Fall Situations Research Articles

Effect of Sensory Integrity and Cognitive Functions on Fall History, Balance and Quality of Life in Elderly Individuals

Abstract Objectives The aim of this study is to examine the effects of sensory integrity and cognitive functions on fall history, balance and quality of life in elderly individuals. Materials and Methods After obtaining the sociodemographic information of the elderly individuals included in the study; Pattern Copying, Kinesthesia and Right-Left Discrimination, which is a subtest of the Ayres Southern California Sensory Integrity Test to evaluate sensory integrity, the Standing Balance Test for balance abilities, the Hodkinson Mental Test for cognitive abilities, the history of falling in the last year for fall situations, and the EuroQol General Quality of Life Scale (EQ-5D-3L) were used to assess the quality of life and life satisfaction. Results The study involved 44 volunteers with a mean age of 69.52±4.25 years. Cognitive ability, standing balance, and life satisfaction of elderly individuals with a fall history were worse than those without a fall history. The study also found that cognition and life satisfaction were affected by fall history. Older individuals who were more successful in pattern copying had better balance. A moderate correlation was found between cognitive ability and pattern copying (r=0.591), balance (rdominant=0.385; rnondominant=0.403), and quality of life (r=−0.399) (p<0.05). Conclusion The results suggest that elderly individuals need sensory integration and cognitively supportive treatment approaches to improve balance functions and reduce fall risk.

A graph-based approach can improve keypoint detection of complex poses: a proof-of-concept on injury occurrences in alpine ski racing

For most applications, 2D keypoint detection works well and offers a simple and fast tool to analyse human movements. However, there remain many situations where even the best state-of-the-art algorithms reach their limits and fail to detect human keypoints correctly. Such situations may occur especially when individual body parts are occluded, twisted, or when the whole person is flipped. Especially when analysing injuries in alpine ski racing, such twisted and rotated body positions occur frequently. To improve the detection of keypoints for this application, we developed a novel method that refines keypoint estimates by rotating the input videos. We select the best rotation for every frame with a graph-based global solver. Thereby, we improve keypoint detection of an arbitrary pose estimation algorithm, in particular for ‘hard’ keypoints. In the current proof-of-concept study, we show that our approach outperforms standard keypoint detection results in all categories and in all metrics, in injury-related out-of-balance and fall situations by a large margin as well as previous methods, in performance and robustness. The Injury Ski II dataset was made publicly available, aiming to facilitate the investigation of sports accidents based on computer vision in the future.

Control method of elderly-assistant robot for preventing elderly fall

The elderly are prone to fall during walking, which not only leads to injury but also can cause permanent disability or even death. Thus, it is crucial for the walking-assistant robot to prevent the elderly from fall during walking. However, the factors affecting the steadiness of the elderly are not considered comprehensively in the existing studies of elderly fall prevention model and control schemes. In this paper, a novel control method of the elderly-assistant robot for preventing elderly fall is proposed by considering the two main factors affecting steadiness of the elderly, namely, position and velocity of the center of mass. Initially, the dynamics of the robot system for preventing elderly fall are derived respectively from different elderly fall situations. Then, the elderly-assistant robot is controlled to stop at an appropriate position, and to minimize the tilt angle as small as possible to avoid overturned. The control system ensures that the position-velocity coordinates of the center of mass of the elderly reach the dynamic steadiness region which guarantees human balance. Finally, the results of the simulation and the experiments show that the proposed control method of the elderly-assistant robot can be effectively used to prevent the elderly fall during walking.

Deep learning-based 2D keypoint detection in alpine ski racing – A performance analysis of state-of-the-art algorithms applied to regular skiing and injury situations

ObjectivesIn this study, we examined the practicability of deep learning-based 2D keypoint detection applied to regular skiing and injury situations (i.e., out-of-balance situations and fall situations) on an alpine ski racing track. MethodsWe therefore created a regular skiing- and injury situation-specific dataset (hereinafter called "Injury Ski Dataset"), on which the state-of-the-art keypoint detection algorithms OpenPose, Mask-R-CNN, AlphaPose and DCPose were compared. The performance of each keypoint detector was evaluated by calculating the mean per joint position error (MPJPE) and the percentage of correct keypoints (PCK). Failure cases and common error patterns were further investigated by a visual analysis. ResultsWe observed the best results for regular skiing, with 81%–92% of all keypoints detected correctly at an MPJPE of 9 (2) to 14 (3) pixels. In injury situations, self-occlusions and rare poses became more likely, similar to occlusions due to snow spray and motion blur. As a result, the performance in out-of-balance situations decreased to 68%–80% (PCK), while in fall situations, only 35%–54% of all keypoints were detected correctly, with mean errors of 26–36 pixels. Among all algorithms, AlphaPose was the most robust and achieved the best results. ConclusionsPCK and MPJPE for regular skiing were in the range of manual annotation errors and can be considered low enough for further biomechanical analysis. For fall situations, keypoint detection should be further improved. Regarding the development of a deep learning tool for injury analysis in alpine skiing in the future, we propose to fine-tune a well-performing keypoint detector, such as AlphaPose, on a ski- and injury-specific dataset, such as ours.

Symmetrically Stacked Long Short-Term Memory Networks for Fall Event Recognition Using Compact Convolutional Neural Networks-Based Tracker

In recent years, the advancement of pattern recognition algorithms, specifically the deep learning-related techniques, have propelled a tremendous amount of researches in fall event recognition systems. It is important to detect a fall incident as early as possible, whereby a slight delay in providing immediate assistance can cause severe unrecoverable injuries. One of the main challenges in fall event recognition is the imbalanced training data between fall and no-fall events, where a real-life fall incident is a sporadic event that occurs infrequently. Most of the recent techniques produce a lot of false alarms, as it is hard to train them to cover a wide range of fall situations. Hence, this paper aims to detect the exact fall frame in a video sequence, as such it will not be dependent on the whole clip of the video sequence. Our proposed approach consists of a two-stage module where the first stage employs a compact convolutional neural network tracker to generate the object trajectory information. Features of interest will be sampled from the generated trajectory paths, which will be fed as the input to the second stage. The next stage network then models the temporal dependencies of the trajectory information using symmetrical Long Short-Term Memory (LSTM) architecture. This two-stage module is a novel approach as most of the techniques rely on the detection module rather than the tracking module. The simulation experiments were tested using Fall Detection Dataset (FDD). The proposed approach obtains an expected average overlap of 0.167, which is the best performance compared to Multi-Domain Network (MDNET) and Tree-structured Convolutional Neural Network (TCNN) trackers. Furthermore, the proposed 3-layers of stacked LSTM architecture also performs the best compared to the vanilla recurrent neural network and single-layer LSTM. This approach can be further improved if the tracker model is firstly pre-tuned in offline mode with respect to a specific type of object of interest, rather than a general object.

A Wearable Fall Detection System Based on Body Area Networks

Falls can have serious consequences for people, leading to restrictions in mobility or, in the worst case, to traumatic-based cases of death. To provide rapid assistance, a portable fall detection system has been developed that is capable of detecting fall situations and, if necessary, alerting emergency services without any user interaction. The prototype is designed to facilitate reliable fall detection and to classify several fall types and human activities. This solution represents a life-saving service for every person that will significantly improve assistance in the case of fall events, which are a part of daily life. Additionally, this approach facilitates independent system operation, since the system does not depend on sensor or network units located within a building structure. This article also introduces fall analysis. To guarantee functional safety, a hazard analysis method named system-theoretic accident model and processes (STAMP) is applied.

Implementation of a real-time fall detection system for elderly Korean farmers using an insole-integrated sensing device

Korea’s rural communities suffer from serious health issues from a rise in fall accidents due to an aging population. This study proposes a real-time fall detection system using an insole-integrated sensor device that detects fall accidents during farming activities and enables rapid responses to these emergencies. This study developed an insole-integrated fall detection device comprising a three-axial accelerometer and force sensitive resistor pressure sensors. Threshold values were set using data captured by the sensors to distinguish fall and emergency situations from other normal activities. In addition, a real-time monitoring system was implemented so that accident-related information may be automatically communicated to a third party or the monitoring center when a fall or emergency is recognized. A reliability test of the system showed 94.8% accuracy, 98.9% sensitivity, and 91.3% specificity. Furthermore, the detection rate for emergency conditions showed a high accuracy of 100%. Based on the results of this study, the fall detection system based on the insole-integrated sensor device may be implemented as a useful tool to address the issue of fall accidents involving elderly farmers in agricultural activities.

Falls in people with multiple sclerosis: experiences of 115 fall situations

Objectives: The aim was to describe falls and the perceived causes, experienced by people with multiple sclerosis shortly after falling. Design: A qualitative study using content analysis and quantitative data to illustrate where and why people report falls most commonly. Semi-structured telephone interviews were performed. Interviews were conducted shortly (0–10 days) after a fall. Subjects: In all, 67 informants who had reported at least one fall during the previous three-month period and who used a walking aid participated. Results: A total of 57 (85%) informants fell at least once during eight months resulting in 115 falls; 90 (78%) falls happened indoors, most commonly in the kitchen (n = 20; 17%) or bathroom (n = 16; 14%). Informants fell during everyday activities and walking aids had been used in more than a third of the reported falls. The falls were influenced of both intrinsic and extrinsic factors. Two categories emerged from the analysis: ‘activities when falling’ and ‘influencing factors’. The category contained three (basic activities of daily living, instrumental activities of daily living and leisure and work) and six (multiple sclerosis–related symptoms, fluctuating body symptoms, being distracted, losing body control, challenging surrounding and involvement of walking aid) subcategories, respectively. Conclusion: The majority of falls occurs indoors and in daily activities. Several factors interacted in fall situations and should be monitored and considered to reduce the gap between the person’s capacity and the environmental demands that cause fall risk. Fluctuation of bodily symptoms between and within a day is a variable not earlier targeted in multiple sclerosis fall risk research.

Accelerometer and Camera-Based Strategy for Improved Human Fall Detection.

In this paper, we address the problem of detecting human falls using anomaly detection. Detection and classification of falls are based on accelerometric data and variations in human silhouette shape. First, we use the exponentially weighted moving average (EWMA) monitoring scheme to detect a potential fall in the accelerometric data. We used an EWMA to identify features that correspond with a particular type of fall allowing us to classify falls. Only features corresponding with detected falls were used in the classification phase. A benefit of using a subset of the original data to design classification models minimizes training time and simplifies models. Based on features corresponding to detected falls, we used the support vector machine (SVM) algorithm to distinguish between true falls and fall-like events. We apply this strategy to the publicly available fall detection databases from the university of Rzeszow's. Results indicated that our strategy accurately detected and classified fall events, suggesting its potential application to early alert mechanisms in the event of fall situations and its capability for classification of detected falls. Comparison of the classification results using the EWMA-based SVM classifier method with those achieved using three commonly used machine learning classifiers, neural network, K-nearest neighbor and naïve Bayes, proved our model superior.

Fall Detection and Damage Reduction in Biped Humanoid Robots

The appropriate management of fall situations, i.e., fast instability detection, avoidance of unintentional falls, falling without damaging the body, fast recovery of the standing position after a fall-is an essential ability for biped humanoid robots. This issue is especially important for biped humanoid robots carrying out demanding movements such as walking on irregular surfaces, running, or playing a given sport. In this paper, we tackle the detection of instability, and the management of falls in biped humanoids using an integrated framework. In this framework, after instability is detected, a fall can be avoided, or at the very least, low-damage falling sequences can be triggered, depending on the degree of the detected instability. The proposed fall detection and fall avoidance methodologies have been validated in real-world experiments with biped humanoid robots (525 collision experiments were carried out). The obtained results show the robustness of the fall detection methodology. In addition, they show under which conditions falls can be avoided, and the reduction of the fall damage when a fall occurs. Besides, results also suggest that the proposed fall avoidance methodology still needs improvements. Moreover, the methodology is more suitable for small-sized and light-weight robots which are more mechanically robust to falling and can bear the required tests.

Falls in Hip Fracture Patients: Relation between Fall Situations and ADL Capability before Injury

Background: Most hip fractures are due to falls among frail elderly people. In order to formulate strategies to prevent falls leading to hip fracture, the characteristics of the falls of patients should be clarified in more detail. We investigate the fall situations of hip fracture patients. Methods: The fall conditions of eighty female and seventeen male hip fracture patients with an average age of 82.3 y.o. who were admitted to the hospital of the National Center for Geriatrics and Gerontology, Japan over a two-year period were investigated by questionnaire concerning 1) the time of falling, 2) the place of falling, 3) what they were doing at the moment of falling, and 4) the cause of falling. Their Barthel Index, representing patients’ ADL capability before injuries, was obtained as well, and the relation between falling situations and the Barthel Index score was analyzed. Results: The most frequent time, place, cause of the fall and activities at the time of the fall were in the evening (from 5 to 9 P.M.), in their own room, and involved their loss of balance and walking, respectively. Those who fell in the following situations had lower ADL capacity before injury: occurring at evening and night, either in the toilet or other places indoors and at falling, from losing their balance or slipping. Conclusion: Fall situations causing hip fractures differed considerably with the hip fracture patients’ capacities before injury. These differences must be taken into careful consideration in order to formulate effective strategies in hip fracture prevention.

Development of childhood fall motion database and browser based on behavior measurements

This paper describes the development of a fall motion database and a browser designed to facilitate investigations into fall-related injury risk. First, child-related daily activities were collected at a “sensor home”, which is a model of a normal living environment equipped with an embedded video-surveillance system and within which child test subjects were equipped with wearable acceleration-gyro sensors. As of this report, measurements have been conducted for 19 children (months age: mean=23.8, standard deviation=10.5), and data has been obtained on 105 fall incidents. During our research, falls were detected from the accumulated sensor data using a detection algorithm developed by the authors, and then video clips of detected falls were extracted from the recorded video streams automatically. The extracted video clips were then used for fall motion analysis. A computer vision (CV) algorithm, which was developed to automate fall motion analysis, facilitates accumulation of fall motion data into the abovementioned database, and the associated database browser allows users to perform conditional searches of fall data by inputting search conditions, such as child attributes and specific fall situations. Before this study, there was no database which contains child's actual fall motion data, and it has the potential to facilitate injury risk reduction related to falls in daily living environments.

Agreement of Fall Classifications Among Staff in U.S. Hospitals

Patient falls in hospitals are a performance measure endorsed by the National Quality Forum. Agreement of staff classifications of fall situations is not documented. The aims of this study were to (a) investigate how experts classify fall scenarios according to the fall definition of the National Quality Forum; (b) investigate how nursing staff classifies the same fall scenarios; and (c) assess the extent to which fall classifications differ among units, hospitals, and individuals. Twenty video scenarios of falls were embedded in an online video survey. A panel of 24 experts and 6,342 hospital staff members from 362 units in 170 U.S. hospitals were asked to classify fall and nonfall scenarios. Experts consisted of nurses, physicians, physical therapists, and statisticians. Hospital staff were registered nurses (78%), unlicensed staff (15%), and other staff (7%). Experts unambiguously classified 14 out of 20 scenarios according to the National Quality Forum fall definition, whereas hospital staff clearly classified 12 scenarios. Experts and hospital staff did not agree on 4 out of 20 scenarios. The sensitivity was 0.90, the specificity was 0.88, and the mean probability for classifying a scenario as a fall was 0.60. Results indicate that the National Quality Forum fall definition needs further refinement to classify all scenarios properly. Although variability between individuals indicates some potential for fall reporting training, variability between units and hospitals does not seem to affect fall reporting.

Development of Database of Children’s Fall Dynamics Using Daily Behavior Observing System

This paper describes the development of a fall database for biomechanical simulation. First, data on children’s daily activities were collected at a “sensor home,” which is a imitation daily living space. The sensor-based home comprises a video-surveillance system embedded into a daily-living environment and a wearable acceleration-gyro sensor. Falls were then detected from sensor data using a fall detection algorithm that we developed, and videos of detected falls were extracted from long-time recorded video. Extracted videos were used for fall motion analysis. A new Computer Vision (CV) algorithm was developed to automate fall motion analysis. Using the CV algorithm, fall motion data were accumulated into a database. The database allows a user to perform conditional searches for fall data by inputting search conditions, such as a child’s attributes, and fall situations.

A vulnerabilidade do idoso para as quedas: análise dos incidentes críticos

Exploratory and descriptive study that aims to characterize the situations of falls in aged people in a community of Programa de Saude da Familia (PSF) from Goiânia. The informations were collected by interview, using the technique of critical incident, and analyzed by qualitative form, also according to the technique and to the elements which characterize the critical incident: situation, behavior and consequence. There were 20 eldery between 60 and 86 years old, who had participated of the study. The fall situations occurred more frequently (70%) at domiciliary circles, demonstrating the vulnerability of the aged in the domicile. The most frequent behaviors were about independence for the self care, represented by the popular knowledge practices. Pain, wound and alterations in physical mobility prevailed as consequences of that. The results infotmations reveal the levels of critical requirement necessary to establish planning of action to prevent the falls in the aged people. It is up to the Family Health Team to evaluate ways to intervene in the community from the knowledge of this reality in order to reduce this event of significant proportion in aged population.

Simulation of hip fracture in sideways fall using a 3D finite element model of pelvis–femur–soft tissue complex with simplified representation of whole body

Hip fractures due to sideways falls are a worldwide health problem, especially among the elderly population. The objective of this study was to simulate a real life sideways fall leading to hip fracture. To achieve this a computed tomography (CT) scan based three-dimensional (3D) finite element (FE) model of the pelvis–femur complex was developed using a wide range of mechanical properties in the bone of the complex. For impact absorption through large deformation, surrounding soft tissue was also included in the FE model from CT scan data. To incorporate the inertia effect, the whole body was represented by a spring-mass-dashpot system. For trochanteric soft tissue thickness of 14 mm, body weight of 77.47 kg and average hip impact velocity of 3.17 m/s, this detailed FE model could approximately simulate a sideways fall configuration and examine femoral fracture situation. At the contact surface, the peak impact load was 8331 N. In spite of the presence of 14 mm thick trochanteric soft tissue, within the trochanteric zone the most compressive peak principal strain was 3.5% which exceeds ultimate compressive strain. The modeled trochanteric fracture was consistent with clinical findings and with the findings of previous studies. Further, this detailed FE model may be used to find the effect of trochanteric soft tissue thickness variations on peak impact force, peak strain in sideways fall, and to simulate automobile side impact and backward fall situations.

Restriction of foot supination by ankle braces in sudden fall situations

Objective. To investigate the restriction of foot supination in different shoe orthotic combinations for unexpected ankle turns. Design. A supination platform was used to experimentally induce sudden ankle turns. Background. This study expanded on previous investigations of the effectiveness of different ankle braces and the influence of shoe material on foot supination. Methods. For 21 male subjects pressure distribution data, achilles tendon angle, and supination velocity were collected. A comparison between a shoe with cut upper material and an intact shoe was done to show the influence of upper material on supination. The effectiveness of four different orthotics inside of an intact shoe were tested in comparison to the same shoe without ankle brace. Additionally, subjects rated their perceived supination movement of the foot. Results. Between experimental conditions no large differences for the pressure distribution patterns were found. Three of the ankle braces reduced foot supination as well as supination velocities. More than two times lower supination values were measured for a footwear condition in which the upper material was cut away. Already during the free-fall most ankle braces caused a substantial reduction of foot inversion. Conclusion. The experiments demonstrated the influence of the upper material of a shoe and different orthotics on foot supination. Biomechanical measurements are necessary because of limited capabilities of subjects in detecting subtalar angular motions.

Fall risk index for elderly patients based on number of chronic disabilities

The present study was designed to identify prospectively the individual chronic characteristics associated with falling among elderly persons and to test the hypothesis that risk of falling increases as the number of chronic disabilities increases. Seventy-nine consecutive admissions to three intermediate care facilities were evaluated. Twenty-five of the subjects became recurrent fallers. The nine risk factors included in the fall risk index were mobility score, morale score, mental status score, distant vision, hearing, postural blood pressure, results of back examination, postadmission medications, and admission activities of daily living score. A subject's fall risk score was the number of index factors present. The proportions of recurrent falters increased from 0 percent (0 of 30) in those with 0 to three risk factors, to 31 percent (11 of 35) in those with four to six factors, to 100 percent (14 of 14) in those with seven or more factors. Falling, at least among some elderly persons, appears to result from the accumulated effect of multiple specific disabilities. Some of these disabilities may be remediable. The mobility test, the best single predictor of recurrent falling, may be useful clinically because it is simple, recreates fall situations, and provides a dynamic, integrated assessment of mobility.