A Bayesian belief network-based analytics methodology for early-stage risk detection of novel diseases.

During a pandemic, medical specialists have substantial challenges in discovering and validating new disease risk factors and designing effective treatment strategies. Traditionally, this approach entails several clinical studies and trials that might last several years, during which strict preventive measures are enforced to manage the outbreak and limit the death toll. Advanced data analytics technologies, on the other hand, could be utilized to monitor and expedite the procedure. This research integrates evolutionary search algorithms, Bayesian belief networks, and innovative interpretation techniques to provide a comprehensive exploratory-descriptive-explanatory machine learning methodology to assist clinical decision-makers in responding promptly to pandemic scenarios. The proposed approach is illustrated through a case study in which the survival of COVID-19 patients is determined using inpatient and emergency department (ED) encounters from a real-world electronic health record database. Following an exploratory phase in which genetic algorithms are used to identify a set of the most critical chronic risk factors and their validation using descriptive tools based on the concept of Bayesian Belief Nets, the framework develops and trains a probabilistic graphical model to explain and predict patient survival (with an AUC of 0.92). Finally, a publicly available online, probabilistic decision support inference simulator was constructed to facilitate what-if analysis and aid general users and healthcare professionals in interpreting model findings. The results widely corroborate intensive and expensive clinical trial research assessments.

Predicting and Mitigating Freshmen Student Attrition: A Local-Explainable Machine Learning Framework.

With the emergence of novel methods for improving machine learning (ML) transparency, traditional decision-support-focused information systems seem to need an upgrade in their approach toward providing more actionable insights for practitioners. Particularly, given the complex decision-making process of humans, using insights obtained from group-level interpretation of ML models for designing individual interventions may lead to mixed results. The present study proposes a hybrid ML framework by integrating established predictive and explainable ML approaches for decision support systems involving the prediction of human decisions and designing individualized interventions accordingly. The proposed framework is aimed at providing actionable insights for designing individualized interventions. It was showcased in the context of college students' attrition problem using a large and feature-rich integrated data set of freshman college students containing information about their demographics, educational, financial, and socioeconomic factors. A comparison of feature importance scores at the group- vs. individual-level revealed that while group-level insights might be useful for adjusting long-term strategies, using them as a one-size-fits-all strategy to design and implement individual interventions is subject to suboptimal outcomes.

An explanatory machine learning framework for studying pandemics: The case of COVID-19 emergency department readmissions.

One of the major challenges that confront medical experts during a pandemic is the time required to identify and validate the risk factors of the novel disease and to develop an effective treatment protocol. Traditionally, this process involves numerous clinical trials that may take up to several years, during which strict preventive measures must be in place to control the outbreak and reduce the deaths. Advanced data analytics techniques, however, can be leveraged to guide and speed up this process. In this study, we combine evolutionary search algorithms, deep learning, and advanced model interpretation methods to develop a holistic exploratory-predictive-explanatory machine learning framework that can assist clinical decision-makers in reacting to the challenges of a pandemic in a timely manner. The proposed framework is showcased in studying emergency department (ED) readmissions of COVID-19 patients using ED visits from a real-world electronic health records database. After an exploratory feature selection phase using genetic algorithm, we develop and train a deep artificial neural network to predict early (i.e., 7-day) readmissions (AUC = 0.883). Lastly, a SHAP model is formulated to estimate additive Shapley values (i.e., importance scores) of the features and to interpret the magnitude and direction of their effects. The findings are mostly in line with those reported by lengthy and expensive clinical trial studies.

An explanatory analytics framework for early detection of chronic risk factors in pandemics.

Timely decision-making in national and global health emergencies such as pandemics is critically important from various aspects. Especially, early identification of risk factors of contagious viral diseases can lead to efficient management of limited healthcare resources and saving lives by prioritizing at-risk patients. In this study, we propose a hybrid artificial intelligence (AI) framework to identify major chronic risk factors of novel, contagious diseases as early as possible at the time of pandemics. The proposed framework combines evolutionary search algorithms with machine learning and the novel explanatory AI (XAI) methods to detect the most critical risk factors, use them to predict patients at high risk of mortality, and analyze the risk factors at the individual level for each high-risk patient. The proposed framework was validated using data from a repository of electronic health records of early COVID-19 patients in the US. A chronological analysis of the chronic risk factors identified using our proposed approach revealed that those factors could have been identified months before they were determined by clinical studies and/or announced by the United States health officials.

A Machine Learning Framework for Assessing the Risk of Venous Thromboembolism in Patients Undergoing Hip or Knee Replacement.

Venous thromboembolism (VTE) is a well-recognized complication that is prevalent in patients undergoing major orthopedic surgery (e.g., total hip arthroplasty and total knee arthroplasty). For years, to identify patients at high risk of developing VTE, physicians have relied on traditional risk scoring systems, which are too simplistic to capture the risk level accurately. In this paper, we propose a data-driven machine learning framework to identify such high-risk patients before they undergo a major hip or knee surgery. Using electronic health records of more than 392,000 patients who undergone a major orthopedic surgery, and following a guided feature selection using the genetic algorithm, we trained a fully connected deep neural network model to predict high-risk patients for developing VTE. We identified several risk factors for VTE that were not previously recognized. The best FCDNN model trained using the selected features yielded an area under the ROC curve (AUC) of 0.873, which was remarkably higher than the best AUC obtained by including only risk factors previously known in the medical literature. Our findings suggest several interesting and important insights. The traditional risk scoring tables that are being widely used by physicians to identify high-risk patients are not considering a comprehensive set of risk factors, nor are they as powerful as cutting-edge machine learning methods in distinguishing low- from high-risk patients.

Miscommunication in the age of communication: A crowdsourcing framework for symptom surveillance at the time of pandemics.

OBJECTIVE: There was a significant delay in compiling a complete list of the symptoms of COVID-19 during the 2020 outbreak of the disease. When there is little information about the symptoms of a novel disease, interventions to contain the spread of the disease would be suboptimal because people experiencing symptoms that are not yet known to be related to the disease may not limit their social activities. Our goal was to understand whether users' social media postings about the symptoms of novel diseases could be used to develop a complete list of the disease symptoms in a shorter time. MATERIALS AND METHODS: We used the Twitter API to download tweets that contained 'coronavirus', 'COVID-19', and 'symptom'. After data cleaning, the resulting dataset consisted of over 95,000 unique, English tweets posted between January 17, 2020 and March 15, 2020 that contained references to the symptoms of COVID-19. We analyzed this data using network and time series methods. RESULTS: We found that a complete list of the symptoms of COVID-19 could have been compiled by mid-March 2020, before most states in the U.S. announced a lockdown and about 75 days earlier than the list was completed on CDC's website. DISCUSSION & CONCLUSION: We conclude that national and international health agencies should use the crowd-sourced intelligence obtained from social media to develop effective symptom surveillance systems in the early stages of pandemics. We propose a high-level framework that facilitates the collection, analysis, and dissemination of information that are posted in various languages and on different social media platforms about the symptoms of novel diseases.

A novel diffusion-based model for estimating cases, and fatalities in epidemics: The case of COVID-19.

While the COVID-19 pandemic is still ongoing in a majority of countries, a wealth of literature published in reputable journals attempted to model the spread of the disease. A vast majority of these studies dealt with compartmental models such as susceptible-infected-recovered (SIR) model. Although these models are rather simple, intuitive, and insightful, we argue that they do not necessarily provide a good enough fit to the reported data, which are usually reported in the form of daily fatalities and cases during pandemics. This study proposes an alternative analytics approach that relies on diffusion models to predict the number of cases and fatalities in epidemics. After evaluating several of the well-known and widely used diffusion models in business literature, including ADBUDG, Gompertz, and Bass models, we developed and used a modified/improved version of the original Bass diffusion model to address the shortcomings of the ordinary compartmental models such as SIR and demonstrated its applicability on the portrayal of the COVID-19 pandemic incident data. The proposed model differentiates itself from other similar models by fitting the data without the need for preprocessing, requiring no initial conditions and assumptions, not involving in heavy parameterization, and also properly addressing the pressing issues such as undocumented cases, length of infectious or recovery periods.

No Place Like Home: Cross-National Data Analysis of the Efficacy of Social Distancing During the COVID-19 Pandemic.

BACKGROUND: In the absence of a cure in the time of a pandemic, social distancing measures seem to be the most effective intervention to slow the spread of disease. Various simulation-based studies have been conducted to investigate the effectiveness of these measures. While those studies unanimously confirm the mitigating effect of social distancing on disease spread, the reported effectiveness varies from 10% to more than 90% reduction in the number of infections. This level of uncertainty is mostly due to the complex dynamics of epidemics and their time-variant parameters. However, real transactional data can reduce uncertainty and provide a less noisy picture of the effectiveness of social distancing. OBJECTIVE: The aim of this paper was to integrate multiple transactional data sets (GPS mobility data from Google and Apple as well as disease statistics from the European Centre for Disease Prevention and Control) to study the role of social distancing policies in 26 countries and analyze the transmission rate of the coronavirus disease (COVID-19) pandemic over the course of 5 weeks. METHODS: Relying on the susceptible-infected-recovered (SIR) model and official COVID-19 reports, we first calculated the weekly transmission rate (ß) of COVID-19 in 26 countries for 5 consecutive weeks. Then, we integrated these data with the Google and Apple mobility data sets for the same time frame and used a machine learning approach to investigate the relationship between the mobility factors and ß values. RESULTS: Gradient boosted trees regression analysis showed that changes in mobility patterns resulting from social distancing policies explain approximately 47% of the variation in the disease transmission rates. CONCLUSIONS: Consistent with simulation-based studies, real cross-national transactional data confirms the effectiveness of social distancing interventions in slowing the spread of COVID-19. In addition to providing less noisy and more generalizable support for the idea of social distancing, we provide specific insights for public health policy makers regarding locations that should be given higher priority for enforcing social distancing measures.

Using predictive analytics to identify drug-resistant epilepsy patients.

Epilepsy is one of the most common brain disorders that greatly affects patients' quality of life and poses serious risks to their health. While the majority of the patients positively respond to the existing anti-epilepsy drugs, others who developed the refractory type of epilepsy show resistance against drug therapy and need to undergo advance treatments such as surgery. Given that identifying such patients is not a straightforward process and requires long courses of trial and error with anti-epilepsy drugs, this study aims at predicting those at-risk patients using clinical and demographic data obtained from electronic medical records. Specifically, the study employs several predictive analytics machine-learning methods, equipped with a novel approach for data balancing, to identify drug-resistant patients using their comorbidities and demographic information along with the initial epilepsy-related diagnosis made by their physician. The promising results we obtained highlight the potential use of machine-learning techniques in facilitating medical decisions and suggest the possibility of extending the proposed approach for developing a clinical decision support system for medical professionals.

Examining the effect of prescription sequence on developing adverse drug reactions: The case of renal failure in diabetic patients.

OBJECTIVES: While the effect of medications in development of Adverse Drug Reactions (ADRs) have been widely studied in the past, the literature lacks sufficient coverage in investigating whether the sequence in which [ADR-prone] drugs are prescribed (and administered) can increase the chances of ADR development. The present study investigates this potential effect by applying emergent sequential pattern mining techniques to electronic health records. MATERIALS AND METHODS: Using longitudinal medication and diagnosis records from more than 377,000 diabetic patients, in this study, we assessed the possible effect of prescription sequences in developing acute renal failure as a prevalent ADR among this group of patients. Relying on emergent sequential pattern mining, two statistical case-control approaches were designed and employed for this purpose. RESULTS: The results taken from the two employed approaches (i.e. 76.7% total agreement and 68.4% agreement on the existence of some significant effect) provide evidence for the potential effect of prescription sequence on ADRs development evidenced by the discovery that certain sequential patterns occurred more frequently in one group of patients than the other. CONCLUSION: Given the significant effects shown by our data analyses, we believe that design and implementation of automated clinical decision support systems to constantly monitor patients' medication transactions (and the sequence in which they are administered) and make appropriate alerts to prevent certain possible ADRs, may decrease ADR occurrences and save lives and money.

A chronological pharmacovigilance network analytics approach for predicting adverse drug events.

Objectives: This study extends prior research by combining a chronological pharmacovigilance network approach with machine-learning (ML) techniques to predict adverse drug events (ADEs) based on the drugs' similarities in terms of the proteins they target in the human body. The focus of this research, though, is particularly centered on predicting the drug-ADE associations for a set of 8 common and high-risk ADEs. Materials and methods: large collection of annotated MEDLINE biomedical articles was used to construct a drug-ADE network, and the network was further equipped with information about drugs' target proteins. Several network metrics were extracted and used as predictors in ML algorithms to predict the existence of network edges (ie, associations or relationships). Results: Gradient boosted trees (GBTs) as an ensemble ML algorithm outperformed other prediction methods in identifying the drug-ADE associations with an overall accuracy of 92.8% on the validation sample. The prediction model was able to predict drug-ADE associations, on average, 3.84 years earlier than they were actually mentioned in the biomedical literature. Conclusion: While network analysis and ML techniques were used in separation in prior ADE studies, our results showed that they, in combination with each other, can boost the power of one another and predict better. Moreover, our results highlight the superior capability of ensemble-type ML methods in capturing drug-ADE patterns compared to the regular (ie, singular), ML algorithms.