Computer methods and programs in biomedicine update最新文献

Background and Objective

Hypertension is critical risk factor of fatal cardiovascular diseases and multiple organ damage. Early detection of hypertension even at pre-hypertension stage is helpful in preventing the forthcoming complications. Electrocardiogram (ECG) has been attempted to observe the changes in electrical activities of the hearts of hypertensive patients. To automate the ECG assessment in the detection of hypertension, an interpretable hybrid model is proposed in this paper.

Methods

The proposed hybrid framework consists of one dimensional - Convolutional Neural Network architecture with four blocks of convolutional layers, maxpooling followed by dropout layers fused with Support Vector Machine classifier in the final layer. The implemented hybrid model is made explainable and interpretable using Local Interpretable Model-agnostic Explanations (LIME) method. The developed hybrid model is trained and tested for patient-wise classification of ECGs using online Physionet datasets and hospital data.

Results

The proposed method achieved highest accuracy of 81.81% in patient-wise ECG classification of online datasets, and highest accuracy of 93.33% in patient-wise ECG classification of hospital datasets as normotensive and hypertensive. The visualization of results showed only one normotensive patient's ECG is misclassified (predicted) as hypertensive, with identification of patient number, among the 15 patients (8 normotensive and 7 hypertensive) ECGs tested. In addition, the LIME results demonstrated an explanation to the predictions of hybrid model by highlighting the features and location of ECG waveform responsible for it, thus making the decision of hybrid model more interpretable.

Conclusion

Furthermore, our developed system is implemented as an assisting automated software tool called, HANDI (Hypertensive And Normotensive patient Detection with Interpretability) for real-time validation in clinics for early capture of hypertensive and proper monitoring of the patients.

Background:

In December 2020, the COVID-19 disease was confirmed in 1,665,775 patients and caused 45,784 deaths in Spain. At that time, health decision support systems were identified as crucial against the pandemic.

Methods:

This study applies Deep Learning techniques for mortality prediction of COVID-19 patients. Two datasets with clinical information were used. They included 2,307 and 3,870 COVID-19 infected patients admitted to two Spanish hospitals. Firstly, we built a sequence of temporal events gathering all the clinical information for each patient, comparing different data representation methods. Next, we used the sequences to train a Recurrent Neural Network (RNN) model with an attention mechanism exploring interpretability. We conducted an extensive hyperparameter search and cross-validation. Finally, we ensembled the resulting RNNs to enhance sensitivity.

Results:

We assessed the performance of our models by averaging the performance across all the days in the sequences. Additionally, we evaluated day-by-day predictions starting from both the hospital admission day and the outcome day. We compared our models with two strong baselines, Support Vector Classifier and Random Forest, and in all cases our models were superior. Furthermore, we implemented an ensemble model that substantially increased the system’s sensitivity while producing more stable predictions.

Conclusions:

We have shown the feasibility of our approach to predicting the clinical outcome of patients. The result is an RNN-based model that can support decision-making in healthcare systems aiming at interpretability. The system is robust enough to deal with real-world data and can overcome the problems derived from the sparsity and heterogeneity of data.

Background

With the global spread of coronavirus disease 2019 (COVID-19), understanding the mechanisms and characteristics of epidemic waves has become necessary to control its spread. The sixth epidemic wave of COVID-19 in Sapporo, Japan, was analyzed using a new mathematical model called the SI_UI_CI_CPR_UR_C model. The main objectives are (1) introducing the SI_UI_CI_CPR_UR_C model, (2) introducing algorisms by which parameters and initial states were estimated, and (3) estimating values of parameters and initial states, and analyzing the epidemic wave.

Methods

Reported numbers of daily new confirmed infected cases, currently infected cases, and cumulative numbers of recovered or fatal cases were collected from the official website of the city of Sapporo. The SI_UI_CI_CPR_UR_C model, based on susceptible-infectious-removed and infection-period-structured models, was employed. Parameter values and initial states of variables were estimated using the least squares method, Visual Basic for Applications, and Solver in Microsoft Excel.

Results

The peak time of transmission rate was estimated to be 5.8 to 6.0 days after infection, the peak time of infection confirmation rate was 8.0 to 8.1 days after infection, and the ultimate confirmation ratio of infection was 0.65 to 0.85. It was also estimated that almost all individuals in Sapporo were susceptible to the Omicron variant of the severe acute respiratory syndrome-coronavirus 2.

Conclusion

The sixth epidemic wave of COVID-19 was analyzed with the SI_UI_CI_CPR_UR_C model, with which crucial parameters and initial states were estimated. Furthermore, the results indicate that vaccination against the Wuhan strain and the previous infection were insufficient to induce a level of immunity required to prevent infection by the Omicron variant. Further improvement of mathematical modeling for infectious diseases is required to control emerging infectious diseases in the future, even if the threat of COVID-19 is overcome.

Background

Unawareness of Huntington disease is prevalent where patients might have a denial of illness, less reporting of symptoms such as changes in behavior or cognitive impairment, or poor coping with the disease. Understanding the awareness level of Huntington disease is crucial to provide more suggestions for public health campaigns.

Objective

This study explores the level of awareness of Huntington's disease among users of social media. We will also explore the tweeting behavior during Huntington disease awareness month, and search any missing area related to the awareness by following the framework of Social Media-Based Public Health Campaigns.

Method

We extracted tweets from April 2021-Jun 2021. We used both quantitative and qualitative methods to analyze the data. We used Python programming and various natural language processing tools to process and analyze data for a quantitative investigation. We also carried out a qualitative content analysis to identify themes and subthemes in the data.

Result

We discovered that the most popular hashtag is #LetsTalkAboutHD, and after looking over the data, it seemed to us that the word "support" was used more than 54 times during that time. According to the findings of our analysis of the twitter distribution pattern in terms of time, the most tweets were sent between May 13 and May 16, particularly on Wednesday, which was the busiest day. Also, the United States and Alaska had the highest levels of engagement when the pattern of tweets based on geographic location was examined. The most common pattern in the tweets that we separated based on patterns was news, which was followed by research and clinical trials.

Conclusion

Awareness campaigns needs to follow the framework of social media-Based Public Health Campaigns to provide more comprehensive information about Huntington disease and increase the awareness level among patients and families.

Background

Patient decision aids (PDAs) have shown to be effective in facilitating shared decision-making (SDM) in maternity care. However, many PDAs are difficult to use for clients because of high cognitive demand.

Objective

This study aimed to explore how current digital PDAs support clients’ health literacy skills (understanding, appraising, and applying information) and fit their needs for support in SDM in maternity care.

Methods

Clients (n=21) in Dutch maternity care were invited to use five PDAs during think aloud interviews. The interviews were transcribed verbatim, coded with open and axial coding, and analysed using thematic analysis. A framework of health literacy skills for SDM was used to categorize the themes.

Results

Clients reported a need for support to appraise and understand the purpose of PDAs. Most clients adequately used both benefit/harm information about available options and available Value Clarification Methods (VCM), indicating that these main PDA elements supported them to actively process this information in their decision-making process. However, these elements were only appreciated and adequately used when clients understood the pregnancy- and labour related terminology used. A lack of balanced probability information about outcomes of options for mother and child hindered further information use. VCM were only used when presented attributes were relevant for clients.

Conclusions

Clients were in general able to process and use information presented in PDAs in maternity care tested in this study, thus PDAs were aligned with health literacy skills. Adequate understanding of terminology and perceived relevance of specific information elements were important preconditions.

Background:

The field of tissue engineering has provided valuable three-dimensional species-specific models of the human myocardium in the form of human Engineered Cardiac Tissues (hECTs) and similar constructs. However, hECT systems are often bottlenecked by a lack of openly available software that can collect data from multiple tissues at a time, even in multi-tissue bioreactors, which limits throughput in phenotypic and therapeutic screening applications.

Methods:

We developed Rianú, an open-source web application capable of simultaneously tracking multiple hECTs on flexible end-posts. This software is operating system agnostic and deployable on a remote server, accessible via a web browser with no local hardware or software requirements. The software incorporates object-tracking capabilities for multiple objects simultaneously, an algorithm for twitch tracing analysis and contractile force calculation, and a data compilation system for comparative analysis within and amongst groups. Validation tests were performed using in-silico and in-vitro experiments for comparison with established methods and interventions.

Results:

Rianú was able to detect the displacement of the flexible end-posts with a sub-pixel sensitivity of 0.555 px/post (minimum increment in post displacement) and a lower limit of 1.665 px/post (minimum post displacement). Compared to our established reference for contractility assessment, Rianú had a high correlation for all parameters analyzed (ranging from R${}^2$=0.7514 to R${}^2$=0.9695), demonstrating its high accuracy and reliability.

Conclusions:

Rianú provides simultaneous tracking of multiple hECTs, expediting the recording and analysis processes, and simplifies time-based intervention studies. It also allows data collection from different formats and has scale-up capabilities proportional to the number of tissues per field of view. These capabilities will enhance throughput of hECTs and similar assays for in-vitro analysis in disease modeling and drug screening applications.

Background

This study examines the level of eHealth literacy (eHL) of COPD patients and registered nurses (RN) prior to the implementation of a new national telehealth service. The objective was to provide the nurses with an understanding of eHL and to provide knowledge about the patients’ eHL level, socio-demographic characteristics, and digital behaviour for the nurses to be better able to support the patients’ adoption and usage of telehealth.

Method

The eHealth Literacy Questionnaire (eHLQ) was administered in an outpatient clinic in February and March 2020 (N = 42). The staff-eHLQ was administered by web in November 2019 and at a conference in January 2020 (N = 39). The RNs were asked about workplace and experience with telehealth and the patients about gender, age, and educational level as well as their digital health behaviour.

A multiple linear regression analysis tested for relations between the socio-demographic and digital behaviour variables and the eHLQ-scores for the COPD patients.

Results

The RNs’ eHLQ-scores relating to engagement with information, motivation, and experience with digital services signified an insufficient eHL level which may influence their ability to motivate and promote the usage of telehealth to patients.

The patients’ scores were higher than the RNs’ with respect to motivation and experience with digital services but seemed to have an insufficient level in relation to using technology to process information and actively engage with digital services.

Conclusion

The patients need support in relation to processing information and interacting with services. The RNs’ eHLQ-scores being lower than the patients are problematic as it may influence how well they are able to support the adoption of the new telehealth service.

Hypotension frequently occurs in intensive care units (ICUs) and is correlated to worsening patient outcomes. In this study, we propose a machine learning (ML) algorithm that predicts hypotensive events in ICUs by extracting the information from patients' contextual data and physiological signals. The algorithm uses patients’ history including demographics, pre-ICU medication, and pre-existing comorbidities, and only five minutes of prior physiological history to predict hypotension up to 30 min in advance. We show that adding demographic information to the physiological data does not improve the algorithm's predictive performance of 84% sensitivity, 89% positive predictive value (PPV), and 98% specificity. Furthermore, the results show that including features extracted from patients’ pre-ICU medications and comorbidities lowers the learning algorithm’ prediction performance and leads to 2% degradation in its F1-score. The feature importance analysis showed that the ratio of MAP to HR (MAP2HR) and the average of RR intervals on the ECG (RRI), both extracted from physiological signals, have the highest weights in the prediction of hypotension.

Breast cancer disease is recognized as one of the leading causes of death in women worldwide after lung cancer. Breast cancer refers to a malignant neoplasm that develops from breast cells. Developed and less developed countries both are suffering from this extensive cancer. This cancer can be recuperated if it is detected at an early stage. Many researchers have proposed several machine learning techniques to predict breast cancer with the highest accuracy in the past years. In this research work, the Wisconsin Breast Cancer Dataset (WBCD) has been used as a training set from the UCI machine learning repository to compare the performance of the various machine learning techniques. Different kinds of machine learning classifiers such as support vector machine (SVM), Random Forest (RF), K-nearest neighbors(K-NN), Decision tree (DT), Naïve Bayes (NB), Logistic Regression (LR), AdaBoost (AB), Gradient Boosting (GB), Multi-layer perceptron (MLP), Nearest Cluster Classifier (NCC), and voting classifier (VC) have been used for comparing and analyzing breast cancer into benign and malignant tumors. Various matrices such as error rate, Accuracy, Precision, F1-score, and recall have been implemented to measure the model's performance. Each Algorithm's accuracy has been ascertained for finding the best suitable one. Based on the analysis, the result shows that the Voting classifier has the highest accuracy, which is 98.77%, with the lowest error rate. Finally, a web page is developed using a flask micro-framework integrating the best model using react.

Sudden cardiac death, a prominent cause of mortality, often occurs within a narrow time window of less than an hour. This study introduces a novel methodology with the aim of early prediction of sudden cardiac death. The proposed approach involves the extraction of diverse features from the ECG signal, including the calculation of angles between two vectors, the computation of triangle areas formed by consecutive points, the determination of the shortest distance to a 450 line, and their combinations. Additionally, a thresholding technique is proposed to identify the risk period and predict the occurrence of sudden death. To assess the performance of the algorithm, data from the MIT-BH Holter database were utilized. The results obtained demonstrate that the angle feature achieves an average sensitivity of 93.75% with five false alarms, the area feature achieves an average sensitivity of 88.75% with nine false alarms, the shortest distance feature achieves an average sensitivity of 86.25% with 12 false alarms, and the combined feature achieves an average sensitivity of 96.25% with three false alarms. Remarkably, unlike existing methodologies in the literature, this method exhibits high accuracy in predicting the development of the risk of sudden cardiac death (SCD) even up to 30 min prior to onset. As a consequence, it plays a critical role in diagnosing patients' conditions and facilitating timely interventions. Moreover, the results confirm the feasibility of predicting cardiac arrest solely based on geometric features derived from variations in heart rate variability (HRV) dynamics.