Health Information management
Saba Paydar; GholamAli Reisi Ardali; Hossein Raeisi
Abstract
AbstractIntroduction: Nowadays, hospitals face challenges such as overcrowding in the emergency department and increased chaos and disruption in staff work, which leads to increased patient dissatisfaction. With the advancement of artificial intelligence and the expansion of data mining, predicting patient ...
Read More
AbstractIntroduction: Nowadays, hospitals face challenges such as overcrowding in the emergency department and increased chaos and disruption in staff work, which leads to increased patient dissatisfaction. With the advancement of artificial intelligence and the expansion of data mining, predicting patient admission has become very important. The aim of this research is to predict patient admission in the emergency department of Imam Ali (AS) Hospital in Shahrekord.Methods: In this research, 2180 patient records from the emergency department of the hospital were examined. Initial patient information, including personal details, vital signs, and triage level which were recorded by nurses, were extracted. Using pairwise comparison matrix, the effective features were selected by experts. Then, using naive Bayes, decision tree, random forest, and support vector machine algorithms, the data was classified. Results: Out of the 15 collected features, 9 features were selected by experts, and the results showed that the random forest algorithm had the best performance in predicting patient admission in this case study, with an accuracy of 92/2%Conclusion: These results demonstrate the importance of using artificial intelligence and data mining methods in hospital management and patient admission prediction. It can serve as a helpful tool in decision-making processes. Key words: Data mining, Forecasting, Patient admission, Hospital Emergency Services
Leila Baradaran-Sorkhabi; Farhad Soleimanian-Gharehchopogh; Jafar Shahmfar
Abstract
Introduction: Data mining seems to be a good tool for showing underlying knowledge of Medical Big Data (MBD). Understanding characteristics of data and possible challenges are the first steps of the journey. This study endeavors to inspect reasons, effects, and solutions of challenges as well as benefits ...
Read More
Introduction: Data mining seems to be a good tool for showing underlying knowledge of Medical Big Data (MBD). Understanding characteristics of data and possible challenges are the first steps of the journey. This study endeavors to inspect reasons, effects, and solutions of challenges as well as benefits of MBD mining.Methods: In so doing, PubMed, ScienceDirect, Springer, and Google Scholar databases were scrutinized using two groups of keywords for benefits and challenges in the years 2011-202. The search language was English. Single-purpose studies were excluded and those studies that were focused on MBD mining were included. Then, challenge was examined separately and the results were categorized.Results: Extracted knowledge from MBD enhances quality of care. However, low-quality performance in gathering and storing the data, properties of big data, and inherent structure of medical data cause many problems for mining methods. Inconsistency, veracity, privacy, and security issues are the major challenging problems. Standardization and enhancing quality of data gathering, storing, and representing tasks are the effective problem prevention strategies. Designing and using appropriate frameworks, algorithms, and structures as well as utilizing machine learning and artificial intelligence techniques are the most effective solutions for dealing with the challenges.Conclusion: MBD was appeared and expanded when the world was not ready for it. Thus, it caused many challenges for mining methods. Some of them are traceable, preventable, and manageable. However, some challenges need novel and intelligent methods that are able to handle MBD.
Elham Pourjani; Sara Najafzadeh; Nader Jafarnia-Dabanloo
Abstract
Introduction: Imputation of missing values in a medical data set is one of the important challenges in data mining. Therefore, this study was performed with the aim of imputation the missing values of some features of the diabetes and breast cancer datasets.Methods: In this descriptive study, a breast ...
Read More
Introduction: Imputation of missing values in a medical data set is one of the important challenges in data mining. Therefore, this study was performed with the aim of imputation the missing values of some features of the diabetes and breast cancer datasets.Methods: In this descriptive study, a breast cancer dataset consisting of 699 specimens including 458 benign and 241 malignant specimens, along with a diabetes dataset consisting of 768 specimens including 500 non-diabetic specimens and 268 other specimens with diabetes, were used. For the purpose of the imputation of missing values in these two datasets, a model based on a two-layer perceptron neural network was developed, and for the purpose of assessment, support vector machine (SVM) and t test were used.Results: The mean squared errors (MSEs) obtained in the two-layer perceptron neural network model, in the diabetes dataset about 0.03 and in the breast cancer dataset about 0.04, were less than the MSEs obtained in the imputation method with the mean value. The values imputed by the model were closer to the actual value than the values imputed with the mean value. Accuracy and sensitivity of disease classification in the case of missing values imputed by the perceptron neural network increased in comparison with the two conventional methods of mean value and the method of deleting missing values, about 2, 4, 2, and 4 percent in the diabetes dataset, and about 1, 3, 2, 5 percent in the dataset breast cancer, respectively. There was a significant difference between the two methods of imputation of missing values with the mean value and imputation by the model.Conclusion: The imputation of the missing values in the medical data set by the two-layer perceptron neural network showed better results in the classification of the disease than the two methods of imputation with the mean value and the method of deleting missing values.
Mohammadhassan Ahmadi; Mohammadreza Ramezanpour; Reihaneh Khorsand
Abstract
Introduction: The incidence of liver diseases in a person can lead to susceptibility to liver cancer in long-term, which is one of the deadliest forms of cancer in the world, which can be prevented. Early diagnosis of liver diseases is essential for their treatment. The purpose of this study was to classify ...
Read More
Introduction: The incidence of liver diseases in a person can lead to susceptibility to liver cancer in long-term, which is one of the deadliest forms of cancer in the world, which can be prevented. Early diagnosis of liver diseases is essential for their treatment. The purpose of this study was to classify the status of liver patients based on laboratory parameters using the data mining approach.Methods: In this descriptive study, particle swarm optimization (PSO) algorithms and adaptive neuron-fuzzy inference system (ANFIS) were used to diagnose liver disorders in healthy individuals and patients. The data were taken from University of California-Irvine (UCI) database. Accuracy, sensitivity, and precision criteria were used to evaluate the proposed method.Results: The combination of ANFIS and PSO algorithm with average accuracy of 14.99 percent was able to detect liver disorders in Indian Liver Patient Dataset (ILPD).Conclusion: The results of this study indicate the high abilities of ANFIS in liver disorders detection. The proposed model has minimum error, and maximum accuracy and precision compared to other models. The application of this model is proposed in the detection of liver diseases.
Raana Mahdavi; Samin Fatehi-Raviz; Hossein Rahmani
Abstract
Introduction: In recent years, the infertility ratio in young couples has been increased a lot in Iran. From the other side, it has been shown that data mining techniques are capable of extracting novel patterns from medical data. In this study, we proposed a comprehensive system called Prediction of ...
Read More
Introduction: In recent years, the infertility ratio in young couples has been increased a lot in Iran. From the other side, it has been shown that data mining techniques are capable of extracting novel patterns from medical data. In this study, we proposed a comprehensive system called Prediction of the best Infertility treatment using Outlier Detection and Ensemble Methods (PIODEM) for predicting of the best infertility treatment method for infertile couples.Methods: This descriptive-correlation study used the information of 527 infertile couples, which collected from Avicenna specialized infertility center, Tehran, Iran. PIODEM consists of three steps: First, PIODEM uses the discriminant analysis to find effective factors for choosing the best infertility treatment. Second, PIODEM detects the outlier samples, and applies a correlation between these samples and the choice of treatment method. Third, it uses ensemble methods to increase the precision of classifiers.Results: The PIODEM system succeeded in discovering affective factors such as male-partner’s age, infertility duration, immotile sperm, decreasing of sperm concentration decrease, total sperm count, morphology, sperm motility, sperm with rapid progressive-a motility, and sperm with slow progressive-b motility. Additionally, PIODEM indicates that if one of four features of sperm concentration, toxoplasma immunoglobulin M (IgM), triiodothyronine (T3) hormone, and thyroid peroxidase (TPO) was an outlier, then the prediction of treatment would be more accurate. Finally, using ensemble methods increased the F-measure of PIODEM system by up to 76%.Conclusion: The PIODEM system is able to discover effective factors in the choice of treatment method, using differential analysis and analysis of pert data. This system offers patient information as input for the treatment method.
Elham Mirzakazemi; Mohammad Ghamgosar-Naseri
Volume 14, Issue 4 , November 2017, , Pages 144-149
Abstract
Introduction: After applying breast cancer treatment methods, there is a possibility of recurrence of the disease. The aim of the present study was using data mining techniques in order to provide predicting models for breast cancer recurrence.Methods: 18 features of 809 patients were used in the current ...
Read More
Introduction: After applying breast cancer treatment methods, there is a possibility of recurrence of the disease. The aim of the present study was using data mining techniques in order to provide predicting models for breast cancer recurrence.Methods: 18 features of 809 patients were used in the current descriptive study. The study consisted of two phases, preprocessing phase and model learning. Expectation Maximization (EM) and Classification and Regression (C and R) were used for the analysis of the first phase. In order to analyze the second phase, the five algorithm model including Neural Network, C and R, the decision tree algorithm C5.0, Bayes Net, and Support Vector Machine (SVM) was used.Results: The accuracy of the EM and C and R algorithms was 0.641 and 0.420, respectively, in the preprocessing phase. The accuracy of Neural Network, C and R, the decision tree algorithm C5.0, Bayes Net, and SVM algorithms was 0.858, 0.865, 0.870, 0.883, and 0.998, respectively, for the model learning phase.Conclusion: According to the findings, the model with the application of EM algorithm in the first phase and SVM algorithm in the second phase had the highest functionality. It was also important in determining the treatment process.
Majid Zarabian; Masoud Abessi
Volume 14, Issue 1 , May 2017, , Pages 16-25
Abstract
Introduction: Ischemic Heart Disease (IHD) is one of the costly and controversial topics in the field of healthcare in Iran. Due to limitation in hospital resources for patient care, studying patient’s length of stay (LOS) is very important in hospital management. This study presents suitable models ...
Read More
Introduction: Ischemic Heart Disease (IHD) is one of the costly and controversial topics in the field of healthcare in Iran. Due to limitation in hospital resources for patient care, studying patient’s length of stay (LOS) is very important in hospital management. This study presents suitable models for estimating the LOS of IHD patients and its influencing factors.Methods: In this applied research, LOS of 6524 IHD in-patients admitted to 16 hospitals in Tehran, Iran, between October 2013 and March 2014 and the remedial proceedings provided for them were recorded using treatment services tariff codes. After data collection using a predesigned form and data cleaning, LOS fort models were created using data mining algorithms of artificial neural network (ANN), support vector machines (SVM), chi-squared automatic interaction detection (CHAID), classification and regression trees (CART), and ensemble model in SPSS Clementine software.Results: The average and standard deviation of LOS in this study was 7.727 ± 9.608 days. The linear correlation of models with actual LOS and their relative error were above 0.7 and below 0.5, respectively. The most accurate models were the ensemble model and SVM.Conclusion: According to the proposed models, ischemic patients who required rehabilitation, consultation, radiation therapy, and computerized tomography (CT) scan have longer LOS. Moreover, type of IHD disease and especially California remedial codes are important factors in estimating patients’ LOS.
Ehsan Nabovati; Amir Abas Azizi; Ebrahim Abbasi; Hassan Vakili-Arki; Javad Zarei; Amir Reza Razavi
Volume 10, Issue 6 , December 2012, , Pages 789-799
Abstract
Introduction: In the past decades, machine learning algorithms have become a useful tool for data mining within huge amounts of health data to create prediction models. Burn is one of the diseases that predicting of its outcome has high importance. The aim of this study was to survey two widely used ...
Read More
Introduction: In the past decades, machine learning algorithms have become a useful tool for data mining within huge amounts of health data to create prediction models. Burn is one of the diseases that predicting of its outcome has high importance. The aim of this study was to survey two widely used machine learning algorithms; neural network and decision tree, and compare them with logistic regression method to predict the outcome of burn patients. Methods: In this retrospective observational study, following preprocessing of the data and determining the outcome of patient (live or death), two well-known machine learning algorithms (neural network and decision tree) and logistic regression method were used to create prediction models using data from 4804 burn patients hospitalized in Taleghani Burn Center in Ahvaz during the years 2001-2007. The preprocessing of the data was performed using SPSS (Version16.0), and in the modeling phase, Clementine (Version 12.0) software was used. Moreover, 10-fold cross validation technique was used to validate the model and criteria for evaluating the performance of models were measured and compared. Results: The results showed that the neural network algorithm with accuracy of 97% resulted the most accurate model on the studied data. The decision tree model with 95% accuracy was in the second place and the logistic regression model with an accuracy of 90% was the least accurate. Moreover other evaluating criteria such as sensitivity, specificity, PPV, NPV and AUC showed that performance of the neural network model was better than the others. Conclusion: The current study shows that machine learning algorithms compared with statistical methods create more accurate models. In analyzing the current data, the model created by artificial neural network is more accurate than the other machine learning algorithm, decision tree. Keywords: Data Mining; Machine Learning; Forecasting; Decision Tree; Artificial Neural Network; Burns