Factors associated with fall risk of community-dwelling older people: A decision tree analysis

A scoping review of machine learning models to predict risk of falls in elders, without using sensor data

OBJECTIVES

This scoping review assesses machine learning (ML) tools that predicted falls, relying on information in health records without using any sensor data. The aim was to assess the available evidence on innovative techniques to improve fall prevention management.

METHODS

Studies were included if they focused on predicting fall risk with machine learning in elderly populations and were written in English. There were 13 different extracted variables, including population characteristics (community dwelling, inpatients, age range, main pathology, ethnicity/race). Furthermore, the number of variables used in the final models, as well as their type, was extracted.

RESULTS

A total of 6331 studies were retrieved, and 19 articles met criteria for data extraction. Metric performances reported by authors were commonly high in terms of accuracy (e.g., greater than 0.70). The most represented features included cardiovascular status and mobility assessments. Common gaps identified included a lack of transparent reporting and insufficient fairness assessments.

CONCLUSIONS

This review provides evidence that falls can be predicted using ML without using sensors if the amount of data and its quality is adequate. However, further studies are needed to validate these models in diverse groups and populations.

Efficacy of a Waist-Mounted Sensor in Predicting Prospective Falls Among Older People Residing in Community Dwellings: A Prospective Cohort Study

Falls pose a significant health risk for older people, necessitating accurate predictive tools for fall prevention. This study evaluated the sensitivity of a wearable waist-belt sensor, the Booguu Aspire system, in predicting prospective fall incidents among 37 community-dwelling older people in Hong Kong. A prospective cohort design was employed, involving two analytical groups: the overall cohort and a subset with cognitive performance data available, measured using the Montreal Cognitive Assessment (MoCA). Participants were categorized into moderate- or high-risk groups for falls using the sensor and further assessed with physical function tests, including the Single Leg Stand Test (SLST), 6 Meter Walk Test (6MWT), and Five Times Sit to Stand Test (5STS). Fall incidents were monitored for 12 months through quarterly follow-up phone calls. Statistical analyses showed no significant differences in physical performance between high- and moderate-risk groups and no significant correlations between sensor-based fall risk ratings and physical function test outcomes. The SLST, 6MWT, 5STS, and MoCA tests classified sensor-determined fall risk ratings with accuracies of 51.4%, 64.9%, 59.5%, and 50%. The sensor showed low sensitivity, with 13.51% true positives for fallers and a 20% sensitivity for high-risk individuals. ROC analysis yielded an Area Under the Curve of 0.688. Our findings indicate that the wearable waist-belt Sensor System may not be a sensitive tool in predicting prospective fall incidents. The algorithm for fall risk classification in the wearable sensor merits further exploration.

A Machine Learning-Based Prediction Model for the Probability of Fall Risk Among Chinese Community-Dwelling Older Adults

Fall is a common adverse event among older adults. This study aimed to identify essential fall factors and develop a machine learning-based prediction model to predict the fall risk category among community-dwelling older adults, leading to earlier intervention and better outcomes. Three prediction models (logistic regression, random forest, and naive Bayes) were constructed and evaluated. A total of 459 people were involved, including 156 participants (34.0%) with high fall risk. Seven independent predictors (frail status, age, smoking, heart attack, cerebrovascular disease, arthritis, and osteoporosis) were selected to develop the models. Among the three machine learning models, the logistic regression model had the best model fit, with the highest area under the curve (0.856) and accuracy (0.797) and sensitivity (0.735) in the test set. The logistic regression model had excellent discrimination, calibration, and clinical decision-making ability, which could aid in accurately identifying the high-risk groups and taking early intervention with the model.

Knowledge mapping and research trends of accidental falls in patients with Parkinson's disease from 2003 to 2023: a bibliometric analysis

Background

Recent years have witnessed a rapid growth in research on accidental falls in patients with Parkinson's Disease (PD). However, a comprehensive and systematic bibliometric analysis is still lacking. This study aims to systematically analyze the current status and development trends of research related to accidental falls in patients with PD using bibliometric methods.

Methods

We retrieved literature related to accidental falls in patients with PD published between January 1, 2003, and December 31, 2023, from the Web of Science Core Collection (WoSCC) database. Statistical analysis and knowledge mapping of the literature were conducted using VOSviewer, CiteSpace, and Microsoft Excel software.

Results

A total of 3,195 publications related to accidental falls in patients with PD were retrieved. These articles were authored by 13,202 researchers from 3,834 institutions across 87 countries and published in 200 academic journals. Over the past 20 years, the number of published articles and citations has increased annually. The United States and the United Kingdom have the highest number of publications in this field, while Harvard University and Tel Aviv University are the most influential institutions. The Parkinsonism & Related Disorders journal published the highest number of articles, while the Movement Disorders journal had the highest number of citations. The most prolific author is Bloem, Bastiaan R, while the most cited author is Hausdorff, Jeffrey. The main research areas of these publications are Neurosciences, Biomedical, Electrical & Electronic, and Biochemistry & Molecular Biology. Currently, high-frequency keywords related to accidental falls in patients with PD include risk factors, clinical manifestations, and interventions. Prediction and prevention of accidental falls in such patients is a research topic with significant potential and is currently a major focus of research.

Conclusion

This study used bibliometric and knowledge mapping analysis to reveal the current research status and hotspots in the field of accidental falls in patients with PD. It also points out directions for future research. This study can provide theoretical support and practical guidance for scholars to further conduct related research.

Effect of number of medications on the risk of falls among community-dwelling older adults: A 3-year follow-up of the SONIC study

AIM

This study examined the association between the number of prescribed medications and falls among community-dwelling older adults.

METHODS

We conducted a geriatric comprehensive health-checkup on community-dwelling adults aged 69-91 years who participated in the Septuagenarians, Octogenarians, and Nonagenarians Investigation with Centenarians study. The final analysis of this study included 1,076 participants with complete data. The participants were divided into four groups based on the number of medications at baseline: 0, 1, 2-4, and ≥5. At the 3-year follow-up, the participants were asked whether they had fallen in the past year. Multivariable logistic regression analysis was performed to assess the relationship between the number of medications taken and falls after adjusting for confounding factors.

RESULTS

The prevalence rates of falls were 10.5%, 18.2%, 18.3%, and 19.8% in the no-medication, one-medication, comedication, and polypharmacy groups, respectively. In the one-medication prescription group, 59% of prescriptions were for fall-risk-increasing drugs (FRID). Multivariable analysis showed a significantly higher incidence of falls in the one-medication group (adjusted odds ratio [OR], 1.91; 95% confidence interval [CI], 1.04-3.54), co-medication (OR, 1.89; 95% CI, 1.09-3.29), and polypharmacy groups (OR, 1.94; 95% CI, 1.09-3.45) than in the no-medication group.

CONCLUSIONS

The study showed that polypharmacy, as well as just taking one medication, can affect the occurrence of falls. This suggests that in addition to the number of medications and polypharmacy, the type of medication, such as FRID, affects the risk of falls. Therefore, pharmacotherapy should consider the risk of falls in older adults when prescribing medications. Geriatr Gerontol Int 2024; 24: 306-310.