RESUMEN

Objectives: Artificial intelligence is evolving and significantly impacting health care, promising to transform access to medical information. With the rise of medical misinformation and frequent internet searches for health-related advice, there is a growing demand for reliable patient information. This study assesses the effectiveness of ChatGPT in providing information and treatment options for chronic rhinosinusitis (CRS). Methods: Six inputs were entered into ChatGPT regarding the definition, prevalence, causes, symptoms, treatment options, and postoperative complications of CRS. International Consensus Statement on Allergy and Rhinology guidelines for Rhinosinusitis was the gold standard for evaluating the answers. The inputs were categorized into three categories and Flesch-Kincaid readability, ANOVA and trend analysis tests were used to assess them. Results: Although some discrepancies were found regarding CRS, ChatGPT's answers were largely in line with existing literature. Mean Flesch Reading Ease, Flesch-Kincaid Grade Level and passive voice percentage were (40.7%, 12.15%, 22.5%) for basic information and prevalence category, (47.5%, 11.2%, 11.1%) for causes and symptoms category, (33.05%, 13.05%, 22.25%) for treatment and complications, and (40.42%, 12.13%, 18.62%) across all categories. ANOVA indicated no statistically significant differences in readability across the categories (p-values: Flesch Reading Ease = 0.385, Flesch-Kincaid Grade Level = 0.555, Passive Sentences = 0.601). Trend analysis revealed readability varied slightly, with a general increase in complexity. Conclusion: ChatGPT is a developing tool potentially useful for patients and medical professionals to access medical information. However, caution is advised as its answers may not be fully accurate compared to clinical guidelines or suitable for patients with varying educational backgrounds.Level of evidence: 4.

RESUMEN

The goal of this work is to develop a Machine Learning model to predict the need for both invasive and non-invasive mechanical ventilation in intensive care unit (ICU) patients. Using the Philips eICU Research Institute (ERI) database, 2.6 million ICU patient data from 2010 to 2019 were analyzed. This data was randomly split into training (63%), validation (27%), and test (10%) sets. Additionally, an external test set from a single hospital from the ERI database was employed to assess the model's generalizability. Model performance was determined by comparing the model probability predictions with the actual incidence of ventilation use, either invasive or non-invasive. The model demonstrated a prediction performance with an AUC of 0.921 for overall ventilation, 0.937 for invasive, and 0.827 for non-invasive. Factors such as high Glasgow Coma Scores, younger age, lower BMI, and lower PaCO2 were highlighted as indicators of a lower likelihood for the need for ventilation. The model can serve as a retrospective benchmarking tool for hospitals to assess ICU performance concerning mechanical ventilation necessity. It also enables analysis of ventilation strategy trends and risk-adjusted comparisons, with potential for future testing as a clinical decision tool for optimizing ICU ventilation management.

RESUMEN

In China, several problems were common in the telemedicine systems, such as the poor network stability and difficult interconnection. A new telemedicine system jointly driven by multinetwork integration and remote control has been designed to address these problems. A multilink aggregation algorithm and an overlay network for telemedicine system (ONTMS) were developed to improve network stability, and a non-intervention remote control method was designed for Internet of Things (IoT) devices/systems. The authors monitored the network parameters, and distributed the questionnaire to participants, for evaluating the telemedicine system and services. Under a detection bandwidth of 8 Mbps, the aggregation parameters of Unicom 4G, Telecom 4G, and China Mobile 4G were optimal, with an uplink bandwidth, delay, and packet loss ratio (PLR) of 7.93 Mbps, 58.80 ms, and 0.06%, respectively. These parameters were significantly superior to those of China Mobile 4G, the best single network (p < 0.001). Through the ONTMS, the mean round-trip transporting delay from Beijing to Sanya was 76 ms, and the PLR was 0 at vast majority of time. A total of 1988 participants, including 1920 patients and 68 doctors, completed the questionnaires. More than 97% of participants felt that the audio and video transmission and remote control were fluent and convenient. 96% of patients rated the telemedicine services with scores of 4 or 5. This system has shown robust network property and excellent interaction ability, and satisfied the needs of patients and doctors.

RESUMEN

Observational medical data present unique opportunities for analysis of medical outcomes and treatment decision making. However, because these datasets do not contain the strict pairing of randomized control trials, matching techniques are to draw comparisons among patients. A key limitation to such techniques is verification that the variables used to model treatment decision making are also relevant in identifying the risk of major adverse events. This article explores a deep mixture of experts approach to jointly learn how to match patients and model the risk of major adverse events in patients. Although trained with information regarding treatment and outcomes, after training, the proposed model is decomposable into a network that clusters patients into phenotypes from information available before treatment. This model is validated on a dataset of patients with acute myocardial infarction complicated by cardiogenic shock. The mixture of experts approach can predict the outcome of mortality with an area under the receiver operating characteristic curve of 0.85 ± 0.01 while jointly discovering five potential phenotypes of interest. The technique and interpretation allow for identifying clinically relevant phenotypes that may be used both for outcomes modeling as well as potentially evaluating individualized treatment effects.

RESUMEN

Background: Patient-reported outcomes (PRO) allow clinicians to measure health-related quality of life (HRQOL) and understand patients' treatment priorities, but obtaining PRO requires surveys which are not part of routine care. We aimed to develop a preliminary natural language processing (NLP) pipeline to extract HRQOL trajectory based on deep learning models using patient language. Materials and methods: Our data consisted of transcribed interviews of 100 patients undergoing surgical intervention for low-risk thyroid cancer, paired with HRQOL assessments completed during the same visits. Our outcome measure was HRQOL trajectory measured by the SF-12 physical and mental component scores (PCS and MCS), and average THYCA-QoL score.We constructed an NLP pipeline based on BERT, a modern deep language model that captures context semantics, to predict HRQOL trajectory as measured by the above endpoints. We compared this to baseline models using logistic regression and support vector machines trained on bag-of-words representations of transcripts obtained using Linguistic Inquiry and Word Count (LIWC). Finally, given the modest dataset size, we implemented two data augmentation methods to improve performance: first by generating synthetic samples via GPT-2, and second by changing the representation of available data via sequence-by-sequence pairing, which is a novel approach. Results: A BERT-based deep learning model, with GPT-2 synthetic sample augmentation, demonstrated an area-under-curve of 76.3% in the classification of HRQOL accuracy as measured by PCS, compared to the baseline logistic regression and bag-of-words model, which had an AUC of 59.9%. The sequence-by-sequence pairing method for augmentation had an AUC of 71.2% when used with the BERT model. Conclusions: NLP methods show promise in extracting PRO from unstructured narrative data, and in the future may aid in assessing and forecasting patients' HRQOL in response to medical treatments. Our experiments with optimization methods suggest larger amounts of novel data would further improve performance of the classification model.

Asunto(s)

RESUMEN

The HeartPy Python toolkit for analysis of noisy signals from heart rate measurements is an excellent tool to use in conjunction with novel wearable sensors. Nevertheless, most of the work to date has focused on applying the toolkit to data measured with commercially available sensors. We demonstrate the application of the HeartPy functions to data obtained with a novel graphene-based heartbeat sensor. We produce the sensor by laser-inducing graphene on a flexible polyimide substrate. Both graphene on the polyimide substrate and graphene transferred onto a PDMS substrate show piezoresistive behavior that can be utilized to measure human heartbeat by registering median cubital vein motion during blood pumping. We process electrical resistance data from the graphene sensor using HeartPy and demonstrate extraction of several heartbeat parameters, in agreement with measurements taken with independent reference sensors. We compare the quality of the heartbeat signal from graphene on different substrates, demonstrating that in all cases the device yields results consistent with reference sensors. Our work is a first demonstration of successful application of HeartPy to analysis of data from a sensor in development.

Asunto(s)

RESUMEN

The article presents analysis of the Arkhangelsk City Clinical Hospital № 7 capacity and medical personnel staffing. The new coronavirus infection incidence, cases of temporary disability due to disease and quarantine of COVID-19 were analyzed in the Solombalsky district of Arkhangelsk. The issues of primary health care support and readiness of medical personnel to work with medical information systems in in conditions of COVID-19 pandemic were studied. The management decision-makings improving quality and accessibility of medical services using information technologies are demonstrated.

Asunto(s)

RESUMEN

Clinical guidelines are recommendations of how to diagnose, treat, and manage a patient's medical condition. Health organizations must measure adherence to clinical guidelines to enhance the quality of service, but due to the complexity of the medical environment, there is no simple way of measuring adherence to clinical guidelines. This scoping review will systematically assess the criteria used to measure adherence to clinical guidelines in the past 20 years and explore the suitability of using process mining techniques. We will use a workflow protocol based on declarative and temporal constraints to translate the narrative text rules in the publications into a high-level process model. This approach will enable us to explore the main patterns and gaps identified when measuring adherence to clinical guidelines and how they affect the adoption of process mining techniques. The main contributions of this paper are a) a comprehensive analysis of the criteria used for measuring adherence, considering a diverse set of medical conditions b) a framework that will classify the level of complexity of the rules used to measure adherence based on declarative and temporal constraints c) list of key trends and gaps identified in the literature and how they relate to the use of process mining techniques in healthcare.

Asunto(s)

RESUMEN

Kidney disease is a major public health concern that has only recently emerged. Toxins are removed from the body by the kidneys through urine. In the early stages of the condition, the patient has no problems, but recovery is difficult in the later stages. Doctors must be able to recognize this condition early in order to save the lives of their patients. To detect this illness early on, researchers have used a variety of methods. Prediction analysis based on machine learning has been shown to be more accurate than other methodologies. This research can help us to better understand global disparities in kidney disease, as well as what we can do to address them and coordinate our efforts to achieve global kidney health equity. This study provides an excellent feature-based prediction model for detecting kidney disease. Various machine learning algorithms, including k-nearest neighbors algorithm (KNN), artificial neural networks (ANN), support vector machines (SVM), naive bayes (NB), and others, as well as Re-cursive Feature Elimination (RFE) and Chi-Square test feature-selection techniques, were used to build and analyze various prediction models on a publicly available dataset of healthy and kidney disease patients. The studies found that a logistic regression-based prediction model with optimal features chosen using the Chi-Square technique had the highest accuracy of 98.75 percent. White Blood Cell Count (Wbcc), Blood Glucose Random (bgr), Blood Urea (Bu), Serum Creatinine (Sc), Packed Cell Volume (Pcv), Albumin (Al), Hemoglobin (Hemo), Age, Sugar (Su), Hypertension (Htn), Diabetes Mellitus (Dm), and Blood Pressure (Bp) are examples of these traits.

RESUMEN

BACKGROUND AND OBJECTIVE: Increasing requirements for documentation, cross-sectoral communication and quality management are leading to increased organizational effort in emergency medical services (EMS). On the one hand, the use of digital information systems in prehospital settings can help to support emergency physicians and paramedics in these tasks and on the other hand, it opens new treatment options such as telemedical care for patients. This work attempts to provide a comprehensive picture of the current use of digital systems for ambulance services in Germany. To do so, the study investigated how widespread various information and communication systems currently are at local EMS stations and ambulances, how they are used by emergency personnel, how they are assessed by users and what challenges currently exist for further expansion and greater acceptance of the users. MATERIAL AND METHODS: The cross-sectional study was conducted as a nationwide, exploratory online survey among emergency physicians and paramedic professionals in July and August 2020 covering 24 different questions. Participation was called for on the Internet, at EMS stations and in hospital emergency departments. Subsequent data analysis was performed using descriptive statistical methods. Solutions considered included digital documentation and hospital prenotification, interdisciplinary care capacity notification, real-time telehealth services and digital radio units. RESULTS: In total, 821 responses of participants from 481 different EMS stations from 382 cities nationwide were included in the evaluation. The availability of the 16 systems surveyed varies significantly throughout Germany, depending on the federal state and application. While basic equipment such as radio units or navigation devices are available on almost all surveyed ambulances, the share which has real-time telehealth applications at their disposal is just 6%. A proportion of 72% reported the usage of any type of digital documentation and 41% used a digital tool for prenotification of emergency rooms in at least one hospital. The emergency staff surveyed were generally open to new technologies and resulting possibilities, such as having an electronic patient care record or transmitting patient data digitally to emergency room. Almost all participants see a benefit in the use of information technology in ambulance service, although slightly more than half considered current implementation as unsatisfactory. Challenges are particularly evident with regard to reliability, hardware, useability and interoperability with third parties, such as dispatch centres and hospitals. CONCLUSION: Although information technology systems in German EMS are no longer in their infancy, there is still a long way to go before prehospital emergency care can be considered as extensively and adequately digitalized. A more holistic perspective and networked implementation of all systems and processes involved in emergency response operations can help improve and further spread digital solutions for prehospital emergency care. Incorporating field experience into the development process could contribute to increasing functionality and user acceptance.

Asunto(s)

RESUMEN

As a side-effect of the Covid-19 pandemic, significant decreases in medical procedures for noncommunicable diseases have been observed. This calls for a decision support assisting in the analysis of opportunities to relocate procedures among hospitals in an efficient or, preferably, optimal manner. In the current paper we formulate corresponding decision problems and develop linear (mixed integer) programming models for them. Since solving mixed integer programming problems is NP-complete, we verify experimentally their usefulness using real-world data about urological procedures. We show that even for large models, with millions of variables, the problems' instances are solved in perfectly acceptable time.

RESUMEN

Colonoscopies reduce the incidence of colorectal cancer through early recognition and resecting of the colon polyps. However, the colon polyp miss detection rate is as high as 26% in conventional colonoscopy. The search for methods to decrease the polyp miss rate is nowadays a paramount task. A number of algorithms or systems have been developed to enhance polyp detection, but few are suitable for real-time detection or classification due to their limited computational ability. Recent studies indicate that the automated colon polyp detection system is developing at an astonishing speed. Real-time detection with classification is still a yet to be explored field. Newer image pattern recognition algorithms with convolutional neuro-network (CNN) transfer learning has shed light on this topic. We proposed a study using real-time colonoscopies with the CNN transfer learning approach. Several multi-class classifiers were trained and mAP ranged from 38% to 49%. Based on an Inception v2 model, a detector adopting a Faster R-CNN was trained. The mAP of the detector was 77%, which was an improvement of 35% compared to the same type of multi-class classifier. Therefore, our results indicated that the polyp detection model could attain a high accuracy, but the polyp type classification still leaves room for improvement.

Asunto(s)

RESUMEN

OBJECTIVE: Hysterectomy is the most common procedure in women. We wanted to make an assessment of the hysterectomy in France in 2019. We also assessed the variations over time in the indications and the surgical approch for hysterectomy, this with regard to the various events that may have been at the origin of the modification practices. METHODS: We used the Medical Information Systems Program in Medicine, Surgery, Obstetrics and Dentistry to extract all acts relating to a hysterectomy regardless of its route of approach from 2009 to 2019. RESULTS: Hysterectomy is a frequent procedure which was performed in nearly 60,000 women in France in 2019. The most frequently used surgical approach is now laparoscopy, performed in 30% of hysterectomies, followed by laparotomic (29%), then vaginal approaches (26%) and coelio-vaginal (15%). Laparoscopic procedures are performed more often in public than private hospitals. Adnexectomy is associated with 41% of hysterectomies. A decrease in the number of hysterectomies was observed between 2008 and 2019, from approximately 72,000 in 2008 to approximately 60,000 in 2019. This decrease occurs during a period in which new therapies have emerged as well as new recommendations. CONCLUSION: The evolution of the number of hysterectomies is correlated with the development of therapeutic alternatives for pathologies for which a hysterectomy has traditionally been performed.

Asunto(s)

RESUMEN

Big data have shown their great potential value to serve many aspects of human life. Due to complexity of the medical and healthcare big data in real life, traditional big data analysis methods are difficult to be dealt with. Therefore, a single method is unable to analyze and manage heterogeneous big data sources. To utilize data fully from the perspective of decision-making, we propose a novel framework which guides the healthcare big data to be smartly and proactively processed for decision-making without user interventions. The framework contains five stages, which are intelligent data cleaning, customized data fusion, analysis mapping, exploratory visualization analysis, and generation of decision-making reports. It also enables learning from the data and correlating them with the existing human knowledge. Subsequently, a smart big data-driven application exhibits innovative management in intelligent healthcare. The proposed framework provides the guidelines of the best practices of big data-driven analysis for intelligent healthcare according to our practical applications. The platform provides the appropriate reference for the big data-driven innovation of management in intelligent healthcare.

Asunto(s)

RESUMEN

An automated vendor-independent system for dose monitoring in computed tomography (CT) medical examinations involving ionizing radiation is presented in this paper. The system provides precise size-specific dose estimates (SSDE) following the American Association of Physicists in Medicine regulations. Our dose management can operate on incomplete DICOM header metadata by retrieving necessary information from the dose report image by using optical character recognition. For the determination of the patient's effective diameter and water equivalent diameter, a convolutional neural network is employed for the semantic segmentation of the body area in axial CT slices. Validation experiments for the assessment of the SSDE determination and subsequent stages of our methodology involved a total of 335 CT series (60 352 images) from both public databases and our clinical data. We obtained the mean body area segmentation accuracy of 0.9955 and Jaccard index of 0.9752, yielding a slice-wise mean absolute error of effective diameter below 2 mm and water equivalent diameter at 1 mm, both below 1%. Three modes of the SSDE determination approach were investigated and compared to the results provided by the commercial system GE DoseWatch in three different body region categories: head, chest, and abdomen. Statistical analysis was employed to point out some significant remarks, especially in the head category.

Asunto(s)

RESUMEN

Along with digitization, automatic data-driven decision support systems become increasingly popular. Mortality prediction is a vital part of that decision process. With more data available, sophisticated machine learning models like (Artificial) Neural Networks (NNs) can be applied and promise favorable performance. We evaluate the reproducibility of a published mortality prediction approach using NNs along with the possibility to generalize it to a bigger and more generic dataset. We describe an extensive preprocessing pipeline, as well as the evaluation of different sampling techniques and NN architectures. Through training on a loss function that optimizes both, precision and recall, in combination with a good set of hyperparameters and a set of new features, we use a NN to predict in-hospital mortality with accuracy, sensitivity, and area under the receiver operating characteristic score of greater than 0.8.

Asunto(s)

RESUMEN

Ballistocardiogram (BCG) is a graphical representation of the subtle oscillations in body movements caused by cardiovascular activity. Although BCGs cause less burden to the user, electrocardiograms (ECGs) are still commonly used in the clinical scene due to BCG sensors' noise sensitivity. In this paper, a robust method for sleep time BCG measurement and a mathematical model for predicting sleep stages using BCG are described. The novel BCG measurement algorithm can be described in three steps: preprocessing, creation of heartbeat signal template, and template matching for heart rate variability detection. The effectiveness of this algorithm was validated with 99 datasets from 36 subjects, with photoplethysmography (PPG) to compute ground truth heart rate variability (HRV). On average, 86.9% of the inter-beat intervals were detected and the mean error was 8.5ms. This shows that our method successfully extracted beat-to-beat intervals from BCG during sleep, making its usability comparable to those of clinical ECGs. Consequently, compared to other conventional BCG systems, even more accurate sleep heart rate monitoring with a smaller burden to the patient is available. Moreover, the accuracy of the sleep stages mathematical model, validated with 100 datasets from 25 subjects, is 80%, which is higher than conventional five-stage sleep classification algorithms (max: 69%). Although, in this paper, we applied the mathematical model to heart rate interval features from BCG, theoretically, this sleep stage prediction algorithm can also be applied to ECG-extracted heart rate intervals.

Asunto(s)

RESUMEN

Globally, cervical cancer remains as the foremost prevailing cancer in females. Hence, it is necessary to distinguish the importance of risk factors of cervical cancer to classify potential patients. The present work proposes a cervical cancer prediction model (CCPM) that offers early prediction of cervical cancer using risk factors as inputs. The CCPM first removes outliers by using outlier detection methods such as density-based spatial clustering of applications with noise (DBSCAN) and isolation forest (iForest) and by increasing the number of cases in the dataset in a balanced way, for example, through synthetic minority over-sampling technique (SMOTE) and SMOTE with Tomek link (SMOTETomek). Finally, it employs random forest (RF) as a classifier. Thus, CCPM lies on four scenarios: (1) DBSCAN + SMOTETomek + RF, (2) DBSCAN + SMOTE+ RF, (3) iForest + SMOTETomek + RF, and (4) iForest + SMOTE + RF. A dataset of 858 potential patients was used to validate the performance of the proposed method. We found that combinations of iForest with SMOTE and iForest with SMOTETomek provided better performances than those of DBSCAN with SMOTE and DBSCAN with SMOTETomek. We also observed that RF performed the best among several popular machine learning classifiers. Furthermore, the proposed CCPM showed better accuracy than previously proposed methods for forecasting cervical cancer. In addition, a mobile application that can collect cervical cancer risk factors data and provides results from CCPM is developed for instant and proper action at the initial stage of cervical cancer.

Asunto(s)

RESUMEN

The development of machine learning solutions in medicine is often hindered by difficulties associated with sharing patient data. Distributed learning aims to train machine learning models locally without requiring data sharing. However, the utility of distributed learning for rare diseases, with only a few training examples at each contributing local center, has not been investigated. The aim of this work was to simulate distributed learning models by ensembling with artificial neural networks (ANN), support vector machines (SVM), and random forests (RF) and evaluate them using four medical datasets. Distributed learning by ensembling locally trained agents improved performance compared to models trained using the data from a single institution, even in cases where only a very few training examples are available per local center. Distributed learning improved when more locally trained models were added to the ensemble. Local class imbalance reduced distributed SVM performance but did not impact distributed RF and ANN classification. Our results suggest that distributed learning by ensembling can be used to train machine learning models without sharing patient data and is suitable to use with small datasets.

Asunto(s)