Reduction of false arterial blood pressure alarms using signal quality assessement and relationships between the electrocardiogram and arterial blood pressure

The paper presents an algorithm for reducing false alarms related to changes in arterial blood pressure (ABP) in intensive care unit (ICU) monitoring. The algorithm assesses the ABP signal quality, analyses the relationship between the electrocardiogram and ABP using a fuzzy logic approach and post-processes (accepts or rejects) ABP alarms produced by a commerical monitor. The algorithm was developed and evaluated using unrelated sets of data from the MIMIC database. By rejecting 98.2% (159 of 162) of the false ABP alarms produced by the monitor using the test set of data, the algorithm was able to reduce the false ABP alarm rate from 26.8% to 0.5% of ABP alarms, while accepting 99.8% (441 of 442) of true ABP alarms. The results show that the algorithm is effective and practical, and its use in future patient monitoring systems is feasible.     

利用脉搏波传播时间计算动脉血压的研究

本研究的目的是研究动脉血压和脉搏波传播时间的关系,探讨通过脉搏波传播时间计算动脉血压的可靠性。采用麻省理工学院MIMIC数据库,通过心电和光电容积脉搏波计算得到脉搏波传播时间,通过有创动脉血压获得平均动脉压,使用线性回归方法分段求得脉搏波传播时间和平均动脉压之间的线性方程,应用该方程结合脉搏波传播时间计算动脉血压,并与实际血压比较评价算法的效果。结果表明,脉搏波传播时间和动脉血压存在负相关关系,在一定时间范围内,可通过脉搏波传播时间计算平均动脉压,均方根误差小于5 mmHg。对临床采集数据的分析同样说明,通过脉搏波传播时间计算动脉血压是可行的。     

An arrhythmia classification system based on the RR-interval signal

OBJECTIVE: This paper proposes a knowledge-based method for arrhythmic beat classification and arrhythmic episode detection and classification using only the RR-interval signal extracted from ECG recordings. METHODOLOGY: A three RR-interval sliding window is used in arrhythmic beat classification algorithm. Classification is performed for four categories of beats: normal, premature ventricular contractions, ventricular flutter/fibrillation and 2 degrees heart block. The beat classification is used as input of a knowledge-based deterministic automaton to achieve arrhythmic episode detection and classification. Six rhythm types are classified: ventricular bigeminy, ventricular trigeminy, ventricular couplet, ventricular tachycardia, ventricular flutter/fibrillation and 2 degrees heart block. RESULTS: The method is evaluated by using the MIT-BIH arrhythmia database. The achieved scores indicate high performance: 98% accuracy for arrhythmic beat classification and 94% accuracy for arrhythmic episode detection and classification. CONCLUSION: The proposed method is advantageous because it uses only the RR-interval signal for arrhythmia beat and episode classification and the results compare well with more complex methods.     

Predictive combinations of monitor alarms preceding in-hospital code blue events

Bedside monitors are ubiquitous in acute care units of modern healthcare enterprises. However, they have been criticized for generating an excessive number of false positive alarms causing alarm fatigue among care givers and potentially compromising patient safety. We hypothesize that combinations of regular monitor alarms denoted as SuperAlarm set may be more indicative of ongoing patient deteriorations and hence predictive of in-hospital code blue events. The present work develops and assesses an alarm mining approach based on finding frequent combinations of single alarms that are also specific to code blue events to compose a SuperAlarm set. We use 4-way analysis of variance (ANOVA) to investigate the influence of four algorithm parameters on the performance of the data mining approach. The results are obtained from millions of monitor alarms from a cohort of 223 adult code blue and 1768 control patients using a multiple 10-fold cross-validation experiment setup. Using the optimal setting of parameters determined in the cross-validation experiment, final SuperAlarm sets are mined from the training data and used on an independent test data set to simulate running a SuperAlarm set against live regular monitor alarms. The ANOVA shows that the content of a SuperAlarm set is influenced by a subset of key algorithm parameters. Simulation of the extracted SuperAlarm set shows that it can predict code blue events one hour ahead with sensitivity between 66.7% and 90.9% while producing false SuperAlarms for control patients that account for between 2.2% and 11.2% of regular monitor alarms depending on user-supplied acceptable false positive rate. We conclude that even though the present work is still preliminary due to the usage of a moderately-sized database to test our hypothesis it represents an effort to develop algorithms to alleviate the alarm fatigue issue in a unique way.     

Design,implementation and evaluation of amicrocomputer-based portable arrhythmia monitor

The design of a portable, battery-operated microcomputer-based monitor for ambulatory ECG recording and analysis is described. Designed for real-time cardiac arrhythmia analysis, it is suitable for use on ambulator, patients for several weeks, and is about the size and weight of a Holter recorder. The device differs from a Holter recorder in that is does not store normal complexes but recognises and alarms on significant arrhythmias. It sotres 16 s of the arrhythmic event, which it can transmit by telephone to a central receiving station for immediate appraisal by a cardiologist. The monitor uses a CMOS microcomputer and has 2kbytes of program memory and 2kbytes of data memory. The arrhythmia monitor program recognises tachycardia, bradycardia, asystole, dropped beats, and PVCs. The alarm limits are physician programmable. The performance of the monitor was evaluated with standard annotated ECG tapes provided by MIT/BIH. This device should be useful for applications such as antiarrhythmic drug studies, for pacemaker and postsurgery evaluations, and for detecting premonitory as well as life-threatening arrhythmias.     

A practical reconstructed phase space approach for ECG arrhythmias classification

Abstract

Atrial and ventricular arrhythmias are symptoms of the main common causes of rapid death. The severity of these arrhythmias depends on their occurrence either within the atria or ventricles. These abnormalities of the heart activity may cause an immediate death or cause damage of the heart. In this paper, a new algorithm is proposed for the classification of life threatening cardiac arrhythmias including atrial fibrillation (AF), ventricular tachycardia (VT) and ventricular fibrillation (VF). The proposed technique uses a simple signal processing technique for analysing the non-linear dynamics of the ECG signals in the time domain. The classification algorithm is based upon the distribution of the attractor in the reconstructed phase space (RPS). The behaviour of the ECG signal in the reconstructed phase space is used to determine the classification features of the whole classifier. It is found that different arrhythmias occupy different regions in the reconstructed phase space. Three regions in the RPS are found to be more representative of the considered arrhythmias. Therefore, only three simple features are extracted to be used as classification parameters. To evaluate the performance of the presented classification algorithm, real datasets are obtained from the MIT database. A learning dataset is used to design the classification algorithm and a testing dataset is used to verify the algorithm. The algorithm is designed to guarantee achieving both 100% sensitivity and 100% specificity. The classification algorithm is validated by using 45 ECG signals spanning the considered life threatening arrhythmias. The obtained results show that the classification algorithm attains a sensitivity ranging from 85.7–100%, a specificity ranging from 86.7–100% and an overall accuracy of 95.55%.     

A fuzzy logic based apnoea monitor for SIDS risk infants

A unit has been designed which monitors newborn infants at risk of Sudden Infant Death Syndrome (SIDS) in a home environment. The unit monitors respiration, electrocardiogram (ECG) and haemoglobin oxygen saturation (SpO₂) in combination, in order to detect any potentially life threatening event at an early stage. Provision is made for the generation of both audible and silent alarms and for the storage of signals and other information before, during and after an alarm episode for diagnostic purposes. An intelligent fuzzy logic algorithm is used to process the signals monitored and to implement several propositions concerning their status in order to determine the probability of an apnoea event and initiate the appropriate action. This has substantially reduced the number of false alarms and of undetected dangerous situations compared with previous units, which greatly improves the reliability and usefulness of such a monitor.     

Integrated monitoring can detect critical events and improve alarm accuracy

A computer-based, integrated monitor system was designed and utilized to collect and interactively manage physiologic data (13 variables and 3 waveforms) from six routinely used operating room monitors. Various approaches were developed to reduce false alarms, classify waveforms, and recognize events. False alarms: false alarms in ECG heart rate detection were reduced from 37.3% to 2.6% (p=0.005) of total alarms using multi-variable analysis and rate-of-change limits. Waveform classification: using artificial neural networks (AN), CO2 waveforms were classified into (a) spontaneous, (b) mechanical, and (c) mechanical/with spontaneous breathing attempts. The system properly classified 47 of 71 spontaneous, 65 of 67 mechanical, and 37 of 44 mechanical breaths/with spontaneous breathing attempts. Another ANN was used for detection of elevated and depressed ST segments in the ECG signal. All ST segment elevations and depressions of 0.1 mV were correctly identified. Event recognition: an algorithm developed to identify endotracheal intubation correctly recognized 13 of 17 intubations. This resulted in a 42% reduction in low end-tidal-CO2 false alarms.     

A fuzzy clustering neural network architecture for classification of ECG arrhythmias

Accurate and computationally efficient means of classifying electrocardiography (ECG) arrhythmias has been the subject of considerable research effort in recent years. This study presents a comparative study of the classification accuracy of ECG signals using a well-known neural network architecture named multi-layered perceptron (MLP) with backpropagation training algorithm, and a new fuzzy clustering NN architecture (FCNN) for early diagnosis. The ECG signals are taken from MIT-BIH ECG database, which are used to classify 10 different arrhythmias for training. These are normal sinus rhythm, sinus bradycardia, ventricular tachycardia, sinus arrhythmia, atrial premature contraction, paced beat, right bundle branch block, left bundle branch block, atrial fibrillation and atrial flutter. For testing, the proposed structures were trained by backpropagation algorithm. Both of them tested using experimental ECG records of 92 patients (40 male and 52 female, average age is 39.75 +/- 19.06). The test results suggest that a new proposed FCNN architecture can generalize better than ordinary MLP architecture and also learn better and faster. The advantage of proposed structure is a result of decreasing the number of segments by grouping similar segments in training data with fuzzy c-means clustering.     

Characterisation of electrocardiogram signals based on blind source separation

Blind source separation assumes that the acquired signal is composed of a weighted sum of a number of basic components corresponding to a number of limited sources. This work poses the problem of ECG signal diagnosis in the form of a blind source separation problem. In particular, a large number of ECG signals undergo two of the most commonly used blind source separation techniques, namely, principal component analysis (PCA) and independent component analysis (ICA), so that the basic components underlying this complex signal can be identified. Given that such techniques are sensitive to signal shift, a simple transformation is used that computes the magnitude of the Fourier transformation of ECG signals. This allows the phase components corresponding to such shifts to be removed. Using the magnitude of the projection of a given ECG signal onto these basic components as features, it was shown that accurate arrhythmia detection and classification were possible. The proposed strategies were applied to a large number of independent 3s intervals of ECG signals consisting of 320 training samples and 160 test samples from the MIT-BIH database. The samples equally represent five different ECG signal types, including normal, ventricular couplet, ventricular tachycardia, ventricular bigeminy and ventricular fibrillation. The intervals analysed were windowed using either a rectangular or a Hamming window. The methods demonstrated a detection rate of sensitivity 98% at specificity of 100% using nearest neighbour classification of features from ICA and a rectangular window. Lower classification rates were obtained using the same classifier with features from either PCA or ICA and a rectangular window. The results demonstrate the potential of the new method for clinical use.     

Risk evaluation of ventricular tachycardia using wavelet transform irregularity of the high-resolution electrocardiogram

A new method for analysis of high-resolution ECG signals using a wavelet transform based on a modified Morlet function is presented. A polynomial filter is used to reduce low-frequency, high-amplitude noise components in the analysed signals. The method is tested on test ECG signals with simulated late potentials and finally verified on two post-infarction patient (PP) groups: 62 PPs with ventricular tachycardia and 44 PPs without arrhythmia. A new quantitative parameter, the irregularity factor, is proposed for discrimination between the study groups. The results show a significant difference in the parameter values for tachycardia patients compared with those for patients without arrhythmia. The sensitivity of the proposed method is 85%, and the specificity is 93%.     

A Case of Catecholaminergic Polymorphic Ventricular Tachycardia

Catecholaminergic polymorphic ventricular tachycardia (CPVT) is a familial cardiac arrhythmia that is related to RYR2 or CASQ2 gene mutation. It occurs in patients with structurally normal heart and causes exercise-emotion-triggered syncope and sudden cardiac death. We experienced a case of CPVT in an 11 year-old female patient who was admitted for sudden cardiovascular collapse. The initial electrocardiogram (ECG) on emergency department revealed ventricular fibrillation. After multiple defibrillations, sinus rhythm was restored. However, recurrent ventricular fibrillation occurred during insertion of nasogastric tube without sedation in coronary care unit. On ECG monitoring, bidirectional ventricular tachycardia occurred with sinus tachycardia and then degenerated into ventricular fibrillation. To our knowledge, there has been no previous case report of CPVT triggered by sinus tachycardia in Korea. Therefore, we report the case as well as a review of the literature.     

Non-linear feature extraction from HRV signal for mortality prediction of ICU cardiovascular patient

Intensive care unit (ICU) patients are at risk of in-ICU morbidities and mortality, making specific systems for identifying at-risk patients a necessity for improving clinical care. This study presents a new method for predicting in-hospital mortality using heart rate variability (HRV) collected from the times of a patient’s ICU stay. In this paper, a HRV time series processing based method is proposed for mortality prediction of ICU cardiovascular patients. HRV signals were obtained measuring R-R time intervals. A novel method, named return map, is then developed that reveals useful information from the HRV time series. This study also proposed several features that can be extracted from the return map, including the angle between two vectors, the area of triangles formed by successive points, shortest distance to 45° line and their various combinations. Finally, a thresholding technique is proposed to extract the risk period and to predict mortality. The data used to evaluate the proposed algorithm obtained from 80 cardiovascular ICU patients, from the first 48 h of the first ICU stay of 40 males and 40 females. This study showed that the angle feature has on average a sensitivity of 87.5% (with 12 false alarms), the area feature has on average a sensitivity of 89.58% (with 10 false alarms), the shortest distance feature has on average a sensitivity of 85.42% (with 14 false alarms) and, finally, the combined feature has on average a sensitivity of 92.71% (with seven false alarms). The results showed that the last half an hour before the patient’s death is very informative for diagnosing the patient’s condition and to save his/her life. These results confirm that it is possible to predict mortality based on the features introduced in this paper, relying on the variations of the HRV dynamic characteristics.     

MEMORY CONJUNCTION ERRORS IN YOUNGER AND OLDER ADULTS: EVENT-RELATED POTENTIAL AND NEUROPSYCHOLOGICAL DATA

In a study/recognition paradigm, new words at test were recombinations of studied syllables (e.g. BARLEY from BARTER and VALLEY), shared one syllable with studied words, or were completely new. False alarm rates followed the gradient of similarity with studied items. Event-related potentials to the three classes of false alarms were indistinguishable. False alarms elicited different brain activity than did hits, arguing against the idea that conjunction errors occur during encoding and are later retrieved liked genuine memories. In Experiment 2, with healthy older adults, neuropsychological tests sensitive to frontal lobe function predicted false alarm rate, but not hit rate. Performance on standardised memory scales sensitive to medial temporal/diencephalic function influenced the pattern of false alarm rates across the three classes of new words. The experiments suggest that false alarms to conjunction lures are not similar to true recollections, but are products of faulty monitoring at retrieval.     

基于自适应学习的心律失常心拍分类方法

心律失常是因心脏疾病引起的心电活动中的异常症状,早期心室收缩（PVC）是由异位心跳引起的常见心律失常形式。通过心电图（ECG）信号检测PVC对于预测可能的心力衰竭具有重要意义。本文提出一种面向PVC心拍分类的心电信号分类算法,重点研究基于自适应学习的PVC异常心拍分类特征提取模型,通过计算心拍关联后验概率,结合领域专家标注信息训练分类器,提高整体分类效果。实验采用MIT-BIH心律失常数据库的ECG数据,研究结果表明所提方法针对非线性流形结构数据,能够有效提升小样本心拍自适应分类器的准确性。     

Investigating the performance of an amplitude-independent algorithm for detecting the hand muscle activity of stroke survivors

Abstract

To make robotic hand devices controlled by surface electromyography (sEMG) signals feasible and practical tools for assisting patients with hand impairments, the problems that prevent these devices from being widely used have to be overcome. The most significant problem is the involuntary amplitude variation of the sEMG signals due to the movement of electrodes during forearm motion. Moreover, for patients who have had a stroke or another neurological disease, the muscle activity of the impaired hand is weak and has a low signal-to-noise ratio (SNR). Thus, muscle activity detection methods intended for controlling robotic hand devices should not depend mainly on the amplitude characteristics of the sEMG signal in the detection process, and they need to be more reliable for sEMG signals that have a low SNR. Since amplitude-independent muscle activity detection methods meet these requirements, this paper investigates the performance of such a method on people who have had a stroke in terms of the detection of weak muscle activity and resistance to false alarms caused by the involuntary amplitude variation of sEMG signals; these two parameters are very important for achieving the reliable control of robotic hand devices intended for people with disabilities. A comparison between the performance of an amplitude-independent muscle activity detection algorithm and three amplitude-dependent algorithms was conducted by using sEMG signals recorded from six hemiparesis stroke survivors and from six healthy subjects. The results showed that the amplitude-independent algorithm performed better in terms of detecting weak muscle activity and resisting false alarms.     

Computer versus physician identification of gastrointestinal alarm features

ObjectiveIt is important for clinicians to inquire about “alarm features” as it may identify those at risk for organic disease and who require additional diagnostic workup. We developed a computer algorithm called Automated Evaluation of Gastrointestinal Symptoms (AEGIS) that systematically collects patient gastrointestinal (GI) symptoms and alarm features, and then “translates” the information into a history of present illness (HPI). Our study’s objective was to compare the number of alarms documented by physicians during usual care vs. that collected by AEGIS.MethodsWe performed a cross-sectional study with a paired sample design among patients visiting adult GI clinics. Participants first received usual care by their physicians and then completed AEGIS. Each individual thus contributed both a physician-documented and computer-generated HPI. Blinded physician reviewers enumerated the positive alarm features (hematochezia, melena, hematemesis, unintentional weight loss, decreased appetite, and fevers) mentioned in each HPI. We compared the number of documented alarms within patient using the Wilcoxon signed-rank test.ResultsSeventy-five patients had both physician and AEGIS HPIs. AEGIS identified more patients with positive alarm features compared to physicians (53% vs. 27%; p < .001). AEGIS also documented more positive alarms (median 1, interquartile range [IQR] 0–2) vs. physicians (median 0, IQR 0–1; p < .001). Moreover, clinicians documented only 30% of the positive alarms self-reported by patients through AEGIS.ConclusionsPhysicians documented less than one-third of red flags reported by patients through a computer algorithm. These data indicate that physicians may under report alarm features and that computerized “checklists” could complement standard HPIs to bolster clinical care.     

Integrating monitor alarms with laboratory test results to enhance patient deterioration prediction

Patient monitors in modern hospitals have become ubiquitous but they generate an excessive number of false alarms causing alarm fatigue. Our previous work showed that combinations of frequently co-occurring monitor alarms, called SuperAlarm patterns, were capable of predicting in-hospital code blue events at a lower alarm frequency. In the present study, we extend the conceptual domain of a SuperAlarm to incorporate laboratory test results along with monitor alarms so as to build an integrated data set to mine SuperAlarm patterns. We propose two approaches to integrate monitor alarms with laboratory test results and use a maximal frequent itemsets mining algorithm to find SuperAlarm patterns. Under an acceptable false positive rate FPR_max, optimal parameters including the minimum support threshold and the length of time window for the algorithm to find the combinations of monitor alarms and laboratory test results are determined based on a 10-fold cross-validation set. SuperAlarm candidates are generated under these optimal parameters. The final SuperAlarm patterns are obtained by further removing the candidates with false positive rate > FPR_max. The performance of SuperAlarm patterns are assessed using an independent test data set. First, we calculate the sensitivity with respect to prediction window and the sensitivity with respect to lead time. Second, we calculate the false SuperAlarm ratio (ratio of the hourly number of SuperAlarm triggers for control patients to that of the monitor alarms, or that of regular monitor alarms plus laboratory test results if the SuperAlarm patterns contain laboratory test results) and the work-up to detection ratio, WDR (ratio of the number of patients triggering any SuperAlarm patterns to that of code blue patients triggering any SuperAlarm patterns). The experiment results demonstrate that when varying FPR_max between 0.02 and 0.15, the SuperAlarm patterns composed of monitor alarms along with the last two laboratory test results are triggered at least once for [56.7–93.3%] of code blue patients within an 1-h prediction window before code blue events and for [43.3–90.0%] of code blue patients at least 1-h ahead of code blue events. However, the hourly number of these SuperAlarm patterns occurring in control patients is only [2.0–14.8%] of that of regular monitor alarms with WDR varying between 2.1 and 6.5 in a 12-h window. For a given FPR_max threshold, the SuperAlarm set generated from the integrated data set has higher sensitivity and lower WDR than the SuperAlarm set generated from the regular monitor alarm data set. In addition, the McNemar’s test also shows that the performance of the SuperAlarm set from the integrated data set is significantly different from that of the SuperAlarm set from the regular monitor alarm data set. We therefore conclude that the SuperAlarm patterns generated from the integrated data set are better at predicting code blue events.     

A fuzzy logic based apnoea monitor for SIDS risk infants

A unit has been designed which monitors newborn infants at risk of Sudden Infant Death Syndrome (SIDS) in a home environment. The unit monitors respiration, electrocardiogram (ECG) and haemoglobin oxygen saturation (SpO2) in combination, in order to detect any potentially life threatening event at an early stage. Provision is made for the generation of both audible and silent alarms and for the storage of signals and other information before, during and after an alarm episode for diagnostic purposes. An intelligent fuzzy logic algorithm is used to process the signals monitored and to implement several propositions concerning their status in order to determine the probability of an apnoea event and initiate the appropriate action. This has substantially reduced the number of false alarms and of undetected dangerous situations compared with previous units, which greatly improves the reliability and usefulness of such a monitor.     

Support vector machine-based arrhythmia classification using reduced features of heart rate variability signal

OBJECTIVE: This paper presents an effective cardiac arrhythmia classification algorithm using the heart rate variability (HRV) signal. The proposed algorithm is based on the generalized discriminant analysis (GDA) feature reduction scheme and the support vector machine (SVM) classifier. METHODOLOGY: Initially 15 different features are extracted from the input HRV signal by means of linear and nonlinear methods. These features are then reduced to only five features by the GDA technique. This not only reduces the number of the input features but also increases the classification accuracy by selecting most discriminating features. Finally, the SVM combined with the one-against-all strategy is used to classify the HRV signals. RESULTS: The proposed GDA- and SVM-based cardiac arrhythmia classification algorithm is applied to input HRV signals, obtained from the MIT-BIH arrhythmia database, to discriminate six different types of cardiac arrhythmia. In particular, the HRV signals representing the six different types of arrhythmia classes including normal sinus rhythm, premature ventricular contraction, atrial fibrillation, sick sinus syndrome, ventricular fibrillation and 2 degrees heart block are classified with an accuracy of 98.94%, 98.96%, 98.53%, 98.51%, 100% and 100%, respectively, which are better than any other previously reported results. CONCLUSION: An effective cardiac arrhythmia classification algorithm is presented. A main advantage of the proposed algorithm, compared to the approaches which use the ECG signal itself is the fact that it is completely based on the HRV (R-R interval) signal which can be extracted from even a very noisy ECG signal with a relatively high accuracy. Moreover, the usage of the HRV signal leads to an effective reduction of the processing time, which provides an online arrhythmia classification system. A main drawback of the proposed algorithm is however that some arrhythmia types such as left bundle branch block and right bundle branch block beats cannot be detected using only the features extracted from the HRV signal.