PDF(2372 KB)
Prediction of Falls in Older Adults Based on Artificial Intelligence: Pathways Linking Lower-Limb Function, Activities of Daily Living, and Falls
SHI Jipeng, HAN Xuejiao, XU Hongqi, ZHU Tianrui, WEI Jinpeng, QUAN Helong, LIN Xiuzhu
Journal of Capital University of Physical Education and Sports ›› 2026, Vol. 38 ›› Issue (2) : 149-160.
PDF(2372 KB)
PDF(2372 KB)
Prediction of Falls in Older Adults Based on Artificial Intelligence: Pathways Linking Lower-Limb Function, Activities of Daily Living, and Falls
Objective: To construct a fall prediction model for the elderly based on artificial intelligence logic, and systematically explore the association mechanism between lower limb function, activities of daily living (ADL) and falls. Methods: Based on the database of CHARLS, 7693 elderly people over 60 years old were selected as the subjects; The prediction performance of six machine learning models was systematically compared by adopting the dual modeling strategy (the first modeling strictly controlled the collinearity problem, and the second modeling added variables according to the theoretical framework), and the contribution of each characteristic variable to the prediction model was quantified by using the SHAP method; Then, through the analysis of mediating effect, we explored the mediating role of activities of daily living in the relationship between lower limb function and falls. Results: 1) The results of dual modeling comparison showed that the modeling scheme based on variance inflation factor screening variables had better prediction performance, and the gradient boosting tree model had the best performance (area under the curve = 0.677, 95% confidence interval [0.643, 0.711]); 2) The results of the analysis of the importance of Shapley's additive explanation showed that activities of daily living were the key variables affecting falls; 3) The results of mediating effect analysis showed that lower limb function had no significant direct effect on falls, but had an indirect effect on falls through the mediating variable of activities of daily living. Conclusion: 1) The fall prediction model of the elderly is constructed by machine learning and dual modeling strategy, and the gradient boosting tree model has the best calibration performance; 2) The results of Shapley additive interpretation analysis and mediating effect test showed that activities of daily living were the key variables to predict falls, and lower limb function indirectly affected falls through activities of daily living, which provided a theoretical basis for the prevention and intervention of falls in the elderly.
artificial intelligence / machine learning / elder adults / fall prediction / lower-limb function / activities of daily living / mediation analysis
| [1] |
中共中央, 国务院. “健康中国2030”规划纲要[J]. 中华人民共和国国务院公报, 2016(32): 5-20.
|
| [2] |
Falls[Z/OL]. (2021-04-26)[2024-10-20]. https://www.who.int/news-room/fact-sheets/detail/falls.
|
| [3] |
This study aims to clarify the risk factors for falls to prevent severe consequences in older adults.
|
| [4] |
Falls in older adults are a reasonably common occurrence and about 10% of these experience multiple falls annually. These falls may be serious and may cause significant morbidity and mortality. These can also threaten the independence of older people and may be responsible for an individual's loss of independence and socioeconomic consequences. These falls may add extra burden to the health care and to direct and indirect costs.An extensive search of literature was done on the important data bases of PubMed, SCOPUS, and Google Scholar on this topic and all the useful information was derived from the relevant articles for this review.We found that the falls in older individuals are often multi factorial and hence a multidisciplinary approach is required to prevent and manage these falls. The risk factors leading to the falls could be divided into extrinsic, intrinsic and situational factors. The commonest and serious injuries are to the head and fractures, due to fragility of bones.The falls in elderly are on rise and taking the shape of an epidemic. Prevention of these falls is far better than the management. Safe living environment of the elderly people helps in prevention of these falls. The management of the falls should focus on the causative factors, apart from treating the injuries caused by the falls.© Indian Orthopaedics Association 2020.
|
| [5] |
The prevalence of falls among older adults living in the community is ~30% each year. The impacts of falls are not only confined to the individual but also affect families and the community. Injury from a fall also imposes a heavy financial burden on patients and their families. Currently, there are different reports on the risk factors for falls among older adults in the community. A retrospective analysis was used in this study to identify risk factors for falls in community-dwelling older adults. This research aimed to collect published studies to find risk factors for falls in community-dwelling older adults.
|
| [6] |
To identify the characteristics of elderly persons who develop a fear of falling after experiencing a fall and to investigate the association of this fear with changes in health status over time.A prospective study of falls over a 2-year period (1991-92). Falls were ascertained using bimonthly postcards plus telephone interview with a standardized (World Health Organisation) questionnaire for circumstances, fear of falling and consequences of each reported fall. Each participant underwent a physical exam and subjective health assessment each year form 1990 to 1993.New-Mexico Aging Process Study, USA.487 elderly subjects (> 60 years) living independently in the community.Fear of falling after experiencing a fall.70 (32%) of 219 subjects who experienced a fall during the 2 year study period reported a fear of falling. Women were more likely than men to report fear of falling (74% vs 26%). Fallers who were afraid of falling again had significantly ore balance (31.9% vs 12.8%) and gait disorders (31.9% vs 7.4%) at entry in the study in 1990. Among sex, age, mental status, balance and gait abnormalities, economic resource and physical health, logistic regression analysis show gait abnormalities and poor self-perception of physical health, cognitive status and economic resources to be significantly associated with fear of falling. Subjects who reported a fear of falling experienced a greater increase in balance (P = 0.08), gait (P < 0.01) and cognitive disorders (P = 0.09) over time, resulting in a decrease in mobility level.The study indicated that about one-third of elderly people develop a fear of falling after an incident fall and this issue should be specifically addressed in any rehabilitation programme.
|
| [7] |
With increasing life expectancy, declining mortality, and birth rates, the world's geriatric population is increasing. Falls in the older people are one of the most common and serious problems. Injuries from falls can be fatal or non-fatal and physical or psychological, leading to a reduction in the ability to perform activities of daily living. The aim of this study was to determine the prevalence of falls in the older people through systematic review and meta-analysis.
|
| [8] |
|
| [9] |
Although the validity of the sit-to-stand (STS) test as a measure of lower limb strength has been questioned, it is widely used as such among older adults. The purposes of this study were: 1) to describe five-repetition STS test (FRSTST) performance (time) by adolescents and adults and 2) to determine the relationship of isometric knee extension strength (force and torque), age, gender, weight, and stature with that performance. Participants were 111 female and 70 male (14-85 years) community-dwelling enrollees in the NIH Toolbox Assessment of Neurological and Behavioral Function. The FRSTST was conducted using a standard armless chair. Knee extension force was measured using a belt-stabilized hand-held dynamometer; knee extension torque was measured using a Biodex dynamometer. The mean times for the FRSTST ranged from 6.0 sec (20-29 years) to 10.8 sec (80-85 years). For both the entire sample and a sub-sample of participants 50-85 years, knee extension strength ( = -0.388 to -0.634), age ( = 0.561 and 0.466), and gender ( = 0.182 and 0.276) were correlated significantly with FRSTST times. In all multiple regression models, knee extension strength provided the best explanation of FRSTST performance, but age contributed as well. Bodyweight and stature were less consistent in explaining FRSTST performance. Gender did not add to the explanation of FRSTST performance. Our findings suggest, therefore, that FRSTST time reflects lower limb strength, but that performance should be interpreted in light of age and other factors.
|
| [10] |
Previous research has found that lower limb muscle asymmetries increase with age and are linked to fall and injury risks. However, past studies lack a wide variety of muscle function modes and measures as well as comparison to a comparable younger age group. The purpose of this study was to examine age-related lower limb muscle function asymmetries across a variety of muscle action types and velocities in young and old adults. Lower limb balance, strength, power, and velocity were evaluated with concentric, isometric, isotonic, and eccentric muscle actions during a single-leg stance test and on single- and multi-joint dynamometers in 29 young (age = 21.45 ± 3.02) and 23 old (age = 77.00 ± 4.60) recreationally active men and women. Most (15 of 17) variables showed no statistical (p > 0.05) or functional (10% threshold) limb asymmetry for either age group. There was a significant main effect (p = 0.046; collapsed across groups) found for asymmetry (dominant > non-dominant) for the isotonic peak velocity variable. There was a significant (p = 0.010) group × limb interaction for single-joint concentric peak power produced at a slow (60 deg/s) velocity due to the non-dominant limb of the young group being 12.2% greater than the dominant limb (p < 0.001), whereas the old group was not asymmetrical (p = 0.965). The findings of this investigation indicate there is largely no age-related asymmetry of the lower limbs across a range of muscle function-related variables and modes, with a couple of notable exceptions. Also, the significant asymmetries for the isotonic peak velocity variable perhaps show the sensitivity of this uncommonly used measure in detecting minimally present muscle function imbalances.
|
| [11] |
|
| [12] |
HORTOBÁT,
|
| [13] |
|
| [14] |
|
| [15] |
To review the validated instruments that assess gait, balance, and functional mobility to predict falls in older adults across different settings.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.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.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.© 2022. The Author(s).
|
| [16] |
Falls are the most common cause of injury-related morbidity and mortality in older adults.
|
| [17] |
|
| [18] |
To determine whether the interval over which patients are asked to remember their falls affects fall reporting.Systematic literature review.Community.Individuals being monitored for falls in prospective studies that asked participants to recall falls over varying intervals.Sensitivity and specificity of retrospective recall compared with a criterion-standard prospective assessment using some form of ongoing fall monitoring.Six studies met the inclusion criteria. Recall of falls in the previous year was specific (specificity 91-95%) but less sensitive (sensitivity 80-89%) than the criterion standard of ongoing prospective collection of fall data using fall calendars or postcards. Patients with injurious falls were more likely to recall their falls. Lower Mini-Mental State Examination score was associated with poorer recall of falls in the one study addressing this issue.Whenever accurate data on all falls are critical, such as with interventions to decrease the rate of falls, researchers should gather information on falls every week or every month from study participants. The optimal method of fall monitoring--postcard, calendar, diary, telephone, or some combination of these--remains unknown.
|
| [19] |
|
| [20] |
Identification of older persons at risk for the loss of independence, onset of (co)-morbidity or functional limitations through screening/assessment is of interest for the public health-care system. To date several different measurement instruments for overall physical function are frequently used in practice, but little information about their psychometric properties is available. Objectives andOur aim was to assess instruments with an overall score related to functional status and/or physical performance on content and psychometric properties. Electronic databases (Medline, EMBASE, AMED, Cochrane Library and CINAHL) were searched, using MeSH terms and relevant keywords. Studies, published in English, were included if their primary or secondary purpose was to evaluate the measurement properties of measurement instruments for overall physical function in community-dwelling older persons aged 60 years and older. Reliability, validity, responsiveness and practicability were evaluated, adhering to a specified protocol.In total 78 articles describing 12 different functional assessment instruments were included and data extracted. Seven instruments, including their modified versions, were evaluated for reliability. Nine instruments, including their modified versions, were evaluated with regard to validity.In conclusion, the Short Physical Performance Battery can be recommended most highly in terms of validity, reliability and responsiveness, followed by the Physical Performance Test and Continuous Scale Physical Functional Performance.
|
| [21] |
|
| [22] |
|
| [23] |
We examine the following: (1) the appropriateness of using a data quality (DQ) framework developed for relational databases as a data-cleaning tool for a data set extracted from two EPIC databases, and (2) the differences in statistical parameter estimates on a data set cleaned with the DQ framework and data set not cleaned with the DQ framework.The use of data contained within electronic health records (EHRs) has the potential to open doors for a new wave of innovative research. Without adequate preparation of such large data sets for analysis, the results might be erroneous, which might affect clinical decision-making or the results of Comparative Effectives Research studies.Two emergency department (ED) data sets extracted from EPIC databases (adult ED and children ED) were used as examples for examining the five concepts of DQ based on a DQ assessment framework designed for EHR databases. The first data set contained 70,061 visits; and the second data set contained 2,815,550 visits. SPSS Syntax examples as well as step-by-step instructions of how to apply the five key DQ concepts these EHR database extracts are provided.SPSS Syntax to address each of the DQ concepts proposed by Kahn et al. (2012)1 was developed. The data set cleaned using Kahn's framework yielded more accurate results than the data set cleaned without this framework. Future plans involve creating functions in R language for cleaning data extracted from the EHR as well as an R package that combines DQ checks with missing data analysis functions.
|
| [24] |
|
| [25] |
Multiple imputation (MI) is an advanced technique for handing missing values. It is superior to single imputation in that it takes into account uncertainty in missing value imputation. However, MI is underutilized in medical literature due to lack of familiarity and computational challenges. The article provides a step-by-step approach to perform MI by using R multivariate imputation by chained equation (MICE) package. The procedure firstly imputed m sets of complete dataset by calling mice() function. Then statistical analysis such as univariate analysis and regression model can be performed within each dataset by calling with() function. This function sets the environment for statistical analysis. Lastly, the results obtained from each analysis are combined by using pool() function.
|
| [26] |
|
| [27] |
|
| [28] |
|
| [29] |
|
| [30] |
Logistic regression models encounter challenges with correlated predictors and influential outliers. This study integrates robust estimators, including the Bianco–Yohai estimator (BY) and conditionally unbiased bounded influence estimator (CE), with the logistic Liu (LL), logistic ridge (LR), and logistic KL (KL) estimators. The resulting estimators (LL-BY, LL-CE, LR-BY, LR-CE, KL-BY, and KL-CE) are evaluated through simulations and real-life examples. KL-BY emerges as the preferred choice, displaying superior performance by reducing mean squared error (MSE) values and exhibiting robustness against multicollinearity and outliers. Adopting KL-BY can lead to stable and accurate predictions in logistic regression analysis.
|
| [31] |
Technologies have driven big data collection across many fields, such as genomics and business intelligence. This results in a significant increase in variables and data points (observations) collected and stored. Although this presents opportunities to better model the relationship between predictors and the response variables, this also causes serious problems during data analysis, one of which is the multicollinearity problem. The two main approaches used to mitigate multicollinearity are variable selection methods and modified estimator methods. However, variable selection methods may negate efforts to collect more data as new data may eventually be dropped from modeling, while recent studies suggest that optimization approaches via machine learning handle data with multicollinearity better than statistical estimators. Therefore, this study details the chronological developments to mitigate the effects of multicollinearity and up-to-date recommendations to better mitigate multicollinearity.
|
| [32] |
Clinical prediction models are widely used in health and medical research. The area under the receiver operating characteristic curve (AUC) is a frequently used estimate to describe the discriminatory ability of a clinical prediction model. The AUC is often interpreted relative to thresholds, with "good" or "excellent" models defined at 0.7, 0.8 or 0.9. These thresholds may create targets that result in "hacking", where researchers are motivated to re-analyse their data until they achieve a "good" result.We extracted AUC values from PubMed abstracts to look for evidence of hacking. We used histograms of the AUC values in bins of size 0.01 and compared the observed distribution to a smooth distribution from a spline.The distribution of 306,888 AUC values showed clear excesses above the thresholds of 0.7, 0.8 and 0.9 and shortfalls below the thresholds.The AUCs for some models are over-inflated, which risks exposing patients to sub-optimal clinical decision-making. Greater modelling transparency is needed, including published protocols, and data and code sharing.© 2023. BioMed Central Ltd., part of Springer Nature.
|
| [33] |
As the performance of current fall risk assessment tools is limited, clinicians face significant challenges in identifying patients at risk of falling. This study proposes an automatic fall risk prediction model based on eXtreme gradient boosting (XGB), using a data-driven approach to the standardized medical records. This study analyzed a cohort of 639 participants (297 fall patients and 342 controls) from Chang Gung Memorial Hospital, Chiayi Branch, Taiwan. A derivation cohort of 507 participants (257 fall patients and 250 controls) was collected for constructing the prediction model using the XGB algorithm. A comparative validation of XGB and the Morse Fall Scale (MFS) was conducted with a prospective cohort of 132 participants (40 fall patients and 92 controls). The areas under the curves (AUCs) of the receiver operating characteristic (ROC) curves were used to compare the prediction models. This machine learning method provided a higher sensitivity than the standard method for fall risk stratification. In addition, the most important predictors found (Department of Neuro-Rehabilitation, Department of Surgery, cardiovascular medication use, admission from the Emergency Department, and bed rest) provided new information on in-hospital fall event prediction and the identification of patients with a high fall risk.
|
| [34] |
This study aims to develop a advanced machine learning model to predict the fall risk among community-dwelling elders. This study could present actionable advices for early prevention of fall risk.
|
| [35] |
|
| [36] |
Falls are adverse events which commonly occur in hospitalized patients. Inpatient falls may cause bruises or contusions and even a fractures or head injuries, which can lead to significant physical and economic burdens for patients and their families. Therefore, it is important to predict the risks involved surrounding hospitalized patients falling in order to better provide medical personnel with effective fall prevention measures.
|
| [37] |
The global aging crisis has precipitated significant public health challenges, including rising chronic diseases, economic burdens, and labor shortages, particularly in China. Activities of daily living (ADL) dysfunction, affecting over 40 million Chinese older adults (16% of the aging population), severely compromises independence and quality of life while increasing health care costs and mortality. ADL dysfunction encompasses both basic ADL (BADL) and instrumental ADL (IADL), which assess fundamental self-care and complex environmental interactions, respectively. With projections indicating 65 million cases by 2030, there is an urgent need for tools to predict ADL impairment and enable early interventions.
|
| [38] |
|
| [39] |
Impaired respiratory function secondary to acute or chronic respiratory disease poses a significant clinical and healthcare burden. Intrapulmonary percussive ventilation (IPV) is used in various clinical settings to treat excessive airway secretions, pulmonary atelectasis, and impaired gas exchange. Despite IPV’s wide use, there is a lack of clinical guidance on IPV application which may lead to inconsistency in clinical practice. This scoping review aimed to summarise the clinical application methods and dosage of IPV used by clinicians and researchers to provide guidance. A two-staged systematic search was conducted to retrieve studies that used IPV in inpatient and outpatient settings. MEDLINE, EMBASE, CINAHL, Scopus, and Google scholar were searched from January 1979 till 2022. Studies with patients aged ≥16 years and published in any language were included. Two reviewers independently screened the title and abstract, reviewed full text articles, and extracted data. Search yielded 514 studies. After removing duplicates and irrelevant studies, 25 studies with 905 participants met the inclusion criteria. This is the first scoping review to summarise IPV application methods and dosages from the available studies in intensive care unit (ICU), acute inpatient (non-ICU), and outpatient settings. Some variations in clinical applications and prescribed dosages of IPV were noted. Despite variations, common trends in clinical application and prescription of IPV dosages were observed and summarised to assist clinicians with IPV intervention. Although an evidence-based clinical guideline could not be provided, this review provides detailed information on IPV application and dosages in order to provide clinical guidance and lays a foundation towards developing a clinical practice guideline in the future.
|
| [40] |
|
| [41] |
This study aimed to identify the optimal features of gait parameters to predict the fall risk level in older adults. The study included 746 older adults (age: 63-89 years). Gait tests (20 m walkway) included speed modification (slower, preferred, and faster-walking) while wearing the inertial measurement unit sensors embedded in the shoe-type data loggers on both outsoles. A metric was defined to classify the fall risks, determined based on a set of questions determining the history of falls and fear of falls. The extreme gradient boosting (XGBoost) model was built from gait features to predict the factor affecting the risk of falls. Moreover, the definition of the fall levels was classified into high- and low-risk groups. At all speeds, three gait features were identified with the XGBoost (stride length, walking speed, and stance phase) that accurately classified the fall risk levels. The model accuracy in classifying fall risk levels ranged between 67-70% with 43-53% sensitivity and 77-84% specificity. Thus, we identified the optimal gait features for accurate fall risk level classification in older adults. The XGBoost model could inspire future works on fall prevention and the fall-risk assessment potential through the gait analysis of older adults.
|
| [42] |
|
| [43] |
Walking speed (WS) is an easily assessable and interpretable functional outcome measure with great utility for the physical therapist providing care to older adults. Since WS was proposed as the sixth vital sign, research into its interpretation and use has flourished. The purpose of this scoping review is to identify the current prognostic value of WS for the older adult.
|
| [44] |
|
| [45] |
This study aims to investigate the relationship between muscle strength, working memory, and activities of daily living (ADL) in older adults. Additionally, it seeks to clarify the pathways and effects of working memory in mediating the relationship between muscle strength and ADL.
|
| [46] |
|
| [47] |
|
| [48] |
|
/
| 〈 |
|
〉 |
