Dataset Information

Fall Detection in Individuals With Lower Limb Amputations Using Mobile Phones: Machine Learning Enhances Robustness for Real-World Applications.

ABSTRACT:

Background

Automatically detecting falls with mobile phones provides an opportunity for rapid response to injuries and better knowledge of what precipitated the fall and its consequences. This is beneficial for populations that are prone to falling, such as people with lower limb amputations. Prior studies have focused on fall detection in able-bodied individuals using data from a laboratory setting. Such approaches may provide a limited ability to detect falls in amputees and in real-world scenarios.

Objective

The aim was to develop a classifier that uses data from able-bodied individuals to detect falls in individuals with a lower limb amputation, while they freely carry the mobile phone in different locations and during free-living.

Methods

We obtained 861 simulated indoor and outdoor falls from 10 young control (non-amputee) individuals and 6 individuals with a lower limb amputation. In addition, we recorded a broad database of activities of daily living, including data from three participants' free-living routines. Sensor readings (accelerometer and gyroscope) from a mobile phone were recorded as participants freely carried it in three common locations-on the waist, in a pocket, and in the hand. A set of 40 features were computed from the sensors data and four classifiers were trained and combined through stacking to detect falls. We compared the performance of two population-specific models, trained and tested on either able-bodied or amputee participants, with that of a model trained on able-bodied participants and tested on amputees. A simple threshold-based classifier was used to benchmark our machine-learning classifier.

Results

The accuracy of fall detection in amputees for a model trained on control individuals (sensitivity: mean 0.989, 1.96*standard error of the mean [SEM] 0.017; specificity: mean 0.968, SEM 0.025) was not statistically different (P=.69) from that of a model trained on the amputee population (sensitivity: mean 0.984, SEM 0.016; specificity: mean 0.965, SEM 0.022). Detection of falls in control individuals yielded similar results (sensitivity: mean 0.979, SEM 0.022; specificity: mean 0.991, SEM 0.012). A mean 2.2 (SD 1.7) false alarms per day were obtained when evaluating the model (vs mean 122.1, SD 166.1 based on thresholds) on data recorded as participants carried the phone during their daily routine for two or more days. Machine-learning classifiers outperformed the threshold-based one (P<.001).

Conclusions

A mobile phone-based fall detection model can use data from non-amputee individuals to detect falls in individuals walking with a prosthesis. We successfully detected falls when the mobile phone was carried across multiple locations and without a predetermined orientation. Furthermore, the number of false alarms yielded by the model over a longer period of time was reasonably low. This moves the application of mobile phone-based fall detection systems closer to a real-world use case scenario.

SUBMITTER: Shawen N

PROVIDER: S-EPMC5656773 | biostudies-literature | 2017 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Fall Detection in Individuals With Lower Limb Amputations Using Mobile Phones: Machine Learning Enhances Robustness for Real-World Applications.

Shawen Nicholas N Lonini Luca L Mummidisetty Chaithanya Krishna CK Shparii Ilona I Albert Mark V MV Kording Konrad K Jayaraman Arun A

JMIR mHealth and uHealth 20171011 10

<h4>Background</h4>Automatically detecting falls with mobile phones provides an opportunity for rapid response to injuries and better knowledge of what precipitated the fall and its consequences. This is beneficial for populations that are prone to falling, such as people with lower limb amputations. Prior studies have focused on fall detection in able-bodied individuals using data from a laboratory setting. Such approaches may provide a limited ability to detect falls in amputees and in real-wo ...[more]

PMID: 29021127

Similar Datasets

Project description:Despite their importance to fall prevention research, little is known about the details of real-world fall events experienced by lower limb prosthesis users. This gap can be attributed to the lack of a structured, population-specific fall survey to document these adverse health events. The objective of this project was to develop a survey capable of characterizing the circumstances and consequences of fall events in lower limb prosthesis users. Best practices in survey development, including focus groups and cognitive interviews with diverse samples of lower limb prosthesis users, were used to solicit input and feedback from target respondents, so survey content would be meaningful, clear, and applicable to lower limb prosthesis users. Focus group data were used to develop fall event definitions and construct a conceptual fall framework that guided the creation of potential survey questions and response options. Survey questions focused on the activity, surroundings, situation, mechanics, and consequences of fall events. Cognitive interviews revealed that with minor revisions, survey definitions, questions, and response options were clear, comprehensive, and applicable to the experiences of lower limb prosthesis users. Administration of the fall survey to a national sample of 235 lower limb prosthesis users in a cross-sectional preliminary validation study, found survey questions to function as intended. Revisions to the survey were made at each stage of development based on analysis of participant feedback and data. The structured, 37-question lower limb prosthesis user fall event survey developed in this study offers clinicians and researchers the means to document, monitor, and compare fall details that are meaningful and relevant to lower limb prosthesis users in a standardized and consistent manner. Data that can be collected with the developed survey are essential to establishing specific goals for fall prevention initiatives in lower limb prosthesis users.

Project description:BACKGROUND: The mobility of older individuals has often been only partially assessed, without considering all important aspects such as potential (available) versus effective (used) mobilities and the physical and psychosocial factors that modulate them. This study proposes a new model for evaluating mobility that considers all important aspects, applied here to lower-limb amputees with vascular origin. This model integrates the concepts of potential mobility (e.g. balance, speed of movement), effective mobility (e.g. life habits, movements in living areas) and factors that modulate these two types of mobility (e.g. strength, sensitivity, social support, depression). The main objective was to characterize potential and effective mobility as well as mobility modulators in a small sample of people with lower-limb amputations of vascular origin with different characteristics. The second objective of this pilot study was to assess the feasibility of measuring all variables in the model in a residential context. METHODS: An observational and transversal design was used with a heterogeneous sample of 10 participants with a lower-limb amputation of vascular origin, aged 51 to 83, assessed between eight and 18 months after discharge from an acute care hospital. A questionnaire of participant characteristics and 16 reliable and valid measurements were used. RESULTS: The results show that the potential mobility indicators do not accurately predict effective mobility, i.e., participants who perform well on traditional measures done in the laboratory or clinic are not always those who perform well in the real world. The model generated 4 different profiles (categories) of participants ranging from reduced to excellent potential mobility and low to excellent effective mobility, and characterized the modulating factors. The evaluations were acceptable in terms of the time taken (three hours) and the overall measurements, with a few exceptions, which were modified to optimize the data collected and the classification of the participants. For the population assessed, the results showed that some of the negative modulators (particularly living alone, no rehabilitation, pain, limited social support, poor muscle strength) played an important role in reducing effective mobility. CONCLUSION: The first use of the model revealed interesting data that add to our understanding of important aspects linked to potential and effective mobility as well as modulators. The feasibility of measuring all variables in the model in a residential context was demonstrated. A study with a large number of participants is now warranted to rigorously characterize mobility levels of lower-limb amputees with vascular origin.

Dataset Information

Fall Detection in Individuals With Lower Limb Amputations Using Mobile Phones: Machine Learning Enhances Robustness for Real-World Applications.

Background

Objective

Methods

Results

Conclusions

Publications

Fall Detection in Individuals With Lower Limb Amputations Using Mobile Phones: Machine Learning Enhances Robustness for Real-World Applications.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets