دکتر محمد مهدی سپهری

Breast cancer (BC) is the most common cancer and one of the main causes of death among women. This study was conducted to investigate the relationship between BC and nutrition and lifestyle, as well as compare machine learning models in predicting this disease. We designed a questionnaire related to nutrition and lifestyle with a nutritionist's guidance and provided them to 569 patients. After data gathering, we developed some machine-learning algorithms like logistic regression (LR), K-Nearest Neighbor (KNN), Decision tree (DT), and Support vector machine (SVM) classifiers. To make more accurate models, we used an oversampling method to avoid skewing the model due to the lack of balance in the target classes, a grid search method to adjust the model's hyperparameters and finally random forest to identify each variable's importance. The results of this research showed that the accuracy of the DT model was 0.95, SVM and LR were 0.93, and KNN was 0.86. The results indicated the better performance of DT among other models. Our findings show that it is possible to predict the type of cancerous tumor with relatively high accuracy without using specific information about the tumor itself. In particular, in our study, the decision tree has shown better accuracy compared to other models.

Infertility treatment methods that are used today have a limited (or little) success rate, and patients bear a lot of financial and emotional burden to get pregnant. Recently, artificial intelligence has been proposed to evaluate gametes better and choose the best embryo for transfer to the uterus. This study investigated the financial benefit of using artificial intelligence for infertility treatment.

We aim to evaluate the effectiveness of AI in IVF, comparing AI model performance with standard methods and introducing a novel method to measure financial benefits in healthcare resource allocation.

Achieving 75% accuracy, AI significantly outperformed standard methods, reducing the likelihood of discarding viable embryos. This technology streamlines the IVF process, leading to shorter treatment cycles and a cost reduction of 1500 dollars per cycle.

The integration of AI in IVF represents a paradigm shift, improving success rates, cost-efficiency, and patient experiences. Further research and adoption of AI-driven embryo selection can revolutionize infertility treatments, benefiting both patients and healthcare systems.

Hospitals are critical facilities which have a great role to affects the number of mass casualties after disasters. Hence, it is necessary to adopt strategies to increase hospitals preparedness and to improve their resilience. The present paper tries to propose a strategy to cope with surge of demands under disruptions in a hospital. An optimization model for bed management considering collaboration between hospital wards in order to minimize the waiting times of the patients provided in this research and the objective function under the proposed strategy and without the proposed strategy were compared. The results show that the proposed strategy can reduce the patients waiting time under disruptions. Due to the complexity of the proposed model, a Lagrangian relaxation-based heuristic is developed to solve the model. Computational results show that the proposed algorithm is able to reach desirable gap in a reasonable time.

Noise is a critical concern for practical machine learning, especially medical applications. There exist two kinds of noise, including attributes and class noises. Class noise is potentially more dangerous, so various filtering techniques, particularly prediction-based, have been proposed to control it. Great attention to class noise has made the researchers ignorant that attribute noise, in turn, is harmful. Hence, it is improper to utilize prediction-based filtering to correct class noise without regarding attribute noise.

To tackle this problem, we developed a method to fix class noise in the presence of attribute noise. This method excludes noisy components of attributes, based on the information bottleneck principle, by compressing attributes locally and gradually in successive iterations. It uses heterogeneous ensemble filtering to correct class noise. In the initial iteration, filtering is conservative and progressively, in succeeding iterations, tends to majority vote.

We compared the proposed method's predictive performance with the RF majority-vote filter on three real binary classification problems from the UCI repository, including Breast, Transfusion, and Ionosphere. Random forest, adaptive boosting, support vector machines, and naïve Bayes were used for assessing methods from different viewpoints. Results show that the proposed method performed better than the RF majority-vote filter and seems to open a promising research scope for noise filtering.

Our study revealed that correcting class noise by controlling attribute noise enhances the predictive performance of classifiers.

This study attempted to review the articles and explore the gaps and challenges in the areas of resilience, business continuity, risk, and process safety with the aim of providing several directions for future research to understand different research directions in these areas with different perspectives. In addition, in this study, the relationship of articles in these areas with each other was examined. In this research project, related studies were reviewed and reported to identify presented frameworks, models, and methods for them. In the first phase, the articles were divided into three categories according to their similarity, namely “maximizing the value of business continuity and resilience,” “maximizing process safety and the effect of risk and resilience factors,” and “minimizing risk and effect of uncertainty.” In the second phase, the appropriate conceptual frameworks titled “research house” based on resilience, business continuity, and risk categories were created for each category. In the third phase, 22 closed codes were obtained by carefully reviewing the articles, and their co-occurrence network was investigated. The main findings of this article were categorizing the studied articles, providing conceptual frameworks resulting from article analysis, and presenting a conceptual model.

Entropy is a measure of disorder in a system and is widely used in other scientific and engineering disciplines such as statistical mechanics and information theory. In a chaotic supply chain, the goal is to reduce chaotic behaviors and predict the future of the supply chain. In this case, relationships in a three-tier supply chain are considered in a continuous-time environment. In this paper, the chaotic supply chain is investigated and controlled using the entropy minimization method. Due to the dynamic nature of the supply chain and its chaotic behavior, Poincaré mapping of the system has been prepared by the stroboscopic method. Then, by defining Shannon entropy on the map, the entropy of the system is significantly reduced by the gradient descent algorithm.

Prescribing and consuming drugs more than necessary is considered as polypharmacy, which is both wasteful and harmful. The purpose of this paper is to establish an innovative data mining framework for analyzing physicians’ prescriptions regarding polypharmacy. The approach consists of three main steps: pre-modeling, modeling, and post-modeling. In the first step, after collecting and cleaning the raw data, several novel physicians’ features are extracted. In the modeling step, two popular decision trees, i.e., C4.5 and Classification and Regression Tree (CART), are applied to generate a set of If-Then rules in a tree-shaped structure to detect and describe physicians’ features associated with polypharmacy. In a novel approach, the response surface method (RSM) as a tool for hyper-parameter tuning is simultaneously applied along with correlation-based feature selection (CFS) to enhance the performance of the algorithms. In the post-modeling step, the discovered knowledge is visualized to make the results more perceptible, then is presented to domain experts to evaluate whether they make sense or not. The framework has been applied to a real-world dataset of prescriptions. The results have been confirmed by the experts, which demonstrates the capabilities of the data mining framework in the detection and analysis of polypharmacy.

The rising trend of hospital costs as a significant share of healthcare system costs is one of the challenges facing hospital managers today. Hospital as a complex organization includes many factors such as human resources, patient flow and performance indicators and therefore faces a variety of management processes. Hospital costs as part of the in-hospital cash flow are affected by a large number of variables that change over time and interact with each other. This study aimed to provide a model for hospital costs based on the internal behavior of the system in order to control costs.

The research method of the paper is descriptive-analytical. Considering the complex and dynamic nature of the system, a model was designed and presented using the system dynamics approach. Data were collected using interview methods and reviewing past studies. To run the proposed model the computer software (Vensim DSS 6.4E) was employed. After testing the model, six scenarios were defined based on the presented model and its subsystems (financial flow, patient flow and employed nurses) to reduce costs, which include: reducing the average length of stay, increasing the staff productivity, reducing the intensity of hospital care, reducing clothing consumption, modifying the hospital nutrition process and finally the simultaneous implementation of all the above. This system dynamics model integrates all of these subsystem's effects rather than considering them individually which is the strength of system dynamics modeling.

The first scenario, while reducing the total cost by 3.8%, increased the bed admission ratio by 6.5%. It should be noted that this scenario increased the hoteling cost by 2%. The second scenario resulted in a 10% reduction in total cost. The third scenario saved 9% of the total cost. The fourth and fifth scenarios reduced costs 1.5% and 7.5%, respectively by reducing overhead costs. The results showed that the sixth scenario is the most effective policy. It reduced the total cost and the hoteling cost by 26% and 22%, respectively.

Findings indicate that the hospital will face a reduction in cost compared to the current situation by using any of the scenarios but it will see a further reduction with the simultaneous implementation of the scenarios while controlling the cost of hoteling. Based on the results any development in surgery department capacity must be accompanied by a suitable cost control policy.

The health industry is a competitive and lucrative industry that has attracted many investors. Therefore, hospitals must create competitive advantages to stay in the competitive market. Patient satisfaction with the services provided in hospitals is one of the most basic competitive advantages of this industry. Therefore, identifying and analyzing the factors affecting the increase of patient satisfaction is an undeniable necessity that has been addressed in this study.

Because patient satisfaction characteristics used in hospitals may have a hidden relationship with each other, data mining approaches and tools to analyze patient satisfaction according to the questionnaire used We used the hospital. After preparing the data, the characteristics mentioned in the questionnaire for patients, classification models were applied to the collected and cleared data, and with the feature selection methods, effective characteristics Patients were identified and analyzed for satisfaction or dissatisfaction.

Based on the findings of the present study, it can be concluded that the factors of patient mentality of the physician's expertise and skill, appropriate and patient behavior of the physician and food quality (hoteling) respectively have a higher chance of increasing patient satisfaction with Establish services provided in the hospital.

Comparing the approach used in this study with other studies showed that due to the hidden effects of variables on each other and the relatively large number of variables studied, one of the best options for analyzing patient satisfaction questionnaire data, Use of data mining tools and approaches

Population aging has brought a rise in the prevalence of diabetes and hypertension, leading to more cases of renal failure. Hemodialysis, as a method of renal replacement therapy, by far prevails over peritoneal dialysis (93.5% vs. 6.5%). Although arteriovenous fistula (AVF) is frequently chosen as the vascular access route for chronic hemodialysis; it has limitations including non-maturation. As maintenance of an AVF is much more costly than its creation, foreseeing maturation failure can lead to a wiser allocation of patients to AVF surgery or other alternatives, with potential for significant cost containment. Previous studies have some challenges: they used intraoperative and postoperative parameters (AVF blood flow, diameter, and depth) or parameters that are costly to collect (morphologic and functional vessels characteristics), and they used statistical analysis that puts restrictions on data. In this study, we aim to provide a data mining framework for predicting AVF non-maturation using routinely available preoperative parameters, such as serum metabolic values and inflammatory markers.

We investigated the relationship of routinely available systemic inflammatory markers and baseline metabolic values in 114 end-stage renal disease patients (over 35 years of age undergoing their first radio-cephalic AVF access surgery at wrist level for chronic hemodialysis). In this study, for the first time to our knowledge, we applied predictive analytic tools such as Random Forest for retrospective analysis of prospectively collected data between 2011 and 2018.

Our results showed that a combination of inflammatory markers and serum metabolic values can prognosticate AVF maturation outcomes with an accuracy of 0.723, by the 95% confidence interval of (0.715, 0.731) and AUC of 0.853. Also, a combination of inflammatory markers, including albumin, c-reactive protein, erythrocyte sedimentation rate, hemoglobin, lymphocytes, neutrophils, white blood cells, platelets, and red blood cell distribution width, can prognosticate AVF maturation outcomes with an accuracy of 0.674, by the 0.95 confidence interval of (0.665, 0.684) and AUC of 0.824.

Risk stratification of patients for AVF non-maturation before attempting the first AVF surgery may help prevent multiple surgical failures and costly endovascular interventions by allowing vascular surgeons to make an individualized choice of vascular access method for new patients.

The regionalization is a suitable approach to reduce the cost of health services and to increase the number of patients covered by special services. Since the establishment of the Neonatal Intensive Care Unit (NICU) needs expensive equipment and experts, it is critical to find the optimal number and location for NICU beds and referral networks.

The geographical access to NICU beds was investigated by collecting the annual demand and the distance between cities at first. The demand consisted of the number of neonates that were born under 32 weeks of gestational age or having less than 1500 gram birth weight in one province of Iran. Next, the location of the available hospital has defined on the map. A maximizing coverage model was developed to find the optimal location for NICUs by ArcMap software. Scenarios of reducing NICU centers were built to simulate real situations for policymakers. Coverage and average traveled distances were then calculated for each scenario. The results were compared with the natural journey of pregnant women and the available distribution of resources in the province.

The results revealed that reducing the number of NICU centers has had no direct impact on average traveled distance. A comparison of the optimal result with the natural journey of pregnant women represented a long distance traveled. The data also showed that 64% of neonates were born outside of their residential cities, and 31% of them were born outside of their provinces, although the occupation rate of available NICU was less than 50% on average.

The effect of reducing NICU centers on total coverage and average transportation was studied in this paper. The proposed methodology with the objective of equity in access can be used as a referral model to other resource allocation cases in health care.

Mobile health or MHealth refers to the use of mobile phone in healthcare services to enhance the health level of people. Before using MHealth, it is necessary to study the effective factors in physicians’ adoption and acceptance of technology in the field of thalassemia 

This cross sectional study was conducted using the survey and correlation methods. The statistical population of the study consisted of hematologists who were selected using the convenience sampling method. In this study, 58 questionnaires along with structural equations modeling based on partial least squares were used. SPSS and SMART PLS2 were used for data analysis. P values less than 0.05 were considered as statistically significant.

Based on the outcomes of the model from all theories, the coefficient of variation seems to be positive and the possibility of test is lower than 5%. The results indicated that all factors introduced in the proposed model are significantly effective in MHealth technology adoption.

In this study, using the inputs from hematologists in hospitals and clinics in Tehran, it was aimed to find the factors affecting the hematologists’ decision to use mobile health technology in reducing the complications of blood transfusion in patients with thalassemia who needed blood transfusion. Thus, plans were made determine to priorities and the existing conditions to implement this new system. Also, the strengths and weaknesses of each factor were measured to improve the weaker factors. UTAUT was used to determine the acceptance factors.  After reviewing the results, the use of this model is recommended to physicians.

Quality of Intensive care has got more attention in case of the high cost of healthcare and the potential for harm. Poor-quality care causes high cost and quality improvement initiatives in the ICU lead to an improvement in outcomes as well as a decrease in costs. One of the crucial tools that allow physicians and nurses to monitor change in a quality improvement effort is the development of an electronic database for data collection and reporting. The objective of Intensive Care Registries is to create a high-quality registry of patients through a collaboration of academic health centers performing uniform data collection with the purpose of improving the quality and accuracy of healthcare decisions and provide a data-driven clinical decision support system for critical care medicine.

This article reviews real-world data sources in healthcare and considers registry as the main tool to address health services and outcomes research questions in critical care, and briefly describes objective, inputs and outputs of intensive care registries. As it can be comprehended from library research, the combination of patient clinical care data, quality parameters, and ICU operating costs, integrated into an electronic database, provides a valuable tool for quality improvement and overall efficiency of offered care.

Using Big Data effectively within ICUs for supporting clinical decision making can lead to predict numerous diseases and help to discover new patterns in healthcare. The ability to process multiple high-speed clinical data streams from multiple centers could dramatically improve both healthcare efficiency and patient outcomes.

To gain this goal, developing reliable and standardized health analytics platforms as well as quality improvement processes that translate analytical results into new clinical guidelines, is recommended.

Effective and continuous monitoring is one of the essential requirements of ICUs. In recent years, the advent of new technologies has led to the development of smart ICUs that automatically gather, store and analyze healthcare information. In this paper, we aim to design the information view of an IOT-based framework for smart ICUs to provide preventive and intelligent healthcare. In previous studies conducted to develop architectures for smart ICUs, data processing is done in a large centralized fashion by cloud computers. This approach may take a long time when dealing with a large amount of data. In this paper, we used the fog technique to bring some computation and storage resources to the edge of the network instead of relying on the cloud for everything.

  A reference architecture model and new technologies such as IoT, cloud computing, fog computing and smart sensors were used to design the information view of a smart ICU architecture. This view consists of five layers of data acquisition, transfer, storage, process and presentation layer. The proposed framework gathers patients’ health-related data continuously and provides real-time analysis.

In this paper, training the models, that took a long time, was performed in the cloud. Instead, classifying the new records, which took much less time, was performed in the fog. This greatly increased the speed of operations (2 ms vs 13590 ms). In addition, conducting calculations in the fog intensely reduced the transmission delays (8 ms vs 108 ms for only SPO2 variable).

 New technologies were used to provide the information view of a smart ICU framework. Instead of relying solely on cloud, this paper uses fog technology to bring some computation and storage resources to the edge of the network. This greatly reduced the transmission latency and provided real-time analysis.

The high prevalence of cardiovascular diseases has caused many health problems in countries. Cardiac Rehabilitation Programs (CRPs) is a complementary therapy for Percutaneous Coronary Intervention (PCI) patients. However, PCI patients hardly attend CRPs. This study aims to decipher the reasons why PCI patients rarely participate in CRPs after PCI. The parameters affecting the attendance of the patients at CRPs were identified by using the previous studies and opinions of experts. A questionnaire was designed based on the identified parameters and distributed among PCI patients who were referred to Tehran Heart Center Hospital. According to data mining approach, 184 samples were collected and classified with three algorithms (Decision Trees, k-Nearest Neighbor (kNN), and Naïve Bayes). The obtained results by decision trees were superior with the average accuracy of 82%, while kNN and Naïve Bayes obtained 81.2% and 78%, respectively. Results showed that lack of physician’s advice was the most significant reason for non-participation of PCI patients in CRPs (P< .0001). Other factors were family and friends’ encouragement, paying expenses by insurance, awareness of the benefits of the CRPs, and comorbidity, respectively. Results of the best model can enhance the quality of services, promote health and prevent additional costs for patients.

Surgical suits allocate a large amount of expenses to hospitals; on the other hand, they constitute a huge part of hospital revenues. Patient flow optimization in a surgical suite by omitting or reducing bottlenecks which cause loss of time is one of the key solutions in minimizing the patients’ length of stay[1] (LOS) in the system, lowering the expenses, increasing efficiency, and also enhancing patients’ satisfaction. In this paper, an analytical model based on simulation aiming at patient flow optimization in the surgical suite has been proposed. To achieve such a goal, first, modeling of patients' workflow was created by using discrete-event simulation. Afterward, improvement scenarios were applied in the simulated model of surgical suites. Among defined scenarios, the combination scenario consisting of the omission of the waiting time between the patients’ entrance to the surgical suite and beginning of the admission procedure, being on time for the first operation, and adding a resource to the resources of the transportation and recovery room, was chosen as the best scenario. The results of the simulation indicate that performing this scenario can decrease patients’ LOS in such a system to 22.15%.

Thalassemia is the acute hereditary anemia and the most common hemoglobin disorder in the world. The main treatment for this disease is the persistent blood injection, but the injection of blood can have different complications. These complications affect the quality of life of patients and increase the risk of mortality. Moreover, it increases the use of healthcare services and hospital costs. Predicting the risk of complications before blood transfusion, more appropriate alternative treatment can be selected to prevent or reduce the complications. Moreover, identifying high-risk patients and following them after transfusion provides the possibility of timely interventions. So far, several studies have analyzed the effects of blood transfusion and the risk factors of these complications by statistical methods. However, few studies have attempted to predict these complications. In this study, the risk of post-transfusion complications in thalassemia patients is predicted using machine learning algorithms.

The cross-sectional data were collected from 3489 cases in 12 thalassemia centers in Tehran province and 14 thalassemia centers in Mazandaran province in 2018. A set of different classification models including classic and deep learning techniques were trained and studied on this data set.

The results showed that machine learning methods have good accuracy to predict the risk of post-transfusion complications. According to the results, the deep learning method has improved the results considerably in comparison to other models (precision=0.21, sensitivity=0.77, f1-score=0.33).

In this study, machine learning methods were used to predict the occurrence of post-transfusion complications in thalassemia patients. Finally, the deep learning method produced the best prediction results. Using this method,  of patients who will suffer complications are detected before transfusions. Appropriate alternative methods can be used for treating these patients in order to prevent or reduce transfusion complications.

Evaluation and assessment of hospitals Medical faculty performance plays a vital role in improving organizational performance, patient and client satisfaction, learner satisfaction and increase of the brand in hospitals. So faculty evaluation in terms of all aspects is essential in hospitals.
In this study, a multi-dimensional model from different perspectives (hospital management, research department, students and faculty) is provided for evaluation of faculty in hospitals. For this purpose, the indicators influencing assessment were identified and categorized in four dimensions: education, hospital management, research and clinical. Then, to prioritize and weight factors as well as prioritize the faculty, a multi-criteria decision-making model was developed and was solved using triangular FAHP approach.
The results indicate that timely and active presence in the clinic on patient bedside, sending patient to other medical centers, private and non-governmental, with no scientific reason, active cooperation with implementing quality improvement plans of health care in hospitals and etc. have the highest priority and factors such as active cooperation with the hospital committees, quality of theses, physical presence in the office hours and etc. received the lowest score.
he results show that all three aspects of physicians' performance are important and should be considered in their evaluation And the fuzzy hierarchical analysis method has shown this very well.