Bottom-up subspace clustering suggests a paradigm shift to prevent fall injuries

Understanding the Association of Older Adult Fall Risk Factors by Age and Sex Through Factor Analysis

Our aim was to identify latent factors underlying multiple observed risk factors for older adult falls and to examine their effects on falls by age and sex. We performed exploratory factor analysis on 13 risk factors in the Behavioral Risk Factor Surveillance System. We used log-linear regression models to measure the association between the identified factors and older adults reporting falls. We identified two underlying factors: physical and mental health limitations. These shared a 50% correlation. Physical health limitations were more strongly associated with falls among men (prevalence ratio = 1.68, 95% CI = 1.65-1.71) than women (prevalence ratio = 1.51, 95% CI = 1.49-1.54). As physical health limitations increased, men aged 65-74 had a greater association with falls compared with other age-sex subgroups. Our findings highlight the composite relationship between age, sex, and physical and mental health limitations in association with older adult falls, and support the evidence for individually tailored, multifactorial interventions.

Co-occurrences of fall-related factors in adults aged 60 to 85 years in the United States National Health and Nutrition Examination Survey

A broad set of factors are associated with falling (e.g., age, sex, physical activity, vision, health), but their co-occurrence is understudied. Our objectives were to quantify the number and pattern of co-occurring fall-related factors. Data were obtained from the U.S. National Health and Nutrition Examination Survey (N = 1,957, 60–85 years). Twenty fall-related factors were included (based on previous research), covering a wide range including cognitive, motor, sensory, health, and physical activity measures. The number and pattern of co-occurring fall-related factors were quantified with logistic regression and cluster analyses, respectively. Most participants (59%) had ≥4 fall-risk factors, and each additional risk factor increased the odds of reporting difficulty with falling by 1.28. The identified clusters included: (1) healthy, (2) cognitive and sensory impaired, and (3) health impaired. The mean number of co-occurring fall-related factors was 3.7, 3.8, and 7.2, for clusters 1, 2, and 3, respectively (p<0.001). These observations indicate that co-occurrence of multiple fall-risk factors was common in this national sample of U.S. older adults and the factors tended to aggregate into distinct clusters. The findings support the protective effect of physical activity on fall-risk, the association between gait speed and falls, and the detrimental effect of health-related factors on difficulty with falls (e.g., arthritis, prescription medications). Cluster analyses revealed a complex interplay between sex and BMI that may alter the role of BMI in the etiology of falls. Cluster analyses also revealed a large detrimental effect of health-related factors in cluster 3; it is important to extend current fall interventions (typically focused on balance, flexibility, strength, cognitive, fear factors) to include health-related interventions that target factors such as BMI and arthritis.

Identifying profiles of people aged 65 and over who fall at home and associated falling-induced injuries: The French ChuPADom study

BACKGROUND

Falls in older people are a major public health problem due to associated morbidity and mortality. Their origin is most often multifactorial.

OBJECTIVE, DESIGN AND SUBJECTS

The objective of the present study, called ChuPaDom, was to identify patterns or profiles of home fallers (HF) - understood here as people who fall in their place of residence - among a sample of persons aged 65 and over who were hospitalized after a fall, and to study their association with falling-induced injuries using data from the ChuPAdom study.

METHODS

Multiple correspondence analysis and Hierarchical Clustering on Principal Components were performed. Multivariate logistic regression was used to test for associations between HF profiles and injuries.

RESULTS

The sample comprised 1467 patients (69% female, average age = 84.5 years). Five profiles were identified: youngest seniors who took risks and fell from a raised height, youngest seniors with specific health problems who fell down a stairs, autonomous seniors who fell because they lost their balance or fell from their own height, dependent seniors who fell during low-intensity activities, very old seniors for whom missing data were frequent. Fractures were more frequent among the first profile than the last two profiles constituted with more dependent individuals (p < 0.001).

CONCLUSION

These results highlight the heterogeneity of the circumstances in which older people fall. A greater understanding of these circumstances is needed to implement targeted prevention actions.

Risk Profiles for Injurious Falls in People Over 60: A Population-Based Cohort Study

Background

Although falls in older adults are related to multiple risk factors, these factors have commonly been studied individually. We aimed to identify risk profiles for injurious falls in older adults by detecting clusters of established risk factors and quantifying their impact on fall risk.

Methods

Participants were 2,566 people, aged 60 years and older, from the population-based Swedish National Study on Aging and Care in Kungsholmen. Injurious falls was defined as hospitalization for or receipt of outpatient care because a fall. Cluster analysis was used to identify aggregation of possible risk factors including chronic diseases, fall-risk increasing drugs (FRIDs), physical and cognitive impairments, and lifestyle-related factors. Associations between the clusters and injurious falls over 3, 5, and 10 years were estimated using flexible parametric survival models.

Results

Five clusters were identified including: a "healthy", a "well-functioning with multimorbidity", a "well-functioning, with multimorbidity and high FRID consumption", a "physically and cognitively impaired", and a "disabled" cluster. The risk of injurious falls for all groups was significantly higher than for the first cluster of healthy individuals in the reference category. Hazard ratios (95% confidence intervals) ranged from 1.71 (1.02-2.66) for the second cluster to 12.67 (7.38-21.75) for the last cluster over 3 years of follow-up. The highest risk was observed in the last two clusters with high burden of physical and cognitive impairments.

Conclusion

Risk factors for injurious fall tend to aggregate, representing different levels of risk for falls. Our findings can be useful to tailor and prioritize clinical and public health interventions.

Walking-adaptability assessments with the Interactive Walkway: Between-systems agreement and sensitivity to task and subject variations

The ability to adapt walking to environmental circumstances is an important aspect of walking, yet difficult to assess. The Interactive Walkway was developed to assess walking adaptability by augmenting a multi-Kinect-v2 10-m walkway with gait-dependent visual context (stepping targets, obstacles) using real-time processed markerless full-body kinematics. In this study we determined Interactive Walkway's usability for walking-adaptability assessments in terms of between-systems agreement and sensitivity to task and subject variations. Under varying task constraints, 21 healthy subjects performed obstacle-avoidance, sudden-stops-and-starts and goal-directed-stepping tasks. Various continuous walking-adaptability outcome measures were concurrently determined with the Interactive Walkway and a gold-standard motion-registration system: available response time, obstacle-avoidance and sudden-stop margins, step length, stepping accuracy and walking speed. The same holds for dichotomous classifications of success and failure for obstacle-avoidance and sudden-stops tasks and performed short-stride versus long-stride obstacle-avoidance strategies. Continuous walking-adaptability outcome measures generally agreed well between systems (high intraclass correlation coefficients for absolute agreement, low biases and narrow limits of agreement) and were highly sensitive to task and subject variations. Success and failure ratings varied with available response times and obstacle types and agreed between systems for 85-96% of the trials while obstacle-avoidance strategies were always classified correctly. We conclude that Interactive Walkway walking-adaptability outcome measures are reliable and sensitive to task and subject variations, even in high-functioning subjects. We therefore deem Interactive Walkway walking-adaptability assessments usable for obtaining an objective and more task-specific examination of one's ability to walk, which may be feasible for both high-functioning and fragile populations since walking adaptability can be assessed at various levels of difficulty.

Differences Between Gait on Stairs and Flat Surfaces in Relation to Fall Risk and Future Falls

We used body-worn inertial sensors to quantify differences in semi-free-living gait between stairs and on normal flat ground in older adults, and investigated the utility of assessing gait on these terrains for predicting the occurrence of multiple falls. Eighty-two community-dwelling older adults wore two inertial sensors, on the lower back and the right ankle, during several bouts of walking on flat surfaces and up and down stairs, in between rests and activities of daily living. Derived from the vertical acceleration at the lower back, step rate was calculated from the signal's fundamental frequency. Step rate variability was the width of this fundamental frequency peak from the signal's power spectral density. Movement vigor was calculated at both body locations from the signal variance. Partial Spearman correlations between gait parameters and physiological fall risk factors (components from the Physiological Profile Assessment) were calculated while controlling for age and gender. Overall, anteroposterior vigor at the lower back in stair descent was lower in subjects with longer reaction times. Older adults walked more slowly on stairs, but they were not significantly slower on flat surfaces. Using logistic regression, faster step rate in stair descent was associated with multiple prospective falls over 12 months. No significant associations were shown from gait parameters derived during walking upstairs or on flat surfaces. These results suggest that stair descent gait may provide more insight into fall risk than regular walking and stair ascent, and that further sensor-based investigation into unsupervised gait on different terrains would be valuable.

Disentangling the health benefits of walking from increased exposure to falls in older people using remote gait monitoring and multi-dimensional analysis

Falls and physical deconditioning are two major health problems for older people. Recent advances in remote physiological monitoring provide new opportunities to investigate why walking exercise, with its many health benefits, can both increase and decrease fall rates in older people. In this paper we combine remote wearable device monitoring of daily gait with non-linear multi-dimensional pattern recognition analysis; to disentangle the complex associations between walking, health and fall rates. One week of activities of daily living (ADL) were recorded with a wearable device in 96 independent living older people prior to completing 6 months of exergaming interventions. Using the wearable device data; the quantity, intensity, variability and distribution of daily walking patterns were assessed. At baseline, clinical assessments of health, falls, sensorimotor and physiological fall risks were completed. At 6 months, fall rates, sensorimotor and physiological fall risks were re-assessed. A non-linear multi-dimensional analysis was conducted to identify risk-groups according to their daily walking patterns. Four distinct risk-groups were identified: The Impaired (93% fallers), Restrained (8% fallers), Active (50% fallers) and Athletic (4% fallers). Walking was strongly associated with multiple health benefits and protective of falls for the top performing Athletic risk-group. However, in the middle of the spectrum, the Active risk-group, who were more active, younger and healthier were 6.25 times more likely to be fallers than their Restrained counterparts. Remote monitoring of daily walking patterns may provide a new way to distinguish Impaired people at risk of falling because of frailty from Active people at risk of falling from greater exposure to situations were falls could occur, but further validation is required. Wearable device risk-profiling could help in developing more personalised interventions for older people seeking the health benefits of walking without increasing their risk of falls.

Daily-Life Gait Quality as Predictor of Falls in Older People: A 1-Year Prospective Cohort Study

Falls can have devastating consequences for older people. We determined the relationship between the likelihood of fall incidents and daily-life behavior. We used wearable sensors to assess habitual physical activity and daily-life gait quality (in terms of e.g. stability, variability, smoothness and symmetry), and determined their predictive ability for time-to-first-and-second-falls. 319 older people wore a trunk accelerometer (Dynaport MoveMonitor, McRoberts) during one week. Participants further completed questionnaires and performed grip strength and trail making tests to identify risk factors for falls. Their prospective fall incidence was followed up for six to twelve months. We determined interrelations between commonly used gait characteristics to gain insight in their interpretation and determined their association with time-to-falls. For all data -including questionnaires and tests- we determined the corresponding principal components and studied their predictive ability for falls. We showed that gait characteristics of walking speed, stride length, stride frequency, intensity, variability, smoothness, symmetry and complexity were often moderately to highly correlated (r > 0.4). We further showed that these characteristics were predictive of falls. Principal components dominated by history of falls, alcohol consumption, gait quality and muscle strength proved predictive for time-to-fall. The cross-validated prediction models had adequate to high accuracy (time dependent AUC of 0.66–0.72 for time-to-first-fall and 0.69–0.76 for -second-fall). Daily-life gait quality obtained from a single accelerometer on the trunk is predictive for falls. These findings confirm that ambulant measurements of daily behavior contribute substantially to the identification of elderly at (high) risk of falling.

Eight-Week Remote Monitoring Using a Freely Worn Device Reveals Unstable Gait Patterns in Older Fallers

OBJECTIVES

Develop algorithms to detect gait impairments remotely using data from freely worn devices during long-term monitoring. Identify statistical models that describe how gait performances are distributed over several weeks. Determine the data window required to reliably assess an increased propensity for falling.

METHODS

1085 days of walking data were collected from eighteen independent-living older people (mean age 83 years) using a freely worn pendant sensor (housing a triaxial accelerometer and pressure sensor). Statistical distributions from several accelerometer-derived gait features (encompassing quantity, exposure, intensity, and quality) were compared for those with and without a history of falling.

RESULTS

Participants completed more short walks relative to long walks, as approximated by a power law. Walks less than 13.1 s comprised 50% of exposure to walking-related falls. Daily-life cadence was bimodal and step-time variability followed a log-normal distribution. Fallers took significantly fewer steps per walk and had relatively more exposure from short walks and greater mode of step-time variability.

CONCLUSIONS

Using a freely worn device and wavelet-based analysis tools allowed long-term monitoring of walks greater than or equal to three steps. In older people, short walks constitute a large proportion of exposure to falls. To identify fallers, mode of variability may be a better measure of central tendency than mean of variability. A week's monitoring is sufficient to reliably assess the long-term propensity for falling.

SIGNIFICANCE

Statistical distributions of gait performances provide a reference for future wearable device development and research into the complex relationships between daily-life walking patterns, morbidity, and falls.

New Methods to Monitor Stair Ascents Using a Wearable Pendant Device Reveal How Behavior, Fear, and Frailty Influence Falls in Octogenarians

GOALS

To investigate if the stair negotiation by older people during activities of daily life (ADL) can be accurately identified using a freely worn pendant device. To investigate how usual stair-ascent performances during ADL relate to clinical assessments and prospective falls.

METHODS

ADL were recorded for 30 min by 52 community-dwelling older people (83 ± 4 years) using a small pendant device. Classification accuracy was assessed using annotated video and four-fold cross validation. Correlations between sensor-derived stair-ascent features (comprising intensity, variability, and stability) and a battery of clinical tests (comprising physiological, psychological, health, and follow-up falls) were investigated.

RESULTS

Accurate identification of stair events (99.8%, Kappa 0.92) was possible in both "frail" and "athletic" participants by scaling the barometer threshold to stair cadences. Cautious double-stepping strategy could be identified remotely. Stair-ascent performance was correlated with ascent strategy ( r = -0.67), age ( r = -0.44), concern about falling ( r = -0.43 ), fall-risk scores ( r = -0.41), processing speed ( r = -0.38), and contrast sensitivity ( r = 0.32). Follow-up falls were correlated with ascent stability ( r = -0.35).

CONCLUSION

Remote analysis of stair ascents is feasible. In our healthy older people, outcomes appeared more related to mental rather than physiological factors. The ascent strategies we observed in some older people may have reflected an appropriate behavioral response to increased concerns about falling.

SIGNIFICANCE

Given acceptance of wearable devices is increasing; reduced functional performance and altered strategies for undertaking ADL could soon be routinely tracked to augment health care.

Wearable pendant device monitoring using new wavelet-based methods shows daily life and laboratory gaits are different

Morbidity and falls are problematic for older people. Wearable devices are increasingly used to monitor daily activities. However, sensors often require rigid attachment to specific locations and shuffling or quiet standing may be confused with walking. Furthermore, it is unclear whether clinical gait assessments are correlated with how older people usually walk during daily life. Wavelet transformations of accelerometer and barometer data from a pendant device worn inside or outside clothing were used to identify walking (excluding shuffling or standing) by 51 older people (83 ± 4 years) during 25 min of 'free-living' activities. Accuracy was validated against annotated video. Training and testing were separated. Activities were only loosely structured including noisy data preceding pendant wearing. An electronic walkway was used for laboratory comparisons. Walking was classified (accuracy ≥97 %) with low false-positive errors (≤1.9%, κ ≥ 0.90). Median free-living cadence was lower than laboratory-assessed cadence (101 vs. 110 steps/min, p < 0.001) but correlated (r = 0.69). Free-living step time variability was significantly higher and uncorrelated with laboratory-assessed variability unless detrended. Remote gait impairment monitoring using wearable devices is feasible providing new ways to investigate morbidity and falls risk. Laboratory-assessed gait performances are correlated with free-living walks, but likely reflect the individual's 'best' performance.