A Movement Classification of Polymyalgia Rheumatica Patients Using Myoelectric Sensors.

Gait disorder is common among people with neurological disease and musculoskeletal disorders. The detection of gait disorders plays an integral role in designing appropriate rehabilitation protocols. This study presents a clinical gait analysis of patients with polymyalgia rheumatica to determine impaired gait patterns using machine learning models. A clinical gait assessment was conducted at KATH hospital between August and September 2022, and the 25 recruited participants comprised 18 patients and 7 control subjects. The demographics of the participants follow: age 56 years ± 7, height 175 cm ± 8, and weight 82 kg ± 10. Electromyography data were collected from four strained hip muscles of patients, which were the rectus femoris, vastus lateralis, biceps femoris, and semitendinosus. Four classification models were used-namely, support vector machine (SVM), rotation forest (RF), k-nearest neighbors (KNN), and decision tree (DT)-to distinguish the gait patterns for the two groups. SVM recorded the highest accuracy of 85% among the classifiers, while KNN had 75%, RF had 80%, and DT had the lowest accuracy of 70%. Furthermore, the SVM classifier had the highest sensitivity of 92%, while RF had 86%, DT had 90%, and KNN had the lowest sensitivity of 84%. The classifiers achieved significant results in discriminating between the impaired gait pattern of patients with polymyalgia rheumatica and control subjects. This information could be useful for clinicians designing therapeutic exercises and may be used for developing a decision support system for diagnostic purposes.

Simplified Deep Reinforcement Learning Approach for Channel Prediction in Power Domain NOMA System.

In this work, the impact of implementing Deep Reinforcement Learning (DRL) in predicting the channel parameters for user devices in a Power Domain Non-Orthogonal Multiple Access system (PD-NOMA) is investigated. In the channel prediction process, DRL based on deep Q networks (DQN) algorithm will be developed and incorporated into the NOMA system so that this developed DQN model can be employed to estimate the channel coefficients for each user device in NOMA system. The developed DQN scheme will be structured as a simplified approach to efficiently predict the channel parameters for each user in order to maximize the downlink sum rates for all users in the system. In order to approximate the channel parameters for each user device, this proposed DQN approach is first initialized using random channel statistics, and then the proposed DQN model will be dynamically updated based on the interaction with the environment. The predicted channel parameters will be utilized at the receiver side to recover the desired data. Furthermore, this work inspects how the channel estimation process based on the simplified DQN algorithm and the power allocation policy, can both be integrated for the purpose of multiuser detection in the examined NOMA system. Simulation results, based on several performance metrics, have demonstrated that the proposed simplified DQN algorithm can be a competitive algorithm for channel parameters estimation when compared to different benchmark schemes for channel estimation processes such as deep neural network (DNN) based long-short term memory (LSTM), RL based Q algorithm, and channel estimation scheme based on minimum mean square error (MMSE) procedure.

A Study on the Impact of Integrating Reinforcement Learning for Channel Prediction and Power Allocation Scheme in MISO-NOMA System.

In this study, the influence of adopting Reinforcement Learning (RL) to predict the channel parameters for user devices in a Power Domain Multi-Input Single-Output Non-Orthogonal Multiple Access (MISO-NOMA) system is inspected. In the channel prediction-based RL approach, the Q-learning algorithm is developed and incorporated into the NOMA system so that the developed Q-model can be employed to predict the channel coefficients for every user device. The purpose of adopting the developed Q-learning procedure is to maximize the received downlink sum-rate and decrease the estimation loss. To satisfy this aim, the developed Q-algorithm is initialized using different channel statistics and then the algorithm is updated based on the interaction with the environment in order to approximate the channel coefficients for each device. The predicted parameters are utilized at the receiver side to recover the desired data. Furthermore, based on maximizing the sum-rate of the examined user devices, the power factors for each user can be deduced analytically to allocate the optimal power factor for every user device in the system. In addition, this work inspects how the channel prediction based on the developed Q-learning model, and the power allocation policy, can both be incorporated for the purpose of multiuser recognition in the examined MISO-NOMA system. Simulation results, based on several performance metrics, have demonstrated that the developed Q-learning algorithm can be a competitive algorithm for channel estimation when compared to different benchmark schemes such as deep learning-based long short-term memory (LSTM), RL based actor-critic algorithm, RL based state-action-reward-state-action (SARSA) algorithm, and standard channel estimation scheme based on minimum mean square error procedure.

Special Issue "Advanced Signal Processing in Wearable Sensors for Health Monitoring".

Wearable sensors are becoming very popular recently due to their ease of use and flexibility in recording data from home [...].

Investigating the Combination of Deep Learning for Channel Estimation and Power Optimization in a Non-Orthogonal Multiple Access System.

In a non-orthogonal multiple access (NOMA) system, the successive interference cancellation (SIC) procedure is typically employed at the receiver side, where several user's signals are decoded in a subsequent manner. Fading channels may disperse the transmitted signal and originate dependencies among its samples, which may affect the channel estimation procedure and consequently affect the SIC process and signal detection accuracy. In this work, the impact of Deep Neural Network (DNN) in explicitly estimating the channel coefficients for each user in NOMA cell is investigated in both Rayleigh and Rician fading channels. The proposed approach integrates the Long Short-Term Memory (LSTM) network into the NOMA system where this LSTM network is utilized to predict the channel coefficients. DNN is trained using different channel statistics and then utilized to predict the desired channel parameters that will be exploited by the receiver to retrieve the original data. Furthermore, this work examines how the channel estimation based on Deep Learning (DL) and power optimization scheme are jointly utilized for multiuser (MU) recognition in downlink Power Domain Non-Orthogonal Multiple Access (PD-NOMA) system. Power factors are optimized with a view to maximize the sum rate of the users on the basis of entire power transmitted and Quality of service (QoS) constraints. An investigation for the optimization problem is given where Lagrange function and Karush-Kuhn-Tucker (KKT) optimality conditions are applied to deduce the optimum power coefficients. Simulation results for different metrics, such as bit error rate (BER), sum rate, outage probability and individual user capacity, have proved the superiority of the proposed DL-based channel estimation over conventional NOMA approach. Additionally, the performance of optimized power scheme and fixed power scheme are evaluated when DL-based channel estimation is implemented.

Comparison of Deep Learning Algorithms in Predicting Expert Assessments of Pain Scores during Surgical Operations Using Analgesia Nociception Index.

There are many surgical operations performed daily in operation rooms worldwide. Adequate anesthesia is needed during an operation. Besides hypnosis, adequate analgesia is critical to prevent autonomic reactions. Clinical experience and vital signs are usually used to adjust the dosage of analgesics. Analgesia nociception index (ANI), which ranges from 0 to 100, is derived from heart rate variability (HRV) via electrocardiogram (ECG) signals, for pain evaluation in a non-invasive manner. It represents parasympathetic activity. In this study, we compared the performance of multilayer perceptron (MLP) and long short-term memory (LSTM) algorithms in predicting expert assessment of pain score (EAPS) based on patient's HRV during surgery. The objective of this study was to analyze how deep learning models differed from the medical doctors' predictions of EAPS. As the input and output features of the deep learning models, the opposites of ANI and EAPS were used. This study included 80 patients who underwent operations at National Taiwan University Hospital. Using MLP and LSTM, a holdout method was first applied to 60 training patients, 10 validation patients, and 10 testing patients. As compared to the LSTM model, which had a testing mean absolute error (MAE) of 2.633 ± 0.542, the MLP model had a testing MAE of 2.490 ± 0.522, with a more appropriate shape of its prediction curves. The model based on MLP was selected as the best. Using MLP, a seven-fold cross validation method was then applied. The first fold had the lowest testing MAE of 2.460 ± 0.634, while the overall MAE for the seven-fold cross validation method was 2.848 ± 0.308. In conclusion, HRV analysis using MLP algorithm had a good correlation with EAPS; therefore, it can play role as a continuous monitor to predict intraoperative pain levels, to assist physicians in adjusting analgesic agent dosage. Further studies may consider obtaining more input features, such as photoplethysmography (PPG) and other kinds of continuous variable, to improve the prediction performance.

ECG Recurrence Plot-Based Arrhythmia Classification Using Two-Dimensional Deep Residual CNN Features.

In this paper, an effective electrocardiogram (ECG) recurrence plot (RP)-based arrhythmia classification algorithm that can be implemented in portable devices is presented. Public databases from PhysioNet were used to conduct this study including the MIT-BIH Atrial Fibrillation Database, the MIT-BIH Arrhythmia Database, the MIT-BIH Malignant Ventricular Ectopy Database, and the Creighton University Ventricular Tachyarrhythmia Database. ECG time series were segmented and converted using an RP, and two-dimensional images were used as inputs to the CNN classifiers. In this study, two-stage classification is proposed to improve the accuracy. The ResNet-18 architecture was applied to detect ventricular fibrillation (VF) and noise during the first stage, whereas normal, atrial fibrillation, premature atrial contraction, and premature ventricular contractions were detected using ResNet-50 in the second stage. The method was evaluated using 5-fold cross-validation which improved the results when compared to previous studies, achieving first and second stage average accuracies of 97.21% and 98.36%, sensitivities of 96.49% and 97.92%, positive predictive values of 95.54% and 98.20%, and F1-scores of 95.96% and 98.05%, respectively. Furthermore, a 5-fold improvement in the memory requirement was achieved when compared with a previous study, making this classifier feasible for use in resource-constricted environments such as portable devices. Even though the method is successful, first stage training requires combining four different arrhythmia types into one label (other), which generates more data for the other category than for VF and noise, thus creating a data imbalance that affects the first stage performance.

Pain and Stress Detection Using Wearable Sensors and Devices-A Review.

Pain is a subjective feeling; it is a sensation that every human being must have experienced all their life. Yet, its mechanism and the way to immune to it is still a question to be answered. This review presents the mechanism and correlation of pain and stress, their assessment and detection approach with medical devices and wearable sensors. Various physiological signals (i.e., heart activity, brain activity, muscle activity, electrodermal activity, respiratory, blood volume pulse, skin temperature) and behavioral signals are organized for wearables sensors detection. By reviewing the wearable sensors used in the healthcare domain, we hope to find a way for wearable healthcare-monitoring system to be applied on pain and stress detection. Since pain leads to multiple consequences or symptoms such as muscle tension and depression that are stress related, there is a chance to find a new approach for chronic pain detection using daily life sensors or devices. Then by integrating modern computing techniques, there is a chance to handle pain and stress management issue.

Pedestrian and Vehicle Detection in Autonomous Vehicle Perception Systems-A Review.

Autonomous Vehicles (AVs) have the potential to solve many traffic problems, such as accidents, congestion and pollution. However, there are still challenges to overcome, for instance, AVs need to accurately perceive their environment to safely navigate in busy urban scenarios. The aim of this paper is to review recent articles on computer vision techniques that can be used to build an AV perception system. AV perception systems need to accurately detect non-static objects and predict their behaviour, as well as to detect static objects and recognise the information they are providing. This paper, in particular, focuses on the computer vision techniques used to detect pedestrians and vehicles. There have been many papers and reviews on pedestrians and vehicles detection so far. However, most of the past papers only reviewed pedestrian or vehicle detection separately. This review aims to present an overview of the AV systems in general, and then review and investigate several detection computer vision techniques for pedestrians and vehicles. The review concludes that both traditional and Deep Learning (DL) techniques have been used for pedestrian and vehicle detection; however, DL techniques have shown the best results. Although good detection results have been achieved for pedestrians and vehicles, the current algorithms still struggle to detect small, occluded, and truncated objects. In addition, there is limited research on how to improve detection performance in difficult light and weather conditions. Most of the algorithms have been tested on well-recognised datasets such as Caltech and KITTI; however, these datasets have their own limitations. Therefore, this paper recommends that future works should be implemented on more new challenging datasets, such as PIE and BDD100K.

Non-Invasive Hemodynamics Monitoring System Based on Electrocardiography via Deep Convolutional Autoencoder.

This study evaluates cardiovascular and cerebral hemodynamics systems by only using non-invasive electrocardiography (ECG) signals. The Massachusetts General Hospital/Marquette Foundation (MGH/MF) and Cerebral Hemodynamic Autoregulatory Information System Database (CHARIS DB) from the PhysioNet database are used for cardiovascular and cerebral hemodynamics, respectively. For cardiovascular hemodynamics, the ECG is used for generating the arterial blood pressure (ABP), central venous pressure (CVP), and pulmonary arterial pressure (PAP). Meanwhile, for cerebral hemodynamics, the ECG is utilized for the intracranial pressure (ICP) generator. A deep convolutional autoencoder system is applied for this study. The cross-validation method with Pearson's linear correlation (R), root mean squared error (RMSE), and mean absolute error (MAE) are measured for the evaluations. Initially, the ECG is used to generate the cardiovascular waveform. For the ABP system-the systolic blood pressure (SBP) and diastolic blood pressures (DBP)-the R evaluations are 0.894 ± 0.004 and 0.881 ± 0.005, respectively. The MAE evaluations for SBP and DBP are, respectively, 6.645 ± 0.353 mmHg and 3.210 ± 0.104 mmHg. Furthermore, for the PAP system-the systolic and diastolic pressures-the R evaluations are 0.864 ± 0.003 mmHg and 0.817 ± 0.006 mmHg, respectively. The MAE evaluations for systolic and diastolic pressures are, respectively, 3.847 ± 0.136 mmHg and 2.964 ± 0.181 mmHg. Meanwhile, the mean CVP evaluations are 0.916 ± 0.001, 2.220 ± 0.039 mmHg, and 1.329 ± 0.036 mmHg, respectively, for R, RMSE, and MAE. For the mean ICP evaluation in cerebral hemodynamics, the R and MAE evaluations are 0.914 ± 0.003 and 2.404 ± 0.043 mmHg, respectively. This study, as a proof of concept, concludes that the non-invasive cardiovascular and cerebral hemodynamics systems can be potentially investigated by only using the ECG signal.

Genetic Deep Convolutional Autoencoder Applied for Generative Continuous Arterial Blood Pressure via Photoplethysmography.

Hypertension affects a huge number of people around the world. It also has a great contribution to cardiovascular- and renal-related diseases. This study investigates the ability of a deep convolutional autoencoder (DCAE) to generate continuous arterial blood pressure (ABP) by only utilizing photoplethysmography (PPG). A total of 18 patients are utilized. LeNet-5- and U-Net-based DCAEs, respectively abbreviated LDCAE and UDCAE, are compared to the MP60 IntelliVue Patient Monitor, as the gold standard. Moreover, in order to investigate the data generalization, the cross-validation (CV) method is conducted. The results show that the UDCAE provides superior results in producing the systolic blood pressure (SBP) estimation. Meanwhile, the LDCAE gives a slightly better result for the diastolic blood pressure (DBP) prediction. Finally, the genetic algorithm-based optimization deep convolutional autoencoder (GDCAE) is further administered to optimize the ensemble of the CV models. The results reveal that the GDCAE is superior to either the LDCAE or UDCAE. In conclusion, this study exhibits that systolic blood pressure (SBP) and diastolic blood pressure (DBP) can also be accurately achieved by only utilizing a single PPG signal.

MR Image-Based Attenuation Correction of Brain PET Imaging: Review of Literature on Machine Learning Approaches for Segmentation.

Recent emerging hybrid technology of positron emission tomography/magnetic resonance (PET/MR) imaging has generated a great need for an accurate MR image-based PET attenuation correction. MR image segmentation, as a robust and simple method for PET attenuation correction, has been clinically adopted in commercial PET/MR scanners. The general approach in this method is to segment the MR image into different tissue types, each assigned an attenuation constant as in an X-ray CT image. Machine learning techniques such as clustering, classification and deep networks are extensively used for brain MR image segmentation. However, only limited work has been reported on using deep learning in brain PET attenuation correction. In addition, there is a lack of clinical evaluation of machine learning methods in this application. The aim of this review is to study the use of machine learning methods for MR image segmentation and its application in attenuation correction for PET brain imaging. Furthermore, challenges and future opportunities in MR image-based PET attenuation correction are discussed.

Special Issue "Advanced Signal Processing in Intelligent Systems for Health Monitoring".

Recently, significant developments have been achieved in the field of artificial intelligence, in particular the introduction of deep learning technology that has improved the learning and prediction accuracy to unpresented levels, especially when dealing with big data and high-resolution images. Significant developments have occurred in the area of medical signal processing, measurement techniques, and health monitoring, such as vital biological signs for biomedical systems and noise and vibration of mechanical systems, which are carried out by instruments that generate large data sets. These big data sets, ultimately driven by high population growth, would require Artificial Intelligence techniques to analyse and model. In this Special Issue, papers are presented on the latest signal processing and deep learning techniques used for health monitoring of biomedical and mechanical systems.

Integrations between Autonomous Systems and Modern Computing Techniques: A Mini Review.

The emulation of human behavior for autonomous problem solving has been an interdisciplinary field of research. Generally, classical control systems are used for static environments, where external disturbances and changes in internal parameters can be fully modulated before or neglected during operation. However, classical control systems are inadequate at addressing environmental uncertainty. By contrast, autonomous systems, which were first studied in the field of control systems, can be applied in an unknown environment. This paper summarizes the state of the art autonomous systems by first discussing the definition, modeling, and system structure of autonomous systems and then providing a perspective on how autonomous systems can be integrated with advanced resources (e.g., the Internet of Things, big data, Over-the-Air, and federated learning). Finally, what comes after reaching full autonomy is briefly discussed.

Machine Learning Methods Applied to Predict Ventilator-Associated Pneumonia with Pseudomonas aeruginosa Infection via Sensor Array of Electronic Nose in Intensive Care Unit.

One concern to the patients is the off-line detection of pneumonia infection status after using the ventilator in the intensive care unit. Hence, machine learning methods for ventilator-associated pneumonia (VAP) rapid diagnose are proposed. A popular device, Cyranose 320 e-nose, is usually used in research on lung disease, which is a highly integrated system and sensor comprising 32 array using polymer and carbon black materials. In this study, a total of 24 subjects were involved, including 12 subjects who are infected with pneumonia, and the rest are non-infected. Three layers of back propagation artificial neural network and support vector machine (SVM) methods were applied to patients' data to predict whether they are infected with VAP with Pseudomonas aeruginosa infection. Furthermore, in order to improve the accuracy and the generalization of the prediction models, the ensemble neural networks (ENN) method was applied. In this study, ENN and SVM prediction models were trained and tested. In order to evaluate the models' performance, a fivefold cross-validation method was applied. The results showed that both ENN and SVM models have high recognition rates of VAP with Pseudomonas aeruginosa infection, with 0.9479 ± 0.0135 and 0.8686 ± 0.0422 accuracies, 0.9714 ± 0.0131, 0.9250 ± 0.0423 sensitivities, and 0.9288 ± 0.0306, 0.8639 ± 0.0276 positive predictive values, respectively. The ENN model showed better performance compared to SVM in the recognition of VAP with Pseudomonas aeruginosa infection. The areas under the receiver operating characteristic curve of the two models were 0.9842 ± 0.0058 and 0.9410 ± 0.0301, respectively, showing that both models are very stable and accurate classifiers. This study aims to assist the physician in providing a scientific and effective reference for performing early detection in Pseudomonas aeruginosa infection or other diseases.

Design and Evaluation of a Real Time Physiological Signals Acquisition System Implemented in Multi-Operating Rooms for Anesthesia.

With critical importance of medical healthcare, there exist urgent needs for in-depth medical studies that can access and analyze specific physiological signals to provide theoretical support for practical clinical care. As a consequence, obtaining the valuable medical data with minimal cost and impacts on hospital work comes as the first concern of researchers. Anesthesia plays a widely recognized role in surgeries, which attracts people to undertake relevant research. In this paper, a real-time physiological medical signal data acquisition system (PMSDA) for the multi-operating room applications is proposed with high universality of the hospital practical settings and research requirements. By utilizing a wireless communication approach, it provides an easily accessible network platform for collection of physiological medical signals such as photoplethysmogram (PPG), electrocardiograph (ECG) and electroencephalogram (EEG) during the surgery. In addition, the raw data is stored on a server for safe backup and further analysis of depth of anesthesia (DoA). Results show that the PMSDA exhibits robust, high quality performance and efficiently reduces costs compared to previously manual methods and allows seamless integration into hospital environment, independent of its routine work. Overall, it provides a pragmatic and flexible surgery-data acquisition system model with low impact and resource cost applicable to research in critical and practical medical circumstances.

Arrhythmia Evaluation in Wearable ECG Devices.

This study evaluates four databases from PhysioNet: The American Heart Association database (AHADB), Creighton University Ventricular Tachyarrhythmia database (CUDB), MIT-BIH Arrhythmia database (MITDB), and MIT-BIH Noise Stress Test database (NSTDB). The ANSI/AAMI EC57:2012 is used for the evaluation of the algorithms for the supraventricular ectopic beat (SVEB), ventricular ectopic beat (VEB), atrial fibrillation (AF), and ventricular fibrillation (VF) via the evaluation of the sensitivity, positive predictivity and false positive rate. Sample entropy, fast Fourier transform (FFT), and multilayer perceptron neural network with backpropagation training algorithm are selected for the integrated detection algorithms. For this study, the result for SVEB has some improvements compared to a previous study that also utilized ANSI/AAMI EC57. In further, VEB sensitivity and positive predictivity gross evaluations have greater than 80%, except for the positive predictivity of the NSTDB database. For AF gross evaluation of MITDB database, the results show very good classification, excluding the episode sensitivity. In advanced, for VF gross evaluation, the episode sensitivity and positive predictivity for the AHADB, MITDB, and CUDB, have greater than 80%, except for MITDB episode positive predictivity, which is 75%. The achieved results show that the proposed integrated SVEB, VEB, AF, and VF detection algorithm has an accurate classification according to ANSI/AAMI EC57:2012. In conclusion, the proposed integrated detection algorithm can achieve good accuracy in comparison with other previous studies. Furthermore, more advanced algorithms and hardware devices should be performed in future for arrhythmia detection and evaluation.

A comparison of different methods to maximise signal extraction when using central venous pressure to optimise atrioventricular delay after cardiac surgery.

Objective: Our group has shown that central venous pressure (CVP) can optimise atrioventricular (AV) delay in temporary pacing (TP) after cardiac surgery. However, the signal-to-noise ratio (SNR) is influenced both by the methods used to mitigate the pressure effects of respiration and the number of heartbeats analysed. This paper systematically studies the effect of different analysis methods on SNR to maximise the accuracy of this technique. Methods: We optimised AV delay in 16 patients with TP after cardiac surgery. Transitioning rapidly and repeatedly from a reference AV delay to different tested AV delays, we measured pressure differences before and after each transition. We analysed the resultant signals in different ways with the aim of maximising the SNR: (1) adjusting averaging window location (around versus after transition), (2) modifying window length (heartbeats analysed), and (3) applying different signal filtering methods to correct respiratory artefact. Results: (1) The SNR was 27 % higher for averaging windows around the transition versus post-transition windows. (2) The optimal window length for CVP analysis was two respiratory cycle lengths versus one respiratory cycle length for optimising SNR for arterial blood pressure (ABP) signals. (3) Filtering with discrete wavelet transform improved SNR by 62 % for CVP measurements. When applying the optimal window length and filtering techniques, the correlation between ABP and CVP peak optima exceeded that of a single cycle length (R = 0.71 vs. R = 0.50, p < 0.001). Conclusion: We demonstrated that utilising a specific set of techniques maximises the signal-to-noise ratio and hence the utility of this technique.

Disease specific mortality in patients with low risk bladder cancer and the impact of cystoscopic surveillance.

PURPOSE: We determined the risk of disease specific mortality in patients with primary, low risk, noninvasive (G1pTa) bladder cancer and compared it to disease specific mortality in age and gender matched general populations. MATERIALS AND METHODS: We identified all patients with primary low risk cancer at our institution. We excluded those with adverse pathological features and then matched histopathology, pharmacy, hospital episode and Cancer Registry records. We reviewed case notes on patients with subsequent muscle invasion (progression) or disease specific mortality. Patients underwent post-resection surveillance and treatment using standard regimens. National and regional disease specific mortality rates were calculated from appropriate data. RESULTS: A total of 699 patients met study inclusion criteria. Median followup was 61 months (IQR 24-105). Of the patients 17 (2.4%) died of bladder cancer, including 13 of 14 with progression to muscle invasion and 4 of 19 with grade progression to high grade, nonmuscle invasive disease. On Cox regression analyses low grade dysplasia in the initial resection specimen and tumor weight were associated with disease specific mortality (p <0.003). Disease specific mortality in these patients was 5 times the background rate in matched populations. Limitations of this study include its retrospective nature and the low frequency of adverse events. CONCLUSIONS: Patients with low risk bladder cancer rarely progress to muscle invasion but they are at higher risk for disease specific mortality than the general population. Current surveillance regimens appear ineffective for detecting progression in time to alter prognosis.

Hip fracture risk assessment: artificial neural network outperforms conditional logistic regression in an age- and sex-matched case control study.

BACKGROUND: Osteoporotic hip fractures with a significant morbidity and excess mortality among the elderly have imposed huge health and economic burdens on societies worldwide. In this age- and sex-matched case control study, we examined the risk factors of hip fractures and assessed the fracture risk by conditional logistic regression (CLR) and ensemble artificial neural network (ANN). The performances of these two classifiers were compared. METHODS: The study population consisted of 217 pairs (149 women and 68 men) of fractures and controls with an age older than 60 years. All the participants were interviewed with the same standardized questionnaire including questions on 66 risk factors in 12 categories. Univariate CLR analysis was initially conducted to examine the unadjusted odds ratio of all potential risk factors. The significant risk factors were then tested by multivariate analyses. For fracture risk assessment, the participants were randomly divided into modeling and testing datasets for 10-fold cross validation analyses. The predicting models built by CLR and ANN in modeling datasets were applied to testing datasets for generalization study. The performances, including discrimination and calibration, were compared with non-parametric Wilcoxon tests. RESULTS: In univariate CLR analyses, 16 variables achieved significant level, and six of them remained significant in multivariate analyses, including low T score, low BMI, low MMSE score, milk intake, walking difficulty, and significant fall at home. For discrimination, ANN outperformed CLR in both 16- and 6-variable analyses in modeling and testing datasets (p?