enterprise operations. In the experiment, the actual human resource data of a certain company are used as the experimental basic samples to train and test the two types of machine learning tools. The experimental results show that the method proposed in this paper can effectively predict the number of personnel required and can support the planning and allocation of human resources.

Development of Data-driven Metrics for Balance Impairment and Fall Risk Assessment in Older Adults

Ageing incurs a natural decline of postural control which has been linked to an increased risk of falling. Accurate balance assessment is important in identifying postural instability and informing targeted interventions to prevent falls in older adults. Inertial sensor (IMU) technology offers a low-cost means for objective quantification of human movement. This paper describes two studies carried out to advance the use of IMU-based balance assessments in older adults. Study 1 (N=39) presents the development of two new IMU-derived balance measures. Study 2 (N=248) reports a reliability analysis of IMU postural stability measures and validates the novel balance measures through comparison with clinical scales. We also report a statistical fall risk estimation algorithm based on IMU data captured during static balance assessments alongside a method of improving this fall risk estimate by incorporating standard clinical fall risk factor data. Results suggest that both new balance measures are sensitive to balance deficits captured by the Berg Balance Scale (BBS) and Timed Up and Go test. Results obtained from the fall risk classifier models suggest they are more accurate (67.9%) at estimating fall risk status than a model based on BBS (59.2%). While the accuracies of the reported models are lower than others reported in the literature, the simplicity of the assessment makes it a potentially useful screening tool for balance impairments and falls risk. The algorithms presented in this paper may be suitable for implementation on a smartphone and could facilitate unsupervised assessment in the home.

Risk Analysis of Textile Industry Foreign Investment Based on Deep Learning

With the decline of China's economic growth rate and the uproar of antiglobalization, the textile industry, one of the business cards of China's globalization, is facing a huge impact. When the economic model is undergoing transformation, it is more important to prevent enterprises from falling into financial distress. So, the financial risk early warning is one of the important means to prevent enterprises from falling into financial distress. Aiming at the risk analysis of the textile industry's foreign investment, this paper proposes an analysis method based on deep learning. This method combines residual network (ResNet) and long short-term memory (LSTM) risk prediction model. This method first establishes a risk indicator system for the textile industry and then uses ResNet to complete deep feature extraction, which are further used for LSTM training and testing. The performance of the proposed method is tested based on part of the measured data, and the results show the effectiveness of the proposed method.

Predicting Fall Counts Using Wearable Sensors: A Novel Digital Biomarker for Parkinson's Disease

People with Parkinson's disease (PD) experience significant impairments to gait and balance; as a result, the rate of falls in people with Parkinson's disease is much greater than that of the general population. Falls can have a catastrophic impact on quality of life, often resulting in serious injury and even death. The number (or rate) of falls is often used as a primary outcome in clinical trials on PD. However, falls data can be unreliable, expensive and time-consuming to collect. We sought to validate and test a novel digital biomarker for PD that uses wearable sensor data obtained during the Timed Up and Go (TUG) test to predict the number of falls that will be experienced by a person with PD. Three datasets, containing a total of 1057 (671 female) participants, including 71 previously diagnosed with PD, were included in the analysis. Two statistical approaches were considered in predicting falls counts: the first based on a previously reported falls risk assessment algorithm, and the second based on elastic net and ensemble regression models. A predictive model for falls counts in PD showed a mean R² value of 0.43, mean error of 0.42 and a mean correlation of 30% when the results were averaged across two independent sets of PD data. The results also suggest a strong association between falls counts and a previously reported inertial sensor-based falls risk estimate. In addition, significant associations were observed between falls counts and a number of individual gait and mobility parameters. Our preliminary research suggests that the falls counts predicted from the inertial sensor data obtained during a simple walking task have the potential to be developed as a novel digital biomarker for PD, and this deserves further validation in the targeted clinical population.

Enterprise Risk Assessment Based on Machine Learning

Scientific risk assessment is an important guarantee for the healthy development of an enterprise. With the continuous development and maturity of machine learning technology, it has played an important role in the field of data prediction and risk assessment. This paper conducts research on the application of machine learning technology in enterprise risk assessment. According to the existing literature, this paper uses three machine learning algorithms, i.e., random forest (RF), support vector machine (SVM), and AdaBoost, to evaluate enterprise risk. In the specific implementation, the enterprise's risk assessment indexes are first established, which comprehensively describe the various risks faced by the enterprise through a number of parameters. Then, the three types of machine learning algorithms are trained based on historical data to build a risk assessment model. Finally, for a set of risk indicators obtained under current conditions, the risk index is output through the risk assessment model. In the experiment, some actual data are used to analyze and verify the method, and the results show that the proposed three types of machine learning algorithms can effectively evaluate enterprise risks.

Performance and Characteristics of Wearable Sensor Systems Discriminating and Classifying Older Adults According to Fall Risk: A Systematic Review

Sensor-based fall risk assessment (SFRA) utilizes wearable sensors for monitoring individuals’ motions in fall risk assessment tasks. Previous SFRA reviews recommend methodological improvements to better support the use of SFRA in clinical practice. This systematic review aimed to investigate the existing evidence of SFRA (discriminative capability, classification performance) and methodological factors (study design, samples, sensor features, and model validation) contributing to the risk of bias. The review was conducted according to recommended guidelines and 33 of 389 screened records were eligible for inclusion. Evidence of SFRA was identified: several sensor features and three classification models differed significantly between groups with different fall risk (mostly fallers/non-fallers). Moreover, classification performance corresponding the AUCs of at least 0.74 and/or accuracies of at least 84% were obtained from sensor features in six studies and from classification models in seven studies. Specificity was at least as high as sensitivity among studies reporting both values. Insufficient use of prospective design, small sample size, low in-sample inclusion of participants with elevated fall risk, high amounts and low degree of consensus in used features, and limited use of recommended model validation methods were identified in the included studies. Hence, future SFRA research should further reduce risk of bias by continuously improving methodology.

Detecting subtle mobility changes among older adults: the Quantitative Timed Up and Go test

BACKGROUND

The Quantitative Timed Up and Go (QTUG) test uses wearable sensors, containing a triaxial accelerometer and an add-on triaxial gyroscope, to quantify performance during the TUG test with potential to capture more minor changes in mobility.

AIMS

To examine the responsiveness, minimum detectable change (MDC) and observed effect size of QTUG in a cohort of socially active adults aged 50 years and over participating in a structured community exercise program.

METHODS

54 participants (91% females, mean age 63.6 ± 6.5 years) completed repeated QTUG testing under single- and dual-task conditions. Responsiveness of the QTUG was assessed by correlation of change in standard TUG with QTUG change (Pearson's correlation coefficient). MDC and effect sizes (standardized mean difference and Cohen's d) were also calculated for QTUG.

RESULTS

There was a strong positive correlation between change in the standard TUG and change in QTUG (single task r = 0.91, p < 0.001). MDC in QTUG was calculated as 0.77 (Sd, 1.39; ICC 0.96) seconds (single task) and 2.33 (Sd 2.18; ICC 0.85) seconds (dual task). Several QTUG parameters showed improvements in mean values with small effect sizes (sit -to-stand transition time d = 0.418; walk time d = 0.398; cadence d = 0.306, swing time d = 0.314; step time d = 0.479; gait velocity d = 0.365; time to reach turn d = 0.322) under single-task conditions and with a moderate effect size (d = 0.549) in time taken to turn under the dual-task condition.

CONCLUSION

Initial evidence of QTUG's responsiveness to change in mobility in active middle to older age adults has been demonstrated with small to moderate effect sizes observed in specific QTUG parameters.

Unsupervised Assessment of Balance and Falls Risk Using a Smartphone and Machine Learning

Assessment of health and physical function using smartphones (mHealth) has enormous potential due to the ubiquity of smartphones and their potential to provide low cost, scalable access to care as well as frequent, objective measurements, outside of clinical environments. Validation of the algorithms and outcome measures used by mHealth apps is of paramount importance, as poorly validated apps have been found to be harmful to patients. Falls are a complex, common and costly problem in the older adult population. Deficits in balance and postural control are strongly associated with falls risk. Assessment of balance and falls risk using a validated smartphone app may lessen the need for clinical assessments which can be expensive, requiring non-portable equipment and specialist expertise. This study reports results for the real-world deployment of a smartphone app for self-directed, unsupervised assessment of balance and falls risk. The app relies on a previously validated algorithm for assessment of balance and falls risk; the outcome measures employed were trained prior to deployment on an independent data set. Results for a sample of 594 smartphone assessments from 147 unique phones show a strong association between self-reported falls history and the falls risk and balance impairment scores produced by the app, suggesting they may be clinically useful outcome measures. In addition, analysis of the quantitative balance features produced seems to suggest that unsupervised, self-directed assessment of balance in the home is feasible.

The effects of mechanical noise bandwidth on balance across flat and compliant surfaces

Although the application of sub-sensory mechanical noise to the soles of the feet has been shown to enhance balance, there has been no study on how the bandwidth of the noise affects balance. Here, we report a single-blind randomized controlled study on the effects of a narrow and wide bandwidth mechanical noise on healthy young subjects’ sway during quiet standing on firm and compliant surfaces. For the firm surface, there was no improvement in balance for both bandwidths—this may be because the young subjects could already balance near-optimally or optimally on the surface by themselves. For the compliant surface, balance improved with the introduction of wide but not narrow bandwidth noise, and balance is improved for wide compared to narrow bandwidth noise. This could be explained using a simple model, which suggests that adding noise to a sub-threshold pressure stimulus results in markedly different frequency of nerve impulse transmitted to the brain for the narrow and wide bandwidth noise—the frequency is negligible for the former but significantly higher for the latter. Our results suggest that if a person’s standing balance is not optimal (for example, due to aging), it could be improved by applying a wide bandwidth noise to the feet.

Channel Modeling of an Optical Wireless Body Sensor Network for Walk Monitoring of Elderly

The growing aging of the world population is leading to an aggravation of diseases, which affect the autonomy of the elderly. Wireless body sensor networks (WBSN) are part of the solutions studied for several years to monitor and prevent loss of autonomy. The use of optical wireless communications (OWC) is seen as an alternative to radio frequencies, relevant when electromagnetic interference and data security considerations are important. One of the main challenges in this context is optical channel modeling for efficiently designing high-reliability systems. We propose here a suitable optical WBSN channel model for tracking the elderly during a walk. We discuss the specificities related to the model of the body, to movements, and to the walking speed by comparing elderly and young models, taking into account the walk temporal evolution using the sliding windowing technique. We point out that, when considering a young body model, performance is either overestimated or underestimated, depending on which windowing parameter is fixed. It is, therefore, important to consider the body model of the elderly in the design of the system. To illustrate this result, we then evaluate the minimal power according to the maximal bandwidth for a given quality of service.

Technology Utilization in Fall Prevention

ABSTRACT

Falls, defined as unplanned descents to the floor with or without injury to an individual, remain to be one of the most challenging health conditions. Fall rate is a key quality metric of acute care hospitals, rehabilitation settings, and long-term care facilities. Fall prevention policies with proper implementation have been the focus of surveys by regulatory bodies, including The Joint Commission and the Centers for Medicare and Medicaid Services, for all healthcare settings. Since October 2008, the Centers for Medicare and Medicaid Services has stopped reimbursing hospitals for the costs related to patient falls, shifting the accountability for fall prevention to the healthcare providers. Research shows that almost one-third of falls can be prevented and extensive fall prevention interventions exist. Recently, technology-based applications have been introduced in healthcare to obtain superior patient care outcomes and experience via efficiency, access, and reliability. Several areas in fall prevention deploy technology, including predictive and prescriptive analytics using big data, video monitoring and alarm technology, wearable sensors, exergame and virtual reality, robotics in home environment assessment, and personal coaching. This review discusses an overview of these technology-based applications in various settings, focusing on the outcomes of fall reductions, cost, and other benefits.

Using a Motion Sensor to Categorize Nonspecific Low Back Pain Patients: A Machine Learning Approach

Nonspecific low back pain (NSLBP) constitutes a critical health challenge that impacts millions of people worldwide with devastating health and socioeconomic consequences. In today’s clinical settings, practitioners continue to follow conventional guidelines to categorize NSLBP patients based on subjective approaches, such as the STarT Back Screening Tool (SBST). This study aimed to develop a sensor-based machine learning model to classify NSLBP patients into different subgroups according to quantitative kinematic data, i.e., trunk motion and balance-related measures, in conjunction with STarT output. Specifically, inertial measurement units (IMU) were attached to the trunks of ninety-four patients while they performed repetitive trunk flexion/extension movements on a balance board at self-selected pace. Machine learning algorithms (support vector machine (SVM) and multi-layer perceptron (MLP)) were implemented for model development, and SBST results were used as ground truth. The results demonstrated that kinematic data could successfully be used to categorize patients into two main groups: high vs. low-medium risk. Accuracy levels of ~75% and 60% were achieved for SVM and MLP, respectively. Additionally, among a range of variables detailed herein, time-scaled IMU signals yielded the highest accuracy levels (i.e., ~75%). Our findings support the improvement and use of wearable systems in developing diagnostic and prognostic tools for various healthcare applications. This can facilitate development of an improved, cost-effective quantitative NSLBP assessment tool in clinical and home settings towards effective personalized rehabilitation.

Development and validation of a robotic multifactorial fall-risk predictive model: A one-year prospective study in community-dwelling older adults

BACKGROUND

Falls in the elderly are a major public health concern because of their high incidence, the involvement of many risk factors, the considerable post-fall morbidity and mortality, and the health-related and social costs. Given that many falls are preventable, the early identification of older adults at risk of falling is crucial in order to develop tailored interventions to prevent such falls. To date, however, the fall-risk assessment tools currently used in the elderly have not shown sufficiently high predictive validity to distinguish between subjects at high and low fall risk. Consequently, predicting the risk of falling remains an unsolved issue in geriatric medicine. This one-year prospective study aims to develop and validate, by means of a cross-validation method, a multifactorial fall-risk model based on clinical and robotic parameters in older adults.

METHODS

Community-dwelling subjects aged ≥ 65 years were enrolled. At the baseline, all subjects were evaluated for history of falling and number of drugs taken daily, and their gait and balance were evaluated by means of the Timed "Up & Go" test (TUG), Gait Speed (GS), Short Physical Performance Battery (SPPB) and Performance-Oriented Mobility Assessment (POMA). They also underwent robotic assessment by means of the hunova robotic device to evaluate the various components of balance. All subjects were followed up for one-year and the number of falls was recorded. The models that best predicted falls-on the basis of: i) only clinical parameters; ii) only robotic parameters; iii) clinical plus robotic parameters-were identified by means of a cross-validation method.

RESULTS

Of the 100 subjects initially enrolled, 96 (62 females, mean age 77.17±.49 years) completed the follow-up and were included. Within one year, 32 participants (33%) experienced at least one fall ("fallers"), while 64 (67%) did not ("non-fallers"). The best classifier model to emerge from cross-validated fall-risk estimation included eight clinical variables (age, sex, history of falling in the previous 12 months, TUG, Tinetti, SPPB, Low GS, number of drugs) and 20 robotic parameters, and displayed an area under the receiver operator characteristic (ROC) curve of 0.81 (95% CI: 0.72-0.90). Notably, the model that included only three of these clinical variables (age, history of falls and low GS) plus the robotic parameters showed similar accuracy (ROC AUC 0.80, 95% CI: 0.71-0.89). In comparison with the best classifier model that comprised only clinical parameters (ROC AUC: 0.67; 95% CI: 0.55-0.79), both models performed better in predicting fall risk, with an estimated Net Reclassification Improvement (NRI) of 0.30 and 0.31 (p = 0.02), respectively, and an estimated Integrated Discrimination Improvement (IDI) of 0.32 and 0.27 (p<0.001), respectively. The best model that comprised only robotic parameters (the 20 parameters identified in the final model) achieved a better performance than the clinical parameters alone, but worse than the combination of both clinical and robotic variables (ROC AUC: 0.73, 95% CI 0.63-0.83).

CONCLUSION

A multifactorial fall-risk assessment that includes clinical and hunova robotic variables significantly improves the accuracy of predicting the risk of falling in community-dwelling older people. Our data suggest that combining clinical and robotic assessments can more accurately identify older people at high risk of falls, thereby enabling personalized fall-prevention interventions to be undertaken.

Unit-level variation in bed alarm use in US hospitals

Bed and chair alarms are widely used in hospitals, despite lack of effectiveness and unintended negative consequences. In this cross-sectional, observational study, we examined alarm prevalence and contributions of patient- and unit-level factors to alarm use on 59 acute care nursing units in 57 US hospitals participating in the National Database of Nursing Quality Indicators®. Nursing unit staff reported data on patient-level fall risk and fall prevention measures for 1,489 patients. Patient-level propensity scores for alarm use were estimated using logistic regression. Expected alarm use on each unit, defined as the mean patient propensity-for-alarm score, was compared with the observed rate of alarm use. Over one-third of patients assessed had an alarm in the "on" position. Patient characteristics associated with higher odds of alarm use included recent fall, need for ambulation assistance, poor mobility judgment, and altered mental status. Observed rates of unit alarm use ranged from 0% to 100% (median 33%, 10th percentile 5%, 90th percentile 67%). Expected alarm use varied less (median 31%, 10th percentile 27%, and 90th percentile 45%). Only 29% of variability in observed alarm use was accounted for by expected alarm use. Unit assignment was a stronger predictor of alarm use than patient-level fall risk variables. Alarm use is common, varies widely across hospitals, and cannot be fully explained by patient fall risk factors; alarm use is driven largely by unit practices. Alarms are used too frequently and too indiscriminately, and guidance is needed for optimizing alarm use to reduce noise and encourage mobility in appropriate patients.

Timed Up and Go and Six-Minute Walking Tests with Wearable Inertial Sensor: One Step Further for the Prediction of the Risk of Fall in Elderly Nursing Home People

Assessing the risk of fall in elderly people is a difficult challenge for clinicians. Since falls represent one of the first causes of death in such people, numerous clinical tests have been created and validated over the past 30 years to ascertain the risk of falls. More recently, the developments of low-cost motion capture sensors have facilitated observations of gait differences between fallers and nonfallers. The aim of this study is twofold. First, to design a method combining clinical tests and motion capture sensors in order to optimize the prediction of the risk of fall. Second to assess the ability of artificial intelligence to predict risk of fall from sensor raw data only. Seventy-three nursing home residents over the age of 65 underwent the Timed Up and Go (TUG) and six-minute walking tests equipped with a home-designed wearable Inertial Measurement Unit during two sets of measurements at a six-month interval. Observed falls during that interval enabled us to divide residents into two categories: fallers and nonfallers. We show that the TUG test results coupled to gait variability indicators, measured during a six-minute walking test, improve (from 68% to 76%) the accuracy of risk of fall’s prediction at six months. In addition, we show that an artificial intelligence algorithm trained on the sensor raw data of 57 participants reveals an accuracy of 75% on the remaining 16 participants.

A Practical Sensor-Based Methodology for the Quantitative Assessment and Classification of Chronic Non Specific Low Back Patients (NSLBP) in Clinical Settings

The successful clinical application of patient-specific personalized medicine for the management of low back patients remains elusive. This study aimed to classify chronic nonspecific low back pain (NSLBP) patients using our previously developed and validated wearable inertial sensor (SHARIF-HMIS) for the assessment of trunk kinematic parameters. One hundred NSLBP patients consented to perform repetitive flexural movements in five different planes of motion (PLM): 0° in the sagittal plane, as well as 15° and 30° lateral rotation to the right and left, respectively. They were divided into three subgroups based on the STarT Back Screening Tool. The sensor was placed on the trunk of each patient. An ANOVA mixed model was conducted on the maximum and average angular velocity, linear acceleration and maximum jerk, respectively. The effect of the three-way interaction of Subgroup by direction by PLM on the mean trunk acceleration was significant. Subgrouping by STarT had no main effect on the kinematic indices in the sagittal plane, although significant effects were observed in the asymmetric directions. A significant difference was also identified during pre-rotation in the transverse plane, where the velocity and acceleration decreased while the jerk increased with increasing asymmetry. The acceleration during trunk flexion was significantly higher than that during extension, in contrast to the velocity, which was higher in extension. A Linear Discriminant Analysis, utilized for classification purposes, demonstrated that 51% of the total performance classifying the three STarT subgroups (65% for high risk) occurred at a position of 15° of rotation to the right during extension. Greater discrimination (67%) was obtained in the classification of the high risk vs. low-medium risk. This study provided a smart "sensor-based" practical methodology for quantitatively assessing and classifying NSLBP patients in clinical settings. The outcomes may also be utilized by leveraging cost-effective inertial sensors, already available in today's smartphones, as objective tools for various health applications towards personalized precision medicine.

Gait Speed Characteristics and Their Spatiotemporal Determinants in Nursing Home Residents: A Cross-Sectional Study

BACKGROUND AND PURPOSE

Low and slowing gait speeds among nursing home residents are linked to a higher risk of disability, cognitive impairment, falls, and mortality. A better understanding of the spatiotemporal parameters of gait that influence declining mobility could lead to effective rehabilitation and preventative intervention. The aims of this study were to objectively quantify the spatiotemporal characteristics of gait in the nursing home setting and define the relationship between these parameters and gait speed.

METHODS

One hundred nursing home residents were enrolled into the study and completed 3 habitual gait speed trials over a distance of 3.66 m. Trials were performed using an instrumented gait analysis. The manner in which the spatiotemporal parameters predicted gait speed was examined by univariate and multivariable regression modeling.

RESULTS

The nursing home residents had a habitual mean (SD) gait speed of 0.63 (0.19) m/s, a stride length of 0.83 (0.15) m, a support base of 0.15 (0.06) m, and step time of 0.66 (0.12) seconds. Multivariable linear regression revealed stride length, support base, and step time predicted gait speed (R = 0.89, P < .05). Step time had the greatest influence on gait speed, with each 0.1-second decrease in step time resulting in a 0.09 m/s (95% confidence interval, 0.08-0.10) increase in habitual gait speed.

CONCLUSIONS

This study revealed step time, stride length, and support base are the strongest predictors of gait speed among nursing home residents. Future research should concentrate on developing and evaluating intervention programs that were specifically designed to focus on the strong predictors of gait speed in nursing home residents. We would also suggest that routine assessments of gait speed, and if possible their spatiotemporal characteristics, be done on all nursing home residents in an attempt to identify residents with low or slowing gait speed.

Reliability, Validity and Utility of Inertial Sensor Systems for Postural Control Assessment in Sport Science and Medicine Applications: A Systematic Review

BACKGROUND

Recent advances in mobile sensing and computing technology have provided a means to objectively and unobtrusively quantify postural control. This has resulted in the rapid development and evaluation of a series of wearable inertial sensor-based assessments. However, the validity, reliability and clinical utility of such systems is not fully understood.

OBJECTIVES

This systematic review aims to synthesise and evaluate studies that have investigated the ability of wearable inertial sensor systems to validly and reliably quantify postural control performance in sports science and medicine applications.

METHODS

A systematic search strategy utilising the PRISMA guidelines was employed to identify eligible articles through ScienceDirect, Embase and PubMed databases. In total, 47 articles met the inclusion criteria and were evaluated and qualitatively synthesised under two main headings: measurement validity and measurement reliability. Furthermore, studies that investigated the utility of these systems in clinical populations were summarised and discussed.

RESULTS

After duplicate removal, 4374 articles were identified with the search strategy, with 47 papers included in the final review. In total, 28 studies investigated validity in healthy populations, and 15 studies investigated validity in clinical populations; 13 investigated the measurement reliability of these sensor-based systems.

CONCLUSIONS

The application of wearable inertial sensors for sports science and medicine postural control applications is an evolving field. To date, research has primarily focused on evaluating the validity and reliability of a heterogeneous set of assessment protocols, in a laboratory environment. While researchers have begun to investigate their utility in clinical use cases such as concussion and musculoskeletal injury, most studies have leveraged small sample sizes, are of low quality and use a variety of descriptive variables, assessment protocols and sensor-mounting locations. Future research should evaluate the clinical utility of these systems in large high-quality prospective cohort studies to establish the role they may play in injury risk identification, diagnosis and management. This systematic review was registered with the International Prospective Register of Systematic Reviews on 10 August 2018 (PROSPERO registration: CRD42018106363): https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=106363 .

Internet of Things (IoT)-Enabled Elderly Fall Verification, Exploiting Temporal Inference Models in Smart Homes

Everyday life of the elderly and impaired population living in smart homes is challenging because of possible accidents that may occur due to daily activities. In such activities, persons often lean over (to reach something) and, if they not cautious, are prone to falling. To identify fall incidents, which could stochastically cause serious injuries or even death, we propose specific temporal inference models; namely, CM-I and CM-II. These models can infer a fall incident based on classification methods by exploiting wearable Internet of Things (IoT) altimeter sensors adopted by seniors. We analyzed real and synthetic data of fall and lean over incidents to test the proposed models. The results are promising for incorporating such inference models to assist healthcare for fall verification of seniors in smart homes. Specifically, the CM-II model achieved a prediction accuracy of 0.98, which is the highest accuracy when compared to other models in the literature under the McNemar's test criterion. These models could be incorporated in wearable IoT devices to provide early warning and prediction of fall incidents to clinical doctors.

Digital assessment of falls risk, frailty, and mobility impairment using wearable sensors

Falls are among the most frequent and costly population health issues, costing $50bn each year in the US. In current clinical practice, falls (and associated fall risk) are often self-reported after the “first fall”, delaying primary prevention of falls and development of targeted fall prevention interventions. Current methods for assessing falls risk can be subjective, inaccurate, have low inter-rater reliability, and do not address factors contributing to falls (poor balance, gait speed, transfers, turning). 8521 participants (72.7 ± 12.0 years, 5392 female) from six countries were assessed using a digital falls risk assessment protocol. Data consisted of wearable sensor data captured during the Timed Up and Go (TUG) test along with self-reported questionnaire data on falls risk factors, applied to previously trained and validated classifier models. We found that 25.8% of patients reported a fall in the previous 12 months, of the 74.6% of participants that had not reported a fall, 21.5% were found to have a high predicted risk of falls. Overall 26.2% of patients were predicted to be at high risk of falls. 29.8% of participants were found to have slow walking speed, while 19.8% had high gait variability and 17.5% had problems with transfers. We report an observational study of results obtained from a novel digital fall risk assessment protocol. This protocol is intended to support the early identification of older adults at risk of falls and inform the creation of appropriate personalized interventions to prevent falls. A population-based approach to management of falls using objective measures of falls risk and mobility impairment, may help reduce unnecessary outpatient and emergency department utilization by improving risk prediction and stratification, driving more patients towards clinical and community-based falls prevention activities.

Deep Learning for Fall Risk Assessment With Inertial Sensors: Utilizing Domain Knowledge in Spatio-Temporal Gait Parameters

Fall risk assessment is essential to predict and prevent falls in geriatric populations, especially patients with life-long conditions like neurological disorders. Inertial sensor-based pervasive gait analysis systems have become viable means to facilitate continuous fall risk assessment in non-hospital settings. However, a gait analysis system is not sufficient to detect the characteristics leading to increased fall risk, and powerful inference models are required to detect the underlying factors specific to fall risk. Machine learning models and especially the recently proposed deep learning methods offer the needed predictive power. Deep neural networks have the potential to produce models that can operate directly on the raw data, thus alleviating the need for feature engineering. However, the domain knowledge inherent in the well-established spatio-temporal gait parameters are still valuable to help a model achieve high inference accuracies. In this study, we explore deep learning methods, specifically long short-term memory (LSTM) neural networks, for the problem of fall risk assessment. We utilize sequences of spatio-temporal gait parameters extracted by an inertial sensor-based gait analysis system as input features. To quantify the performance of the proposed approach, we compare it with more traditional machine learning methods. The proposed LSTM model, trained with a gait dataset collected from 60 neurological disorder patients, achieves a superior classification accuracy of 92.1% on a separate test dataset collected from 16 patients. This study serves as one of the first attempts to employ deep learning approaches in this domain and the results demonstrate their potential.

A modelling-based economic evaluation of primary-care-based fall-risk screening followed by fall-prevention intervention: a cohort-based Markov model stratified by older age groups

Background

fall-risk assessment with fall-prevention intervention referral for at-risk groups to avoid falls could be cost-effective from a care-payer perspective.

Aims

to model the cost-effectiveness of a fall-risk assessment (QTUG compared to TUG) with referral to one of four fall-prevention interventions (Otago, FaME, Tai Chi, home safety assessment and modification) compared to no care pathway, when the decision to screen is based on older age in a primary care setting for community-dwelling people.

Methods

a cohort-based, decision analytic Markov model was stratified by five age groupings (65–70, 70–75, 65–89, 70–89 and 75–89) to estimate cost per quality-adjusted life years (QALYs). Costs included fall-risk assessment, fall-prevention intervention and downstream resource use (e.g. inpatient and care home admission). Uncertainty was explored using univariate, bivariate and probabilistic sensitivity analyses.

Results

screening with QTUG dominates (>QALYs; <costs) screening with TUG irrespective of subsequent fall-prevention intervention. The QTUG-based care pathways relative to no care pathway have a high probability of cost-effectiveness in those aged 75–89 (>85%), relative to those aged 70–74 (~10 < 30%) or 65–69 (<10%). In the older age group, only a 10% referral uptake is required for the QTUG with FaME or Otago modelled care pathways to remain cost-effective.

Conclusion

the highest probability of cost-effectiveness observed was a care pathway incorporating QTUG with FaME in those aged 75–89. Although the model does not fully represent current NICE Falls guidance, decision makers should still give careful consideration to implementing the aforementioned care pathway due to the modelled high probability of cost-effectiveness.

Development of a Caregivers' Support Platform (Connected Health Sustaining Home Stay in Dementia): Protocol for a Longitudinal Observational Mixed Methods Study

Background

Dementia disease is a chronic condition that leads a person with dementia (PwD) into a state of progressive deterioration and a greater dependence in performing their activities of daily living (ADL). It is believed nowadays that PwDs and their informal caregivers can have a better life when provided with the appropriate services and support. Connected Health (CH) is a new technology-enabled model of chronic care delivery where the stakeholders are connected through a health portal, ensuring continuity and efficient flow of information. CH has demonstrated promising results regarding supporting informal home care and Aging in Place, and it has been increasingly considered by researchers and health care providers as a method for dementia home care management.

Objective

This study aims to describe the development and implementation protocol of a CH platform system to support informal caregivers of PwDs at home.

Methods

This is a longitudinal observational mixed methods study where quantitative and qualitative data will be combined for determining the utility of the CH platform for dementia home care. Dyads, consisting of a PwD and their informal caregiver living in the community, will be divided into 2 groups: the intervention group, which will receive the CH technology package at home, and the usual care group, which will not have any CH technology at all. Dyads will be followed up for 12 months during which they will continue with their traditional care plan, but in addition, the intervention group will receive the CH package for their use at home during 6 months (months 3 to 9 of the yearly follow-up). Further comprehensive assessments related to the caregiver’s and PwD’s emotional and physical well-being will be performed at the initial assessment and at 3, 6, 9, and 12 months using international and standardized validated questionnaires and semistructured individual interviews.

Results

This 3-year funded study (2016-2019) is currently in its implementation phase and is expected to finish by December 2019. We believe that CH can potentially change the PwD current care model, facilitating a proactive and preventive model, utilizing self-management–based strategies, and enhancing caregivers’ involvement in the management of health care at home for PwDs.

Conclusions

We foresee that our CH platform will provide knowledge and promote autonomy for the caregivers, which may empower them into greater control of the care for PwDs, and with it, improve the quality of life and well-being for the person they are caring for and for themselves through a physical and cognitive decline predictive model. We also believe that facilitating information sharing between all the PwDs’ care stakeholders may enable a stronger relationship between them, facilitate a more coordinated care plan, and increase the feelings of empowerment in the informal caregivers.

International Registered Report Identifier (IRRID)

DERR1-10.2196/13280

Remote timed up and go evaluation from activities of daily living reveals changing mobility after surgery

BACKGROUND

Mobility impairment is common in older adults and negatively influences the quality of life. Mobility level may change rapidly following surgery or hospitalization in the elderly. The timed up and go (TUG) is a simple, frequently used clinical test for functional mobility; however, TUG requires supervision from a trained clinician, resulting in infrequent assessments. Additionally, assessment by TUG in clinic settings may not be completely representative of the individual's mobility in their home environment.

OBJECTIVE

In this paper, we introduce a method to estimate TUG from activities detected in free-living, enabling continuous remote mobility monitoring without expert supervision. The method is used to monitor changes in mobility following total hip arthroplasty (THA).

METHODS

Community-living elderly (n  =  239, 65-91 years) performed a standardized TUG in a laboratory and wore a wearable pendant device that recorded accelerometer and barometric sensor data for at least three days. Activities of daily living (ADLs), including walks and sit-to-stand transitions, and their related mobility features were extracted and used to develop a regularized linear model for remote TUG test estimation. Changes in the remote TUG were evaluated in orthopaedic patients (n  =  15, 55-75 years), during 12-weeks period following THA.

MAIN RESULTS

In leave-one-out-cross-validation (LOOCV), a strong correlation (ρ  =  0.70) was observed between the new remote TUG and standardized TUG times. Test-retest reliability of 3-days estimates was high (ICC  =  0.94). Compared to week 2 post-THA, remote TUG was significantly improved at week 6 (11.7  ±  3.9 s versus 8.0  ±  1.8 s, p   <  0.001), with no further change at 12-weeks (8.1  ±  3.9 s, p   =  0.37).

SIGNIFICANCE

Remote TUG can be estimated in older adults using 3-days of ADLs data recorded using a wearable pendant. Remote TUG has discriminatory potential for identifying frail elderly and may provide a convenient way to monitor changes in mobility in unsupervised settings.

Mobile Technology in Running Science and Medicine: Are We Ready?

Since the running revolution of the 1970s, one of the major challenges has been the burden of running-related injuries (RRIs). Researchers, sports medicine practitioners, and strength and conditioning coaches are striving to develop an understanding of which factors may increase an individuals risk of developing RRIs, which strategies can be used to ensure optimal rehabilitation and recovery from an injury, and how to best optimize athletic performance. This Viewpoint explores these factors to demonstrate how recent advances in mobile technology may allow us to uncover novel insights related to the science and medicine of running. J Orthop Sports Phys Ther 2019;49(3):122-125. doi:10.2519/jospt.2019.0604.

Implicit Irregularity Detection Using Unsupervised Learning on Daily Behaviors

The irregularity detection of daily behaviors for the elderly is an important issue in homecare. Plenty of mechanisms have been developed to detect the health condition of the elderly based on the explicit irregularity of several biomedical parameters or some specific behaviors. However, few research works focus on detecting the implicit irregularity involving the combination of diverse behaviors, which can assess the cognitive and physical wellbeing of elders but cannot be directly identified based on sensor data. This paper proposes an Implicit IRregularity Detection (IIRD) mechanism that aims to detect the implicit irregularity by developing the unsupervised learning algorithm based on daily behaviors. The proposed IIRD mechanism identifies the distance and similarity between daily behaviors, which are important features to distinguish the regular and irregular daily behaviors and detect the implicit irregularity of elderly health condition. Performance results show that the proposed IIRD outperforms the existing unsupervised machine-learning mechanisms in terms of the detection accuracy and irregularity recall.

Application of Wearable Inertial Sensors and A New Test Battery for Distinguishing Retrospective Fallers from Non-fallers among Community-dwelling Older People

Considering the challenge of population ageing and the substantial health problem among the elderly population from falls, the purpose of this study was to verify whether it is possible to distinguish accurately between older fallers and non-fallers, based on data from wearable inertial sensors collected during a specially designed test battery. A comprehensive but practical test battery using 5 wearable inertial sensors for multifactorial fall risk assessment was designed. This was followed by an experimental study on 196 community-dwelling Korean older women, categorized as fallers (N1 = 82) and non-fallers (N2 = 114) based on prior history of falls. Six machine learning models (logistic regression, naïve bayes, decision tree, random forest, boosted tree and support vector machine) were proposed for faller classification. Results indicated that compared with non-fallers, fallers performed significantly worse on the test battery. In addition, the application of sensor data and support vector machine for faller classification achieved an overall accuracy of 89.4% with 92.7% sensitivity and 84.9% specificity. These findings suggest that wearable inertial sensor based systems show promise for elderly fall risk assessment, which could be implemented in clinical practice to identify "at-risk" individuals reliably to promote proactive fall prevention.

Longitudinal assessment of falls in patients with Parkinson's disease using inertial sensors and the Timed Up and Go test

Objective

To examine the predictive validity of a TUG test for falls risk, quantified using body-worn sensors (QTUG) in people with Parkinson's Disease (PD). We also sought to examine the inter-session reliability of QTUG sensor measures and their association with the Unified Parkinson's Disease Rating Scale (UPDRS) motor score.

Approach

A six-month longitudinal study of 15 patients with Parkinson's disease. Participants were asked to complete a weekly diary recording any falls activity for six months following baseline assessment. Participants were assessed monthly, using a Timed Up and Go test, quantified using body-worn sensors, placed on each leg below the knee.

Main results

The results suggest that the QTUG falls risk estimate recorded at baseline is 73.33% (44.90, 92.21) accurate in predicting falls within 90 days, while the Timed Up and Go time at baseline was 46.67% (21.27, 73.41) accurate. The Timed Up and Go time and QTUG falls risk estimate were strongly correlated with UPDRS motor score. Fifty-two of 59 inertial sensor parameters exhibited excellent inter-session reliability, five exhibited moderate reliability, while two parameters exhibited poor reliability.

Significance

The results suggest that QTUG is a reliable tool for the assessment of gait and mobility in Parkinson's disease and, furthermore, that it may have utility in predicting falls in patients with Parkinson's disease.