RESUMEN

There is no effective fall risk screening tool for the elderly that can be integrated into clinical practice. Developing a system that can be easily used in primary care services is a current need. Current studies focus on the use of multiple sensors or activities to achieve higher accuracy. However, multiple sensors and activities reduce the availability of these systems. This study aims to develop a system to perform fall prediction for the elderly by using signals recorded from a single sensor during a short-term activity. A total of 168 features in the time and frequency domains were created using acceleration signals obtained from 71 elderly people. The features were weighted based on the ReliefF algorithm, and the artificial neural networks model was developed using the most important features. The best classification result was obtained using the 17 most important features of those weighted for K = 20 nearest neighbors. The highest accuracy was 82.2% (82.9% Sensitivity, 81.6% Specificity). The partially high accuracy obtained in our study shows that falling can be detected early with a sensor and a simple activity by determining the right features and can be easily applied in the assessment of the elderly during routine follow-ups.

RESUMEN

BACKGROUND: Early detection of patients at risk of falling is crucial. This study was designed to develop and internally validate a novel risk score to classify patients at risk of falls. METHODS: A total of 334 older people from a fall clinic in a medical center were selected. Least absolute shrinkage and selection operator (LASSO) regression was used to minimize the potential concatenation of variables measured from the same patient and the overfitting of variables. A logistic regression model for 1-year fall prediction was developed for the entire dataset using newly identified relevant variables. Model performance was evaluated using the bootstrap method, which included measures of overall predictive performance, discrimination, and calibration. To streamline the assessment process, a scoring system for predicting 1-year fall risk was created. RESULTS: We developed a new model for predicting 1-year falls, which included the FRQ-Q1, FRQ-Q3, and single-leg standing time (left foot). After internal validation, the model showed good discrimination (C statistic, 0.803 [95% CI 0.749-0.857]) and overall accuracy (Brier score, 0.146). Compared to another model that used the total FRQ score instead, the new model showed better continuous net reclassification improvement (NRI) [0.468 (0.314-0.622), P < 0.01], categorical NRI [0.507 (0.291-0.724), P < 0.01; cutoff: 0.200-0.800], and integrated discrimination [0.205 (0.147-0.262), P < 0.01]. The variables in the new model were subsequently incorporated into a risk score. The discriminatory ability of the scoring system was similar (C statistic, 0.809; 95% CI, 0.756-0.861; optimism-corrected C statistic, 0.808) to that of the logistic regression model at internal bootstrap validation. CONCLUSIONS: This study resulted in the development and internal verification of a scoring system to classify 334 patients at risk for falls. The newly developed score demonstrated greater accuracy in predicting falls in elderly people than did the Timed Up and Go test and the 30-Second Chair Sit-Stand test. Additionally, the scale demonstrated superior clinical validity for identifying fall risk.

Asunto(s)

RESUMEN

When legged robots perform complex tasks in unstructured environments, falls are inevitable due to unknown external disturbances. However, current research mainly focuses on the locomotion control of legged robots without falling. This paper proposes a comprehensive decision-making and control framework to address the falling over of quadruped robots. First, a capturability-based fall prediction algorithm is derived for planar single-contact and 3D multi-contact locomotion with a predefined gait sequence. For safe fall control, a novel contact-implicit trajectory optimization method is proposed to generate both state and input trajectories and contact mode sequences. Specifically, incorporating uncertainty into the system and terrain models enables mitigating the non-smoothness of contact dynamics while improving the robustness of the resulting trajectories. Furthermore, a model-free deep reinforcement learning-based approach is presented to achieve fall recovery after the robot completes a fall. Experimental results demonstrate that the proposed fall prediction algorithm accurately predicts robot falls with up to 95% accuracy approximately 395ms in advance. Compared to classical locomotion controllers, which often struggle to maintain balance under significant pushes or terrain perturbations, the presented framework can autonomously switch to the fall controller approximately 0.06s after the perturbation, effectively preventing falls or achieving recovery with a threefold reduction in touchdown impact velocity. These findings highlight the effectiveness of the proposed framework in enhancing the stability and safety of legged robots in unstructured environments.

RESUMEN

BACKGROUND: Identifying patients at high risk of falling is crucial in implementing effective fall prevention programs. While the integration of information systems is becoming more widespread in the healthcare industry, it poses a significant challenge in analysing vast amounts of data to identify factors that could enhance patient safety. OBJECTIVE: To determine fall-associated factors and develop high-performance prediction tools for at-risk patients in acute and sub-acute care services in Australia. METHODS: A retrospective study of 672,400 patients admitted to acute and sub-acute care services within a large metropolitan tertiary health service in Victoria, Australia, between January 1, 2019, and December 31, 2021. Data were obtained from four sources: the Department of Health Victorian Admitted Episodes Dataset, RiskManTM, electronic health records, and the health workforce dataset. Machine learning techniques, including Random Forest and Deep Neural Network models, were used to analyse the data, predict patient falls, and identify the most important risk factors for falls in this population. Model performance was evaluated using accuracy, F1-score, precision, recall, specificity, Matthew's correlation coefficient, and the area under the receiver operating characteristic curve (AUC). RESULTS: The deep neural network and random forest models were highly accurate in predicting hospital patient falls. The deep neural network model achieved an accuracy of 0.988 and a specificity of 0.999, while the RF achieved an accuracy of 0.989 and a specificity of 1.000. The top 20 variables impacting falls were compared across both models, and 12 common factors were identified. These factors can be broadly classified into three categories: patient-related factors, staffing-related factors, and admission-related factors. Although not all factors are modifiable, they must be considered when planning fall prevention interventions. CONCLUSION: The study demonstrated machine learning's potential to predict falls and identify key risk factors. Further validation across diverse populations and settings is essential for broader applicability.

Asunto(s)

RESUMEN

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.

Asunto(s)

RESUMEN

Objectives: To predict falls by adding an adherence assessment to a static balance ability assessment, and to evaluate fall prediction accuracy. Methods: This study included 416 patients who were admitted to a 45-bed convalescent rehabilitation ward over a 2-year period. The patients were assessed at the time of admission using the Standing Test for Imbalance and Disequilibrium (SIDE) and three additional, newly developed adherence items. Patients were divided into two groups: a group that experienced falls (fall group) and a group that did not experience falls (non-fall group) within 14 days of admission. The sensitivity and specificity of the assessment items for predicting falls were calculated. Results: Sensitivity was 0.86 and specificity was 0.42 when the cutoff was between SIDE levels 0-2a and 2b-4. Combining balance assessment using the SIDE with the memory and instruction adherence items improved fall prediction accuracy such that the sensitivity was 0.75 and the specificity was 0.64. Conclusions: Our analysis suggested that adherence assessment can improve fall risk prediction accuracy.

RESUMEN

BACKGROUND:During stair walking,different muscles work in concert and compensate for each other,and it is unclear whether weakened muscle strength actually affects stair fall risk in older adults.Real-time electromyographic signals from older adults during stair walking are used to reflect high fall risk in older adults during stair walking,which may further improve the accuracy of prediction methods. OBJECTIVE:To investigate the effects of aging on lower limb muscle activation in older adults during stair walking and to analyze the relationship between their muscle activation characteristics and stair fall risk. METHODS:Subjects were divided by age into an older group(n=19)and a younger group(n=18)group and were asked to walk on a 10-step staircase at a natural speed,incorporating surface electromyography acquisition technology,to capture surface electromyography signals during stair walking and calculate the root mean square(RMS)to analyze differences in muscle activation levels.Logistic regression analysis was utilized to establish a predictive model for stair fall risk in older adults by incorporating the lateral femoral and gastrocnemius muscle RMS.The discrimination of the model was evaluated by the receiver operating characteristic curve and area under the curve,and the fit of the model was evaluated using the Hosmer-Lemeshow test RESULTS AND CONCLUSION:Activation of the rectus femoris(Z=-3.464,P=0.001;t=3.379,P=0.002)and lateral gastrocnemius muscle(Z=-2.978,P=0.003;Z=-3.555,P<0.001)was higher in older adults than in younger adults when walking up and down stairs.Activation of the anterior tibialis(Z=-2.350,P=0.019)and medial(Z=2.321,P=0.020)and lateral(t=3.158,P=0.004)gastrocnemius muscles was higher in older adults when ascending stairs than descending stairs.Older adults at risk for falls had less activation of the lateral femoral muscle(Z=-2.613,P=0.009),medial gastrocnemius muscle(Z=-2.286,P=0.022)when walking upstairs,and lateral femoral muscle(Z=-2.368,P=0.018)when walking downstairs than did older adults not at risk for falls.The predictive ability,goodness of fit,and discrimination of the stair fall prediction model for older adults based on surface electromyography were good(P-value of 0.010 for the Omnibus test of the model coefficients,P-value of 0.214 for the Hosmer-Lemeshow test,and the area of the curve of the upper staircase lateral femoral muscle=0.856,P=0.009).(5)The model was modeled with a cut-off value of 38.64 for the upper staircase lateral femoral muscle RMS value and there was a 0.952-fold increase in the risk of staircase falls for each unit decrease in the upper staircase lateral femoral muscle RMS in older adults.

RESUMEN

Background: In recent years, researchers have been advocating for the integration of ambulatory gait monitoring as a complementary approach to traditional fall risk assessments. However, current research relies on dedicated inertial sensors that are fixed on a specific body part. This limitation impacts the acceptance and adoption of such devices. Objective: Our study objective is twofold: (1) to propose a set of step-based fall risk parameters that can be obtained independently of the sensor placement by using a ubiquitous step detection method and (2) to evaluate their association with prospective falls. Methods: A reanalysis was conducted on the 1-week ambulatory inertial data from the StandingTall study, which was originally described by Delbaere et al. The data were from 301 community-dwelling older people and contained fall occurrences over a 12-month follow-up period. Using the ubiquitous and robust step detection method Smartstep, which is agnostic to sensor placement, a range of step-based fall risk parameters can be calculated based on walking bouts of 200 steps. These parameters are known to describe different dimensions of gait (ie, variability, complexity, intensity, and quantity). First, the correlation between parameters was studied. Then, the number of parameters was reduced through stepwise backward elimination. Finally, the association of parameters with prospective falls was assessed through a negative binomial regression model using the area under the curve metric. Results: The built model had an area under the curve of 0.69, which is comparable to models exclusively built on fixed sensor placement. A higher fall risk was noted with higher gait variability (coefficient of variance of stride time), intensity (cadence), and quantity (number of steps) and lower gait complexity (sample entropy and fractal exponent). Conclusions: These findings highlight the potential of our method for comprehensive and accurate fall risk assessments, independent of sensor placement. This approach has promising implications for ambulatory gait monitoring and fall risk monitoring using consumer-grade devices.

RESUMEN

BACKGROUND: Falls in older people are common and morbid. Prediction models can help identifying individuals at higher fall risk. Electronic health records (EHR) offer an opportunity to develop automated prediction tools that may help to identify fall-prone individuals and lower clinical workload. However, existing models primarily utilise structured EHR data and neglect information in unstructured data. Using machine learning and natural language processing (NLP), we aimed to examine the predictive performance provided by unstructured clinical notes, and their incremental performance over structured data to predict falls. METHODS: We used primary care EHR data of people aged 65 or over. We developed three logistic regression models using the least absolute shrinkage and selection operator: one using structured clinical variables (Baseline), one with topics extracted from unstructured clinical notes (Topic-based) and one by adding clinical variables to the extracted topics (Combi). Model performance was assessed in terms of discrimination using the area under the receiver operating characteristic curve (AUC), and calibration by calibration plots. We used 10-fold cross-validation to validate the approach. RESULTS: Data of 35,357 individuals were analysed, of which 4,734 experienced falls. Our NLP topic modelling technique discovered 151 topics from the unstructured clinical notes. AUCs and 95% confidence intervals of the Baseline, Topic-based and Combi models were 0.709 (0.700-0.719), 0.685 (0.676-0.694) and 0.718 (0.708-0.727), respectively. All the models showed good calibration. CONCLUSIONS: Unstructured clinical notes are an additional viable data source to develop and improve prediction models for falls compared to traditional prediction models, but the clinical relevance remains limited.

Asunto(s)

RESUMEN

In Industry 4.0 scenarios, wearable sensing allows the development of monitoring solutions for workers' risk prevention. Current approaches aim to identify the presence of a risky event, such as falls, when it has already occurred. However, there is a need to develop methods capable of identifying the presence of a risk condition in order to prevent the occurrence of the damage itself. The measurement of vital and non-vital physiological parameters enables the worker's complex state estimation to identify risk conditions preventing falls, slips and fainting, as a result of physical overexertion and heat stress exposure. This paper aims at investigating classification approaches to identify risk conditions with respect to normal physical activity by exploiting physiological measurements in different conditions of physical exertion and heat stress. Moreover, the role played in the risk identification by specific sensors and features was investigated. The obtained results evidenced that k-Nearest Neighbors is the best performing algorithm in all the experimental conditions exploiting only information coming from cardiorespiratory monitoring (mean accuracy 88.7±7.3% for the model trained with max(HR), std(RR) and std(HR)).

Asunto(s)

RESUMEN

PURPOSE: Aging impacts muscle strength and elasticity, which in turn influence dynamic balance, walking speed, and physical performance. We aimed to evaluate the relationship between the elasticity of leg muscles and incidence of falls in older adults. METHODS: We conducted a prospective cohort analysis with outpatients from a geriatric clinic. Any history of falls in the past year was recorded. Timed up and go test, muscle thickness, and handgrip strength tests were performed. Elasticities of the gastrocnemius medialis (GM) and rectus femoris (RF) muscles were evaluated using shear wave elastography. Patients self-recorded their falls, and additional phone calls were made to them each month for 6 months. RESULTS: The median age of the patients (n = 55) was 72 years (66-86); and 72% were women. The GM showed significantly lower elasticity in patients with history of falls in the past year than in those without it (8.08 kPa [3.90-16.17] vs. 9.70 kPa [4.99-20.95]; p = 0.028). A similar negative correlation between GM and fall incidence was noted among those with additional falls during the follow-up period (6.96 kPa [3.90-12.41] vs. 9.13 kPa [4.99-20.95]; p = 0.019). GM elasticity was significantly correlated with the timed up and go test score (r = - 0.612, p < 0.001), handgrip strength (r = 0.384, p = 0.015), and muscle thickness (r = 0.232, p = 0.049). No such associations were observed for the RF muscles. CONCLUSION: GM muscle elasticity is associated with alterations in muscle structure that may lead to falls in older adults. Therefore, muscle elasticity may be a fall predictor in older adults.

Asunto(s)

RESUMEN

This cohort study aims to evaluate the predictive validity of multimodal clinical assessment and quantitative measures of in- and off-laboratory mobility for fall-risk estimation in patients with cerebellar ataxia (CA).Occurrence, severity, and consequences of falling were prospectively assessed for 6 months in 93 patients with hereditary (N = 36) and sporadic or secondary (N = 57) forms of CA and 63 healthy controls. Participants completed a multimodal clinical and functional fall risk assessment, in-laboratory gait examination, and a 2-week inertial sensor-based daily mobility monitoring. Multivariate logistic regression analyses were performed to evaluate the predictive capacity of all clinical and in- and off-laboratory mobility measures with respect to fall (1) status (non-faller vs. faller), (2) frequency (occasional vs. frequent falls), and (3) severity (benign vs. injurious fall) of patients. 64% of patients experienced one or recurrent falls and 65% of these severe fall-related injuries during prospective assessment. Mobility impairments in patients corresponded to a mild-to-moderate ataxic gait disorder. Patients' fall status and frequency could be reliably predicted (78% and 81% accuracy, respectively), primarily based on their retrospective fall status. Clinical scoring of ataxic symptoms and in- and off-laboratory gait and mobility measures improved classification and provided unique information for the prediction of fall severity (84% accuracy).These results encourage a stepwise approach for fall risk assessment in patients with CA: fall history-taking readily and reliably informs the clinician about patients' general fall risk. Clinical scoring and instrument-based mobility measures provide further in-depth information on the risk of recurrent and injurious falling.

Asunto(s)

RESUMEN

Nowadays, it becomes of paramount societal importance to support many frail-prone groups in our society (elderly, patients with neurodegenerative diseases, etc.) to remain socially and physically active, maintain their quality of life, and avoid their loss of autonomy. Once older people enter the prefrail stage, they are already likely to experience falls whose consequences may accelerate the deterioration of their quality of life (injuries, fear of falling, reduction of physical activity). In that context, detecting frailty and high risk of fall at an early stage is the first line of defense against the detrimental consequences of fall. The second line of defense would be to develop original protocols to detect future fallers before any fall occur. This paper briefly summarizes the current advancements and perspectives that may arise from the combination of affordable and easy-to-use non-wearable systems (force platforms, 3D tracking motion systems), wearable systems (accelerometers, gyroscopes, inertial measurement units-IMUs) with appropriate machine learning analytics, as well as the efforts to address these challenges.

Asunto(s)

RESUMEN

BACKGROUND: To review the validated instruments that assess gait, balance, and functional mobility to predict falls in older adults across different settings. METHODS: Umbrella review of narrative- and systematic reviews with or without meta-analyses of all study types. Reviews that focused on older adults in any settings and included validated instruments assessing gait, balance, and functional mobility were included. Medical and allied health professional databases (MEDLINE, PsychINFO, Embase, and Cochrane) were searched from inception to April 2022. Two reviewers undertook title, abstract, and full text screening independently. Review quality was assessed through the Risk of Bias Assessment Tool for Systematic Reviews (ROBIS). Data extraction was completed in duplicate using a standardised spreadsheet and a narrative synthesis presented for each assessment tool. RESULTS: Among 2736 articles initially identified, 31 reviews were included; 11 were meta-analyses. Reviews were primarily of low quality, thus at high risk of potential bias. The most frequently reported assessments were: Timed Up and Go, Berg Balance Scale, gait speed, dual task assessments, single leg stance, functional Reach Test, tandem gait and stance and the chair stand test. Findings on the predictive ability of these tests were inconsistent across the reviews. CONCLUSIONS: In conclusion, we found that no single gait, balance or functional mobility assessment in isolation can be used to predict fall risk in older adults with high certainty. Moderate evidence suggests gait speed can be useful in predicting falls and might be included as part of a comprehensive evaluation for older adults.

Asunto(s)

RESUMEN

Standardized fall risk scores have not proven to reliably predict falls in clinical settings. Machine Learning offers the potential to increase the accuracy of such predictions, possibly vastly improving care for patients at high fall risks. We developed a boosting algorithm to predict both recurrent falls and the severity of fall injuries. The model was trained on a dataset including extensive information on fall events of patients who had been admitted to Charité - Universitätsmedizin Berlin between August 2016 and July 2020. The data were recorded according to the German expert standard for fall documentation. Predictive power scores were calculated to define optimal feature sets. With an accuracy of 74% for recurrent falls and 86% for injury severity, boosting demonstrated the best overall predictive performance of all models assessed. Given that our data contain initially rated risk scores, our results demonstrate that well trained ML algorithms possibly provide tools to substantially reduce fall risks in clinical care settings.

Asunto(s)

RESUMEN

Older adults aged 65 and above are at higher risk of falls. Predicting fall risk early can provide caregivers time to provide interventions, which could reduce the risk, potentially avoiding a possible fall. In this paper, we present an analysis of 6-month fall risk prediction in older adults using geriatric assessments, GAITRite measurements, and fall history. The geriatric assessments included were Activities of Daily Living (ADL), Instrumental Activities of Daily Living (IADL), Mini-Mental State Examination (MMSE), Geriatric Depression Scale (GDS), and Short Form 12 (SF12). These geriatric assessments are collected by staff nurses regularly in senior care facilities. From the GAITRite assessments on the residents, we included the Functional Ambulatory Profile (FAP) scores and gait speed to predict fall risk. We used the SHAP (SHapley Additive exPlanations) approach to explain our model predictions to understand which predictor variables contributed to increase or decrease the fall risk for an individual prediction. In case of a high fall risk prediction, predictor variables that contributed the most to elevate the risk could be further examined by the health providers for more personalized health interventions. We used the geriatric assessments, GAITRite measurements, and fall history data collected from 92 older adult residents (age = 86.2 ± 6.4, female = 57) to train machine learning models to predict 6-month fall risk. Our models predicted a 6-month fall with an AUC of 0.80 (95% CI of 0.76-0.85), sensitivity of 0.82 (95% CI of 0.74-0.89), specificity of 0.72 (95% CI of 0.67-0.76), F1 score of 0.76 (95% CI of 0.72-0.79), and accuracy of 0.75 (95% CI of 0.72-0.79). These results show that our early fall risk prediction method performs well in identifying residents who are at higher fall risk, which offers care providers and family members valuable time to perform preventive actions.

RESUMEN

BACKGROUND: Short-term fall prediction models that use electronic health records (EHRs) may enable the implementation of dynamic care practices that specifically address changes in individualized fall risk within senior care facilities. OBJECTIVE: The aim of this study is to implement machine learning (ML) algorithms that use EHR data to predict a 3-month fall risk in residents from a variety of senior care facilities providing different levels of care. METHODS: This retrospective study obtained EHR data (2007-2021) from Juniper Communities' proprietary database of 2785 individuals primarily residing in skilled nursing facilities, independent living facilities, and assisted living facilities across the United States. We assessed the performance of 3 ML-based fall prediction models and the Juniper Communities' fall risk assessment. Additional analyses were conducted to examine how changes in the input features, training data sets, and prediction windows affected the performance of these models. RESULTS: The Extreme Gradient Boosting model exhibited the highest performance, with an area under the receiver operating characteristic curve of 0.846 (95% CI 0.794-0.894), specificity of 0.848, diagnostic odds ratio of 13.40, and sensitivity of 0.706, while achieving the best trade-off in balancing true positive and negative rates. The number of active medications was the most significant feature associated with fall risk, followed by a resident's number of active diseases and several variables associated with vital signs, including diastolic blood pressure and changes in weight and respiratory rates. The combination of vital signs with traditional risk factors as input features achieved higher prediction accuracy than using either group of features alone. CONCLUSIONS: This study shows that the Extreme Gradient Boosting technique can use a large number of features from EHR data to make short-term fall predictions with a better performance than that of conventional fall risk assessments and other ML models. The integration of routinely collected EHR data, particularly vital signs, into fall prediction models may generate more accurate fall risk surveillance than models without vital signs. Our data support the use of ML models for dynamic, cost-effective, and automated fall predictions in different types of senior care facilities.

RESUMEN

The shuffling gait with slowed speed and reduced stride length has been considered classic clinical features in idiopathic Parkinson's disease (PD), and the risk of falling increases as the disease progresses. This raises the possibility that clinical disease severity might mediate the relationship between stride length and speed and the risk of falling in patients with PD. Sixty-one patients with PD patients underwent the clinical scores as well as quantitative biomechanical measures during walking cycles before and after dopamine replacement therapy. Mediation analysis tests whether the direct effect of an independent variable (stride length and speed) on a dependent variable (three-step fall prediction model score) can be explained by the indirect influence of the mediating variable (Unified Parkinson's Disease Rating Scale (UPDRS) total scores). The results demonstrate that decreased stride length, straight walking speed, and turning speed is associated with increased three-step fall prediction model score (r = -0.583, p < 0.0001, r = -0.519, p < 0.0001, and r = -0.462, p < 0.0001, respectively). We further discovered that UPDRS total scores value is negatively correlated with stride length, straight walking, and turning speed (r = -0.651, p < 0.0001, r = -0.555, p < 0.0001, and r = -0.372, p = 0.005, respectively) but positively correlated with the fall prediction model score value (r = 0.527, p < 0.0001). Further mediation analysis shows that the UPDRS total score values serve as mediators between lower stride length, straight walking, and turning speed and higher fall prediction model score values. Our results highlighted the relationship among stride length and speed, clinical disease severity, and risk of falling. As decreased stride length and speed are hallmarks of falls, monitoring the changes of quantitative biomechanical measures along with the use of wearable technology in a longitudinal study can provide a scientific basis for pharmacology, rehabilitation programs, and selecting high-risk candidates for surgical treatment to reduce future fall risk.

RESUMEN

BACKGROUND: Adverse health effects resulting from falls are a major public health concern. Although studies have identified risk factors for falls, none have examined long-term prediction of fall risk. Furthermore, recent evidence suggests that there are additional risk factors, such as psychosocial factors. OBJECTIVE: In this 3-year longitudinal study, we evaluated a predictive model for risk of fall among community-dwelling older adults using machine learning methods. DESIGN: A 3-year follow-up prospective longitudinal study (from 2010 to 2013). SETTING: Twenty-four municipalities in nine of the 47 prefectures (provinces) of Japan. PARTICIPANTS: Community-dwelling individuals aged ≥65 years who were functionally independent at baseline (n = 61,883). METHODS: The baseline survey was conducted from August 2010 to January 2012, and the follow-up survey was conducted from October to December 2013. Both surveys were conducted involving self-reported questionnaires. The measured outcome at the follow-up survey was self-reported multiple falls during the previous year. The 142 variables included in the baseline survey were regarded as candidate predictors. The random-forest-based Boruta algorithm was used to select predictors, and the eXtreme Gradient Boosting algorithm with 10 repetitions of nested k-fold cross-validation was used for modeling and model evaluation. Furthermore, we used shapley additive explanations to gain insight into the behavior of the prediction model. KEY RESULTS: Fourteen out of 142 candidate features were selected as predictors. Among these predictors, experience of falling as of the baseline survey was the most important feature, followed by self-rated health and age. Moreover, sense of coherence was newly identified as a risk factor for falls. CONCLUSIONS: This study suggests that machine learning tools can be adapted to explore new associative factors, make accurate predictions, and provide actionable insights for fall prevention strategies.

Asunto(s)

RESUMEN

Recently, fall risk assessment has been a main focus in fall-related research. Wearable sensors have been used to increase the objectivity of this assessment, building on the traditional use of oversimplified questionnaires. However, it is necessary to define standard procedures that will us enable to acknowledge the multifactorial causes behind fall events while tackling the heterogeneity of the currently developed systems. Thus, it is necessary to identify the different specifications and demands of each fall risk assessment method. Hence, this manuscript provides a narrative review on the fall risk assessment methods performed in the scientific literature using wearable sensors. For each identified method, a comprehensive analysis has been carried out in order to find trends regarding the most used sensors and its characteristics, activities performed in the experimental protocol, and algorithms used to classify the fall risk. We also verified how studies performed the validation process of the developed fall risk assessment systems. The identification of trends for each fall risk assessment method would help researchers in the design of standard innovative solutions and enhance the reliability of this assessment towards a homogeneous benchmark solution.

Asunto(s)