This randomized clinical trial examines a decision support tool that provides lower back pain information and self-management recommendations that are specific to an individual’s characteristics, symptoms, and symptom progression.
Existing accelerometer-based human activity recognition (HAR) benchmark datasets that were recorded during free living suffer from non-fixed sensor placement, the usage of only one sensor, and unreliable annotations. We make two contributions in this work. First, we present the publicly available Human Activity Recognition Trondheim dataset (HARTH). Twenty-two participants were recorded for 90 to 120 min during their regular working hours using two three-axial accelerometers, attached to the thigh and lower back, and a chest-mounted camera. Experts annotated the data independently using the camera’s video signal and achieved high inter-rater agreement (Fleiss’ Kappa =0.96). They labeled twelve activities. The second contribution of this paper is the training of seven different baseline machine learning models for HAR on our dataset. We used a support vector machine, k-nearest neighbor, random forest, extreme gradient boost, convolutional neural network, bidirectional long short-term memory, and convolutional neural network with multi-resolution blocks. The support vector machine achieved the best results with an F1-score of 0.81 (standard deviation: ±0.18), recall of 0.85±0.13, and precision of 0.79±0.22 in a leave-one-subject-out cross-validation. Our highly professional recordings and annotations provide a promising benchmark dataset for researchers to develop innovative machine learning approaches for precise HAR in free living.
Introduction: Accelerometer-based measurements of physical activity types are commonly used to replace self-reports. To advance the field, it is desirable that such measurements allow accurate detection of key daily physical activity types. This study aimed to evaluate the performance of a machine learning classifier for detecting sitting, standing, lying, walking, running, and cycling based on a dual versus single accelerometer setups during free-living. Methods: Twenty-two adults (mean age [SD, range] 38.7 [14.4, 25–68] years) were wearing two Axivity AX3 accelerometers positioned on the low back and thigh along with a GoPro camera positioned on the chest to record lower body movements during free-living. The labeled videos were used as ground truth for training an eXtreme Gradient Boosting classifier using window lengths of 1, 3, and 5 s. Performance of the classifier was evaluated using leave-one-out cross-validation. Results: Total recording time was ∼38 hr. Based on 5-s windowing, the overall accuracy was 96% for the dual accelerometer setup and 93% and 84% for the single thigh and back accelerometer setups, respectively. The decreased accuracy for the single accelerometer setup was due to a poor precision in detecting lying based on the thigh accelerometer recording (77%) and standing based on the back accelerometer recording (64%). Conclusion: Key daily physical activity types can be accurately detected during free-living based on dual accelerometer recording, using an eXtreme Gradient Boosting classifier. The overall accuracy decreases marginally when predictions are based on single thigh accelerometer recording, but detection of lying is poor.
IntroductionPhysical activity among children and adolescents remains insufficient, despite the substantial efforts made by researchers and policymakers. Identifying and furthering our understanding of potential modifiable determinants of physical activity behaviour (PAB) and sedentary behaviour (SB) is crucial for the development of interventions that promote a shift from SB to PAB. The current protocol details the process through which a series of systematic literature reviews and meta-analyses (MAs) will be conducted to produce a best-evidence statement (BESt) and inform policymakers. The overall aim is to identify modifiable determinants that are associated with changes in PAB and SB in children and adolescents (aged 5–19 years) and to quantify their effect on, or association with, PAB/SB.Methods and analysisA search will be performed in MEDLINE, SportDiscus, Web of Science, PsychINFO and Cochrane Central Register of Controlled Trials. Randomised controlled trials (RCTs) and controlled trials (CTs) that investigate the effect of interventions on PAB/SB and longitudinal studies that investigate the associations between modifiable determinants and PAB/SB at multiple time points will be sought. Risk of bias assessments will be performed using adapted versions of Cochrane’s RoB V.2.0 and ROBINS-I tools for RCTs and CTs, respectively, and an adapted version of the National Institute of Health’s tool for longitudinal studies. Data will be synthesised narratively and, where possible, MAs will be performed using frequentist and Bayesian statistics. Modifiable determinants will be discussed considering the settings in which they were investigated and the PAB/SB measurement methods used.Ethics and disseminationNo ethical approval is needed as no primary data will be collected. The findings will be disseminated in peer-reviewed publications and academic conferences where possible. The BESt will also be shared with policy makers within the DE-PASS consortium in the first instance.Systematic review registrationCRD42021282874.
Background Clinical guidelines for nonspecific low back pain (LBP) recommend self-management tailored to individual needs and capabilities as a first-line treatment. Mobile health solutions are a promising method for delivering tailored self-management interventions to patients with nonspecific LBP. However, it is not clear if the effectiveness of such self-management interventions depends on patients’ initial pain characteristics. High pain intensity and long-term symptoms of LBP have been associated with an unfavorable prognosis, and current best evidence indicates that long-term LBP (lasting more than 3 months) requires a more extensive treatment approach compared to more acute LBP. The artificial intelligence–based selfBACK app supports tailored and evidence-based self-management of nonspecific LBP. In a recent randomized controlled trial, we showed that individuals who received the selfBACK app in addition to usual care had lower LBP-related disability at the 3-month follow-up compared to those who received usual care only. This effect was sustained at 6 and 9 months. Objective This study aims to explore if the baseline duration and intensity of LBP influence the effectiveness of the selfBACK intervention in a secondary analysis of the selfBACK randomized controlled trial. Methods In the selfBACK trial, 461 adults (18 years or older) who sought care for nonspecific LBP in primary care or at an outpatient spine clinic were randomized to receive the selfBACK intervention adjunct to usual care (n=232) or usual care alone (n=229). In this secondary analysis, the participants were stratified according to the duration of the current LBP episode at baseline (≤12 weeks vs >12 weeks) or baseline LBP intensity (≤5 points vs >5 points) measured by a 0-10 numeric rating scale. The outcomes were LBP-related disability measured by the Roland-Morris Disability Questionnaire (0- to 24-point scale), average LBP intensity, pain self-efficacy, and global perceived effect. To assess whether the duration and intensity of LBP influenced the effect of selfBACK, we estimated the difference in treatment effect between the strata at the 3- and 9-month follow-ups with a 95% CI. Results Overall, there was no difference in effect for patients with different durations or intensities of LBP at either the 3- or 9-month follow-ups. However, there was suggestive evidence that the effect of the selfBACK intervention on LBP-related disability at the 3-month follow-up was largely confined to people with the highest versus the lowest LBP intensity (mean difference between the intervention and control group −1.8, 95% CI −3.0 to −0.7 vs 0.2, 95% CI −1.1 to 0.7), but this was not sustained at the 9-month follow-up. Conclusions The results suggest that the intensity and duration of LBP have negligible influence on the effectiveness of the selfBACK intervention on LBP-related disability, average LBP intensity, pain self-efficacy, and global perceived effect.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.