Real-world Big Data Research Articles

TeraHeap: Exploiting Flash Storage for Mitigating DRAM Pressure in Managed Big Data Frameworks

Big data analytics frameworks, such as Spark and Giraph, need to process and cache massive datasets that do not always fit on the managed heap. Therefore, frameworks temporarily move long-lived objects outside the heap (off-heap) on a fast storage device. However, this practice results in (1) high serialization/deserialization (S/D) cost and (2) high memory pressure when off-heap objects are moved back for processing. In this article, we propose TeraHeap , a system that eliminates S/D overhead and expensive GC scans for a large portion of objects in analytics frameworks. TeraHeap relies on three concepts: (1) It eliminates S/D by extending the managed runtime (JVM) to use a second high-capacity heap (H2) over a fast storage device. (2) It offers a simple hint-based interface, allowing analytics frameworks to leverage object knowledge to populate H2. (3) It reduces GC cost by fencing the collector from scanning H2 objects while maintaining the illusion of a single managed heap, ensuring memory safety. We implement TeraHeap in OpenJDK8 and OpenJDK17 and evaluate it with fifteen widely used applications in two real-world big data frameworks, Spark and Giraph. We find that for the same DRAM size, TeraHeap improves performance by up to 73% and 28% compared to native Spark and Giraph. Also, it can still provide better performance by consuming up to \(4.6\times\) and \(1.2\times\) less DRAM than native Spark and Giraph, respectively. TeraHeap can also be used for in-memory frameworks and applying it to the Neo4j Graph Data Science library improves its performance by up to 26%. Finally, it outperforms Panthera, a state-of-the-art garbage collector for hybrid DRAM-NVM memories, by up to 69%.

Long-Term Evaluation of Allopurinol Effectiveness in the Thai Population: Insights from 10 Years of Real-World Big Data Analysis

Objective: This study aimed to evaluate the effectiveness and safety profile of allopurinol in Songklanagarind hospital, Thailand. Material and Methods: Electronic medical records (EMRs) of patients having undergone allopurinol treatment for gouty arthritis, or other pertinent indications, were retrospectively reviewed. The primary outcome of this study was the longitudinal measurement of uric acid levels. A linear mixed model (LMM) analysis assessed the statistical significance of uric acid reduction across hospital visits. The safety profile regarding allopurinol-related adverse drug reactions was assessed.Results: A total of 11,277 EMRs were extracted, of which 8,801 EMRs were eligible for further analyses. These records pertain to a cohort of 3,219 unique patients whose visits were arranged chronologically from the earliest to the fifth recorded visit. Overall, uric acid level decreased after allopurinol treatment in all patients. Across subsequent visits, the analysis, using LMM, found a significant reduction in uric acid levels (χ²=1465.4, df=4, p-value<0.001), with the early reduction observed as early as the second visit. The incidence of allopurinol-related adverse reactions was 4.37% (385/8801 cases). Conclusion: The current dosing regimen of allopurinol at Songklanagarind Hospital has demonstrated favorable treatment outcomes and continues to be effective. This data provides insights into pharmacological management and its real-world outcomes.

Learning through Simulation: A Role-Based Learning Exercise to Equip Students in Big Data Requirement Elicitation Skills and Challenges

Bridging the gap between theoretical knowledge and practical application remains challenging in big data education. Our paper proposes a platform where students actively participate in a role-based learning (RBL) exercise designed to simulate real-world big data projects. This RBL exercise assigns each student participant to a business analyst role tasked with eliciting requirements or a customer role that provides those big data requirements. We conducted RBL sessions for postgraduate and undergraduate students learning big data analytics/data warehousing. Through role-playing in a collaborative environment, students face challenges such as unclear objectives, data privacy concerns, and evolving requirements. We evaluated the effectiveness of RBL simulations through brainstorming sessions, which verified that the students achieved the learning outcomes through RBL exercises. We also collected students’ feedback through a survey and found the RBL experience helped us understand the big data requirement elicitation skills, highlighting the significance of communication and collaboration skills. Further, we have evaluated students through a final exam question, and identified that students who participated in the RBL exercise outperformed in the big data requirement elicitation question. In summary, our research demonstrates that this RBL approach offers a valuable learning experience by enabling students to directly experience the complexities of big data requirement elicitation and identify the future requirements or challenges that encourage them to acquire the required skills.

Comparative analysis of adverse events among intravesical drugs in bladder cancer: a real-world study on FAERS database

ABSTRACT Background Intravesical therapy is a commonly utilized treatment for non-muscle invasive bladder cancer (NMIBC). This study focuses on summarizing the signals of all intravesical drugs and aims to highlight the comprehensive differences in adverse events (AEs) between these drugs. Research design and methods We conducted pharmacovigilance data analysis based on the real-world big data from the Food and Drug Administration Adverse Event Reporting System (FAERS) database. Results We elucidated all signals compared with the overall FAERS database or other administration routes for Bacillus Calmette-Guerin (BCG), mitomycin, gemcitabine, valrubicin, and epirubicin. Notably, the distribution of reported AEs associated with intravesical therapy exhibited a noticeable inclination toward male patients. Furthermore, all five drugs demonstrated a disproportionate distribution in local AEs, particularly in renal and urinary disorders. Additionally, specific signals and findings were summarized for each individual drug. Finally, we highlighted the AEs that resulted in serious outcomes for each drug. Conclusion We have compiled an overview of the AEs tied to intravesical drugs whilst considering their individual distinctions. These insightful findings serve to enrich our comprehension of the safety profiles and potential risks linked to intravesical therapy.

A pharmacovigilance study of drug-reduced male semen quality based on the Food and Drug Administration adverse event reporting system database.

Real-world big data studies on drug-reduced male semen quality are few and far between, with most studies based on animal trials, small scale retrospective studies, or a limited number of pre-market clinical trials. This study aimed to identify culprit drugs that reduced male semen quality based on the United States Food and Drug Administration adverse event reporting system. The Medical Dictionary for Regulatory Activities preferred terms and standardized Medical Dictionary for Regulatory Activities queries were used to define reduced male semen quality. Adverse events related to drug-reduced male semen quality were then analyzed by disproportionality analysis using the United States Food and Drug Administration adverse event reporting system data between 2004 and 2023. At the preferred term level, 59 drugs with risk signals were detected to be associated with drug-reduced male semen quality, with the three most frequently reported second-level Anatomical Therapeutic Chemical groups being antineoplastic agents (n=16, 27.12%), psychoanaleptics (n=9, 15.25%), and psycholeptics (n=6, 10.17%). At the standardized Medical Dictionary for Regulatory Activities queries level, the five drugs with the greatest number of cases were finasteride (845 cases, IC025=7.72), dutasteride (163 cases, IC025=7.22), tamsulosin (148 cases, IC025=5.99), testosterone (101 cases, IC025=4.08), and valproic acid (54 cases, IC025=2.44). Additionally, clinical information about drug-reduced male semen quality is absent from the Summary of Product Characteristics of 41 drugs in our study. Using the United States Food and Drug Administration adverse event reporting system database, we offer a list of drugs with risk signals for reducing male semen quality. In the future, there is still a need for more studies on drugs whose effects on male semen quality are not fully understood.

Mining Real-World Big Data to Characterize Adverse Drug Reaction Quantitatively: Mixed Methods Study.

Adverse drug reactions (ADRs), which are the phenotypic manifestations of clinical drug toxicity in humans, are a major concern in precision clinical medicine. A comprehensive evaluation of ADRs is helpful for unbiased supervision of marketed drugs and for discovering new drugs with high success rates. In current practice, drug safety evaluation is often oversimplified to the occurrence or nonoccurrence of ADRs. Given the limitations of current qualitative methods, there is an urgent need for a quantitative evaluation model to improve pharmacovigilance and the accurate assessment of drug safety. In this study, we developed a mathematical model, namely the Adverse Drug Reaction Classification System (ADReCS) severity-grading model, for the quantitative characterization of ADR severity, a crucial feature for evaluating the impact of ADRs on human health. The model was constructed by mining millions of real-world historical adverse drug event reports. A new parameter called Severity_score was introduced to measure the severity of ADRs, and upper and lower score boundaries were determined for 5 severity grades. The ADReCS severity-grading model exhibited excellent consistency (99.22%) with the expert-grading system, the Common Terminology Criteria for Adverse Events. Hence, we graded the severity of 6277 standard ADRs for 129,407 drug-ADR pairs. Moreover, we calculated the occurrence rates of 6272 distinct ADRs for 127,763 drug-ADR pairs in large patient populations by mining real-world medication prescriptions. With the quantitative features, we demonstrated example applications in systematically elucidating ADR mechanisms and thereby discovered a list of drugs with improper dosages. In summary, this study represents the first comprehensive determination of both ADR severity grades and ADR frequencies. This endeavor establishes a strong foundation for future artificial intelligence applications in discovering new drugs with high efficacy and low toxicity. It also heralds a paradigm shift in clinical toxicity research, moving from qualitative description to quantitative evaluation.

Study on adherence to positive airway pressure treatment for patients with obstructive sleep apnea using real-world big data in a telemedicine management system

Study on adherence to positive airway pressure treatment for patients with obstructive sleep apnea using real-world big data in a telemedicine management system

Enhancing safety of construction workers in Korea: an integrated text mining and machine learning framework for predicting accident types

Construction workers face a high risk of various occupational accidents, many of which can result in fatalities. This study aims to develop a prediction model for nine prevalent types of construction accidents, utilizing construction tasks, activities, and tools/materials as input features, through the application of machine learning-based multi-class classification algorithms. 152,867 construction accident summary reports, composed of both structured (construction task, construction activity, accident type) and unstructured data (tools/materials) were used for the study. The study employed several data processing techniques, including keyword extraction through text mining, Boruta feature selection, and SMOTE data resampling enhance model accuracy. Three performance metrics (Multi-class area under the receiver operating characteristic curve (MAUC), Multi-class Matthews Correlation Coefficient (MMCC), Geometric-mean (G-mean)) were used to compare the predictive performance of four machine learning algorithms, including Decision tree, Random forest, Naïve bayes, and XGBoost. Of the four algorithms, XGBoost showed the highest performance in predicting accident type (MAUC: 0.8603, MMCC: 0.3523, G-mean: 0.5009). Furthermore, a Shapley additive explanation (SHAP) analysis was conducted to visualize feature importance. The findings of this study make a valuable contribution to improving construction safety by presenting a prediction model for accident types derived from real-world big data.

Factors predictive of second-line chemotherapy in soft tissue sarcoma: An analysis of the National Genomic Profiling Database.

Of the drugs used in second-line chemotherapy for soft tissue sarcoma (STS), trabectedin is effective for liposarcoma and leiomyosarcoma (L-sarcoma), eribulin for liposarcoma, and pazopanib for non-liposarcoma. The indications for these drugs in STS other than L-sarcoma have not been established. Here we explored the prognosis, mutation profiles, and drug-response factors in STS using real-world big data. Clinicogenomic data on 1761 patients with sarcoma who underwent FoundationOne CDx were obtained from a national database in Japan. Patients with TP53 and KDM2D mutations had a significantly shorter survival period of 253 (95% CI, 99-404) and 330 (95% CI, 20-552) days, respectively, than those without mutations. Non-supervised clustering based on mutation profiles generated 13 tumor clusters. The response rate (RR) to trabectedin was highest in an MDM2-amplification cluster (odds ratio [OR]: 2.2; p = 0.2). The RR was lowest for eribulin in an MDM2-amplification cluster (OR: 0.4; p = 0.03) and highest in a TERT-mutation cluster (OR: 3.0; p = 0.03). The RR was highest for pazopanib in a PIK3CA/PTEN-wild type cluster (OR: 2.1; p = 0.03). In particular, patients harboring mutations in genes regulating the PI3K/Akt/mTOR pathway had a lower RR than patients without mutations (OR: 0.3; p = 0.04). In STS, mutation profiles were more useful in predicting the drug response than histology. The present study demonstrated the potential of tailored therapy guided by mutation profiles established by comprehensive genomic profiling testing in optimizing second-line chemotherapy for STS. The findings of this study will hopefully contribute some valuable insights into enhancing STS treatment strategies and outcomes.

How the National Health Insurance Coverage policy changed the use of lenvatinib for adult patients with advanced hepatocellular carcinoma: a retrospective cohort analysis with real world big data

BackgroundTo establish a long-term mechanism to control the cost burden of drugs, the Chinese government organized seven rounds of price negotiations for the national reimbursement drug list (NRDL) from 2016 to the end of 2022. The study aimed to evaluate the impact of the National Health Insurance Coverage (NHIC) policy on the use of lenvatinib as the first-line treatment for advanced hepatocellular carcinoma (HCC) within a specific medical insurance region from the micro perspective of individual patient characteristics.MethodsThe data of HCC patients that received lenvatinib from September 2019 to August 2022 was retrieved from the Medical and Health Big Data Center and longitudinally analyzed. Contingency table chi-square statistics and binary logistic regression analysis were used to compare the differences in the categorical variables. Interrupted time-series (ITS) regression analysis was performed to evaluate the changes in the utilization of lenvatinib over 36 months. Multiple linear regression was used to analyze the impact of receiving lenvatinib on the total hospitalization expenses of hospitalized patients with advanced HCC.ResultsA total of 12,659 patients with advanced HCC were included in this study. The usage rate of lenvatinib increased from 6.19% to 15.28% over 36 months (P < 0.001). By controlling the other factors, consistent with this, the probability of patients with advanced HCC receiving lenvatinib increased by 2.72-fold after the implementation of the NHIC policy (OR = 2.720, 95% CI:2.396–3.088, P < 0.001). Older, residency in rural areas, lack of fixed income, treatment at hospitals below the tertiary level, and coverage by urban-rural residents’ basic medical insurance (URRBMI) were the factors affecting the use of lenvatinib among patients with advanced HCC (P < 0.05). After the implementation of the NHIC policy, the total hospitalization expenses increased (Beta=-0.040, P < 0.001). However, compared to patients who received lenvatinib, the total hospitalization expenses were higher for those who did not receive the drug (US$5022.07 ± US$5488.70 vs. US$3701.63 ± US$4330.70, Beta = 0.062, P < 0.001).ConclusionsThe NHIC policy has significantly increased the utilization of lenvatinib. In addition, we speculate that establishing multi-level medical insurance systems for economically disadvantaged patients would be beneficial in improving the effectiveness of the NHIC policy in the real world.

Understanding patient needs and predicting outcomes in IgA nephropathy using data analytics and artificial intelligence: a narrative review.

This narrative review explores two case scenarios related to immunoglobulin A nephropathy (IgAN) and the application of predictive monitoring, big data analysis and artificial intelligence (AI) in improving treatment outcomes. The first scenario discusses how online service providers accurately understand consumer preferences and needs through the use of AI-powered big data analysis. The author, a clinical nephrologist, contemplates the potential application of similar methodologies, including AI, in his medical practice to better understand and meet patient needs. The second scenario presents a case study of a 20-year-old man with IgAN. The patient exhibited recurring symptoms, including gross haematuria and tonsillitis, over a 2-year period. Through histological examination and treatment with renin-angiotensin system blockade and corticosteroids, the patient experienced significant improvement in kidney function and reduced proteinuria over 15years of follow-up. The case highlights the importance of individualized treatment strategies and the use of predictive tools, such as AI-based predictive models, in assessing treatment response and predicting long-term outcomes in IgAN patients. The article further discusses the collection and analysis of real-world big data, including electronic health records, for studying disease natural history, predicting treatment responses and identifying prognostic biomarkers. Challenges in integrating data from various sources and issues such as missing data and data processing limitations are also addressed. Mathematical models, including logistic regression and Cox regression analysis, are discussed for predicting clinical outcomes and analysing changes in variables over time. Additionally, the application of machine learning algorithms, including AI techniques, in analysing big data and predicting outcomes in IgAN is explored. In conclusion, the article highlights the potential benefits of leveraging AI-powered big data analysis, predictive monitoring and machine learning algorithms to enhance patient care and improve treatment outcomes in IgAN.

Polygon Simplification for the Efficient Approximate Analytics of Georeferenced Big Data.

The unprecedented availability of sensor networks and GPS-enabled devices has caused the accumulation of voluminous georeferenced data streams. These data streams offer an opportunity to derive valuable insights and facilitate decision making for urban planning. However, processing and managing such data is challenging, given the size and multidimensionality of these data. Therefore, there is a growing interest in spatial approximate query processing depending on stratified-like sampling methods. However, in these solutions, as the number of strata increases, response time grows, thus counteracting the benefits of sampling. In this paper, we originally show the design and realization of a novel online geospatial approximate processing solution called GeoRAP. GeoRAP employs a front-stage filter based on the Ramer-Douglas-Peucker line simplification algorithm to reduce the size of study area coverage; thereafter, it employs a spatial stratified-like sampling method that minimizes the number of strata, thus increasing throughput and minimizing response time, while keeping the accuracy loss in check. Our method is applicable for various online and batch geospatial processing workloads, including complex geo-statistics, aggregation queries, and the generation of region-based aggregate geo-maps such as choropleth maps and heatmaps. We have extensively tested the performance of our prototyped solution with real-world big spatial data, and this paper shows that GeoRAP can outperform state-of-the-art baselines by an order of magnitude in terms of throughput while statistically obtaining results with good accuracy.

Glucose levels are not the same for everyone: a real-world big data study evaluating fasting serum glucose levels by sex and age among children.

Understanding the normal range of laboratory values as pertained to different age groups and males or females is paramount in health care delivery. We aimed to assess the distribution of morning fasting serum glucose levels by age and sex in the general population of children using a large-scale population-based cohort. A retrospective study with real-world de-identified data from a large, state mandated health fund in Israel among children aged 2-18 years old between 2006 and 2019. Age, sex, and BMI differences in mean glucose levels were evaluated. Study included 130,170 venous blood samples from 117,411 children, 53.3 % were female. After adjusting for age boys had higher fasting serum glucose levels than girls, with a mean of 89.21±8.66 mg/dL vs. 87.59±8.35 (p<0.001) [4.95±0.48 mmol/L vs. 4.86±0.46]. Compared to the 15 to 18 year-olds (88.49±7.63 mg/dL) [4.92±0.42 mmol/L], 2 to 5 year-olds had lower glucose levels (84.19±10.65, [4.68±0.59] (p<0.001)), 11 to 14 year-olds had higher glucose (90.40±7.42 [5.02±0.41], (p<0.001)) and 6 to 10 year-olds showed no difference (88.45±8.25) [4.91±0.46]. 33.0 % (n=42,991) had a BMI percentile record the same year as their glucose test result. There was a weak yet significant positive association between blood glucose levels and BMI. Our large cohort indicates that boys have slightly higher fasting serum glucose levels than girls, as do adolescents compared to younger children. This finding is important for the delivery of adequate health care, screening for illness and avoiding unnecessary investigations and tests.

EVOLUTIONARY ALGORITHM-BASED PARETO FRONT EXPLORATION FOR EFFICIENT COST-PERFORMANCE TRADEOFFS IN BIG DATA ANALYTICS

Big data analytics often involves complex decision-making processes that require finding efficient cost-performance tradeoffs. Evolutionary algorithms (EAs) have proven to be effective in solving multi-objective optimization problems by exploring the Pareto front, which represents the optimal tradeoffs between conflicting objectives. In this paper, we propose an evolutionary algorithm-based approach for Pareto front exploration in big data analytics. Our approach employs a novel fitness function that incorporates both cost and performance metrics, allowing the algorithm to simultaneously optimize for both objectives. We introduce several mutation and crossover operators tailored for big data analytics, ensuring effective exploration of the solution space. To validate the effectiveness of our approach, we conduct experiments using real-world big data analytics scenarios. The results demonstrate that our evolutionary algorithm-based approach successfully explores the Pareto front, enabling decision-makers to identify optimal cost-performance tradeoffs in big data analytics.

Bottom-up and top-down paradigms of artificial intelligence research approaches to healthcare data science using growing real-world big data.

As the real-world electronic health record (EHR) data continue to grow exponentially, novel methodologies involving artificial intelligence (AI) are becoming increasingly applied to enable efficient data-driven learning and, ultimately, to advance healthcare. Our objective is to provide readers with an understanding of evolving computational methods and help in deciding on methods to pursue. The sheer diversity of existing methods presents a challenge for health scientists who are beginning to apply computational methods to their research. Therefore, this tutorial is aimed at scientists working with EHR data who are early entrants into the field of applying AI methodologies. This manuscript describes the diverse and growing AI research approaches in healthcare data science and categorizes them into 2 distinct paradigms, the bottom-up and top-down paradigms to provide health scientists venturing into artificial intelligent research with an understanding of the evolving computational methods and help in deciding on methods to pursue through the lens of real-world healthcare data.

Updates in psoriasis diagnosis and treatment status in China: results from the National Psoriasis Center Registry.

Updates in psoriasis diagnosis and treatment status in China: results from the National Psoriasis Center Registry.

Real-world Big-data: Strengths and weaknesses of ASCO's CancerLinQ® discovery multiple myeloma dataset.

None of the authors have any financial or non-financial conflicts of interest to disclose. The opinions and assertions expressed herein are those of the authors and do not necessarily reflect the official policy or position of the US Air Force or the Department of Defense. None

Running Endurance in Women Compared to Men: Retrospective Analysis of Matched Real-World Big Data.

To determine whether the gap in endurance performance between men and women is reduced as distances increase, i.e. if there is a sex difference in endurance, one can analyse the performance of elite runners, all participants, or one can pair women and men during short-distance events and examine the difference over longer distances. The first two methods have caveats, and the last method has never been performed with a large dataset. This was the goal of the present study. A dataset including 38,860 trail running races from 1989 to 2021 in 221 countries was used. It provided information on 1,881,070 unique runners, allowing 7251 pairs of men and women with the same relative level of performance to be obtained, i.e. the same percentage of the winner time of the considered race on short races (25-45km-effort) that were compared during longer races (45-260km-effort). The effect of distance on sex differences in average speed was determined using a gamma mixed model. The gap between sexes decreased as distance increases, i.e. men's speed decreased by 4.02% (confidence interval 3.80-4.25) for every 10km-effort increase, whereas it decreased by 3.25% (confidence interval 3.02-3.46) for women. The men-women ratio decreases from 1.237 (confidence interval 1.232-1.242) for a 25km-effort to 1.031 (confidence interval 1.011-1.052) for a 260km-effort. This interaction was modulated by the level of performance, i.e. the greater the performance level of the runner, the lower the difference in endurance between sexes. This study shows for the first time that the gap between men and women shrinks when trail running distance increases, which demonstrates that endurance is greater in women. Although women narrow the performance gap with men as race distance increases, top male performers still outperform the top women.

ChatGPT and Big Data: Enhancing Text-to-Speech Conversion

Text-to-speech (TTS) conversion is a crucial technology for various applications, including accessibility, education, and entertainment. With the rapid growth of big data, TTS conversion systems face new challenges in terms of data size and diversity. In this paper, we propose to use the state-of-the-art language model ChatGPT to enhance TTS conversion for big data. We first introduce the background of TTS conversion and big data, and then review the existing TTS conversion systems and their limitations. Next, we describe the architecture and training of ChatGPT, and how it can be applied to TTS conversion. Finally, we evaluate the performance of the ChatGPT-based TTS conversion system on a large-scale real-world big data dataset, and compare it with the existing TTS systems. Our experimental results demonstrate that ChatGPT can significantly improve the quality and efficiency of TTS conversion for big data.

Benefits of guideline-directed medical therapy to loop diuretics in management of heart failure.

We sought to compare the outcomes of patients receiving combination therapy of diuretics and neurohormonal blockers, with a matched cohort with monotherapy of loop diuretics, using real-world big data. This study was based on the Diagnosis Procedure Combination database in the Japanese Registry of All Cardiac and Vascular Datasets (JROAD-DPC). After exclusion criteria, we identified 78,685 patients who were first hospitalized with heart failure (HF) between April 2015 and March 2017. Propensity score (PS) was estimated with logistic regression model, with neurohormonal blockers (angiotensin-converting enzyme inhibitor : ACEi or angiotensin receptor blocker : ARB, ?-blockers and mineralocorticoid receptor antagonists : MRA) as the dependent variable and 24 clinically relevant covariates to compare the in-hospital mortality between monotherapy of loop diuretics and combination therapies. On PS-matched analysis, patients with ACEi?/?ARB, ?-blockers, and MRA had lower total in-hospital mortality and in-hospital mortality within 7 days, 14 days and 30 days. In the sub-group analysis, regardless of clinical characteristics including elderly people and cancer, patients treated with a combination of loop diuretics and neurohormonal blockers had significantly lower in-hospital mortality than matched patients. Our data indicate the benefits of guideline-directed medical therapy to loop diuretics in the management of HF. J. Med. Invest. 70 : 41-53, February, 2023.