Fall detection algorithms for real-world falls harvested from lumbar sensors in the elderly population: A machine learning approach

2016 ◽  
Author(s):  
Alan K. Bourke ◽  
Jochen Klenk ◽  
Lars Schwickert ◽  
Kamiar Aminian ◽  
Espen A. F. Ihlen ◽  
...  
Keyword(s):  
Machine Learning ◽  
Real World ◽  
Elderly Population ◽  
Fall Detection ◽  
The Elderly ◽  
Learning Approach ◽  
Detection Algorithms ◽  
Machine Learning Approach

scholarly journals Accelerometer-Based Fall Detection Using Machine Learning: Training and Testing on Real-World Falls

Sensors ◽  
2020 ◽  
Vol 20 (22) ◽  
pp. 6479
Author(s):  
Luca Palmerini ◽  
Jochen Klenk ◽  
Clemens Becker ◽  
Lorenzo Chiari
Keyword(s):  
Machine Learning ◽  
Real World ◽  
Inertial Sensor ◽  
Fall Detection ◽  
Medical Attention ◽  
Support Vector ◽  
Detection Algorithms ◽  
Detection Systems ◽  
Machine Learning Approach ◽  
Extensive Collection

Falling is a significant health problem. Fall detection, to alert for medical attention, has been gaining increasing attention. Still, most of the existing studies use falls simulated in a laboratory environment to test the obtained performance. We analyzed the acceleration signals recorded by an inertial sensor on the lower back during 143 real-world falls (the most extensive collection to date) from the FARSEEING repository. Such data were obtained from continuous real-world monitoring of subjects with a moderate-to-high risk of falling. We designed and tested fall detection algorithms using features inspired by a multiphase fall model and a machine learning approach. The obtained results suggest that algorithms can learn effectively from features extracted from a multiphase fall model, consistently overperforming more conventional features. The most promising method (support vector machines and features from the multiphase fall model) obtained a sensitivity higher than 80%, a false alarm rate per hour of 0.56, and an F-measure of 64.6%. The reported results and methodologies represent an advancement of knowledge on real-world fall detection and suggest useful metrics for characterizing fall detection systems for real-world use.

scholarly journals An Event-Triggered Machine Learning Approach for Accelerometer-Based Fall Detection

Sensors ◽  
2017 ◽  
Vol 18 (2) ◽  
pp. 20 ◽  
Author(s):  
I Putra ◽  
James Brusey ◽  
Elena Gaura ◽  
Rein Vesilo
Keyword(s):  
Machine Learning ◽  
Fall Detection ◽  
Learning Approach ◽  
Machine Learning Approach ◽  
Event Triggered

Risk factors associated with skeletal-related events following denosumab cessation among patients with bone metastases from solid tumors: A real-world machine learning approach.

2021 ◽  
Vol 39 (15_suppl) ◽  
pp. 1567-1567
Author(s):  
Alison Stopeck ◽  
Celestia S. Higano ◽  
David H. Henry ◽  
Basia A. Bachmann ◽  
Marko Rehn ◽  
...  
Keyword(s):  
Machine Learning ◽  
Risk Factors ◽  
Bone Metastases ◽  
Real World ◽  
Treatment Duration ◽  
Learning Approach ◽  
Factors Associated ◽  
Machine Learning Approach ◽  
Skeletal Related Events ◽  
Denosumab Treatment

1567 Background: The anti-RANKL monoclonal antibody denosumab has been shown to be superior to the bisphosphonate zoledronate for the prevention of skeletal-related events (SREs) in patients with incident bone metastases (BM) from solid tumors (ST). Clinical guidelines recommend the use of a bone-targeting agent for SRE prevention for ≥ 2 years. However, real-world treatment patterns in the U.S. suggest that the denosumab treatment duration is often < 1 year. Applying a machine learning approach, we sought to identify risk factors associated with SRE incidence following cessation of denosumab to help inform optimal clinical SRE prevention strategies. Methods: Using the Optum PanTher Electronic Health Record repository, patients diagnosed with incident BM from a primary ST between 1 Jan 2007 and 1 Sep 2019 were evaluated for inclusion in the study. Eligible patients had to receive ≥ 2 consecutive 120 mg denosumab doses on an every 4-week (± 14 days) schedule and have a minimum follow-up ≥ 1 year after the last denosumab dose or an SRE occurring between days 84 and 365 after denosumab cessation. Extreme gradient boosting was used to develop an SRE risk prediction model evaluated on a test dataset. Impact and relative importance of available medical, clinical, and treatment factors on SRE risk following denosumab cessation were extracted from the model using Shapley additive explanations (SHAP). Univariate analyses on risk factors with the highest importance from pooled and tumor-specific models were also conducted. Results: A total of 1,414 patients (breast, n = 563 [40%]; prostate, 421 [30%]; lung, 180 [13%]; other cancers, 250 [17%]) met inclusion criteria, with a median of 253 (min, 88; max, 2726) days of denosumab treatment; 490 (35%) experienced ≥ 1 SRE following denosumab cessation. With a meaningful model performance based on an area under the receiver operating characteristic (AUROC) score of 77%, SHAP identified several significant factors that predicted an increased SRE risk following denosumab cessation, including prior SREs, shorter denosumab treatment duration, and a higher number of clinic visits as the top-ranked factors (Table). Conclusions: A machine learning approach to SRE risk factor identification may help clinicians better assess the individualized patient’s need for denosumab treatment persistence and improve patient outcomes. Results from tumor-specific groups will be presented at the meeting.[Table: see text]

Sign in / Sign up

Export Citation Format

Share Document