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

Diabetes disease seriously threatens people's health and is becoming more common nowadays. Diabetes Mellitus (DM) is a condition caused by high blood sugar levels, inactivity, unhealthy eating, being overweight, and other factors. This research article analyzed and examined various risk prediction models and algorithms for diabetes, including Type 1, Type 2, and Gestational Diabetes. This study develops several Machine Learning (ML) models for predicting diabetes using various datasets. The process involves producing highly informative features called Feature Engineering (FE). We used the Pima Indian Diabetes Dataset (PIDD) to experiment with and examine the effectiveness of ML models' ability to predict diabetes. Using Python programming, we used three classification algorithms, Logistic Regression, Gradient Boost, and Decision Tree, and combined feature selection techniques among the classification techniques, Decision Tree has the highest accuracy rate (91 %), precision (96 %), recall (92 %), and Fi score (94 %).

An infodemic is an information epidemic capable of compromising public health. This manuscript proposes an infoveillance method suitable for listening to web concerns on health to develop adequate infodemiological responses based on the World Health Organization indications. In particular, the case of COVID-19 vaccinations in Italy was investigated. Web interest and concern in COVID-19 vaccines over the past week (January 8–14, 2023) was investigated via the websites Google Trends and Talkwalker by searching for appropriate keywords. Thanks to the analysis of related queries and topics, it was possible to determine and examine the most debated topics relating to specific side effects. Emotional reactions regarding COVID-19 vaccines have been negative in varying percentages between 40 and 70 %, depending on the topic discussed. Feelings of alarm, derision, doubt, and anger were common (about 60 %). The concerns were mainly about the effectiveness against recent COVID-19 variants and alleged side effects such as sudden death, tumors, myocarditis, prion disease, and high ferritin. The most used media among those scrutinized was Twitter (over 90 % of interactions). The male audience participated more and showed more negativity than the female one. The age groups mainly involved were the under-45s. This research discussed the combined use of Google Trends and Talkwalker to conduct rapid infoveillance surveys. The results found showed that the web public has many doubts about COVID-19 vaccines, including the appearance of very rare or unproved side effects. Based on the WHO infodemic management strategy, it is essential that this or similar approaches are adopted by health and government authorities to listen to the community and calibrate appropriate infodemiological responses aimed at preserving public health.

Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) caused the start of the COVID-19 outbreak in the world, including Malaysia and Thailand. This study identifies the trend of the COVID-19 outbreak before and after the vaccination campaign by using the Susceptible-Exposed-Infectious-Recovered (SEIR) and Susceptible-Exposed-Infectious-Recovered-Vaccinated (SEIRV) models. Moreover, we predict the daily reported death and recovery cases using the SEIR model and Holt's linear trend method and then evaluate their performance. The data used in this study is real data from Malaysia and Thailand. The SEIRV model provides a comprehensive view of the efficacy of COVID-19 vaccinations in curbing the COVID-19 outbreak. This research reveals that the SEIR model outperforms Holt's linear trend method in predicting daily reported cases.

Deep learning methods in artificial intelligence are used for brain tumor diagnosis as they handle a huge amount of data. Compared to computerized tomography (CT), Ultrasound, and X-ray imaging, Magnetic Resonance Imaging (MRI) is effectively used for machine vision-based brain tumor diagnosis. However, due to the complex nature of the brain, brain tumor diagnosis is always challenging. This research aims to study the effectiveness of deep transfer learning architectures in brain tumor diagnosis. This paper applies four transfer learning architectures- InceptionV3, VGG19, DenseNet121, and MobileNet. We used a dataset with data from three benchmark databases of figshare, SARTAJ, and Br35H to validate the models. These databases have four classes: pituitary, no tumor, meningioma, and glioma. Image augmentation is applied to make the classes balanced. Experimental results demonstrate that the MobileNet outperforms competing methods by exhibiting an accuracy of 99.60%.

Diabetes is a chronic glucose metabolism disorder with severe clinical consequences. The prevalence of diabetes mellitus, in particular Type 2 Diabetes (T2D), is rising dramatically globally. Several clinical trials provide evidence that lifestyle interventions can prevent or delay the development of T2D, but the impact of lifestyle interventions is seldom investigated using a mathematical model. This study assesses the effects of lifestyle interventions on people by constructing an ordinary differential equation model. In this paper, a general model is developed based on the dynamics of T2D by incorporating a control variable termed as healthy lifestyle. The population is subdivided into five classes: susceptible, affected, treated, healthy lifestyle, and prevented. Sensitivity analysis has been performed to identify the most important parameters, and the stability of the equilibrium point is analyzed. Numerical simulations are conducted using a diabetes data set in Bangladesh to investigate the model's dynamic behavior. The results from this study reveal that maintaining a healthy lifestyle slows disease progression. The sensitivity analysis shows that the healthy lifestyle rate, treatment rate, and diabetes rate from susceptible and healthy lifestyle classes are the most sensitive parameters. Moreover, the study also concludes that diabetes cannot completely be eliminated, but with proper control measures, the burden can be reduced. The findings from the study provide strong reasons to continue implementing lifestyle interventions to prevent the global epidemic and its adverse effects.

The rise of drug resistance has become a major obstacle in treating tuberculosis (TB), significantly contributing to the increasing disease burden. Therefore, it is essential to study the transmission dynamics of the disease, considering the factors that contribute to the strain’s impact on the disease burden, using an epidemiological model. We present a deterministic mathematical model that explores the dynamics of TB with two strains: drug-susceptible and drug-resistant, taking into account exogenous re-infection. We thoroughly analyze to gain insights into the behavior of the model. The qualitative analysis of the model reveals an interesting phenomenon known as “backward bifurcation,” where both stable disease-free and stable endemic equilibria coexist when the associated reproduction number is less than one. In the absence of exogenous re-infection, the model shows the existence of unique positive endemic equilibria. Numerical simulations were conducted, yielding noteworthy results. Increasing the treatment rate for individuals infected with the drug-susceptible strain reduces the number of new cases of drug-susceptible TB while increasing the detection of drug-resistant TB cases. The simulations demonstrate that drug-susceptible and drug-resistant TB strains can coexist when their reproduction numbers exceed one without competitive exclusion occurring. In summary, this study sheds light on the challenges posed by drug resistance in TB treatment and highlights the importance of understanding the disease dynamics through mathematical modeling to develop effective strategies for its control.

Medical infrastructure disruptions during disasters pose a major threat to critically ill patients with advanced chronic kidney disease or end-stage renal disease. There is a need to assess the potential threat to critical care facilities from hazardous events to improve patient access to dialysis treatment. We propose optimization models for patient reallocation and temporary medical facility placement to equitably improve critical care system resilience. We leverage human mobility data in Texas to assess patient access to critical care facilities and dialysis centers under the simulated hazard impacts. The optimization model was formulated as an integer programming and solved by COIN-OR Branch-and-Cut (CBC) solver. The results show (1) the capability of the optimization model in efficient patient reallocation to alleviate disrupted access to dialysis facilities; (2) the importance of large facilities in maintaining the system functionality. The critical care system, particularly the network of dialysis centers, is heavily reliant on a few larger facilities, characteristic of scale-free networks, making it susceptible to targeted disruption, such as capacity failures. (3) Considering equity in the optimization model formulation reduces access loss for vulnerable populations in the simulated scenarios. (4) The proposed temporary facilities placement could improve access for the vulnerable population, thereby improving the equity of access to critical care facilities in disaster. The proposed patient reallocation optimization model and temporary facilities placement offer a data-driven and analytics-based decision support tool tailored to the needs of healthcare organizations across private and public sectors to proactively mitigate the potential loss of access to critical care facilities during disasters.

The first seven vertebrae of our spine are called the cervical spine. It supports the weight of our head, encloses and safeguards our spinal cord, and permits a variety of head motions. The seven cervical vertebrae are joined at the rear of the bone by a kind of joint known as a facet joint. These joints enable us to move our necks forward, backward, and twist. Fractures of the cervical spine are a medical emergency that may lead to lifelong paralysis or even death. If left untreated and undetected, these fractures can worsen over time. Using computed tomography, a cervical spine fracture in individuals can be accurately diagnosed. Given the scarcity of research on the practical use of deep learning methods in detecting spine fractures in persons, it is imperative to address this gap. This study uses a dataset containing fracture and normal cervical spine computed tomography images. This study proposed modified transfer-learning-based MobileNetV2, InceptionV3, and Resnet50V2 models. An ablation study was also conducted to determine the optimal custom layers for models and data augmentation techniques. In addition, evaluation metrics have been used to analyze and compare the model's performance. Among all the approaches, MobileNetV2 with augmentation has achieved the highest accuracy of 99.75%. Furthermore, the best-performing model has been deployed in a smartphone-based Android application.

Monkeypox (Mpox) is a viral disease primarily affecting animals that can occasionally be transmitted to humans. While it can cause a rash and flu-like symptoms, it is generally less severe than smallpox. Awareness of the disease, proper preventive measures, and timely medical care are essential in managing and controlling Mpox outbreaks. We develop a new compartmentalized mathematical model to study the impact of intense awareness on the control and mitigation of the Mpox disease. The model captures the dynamics of humans and hosts (mammals) in transmitting Mpox disease while the human susceptible population was stratified into two subgroups of aware and unaware individuals. The model is analytically analyzed, and the simulation shows the prospect of an effective public awareness campaign program in reducing the cumulative incidence (new cases) of Mpox disease. Furthermore, the results suggest the need for consistency and continuous awareness program, adhering to public health measures to increase awareness of the Mpox disease.

Retinal fundus images play a crucial role in the early detection of eye problems, aiding in timely diagnosis and treatment to prevent vision loss or blindness. With advancements in technology, Convolutional Neural Network (CNN) algorithms have emerged as effective tools for recognition, delineation, and classification tasks. This study proposes a comprehensive review of CNN algorithms used for retinal fundus image segmentation and classification. Our review follows a systematic approach, exploring diverse repositories to identify studies employing CNN to segment and classify retinal fundus images. Utilizing CNNs in the segmentation and classification of retinal fundus images can enhance the precision of segmentation outcomes and alleviate the sole dependence on human experts. This approach enables more accurate segmentation results, reducing the burden on human experts. A total of sixty-two studies are included in our review, analyzing aspects such as database usage and the advantages and disadvantages of the methods employed. The review provides valuable insights, limitations, observations, and future directions in the field. Despite certain limitations, the findings indicate that CNN algorithms consistently achieve high accuracies. The comprehensive examination of the included studies sheds light on the potential of CNN in retinal fundus image analysis.