Search results

The purpose of this study is to develop an efficient prediction model using vital signs and standard medical score systems, which predicts the clinical severity level of the patient in advance based on the quick sequential organ failure assessment (qSOFA) medical score method.

To predict the clinical severity level of the patient in advance, the authors have formulated a training dataset that is constructed based on the qSOFA medical score method. Further, along with the multiple vital signs, different standard medical scores and their correlation features are used to build and improve the accuracy of the prediction model. It is made sure that the constructed training set is suitable for the severity level prediction because the formulated dataset has different clusters each corresponding to different severity levels according to qSOFA score.

From the experimental result, it is found that the inclusion of the standard medical scores and their correlation along with multiple vital signs improves the accuracy of the clinical severity level prediction model. In addition, the authors showed that the training dataset formulated from the temporal data (which includes vital signs and medical scores) based on the qSOFA medical scoring system has the clusters which correspond to each severity level in qSOFA score. Finally, it is found that RAndom k-labELsets multi-label classification performs better prediction of severity level compared to neural network-based multi-label classification.

This paper helps in identifying patient' clinical status.

The purpose of this paper is to develop evident-based predictive no-show models considering patients’ each past appointment status, a time-dependent component, as an independent predictor to improve predictability.

A ten-year retrospective data set was extracted from a pediatric clinic. It consisted of 7,291 distinct patients who had at least two visits along with their appointment characteristics, patient demographics, and insurance information. Logistic regression was adopted to develop no-show models using two-thirds of the data for training and the remaining data for validation. The no-show threshold was then determined based on minimizing the misclassification of show/no-show assignments. There were a total of 26 predictive model developed based on the number of available past appointments. Simulation was employed to test the effective of each model on costs of patient wait time, physician idle time, and overtime.

The results demonstrated the misclassification rate and the area under the curve of the receiver operating characteristic gradually improved as more appointment history was included until around the 20th predictive model. The overbooking method with no-show predictive models suggested incorporating up to the 16th model and outperformed other overbooking methods by as much as 9.4 per cent in the cost per patient while allowing two additional patients in a clinic day.

The challenge now is to actually implement the no-show predictive model systematically to further demonstrate its robustness and simplicity in various scheduling systems.

This paper provides examples of how to build the no-show predictive models with time-dependent components to improve the overbooking policy. Accurately identifying scheduled patients’ show/no-show status allows clinics to proactively schedule patients to reduce the negative impact of patient no-shows.

The incidence of prostate cancer is increasing from the past few decades. Various studies have tried to determine the survival of patients, but metastatic prostate cancer is still not extensively explored. The survival rate of metastatic prostate cancer is very less compared to the earlier stages. The study aims to investigate the survivability of metastatic prostate cancer based on the age group to which a patient belongs, and the difference between the significance of the attributes for different age groups.

Data of metastatic prostate cancer patients was collected from a cancer hospital in India. Two predictive models were built for the analysis-one for the complete dataset, and the other for separate age groups. Machine learning was applied to both the models and their accuracies were compared for the analysis. Also, information gain for each model has been evaluated to determine the significant predictors for each age group.

The ensemble approach gave the best results of 81.4% for the complete dataset, and thus was used for the age-specific models. The results concluded that the age-specific model had the direct average accuracy of 83.74% and weighted average accuracy of 79.9%, with the highest accuracy levels for age less than 60.

The study developed a model that predicts the survival of metastatic prostate cancer based on age. The study will be able to assist the clinicians in determining the best course of treatment for each patient based on ECOG, age and comorbidities.

The purpose of this paper is to formulate a framework to construct a patient-specific risk score and therefore to classify these patients into various risk groups that can be used as a decision support mechanism by the medical decision makers to augment their decision-making process, allowing them to optimally use the limited resources available.

A conventional statistical model (logistic regression) and two machine learning-based (i.e. artificial neural networks (ANNs) and support vector machines) data mining models were employed by also using five-fold cross-validation in the classification phase. In order to overcome the data imbalance problem, random undersampling technique was utilized. After constructing the patient-specific risk score, k-means clustering algorithm was employed to group these patients into risk groups.

Results showed that the ANN model achieved the best results with an area under the curve score of 0.867, while the sensitivity and specificity were 0.715 and 0.892, respectively. Also, the construction of patient-specific risk scores offer useful insights to the medical experts, by helping them find a trade-off between risks, costs and resources.

The study contributes to the existing body of knowledge by constructing a framework that can be utilized to determine the risk level of the targeted patient, by employing data mining-based predictive approach.

Patient treatment trajectory data are used to predict the outcome of the treatment to particular disease that has been carried out in the research. In order to determine the evolving disease on the patient and changes in the health due to treatment has not considered existing methodologies. Hence deep learning models to trajectory data mining can be employed to identify disease prediction with high accuracy and less computation cost.

Multifocus deep neural network classifiers has been utilized to detect the novel disease class and comorbidity class to the changes in the genome pattern of the patient trajectory data can be identified on the layers of the architecture. Classifier is employed to learn extracted feature set with activation and weight function and then merged on many aspects to classify the undetermined sequence of diseases as a new variant. The performance of disease progression learning progress utilizes the precision of the constituent classifiers, which usually has larger generalization benefits than those optimized classifiers.

Deep learning architecture uses weight function, bias function on input layers and max pooling. Outcome of the input layer has applied to hidden layer to generate the multifocus characteristics of the disease, and multifocus characterized disease is processed in activation function using ReLu function along hyper parameter tuning which produces the effective outcome in the output layer of a fully connected network. Experimental results have proved using cross validation that proposed model outperforms methodologies in terms of computation time and accuracy.

Proposed evolving classifier represented as a robust architecture on using objective function to map the data sequence into a class distribution of the evolving disease class to the patient trajectory. Then, the generative output layer of the proposed model produces the progression outcome of the disease of the particular patient trajectory. The model tries to produce the accurate prognosis outcomes by employing data conditional probability function. The originality of the work defines 70% and comparisons of the previous methods the method of values are accurate and increased analysis of the predictions.

As far as the treatment of most complex issues in the design is concerned, approaches based on classical artificial intelligence are inferior compared to the ones based on computational intelligence, particularly this involves dealing with vagueness, multi-objectivity and good amount of possible solutions. In practical applications, computational techniques have given best results and the research in this field is continuously growing. The purpose of this paper is to search for a general and effective intelligent tool for prediction of patient survival after surgery. The present study involves the construction of such intelligent computational models using different configurations, including data partitioning techniques that have been experimentally evaluated by applying them over realistic medical data set for the prediction of survival in pancreatic cancer patients.

On the basis of the experiments and research performed over the data belonging to various fields using different intelligent tools, the authors infer that combining or integrating the qualification aspects of fuzzy inference system and quantification aspects of artificial neural network can prove an efficient and better model for prediction. The authors have constructed three soft computing-based adaptive neuro-fuzzy inference system (ANFIS) models with different configurations and data partitioning techniques with an aim to search capable predictive tools that could deal with nonlinear and complex data. After evaluating the models over three shuffles of data (training set, test set and full set), the performances were compared in order to find the best design for prediction of patient survival after surgery. The construction and implementation of models have been performed using MATLAB simulator.

On applying the hybrid intelligent neuro-fuzzy models with different configurations, the authors were able to find its advantage in predicting the survival of patients with pancreatic cancer. Experimental results and comparison between the constructed models conclude that ANFIS with Fuzzy C-means (FCM) partitioning model provides better accuracy in predicting the class with lowest mean square error (MSE) value. Apart from MSE value, other evaluation measure values for FCM partitioning prove to be better than the rest of the models. Therefore, the results demonstrate that the model can be applied to other biomedicine and engineering fields dealing with different complex issues related to imprecision and uncertainty.

The originality of paper includes framework showing two-way flow for fuzzy system construction which is further used by the authors in designing the three simulation models with different configurations, including the partitioning methods for prediction of patient survival after surgery. Several experiments were carried out using different shuffles of data to validate the parameters of the model. The performances of the models were compared using various evaluation measures such as MSE.

This paper develops a robust hospital readmission prediction framework by combining the feature selection algorithm and machine learning (ML) classifiers. The improved feature selection is proposed by considering the uncertainty in patient's attributes that leads to the output variable.

First, data preprocessing is conducted which includes how raw data is managed. Second, the impactful features are selected through feature selection process. It started with calculating the relational grade of each patient towards readmission using grey relational analysis (GRA) and the grade is used as the target values for feature selection. Then, the influenced features are selected using the Least Absolute Shrinkage and Selection Operator (LASSO) method. This proposed method is termed as Grey-LASSO feature selection. The final task is the readmission prediction using ML classifiers.

The proposed method offered good performances with a minimum feature subset up to 54–65% discarded features. Multi-Layer Perceptron with Grey-LASSO gave the best performance.

The performance of Grey-LASSO is justified in two readmission datasets. Further research is required to examine the generalisability to other datasets.

In designing the feature selection algorithm, the selection on influenced input variables was based on the integration of GRA and LASSO. Specifically, GRA is a part of the grey system theory, which was employed to analyse the relation between systems under uncertain conditions. The LASSO approach was adopted due to its ability for sparse data representation.

The purpose of this study is to present a large-scale real-world comparative study using pre-COVID lockdown data versus post-COVID lockdown data on predicting shipment times of therapeutic supplies in e-pharmacy supply chains and show that our proposed methodology is robust to lockdown effects.

The researchers used organic data of over 5.9 million records of therapeutic shipments, with 2.87 million records collected pre-COVID lockdown and 3.03 million records collected post-COVID lockdown. The researchers built various Machine Learning (ML) classifier models on the two datasets, namely, Random Forest (RF), Extra Trees (XRT), Decision Tree (DT), Multi-Layer Perceptron (MLP), XGBoost (XGB), CatBoost (CB), Linear Stochastic Gradient Descent (SGD) and the Linear Naïve Bayes (NB). Then, the researchers stacked these base models and built meta models on top of them. Further, the researchers performed a detailed comparison of the performances of ML models on pre-COVID lockdown and post-COVID lockdown datasets.

The proposed approach attains performance of 93.5% on real-world post-COVID lockdown data and 91.35% on real-world pre-COVID lockdown data. In contrast, the turn-around times (TAT) provided by therapeutic supply logistics providers are 62.91% accurate compared to reality in post-COVID lockdown times and 73.68% accurate compared to reality pre-COVID lockdown times. Hence, it is clear that while the TAT provided by logistics providers has deteriorated in the post-pandemic business climate, the proposed method is robust to handle pandemic lockdown effects on e-pharmacy supply chains.

The implication of the study provides a novel ML-based framework for predicting the shipment times of therapeutics, diagnostics and vaccines, and it is robust to COVID-19 lockdown effects.

E-pharmacy companies can readily adopt the proposed approach to enhance their supply chain management (SCM) capabilities and build resilience during COVID lockdown times.

The present study is one of the first to perform a large-scale real-world comparative analysis on predicting therapeutic supply shipment times in the e-pharmacy supply chain with novel ML ensemble stacking, obtaining robust results in these COVID lockdown times.

Overcrowding continues to afflict emergency departments (EDs), and its attendant consequences are becoming increasingly severe. The burden of the COVID-19 pandemic is further escalating the situation worldwide. One of the most critical questions is how to adequately quantify what constitutes overcrowding and determine implications for operations management in improving service efficiency. This paper aims to discuss the aforementioned.

The authors propose the time and class complexity measures for ED service systems, taking into account important patient-level and system characteristics. Using an extensive data set from a Canadian ED, the authors investigate the performance of complexity-based measures in predicting service delays.

The authors find that the complexity measure is potentially more important than some well-known crowding metrics. In particular, EDs can improve service efficiency by managing the level of complexity within a desirable interval. Furthermore, complexity exposes how the interplay between demand-side behavioral changes and supply-side responses affects operational performance. Moreover, the results suggest that arrival patterns—the number of patients of each class arriving per time and times between events (arrivals and service completions)—increase the risk of service delays more than the demand volume.

This paper is the first to provide an extensive investigation into the application of the complexity-based measure for ED crowding. The study demonstrates potential values to be gained in ED service systems if complexity measure is incorporated into their operations management decisions.

The study presented here examines the reporting of patient‐to‐patient bullying in a high secure psychiatric hospital housing adult men. It examines official records of bullying recorded on Suspected Bullying Report forms (SBRs), noting motivations for bullying and exploring how information was recorded. Its subsidiary aim was to explore the value of official records, presenting comparison data from three patient bullying surveys completed at the same hospital. All suspected reports of bullying in the hospital documented for a two‐year period were collected. The patient surveys included data from three data‐sets collected at the same hospital over a five‐year period. The prediction that the patient surveys would suggest a higher level of bullying behaviours than official records was confirmed. Results for official records also demonstrated that motivation for bullying was either not recorded or was reported as a typology, bullying was not likely to be recorded as hospital incidents, and perpetrators were unlikely to be offered supportive intervention. The results are discussed with regard to the value of official records as a reporting mechanism.