Healthcare analytics (New York, N.Y.)最新文献

This study leverages machine learning to enhance the diagnostic accuracy of COVID-19 using chest X-rays. The study evaluates various architectures, including efficient neural networks (EfficientNet), multiscale vision transformers (MViT), efficient vision transformers (EfficientViT), and vision transformers (ViT), against a comprehensive open-source dataset comprising 3616 COVID-19, 6012 lung opacity, 10192 normal, and 1345 viral pneumonia images. The analysis, focusing on loss functions and evaluation metrics, demonstrates distinct performance variations among these models. Notably, multiscale models like MViT and EfficientNet tend towards overfitting. Conversely, our vision transformer model, innovatively fine-tuned (FT) on the encoder blocks, exhibits superior accuracy: 95.79% in four-class, 99.57% in three-class, and similarly high performance in binary classifications, along with a recall of 98.58%, precision of 98.87%, F1 score of 98.73%, specificity of 99.76%, and area under the receiver operating characteristic (ROC) curve (AUC) of 0.9993. The study confirms the vision transformer model’s efficacy through rigorous validation using quantitative metrics and visualization techniques and illustrates its superiority over conventional models. The innovative fine-tuning method applied to vision transformers presents a significant advancement in medical image analysis, offering a promising avenue for improving the accuracy and reliability of COVID-19 diagnosis from chest X-ray images.

Traditional randomized clinical trials are regarded as the gold standard for assessing the efficacy of chemotherapy. However, this procedure has drawbacks such as high cost, time consumption, and limited patient exploration of treatment regimens. We develop a multi-objective optimization-based framework to address these limitations and determine the best chemotherapy dosing and treatment duration. The proposed framework uses patient-specific biological parameters to create a mathematical model of cell population dynamics in the patient’s body. The framework employs evolutionary heuristic search methods (simulated annealing and genetic algorithms) and a prescriptive analytics approach to optimize therapy sessions that transition from treatment to relaxation. We carefully adjust the chemotherapy dose during treatment to reduce tumor cells while preserving host cells (such as effector-immune cells). We strategically time the relaxation sessions to aid recovery, considering the ability of tumors and healthy cells to regenerate. We use a combined optimization method to determine the length of the session and the amount of drug to be administered. We compare quadratic and linear optimal control solvers for drug administration while genetic algorithms and simulated annealing are used to optimize session length. This approach is especially important in limited healthcare resources, ensuring efficient allocation while accurately identifying high-risk patients to optimize resource allocation and utilization.

The malaria’s multidrug-resistant strain in Nigeria is prevalent and it poses a significant challenge for disease elimination. The testing for resistance is available but underutilized. Therefore, we develop a mathematical model incorporating the testing as a control strategy. This allows us to make quantifiable predictions about the effects of testing utilization on the malaria prevalence. By fitting the model to data on malaria and using field data reported in the literature, important parameters associated with the disease dynamics are estimated and calculated. First, we analyze the disease-free state of the malaria model and calculate the baseline control reproduction number. Sensitivity analysis is used to investigate the influence of the parameters in curtailing the disease. Numerical simulations are used to explore the behavior of the model solutions involving testing for resistance of the strain and wild strain malaria. We found that the implementation of testing would (a) prevent the increase of malaria prevalence from 30% to 35%, (b) significantly slow down the replacement of the wild strain by the resistant strain, and (c) avert about 6% of treatment failures. We also found that increasing mosquito death rate or reducing mosquito biting rate, mosquito birth rate, transmission to or from mosquitoes would contribute most significantly to the reduction of malaria prevalence in the community. In conclusion, the treatment failure is a significant component of the community malaria epidemic. Testing for multidrug resistance yields a significant reduction in cases with many implications regarding the containment of malaria in Nigeria.

The primary procedures for breast cancer diagnosis involve the assessment of histopathological slide images by skilled patholo-gists. This procedure is prone to human subjectivity and can lead to diagnostic errors with adverse implications for patient health and welfare. Artificial intelligence-based models have yielded promising results in other medical tasks and offer tools for potentially addressing the shortcomings of traditional medical image analysis. The BreakHis breast cancer dataset suffers from insufficient data for the minority class with an imbalance ratio >0.40, which poses challenges for deep learning models. To avoid performance degradation, researchers have explored a variety of data augmentation schemes to generate adequate samples for analysis. This study designed a Deep Convolutional Neural Network (DCGAN) with specific generator and discriminator architectures to mitigate model instability and generate high-quality synthetic data for the minority class. The balanced dataset was passed to the fine-tuned ResNet50 model for breast tumor detection. The study produced high accuracy in diagnosing benign/malignancy at 40X magnification, outperforming the state-of-art. The results demonstrated that deep learning methods can potentially to support effective screening in clinical practice.

COVID-19 has been a significant threat to many countries worldwide. COVID-19 remains a threat even in the presence of vaccination. The study formulates and analyzes a non-linear deterministic mathematical model to investigate the dynamics of COVID-19 in the presence of vaccination. Numerical results show that increasing the treatment rates with a relatively high vaccination rate might subdue the virus in the population. Also, decreasing the vaccine inefficacy increases the vaccine efficacy, and this may result in a population free of the virus. We further show that increasing the vaccination rate as against the vaccine inefficacy, the effective contact rate for COVID-19 and the modification parameter that accounts for increased infectiousness for COVID-19, the virus responsible for COVID-19 can be eradicated from the population. The sensitivity analysis results deduce that hidden factors are driving the model dynamics. These hidden factors must be given special attention and minimized. These factors includes the incubation periods for vaccinated and unvaccinated individuals, the fractions for vaccinated and unvaccinated individuals, and the transition rates for vaccinated and unvaccinated individuals

The Coronavirus Disease 2019 (COVID-19) pandemic has swept almost all parts of the world. With the increasing number of COVID-19 cases worldwide, Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) has mutated into various variants. Given the increasingly dangerous conditions of the pandemic, it is crucial to predict the number of COVID-19 cases. Deep Learning and Long Short-Term Memory (LSTM) have predicted disease progress with reasonable accuracy and small errors. LSTM training is used to predict confirmed cases of COVID-19 based on variants identified using the European Centre for Disease Prevention and Control (ECDC) COVID-19 dataset containing confirmed cases identified from 30 European countries. Tests were conducted using the LSTM and Bidirectional LSTM (BiLSTM) models with the addition of Recurrent Neural Network (RNN) as comparisons on hidden size and layer size. The obtained result showed that in testing hidden sizes 25, 50, 75, and 100, the RNN model provided better results, with the minimum Mean Squared Error (MSE) value of 0.01 and the Root Mean Square Error (RMSE) value of 0.012 for B.1.427/B.1.429 variant with a hidden size of 100. Further testing layer sizes 2, 3, 4, and 5 shows that the BiLSTM model provided better results, with a minimum MSE value of 0.01 and an RMSE of 0.01 for the B.1.427/B.1.429 variant with a hidden size of 100 and layer size of 2.

Tumors result from genetic mutations or environmental factors that prompt cells to divide uncontrollably. This study aims to examine the behavior of tumor-immune cell growth in the presence of chemotherapy drug diffusion at a Caputo fractional order. To accomplish this, we employed the modified homotopy perturbation method to solve a proposed system of nonlinear differential equations. We obtained the analytical solutions to study the spatiotemporal pathological response of tumor-immune cell growth. Our analysis also considered the impact of the Caputo-fractional order on the system's dynamics and compared the results with the classical integer-order scenario. Our findings demonstrated that the proposed method is an effective and precise technique for understanding the intricate interactions of tumor-immune cell growth. Additionally, we revealed that the Caputo-fractional order plays a significant role in the system's behavior and should not be overlooked in future analyses of such systems. In conclusion, this study holds important implications for cancer research by providing insights into the behavior of tumor-immune cell growth in the presence of time-fractional administration of chemotherapy drugs.

Fetal diseases often lead to the death of many babies during pregnancies. Machine learning and deep learning are promising technologies providing efficient and effective detection and treatment of various fetal diseases. We contribute to the medical field by addressing the critical challenge of fetal disease classification, a concern affecting females and infants. This study utilizes 22 features associated with fetal heart rate extracted from 2126 patient records within the Cardiotocography(CTG) datasets. Our classification system offers a cost-effective, efficient, and accurate solution. It classifies fetal diseases into three categories: Normal, Suspect, and Pathological, based on preprocessed data that underwent MinMax Scaling and employed dimensionality reduction techniques, including Principal Component Analysis(PCA) and Autoencoders. By incorporating dimensionality reduction techniques, the computation time has been reduced from 9 to 26 s to just 4 and 15 s, which is less than half of the original computation time. We assessed the performance of 11 standard machine learning algorithms and various performance metrics to identify the best classification model. We have applied the K-fold Cross-Validation technique to validate our model to improve machine learning models and identify the most effective algorithm. When the results are compared, it is observed that Extreme Gradient Boosting (XGBoost) gained the highest accuracy of 0.99% also highest precision 0.93% and outperformed all the other machine learning algorithms.

In this study, we propose a computationally-light and robust neural network for estimating heart rate in remote healthcare systems. We develop a model that can be trained on consumer-grade graphics processing units (GPUs), and can be deployed on edge devices for swift inference. We propose a hybrid model based on convolutional neural network (CNN) and bidirectional long short-term memory (BiLSTM) architectures for estimating heart rate from Electrocardiogram (ECG) and Photoplethysmography (PPG) signals. Considering the sensitive nature of the ECG signals, we ensure a formal privacy guarantee, differential privacy, for the model training. We perform a tight accounting of the overall privacy budget of our training algorithm using the Rényi Differential Privacy technique. We demonstrate that our model outperforms state-of-the-art networks on a benchmark dataset for both ECG and PPG signals despite having a much smaller number of trainable parameters and, consequently, much smaller training and inference times. Our CNN-BiLSTM architecture can also provide excellent heart rate estimation performance even under strict privacy constraints. We develop a prototype Arduino-based data collection system that is low-cost, efficient, and useful for providing access to modern healthcare services to people living in remote areas.

Skin cancer can be detected through visual examination and confirmed through dermoscopic analysis and various diagnostic tests. This is because visual observation enables early detection of unique skin images by artificial intelligence. Promising outcomes are shown by several Convolution Neural Network (CNN)–based skin lesion classification systems that employ tagged skin images. This study suggests a practical approach for identifying skin cancers using dermoscopy pictures, improving specialists' ability to distinguish benign from malignant tumors. The Swarm Intelligence (SI) approach used dermoscopy photographs to locate lesions on the skin areas Region of interest (ROI). The Grasshopper Optimization technique produced the best segmentation outcomes. The Speed-Up Robust Features (SURF) approach is applied to extract features based on these findings. Two groups were created using the ISIC-2017, ISIC-2018, and PH-2 databases to categorize skin tumors. With an estimated accuracy in classification of 98.52%, preciseness of 96.73%, and Matthews Correlation Coefficient (MCC) of 97.04%, the suggested classification and segmentation methodologies have been evaluated for classification efficacy, specificity, sensitivity, F-measure, preciseness, the MCC, the dice coefficient, and Jaccard's index. In every performance indicator, the method we suggest outperformed state-of-the-art methods.