Data Warehouse Research Articles

The primary cause of markedly elevated aminotransferases in hospitalized patients with cirrhosis in ischemic hepatitis.

Marked elevation in aminotransferases (≥1000 IU/l) is typically associated with acute liver injury. Here, we hypothesized that the cause of elevation in aminotransferases ≥1000 in patients with cirrhosis is likely due to a limited number of disorders and may be associated with poor outcomes. We aimed to investigate the most common etiologies of acute elevations in aminotransferases in patients with cirrhosis, and to examine their associated outcomes. From May 2012 to December 2022, all hospitalized patients with cirrhosis and an aspartate aminotransferase or alanine aminotransferase ≥ 1000 IU/l were identified through Medical University of South Carolina's Clinical Data Warehouse. Complete clinical data were abstracted for each patient, and in-hospital mortality was examined. The cohort was made up of 152 patients, who were 57 ± 12 years old, with 51 (34%) women. Underlying liver disease included mainly hepatitis C cirrhosis, alcohol-related cirrhosis, metabolic dysfunction-associated steatohepatitis cirrhosis, autoimmune cirrhosis, primary sclerosing cholangitis cirrhosis, and cryptogenic cirrhosis. The most common cause of marked elevation in aminotransferases in cirrhotic patients was ischemic hepatitis (71%), followed by chemoembolization (7%), autoimmune hepatitis (6%), drug-induced liver injury (3%), post-transjugular intrahepatic portosystemic shunt placement (3%), rhabdomyolysis (3%), and hepatitis C (2%). During hospitalization and over a 1-month follow-up period, the mortality rate in patients with ischemic hepatitis was 73% (79/108), while that for other causes of liver injury was 20% (9/44). Ischemic hepatitis is the leading cause of marked elevation of aminotransferases in patients with cirrhosis, with distinctive clinical characteristics than other etiologies, and significantly poorer outcomes.

Health care utilization at the end of life in Parkinson’s disease: a population-based register study

BackgroundKnowledge of health care utilization at the end of life in Parkinson’s disease (PD) is sparse. This study aims to investigate end of life health care utilization, characterized by emergency room (ER) visits, receipt of specialized palliative care (SPC), and acute hospital deaths in a Swedish population-based PD cohort.MethodsWe conducted a retrospective cohort study on deceased patients (≥ 18 years) with a PD diagnosis during their last year of life (n = 922), based on health care-provider data from Region Stockholm´s data warehouse, for the study period 2015–2021. Univariable and multivariable logistic regression analyses tested associations and adjusted Odds ratios (aORs) were calculated.ResultsDuring the last month of life, approx. half of the cohort had emergency room (ER) visits and risk of frailty (measured by Hospital Frailty Risk Score) significantly predicted these visits (aOR, 3.90 (2.75–5.55)). In total, 120 people (13%) received SPC during their last three months of life, which positively associated with risk for frailty, (aOR, 2.65 (1.43–4.94, p = 0.002). In total, 284 people (31%) died in acute hospital settings. Among community-dwellers, male gender and frailty were strongly associated with acute hospital deaths (aOR, 1.90 (1.15–3.13, p = 0.01) and 3.70 (1.96–6.98, p < 0.0001)).ConclusionsRates of ER visits at end of life and hospital deaths were relatively high in this population-based cohort. Considering a high disease burden, referral to SPC at end of life was relatively low. Sex-specific disparities in health care utilization are apparent. Identifying people with high risk for frailty could assist the planning of optimal end-of-life care for people with PD.

Application of Business Intelligence in the Brazilian Educational System

In this article, we will discuss an application of Business Intelligence (BI) in analyzing data from the Brazilian educational system, specifically working with INEP (The Instituto Brasileiro de Geografia e Estatística — IBGE). This paper then discusses BI implementation's positive contributions and critical challenges in this domain. From the construction of an integrated BI system, our study aimed to offer a helpful tool that was easy to handle and touched on visuals such as dashboards, tables, and graphs to help us understand and manage the educational scenery throughout Brazil. The methodology used has been applied, with case studies and bibliographical research, based on data analysis techniques such as ETL (Extract, Transform, Load) and Data Warehousing. This is all leveraged to interactive dashboards, which should enable data-driven decisions and forge the way for advanced analytics like data mining.

Administrative data for public health: a regional data warehouse applied to dementia epidemiology

Abstract Administrative data can represent valuable sources of information for researchers seeking to advise policymakers. Several countries have established population-based data warehouses (DWH), integrating multiple data sources in single repositories through data linkage techniques. In Lombardy (Italy), a regional population-based DWH was established in the 1980s. It collects data about citizens from multiple sources in the social and health sectors, anonymising them immediately after their acquisition. Lombardy Region signed a specific agreement with the University of Pavia to investigate the epidemiology and access to care among people living with dementia (PLWD), using the DWH. A retrospective cohort study was conducted on hospitalised PLWD, including people aged ≥ 65 years with diagnoses of dementia at discharge from 2002 to 2020. The aim was to study the trends of hospitalisations among PLWD, computing these by calendar year, age, sex and cause of hospitalisation. A total of 340,144 hospital discharges were registered, of which 211,709 (63.0%) were women and 124,195 (37.0%) were men. The rate of hospitalisation decreased over the study period, ranging from 100.6/10,000 in 2002 to 65.1/10,000 in 2020. The average age at hospitalisation increased for men and women from 78.9 and 81.8 years in 2002 to 82.0 years and 84.2 years, respectively, in 2020. Respiratory diseases caused 10.4% of all hospitalisations in 2002 and grew steadily to 26.8% in 2020, becoming the leading cause of hospital admissions since 2017. Results allowed us to elaborate epidemiological measures and to characterise our population in a wide temporal trend, quantifying access to acute care by leading causes of hospitalisation. However, the analysis of administrative data comes along with several challenges, such as data incompleteness and biases. Researchers should take advantage of these data sources while keeping in mind the limitations and constraints related to data quality.

Premature ventricular contraction burden measured by an insertable cardiac monitor and the incidence of ventricular tachycardia and fibrillation

Abstract Background Insertable cardiac monitors (ICM) have the capability to continuously monitor premature ventricular contraction (PVC) burden over time [1] as well as detect ventricular tachycardia and fibrillation (VT/VF). Objective We investigated the association of PVC burden with incidence of VT/VF in a large real-world cohort of patients implanted with ICMs. Methods Patients implanted with an ICM for various reasons for monitoring and PVC detection turned on were included from the ICM manufacturer's de-identified data warehouse. Patients were included if they had at least 90 days of PVC burden follow-up. Tachycardia episodes that were detected by the ICM were first classified as VT/VF, SVT, or oversensing using an artificial intelligence (AI) model that was pre-trained using over 60,000 manually adjudicated ICM detected tachycardia episodes. If the AI model output probability for VT/VF was greater than 0.2, then those episodes were manually adjudicated for incidence of non-induced spontaneous VT/VF. The PVC burden trends recorded by the device were divided into 4 mutually exclusive patient groups: (1) 0% PVC burden on all days, (2) 1-4% PVC burden on ≥1 day, (3) 5-9% PVC burden on ≥1 day, and (4) &amp;gt;10% PVC on ≥1 day. The incidence rate of spontaneous VT/VF was compared between the PVC burden patient groups using a Generalized Estimating Equations model with negative binomial distribution. Time to first spontaneous VT/VF occurrence after the first day of occurrence of PVC burden for the respective PVC burden groups were estimated using Kaplan-Meier analysis and the groups were compared using the Cox proportional hazards model. Results A total of 5,521 patients were included in the analysis. Patients had an average age of 69±15 years and 49.8% being males. There was a total of 33,393 tachycardia episodes from 2,641 patients that were detected by the ICM. After AI model probability-based adjudications, 691 spontaneous VT/VF episodes were identified from 277 patients. Patients with ≥1 day of PVC burden of 1-4%, 5-9%, and ≥10% were associated with 2.9, 7.3, and 7.8 times increased incidence rate of VT/VF episodes during follow-up period relative to patients with 0% PVC burden on all days of follow-up (Figure A). Kaplan Meier curves for incidence of first spontaneous VT/VF episode for the four groups following occurrence of the first day of qualifying PVC burden for the respective groups are shown in figure B. Patients with a day of PVC burden ≥10% were 3.5 times more likely to develop VT/VF in the future compared to patients with 0% PVC burden on all days. Conclusion Days with high PVC burden, as detected by an ICM with continuous PVC detection capability, were associated with increased risk of VT/VF events in a group of real-world patients implanted with ICMs. PVC burden measured by ICMs may be a risk stratification tool for further investigation to prevent sudden cardiac arrest.

Comparing Italian hospital discharge and mortality data: implications for epidemiological research

Abstract Background In Italy, regional governments maintain data warehouses of linked administrative health data (AHD) linked to demographic information including mortality. Legal reporting pathways for death certificate detail bypass the Region, resulting in significant amounts of missing data on cause and location of death in regional AHD warehouses and limiting epidemiological research on mortality and incidence. In this study, we determine for the first time in Italy the degree of agreement between last hospital diagnosis and primary cause of death, and how that agreement changes over time. Previous studies have found agreement ranging from 40% to 60%. Methods The COV-CVD cohort is comprised of 7.3 million adults aged 30+ from the Lombardy region of northwest Italy, representing all Lombardy residents registered with the health system as of 31 December 2019. Concordance was assessed using the first three characters of the reported cause of death (ICD-10 codes) and primary diagnosis at last hospitalization (converted from ICD-9CM). Results In the COV-CVD cohort there were 401,085 deaths between 1 January 2020 and 23 June 2023. Cause and location of death is missing for 50% and 54% of these deaths, respectively. 60% of deceased individuals had a hospitalization in the year before their death. Agreement between primary diagnosis at last hospitalization and cause of death was 27% overall and 40.5% for deaths in hospital. Missing location of death was reduced from 54% to 26%. Conclusions When using AHD from Italy, primary diagnosis at last hospitalization prior to death is not an appropriate proxy for cause of death, even when the death occurred in hospital. Different coding systems (ICD-10, ICD-9CM) further complicate comparisons and definitions in observational studies. Key messages • In Italy, primary diagnosis at last hospitalization before death is not an appropriate proxy for cause of death. • Poor linkage of mortality and health data limits epidemiological research.

Treatment of kidney disease: an analytical decision model at a Swiss university hospital

Abstract Background An observational hospital-based cohort study, where the kidney disease staging is identified as a digital marker, and the association between the standardised staging of Kidney Disease (KD) and the outcomes is assessed along the treatment process. Methods The clinical and administrative data for all patients are stored in the Clinical Data Warehouse. A specific digital marker for an exact KD staging is applied. Three retrospective patient cohorts for exact and imprecise staging and missing diagnosis are defined and compared regarding the in-hospital mortality, all-cause mortality and disease progression. Results In total, 83146 hospitalisations of 40421 patients with KD, treated in 2014-2023 (2016-2023 will be updated) were identified; the number of the first hospitalisations with exact staging was 847 (2%), unprecise 29934 (72%) and missing diagnosis (26%). The demographic and baseline factors distribution between groups was tested with no association (chi-squared test statistic). The median survival time with a 95% CI was 7.94 (7.66, 8.81) years for the exact and 7.47 (7.42, 7.53) for the imprecise staging group. A log-rank and loglikelihood tests indicated a significant difference in survival between the groups, p &amp;lt; 0.0001. The Cox regression model shows a statistically significant difference in Hazard Ratio; KD staging, age, Elixhauser van Walraven index, treatment, nephrotoxic medication, contrast agent application, and AKI were identified as the significant covariates, C-statistic 0.730, Likelihood ratio test, Wald test, log-rank score on 42 df, p &amp;lt; 2e-16. ANOVA analysis confirmed a significant difference in cost distribution between the groups. An economic decision model proposal integrates the findings, combining a profile-level Markov cohort with a Discreet Time-To-Event Simulation DICE. Conclusions The study demonstrates a better outcome regarding all-cause mortality, in-hospital mortality, and costs if a standardised KD staging was applied. Key messages • Intuition (clinical judgement) and standardised kidney disease staging perform differently regarding clinical outcomes and costs. • The study demonstrates a better outcome regarding all-cause mortality, in-hospital mortality, and costs if a standardised KD staging was applied.

Dynamic QTc interval changes observed in real-world patients with implanted insertable cardiac monitors taking class III antiarrhythmic drugs

Abstract Background QT prolongation can lead to sustained ventricular arrhythmias . Class III Antiarrhythmic drugs (AAD) are known to cause QT prolongation. Insertable cardiac monitors (ICM) have the capability to continuously monitor and detect dynamic changes in QT. Purpose To investigate the characteristics of QTc interval changes before and after antiarrhythmic drug loading in real-world ICM patients administered with Sotalol, Amiodarone, and Dofetilide in hospital. Methods The QT detection algorithm detects the T-waves and determines the QTc intervals for every beat, thus providing continuous long-term QTc trends from ICM ECG. Real-world patients from de-identified Electronic Health Record (EHR) database linked with ICM device data warehouse with implanted ICM being administered sotalol, amiodarone or dofetilide AAD for the first time in the hospital were analyzed. The patients were broken down into groups based on the AAD taken. Patients not taking AAD were selected as control. QTc intervals 7-days pre and post administered AAD are reported, and a random 14-day period was used for control patients. The paired difference in median QTc 7 days pre and post administered AAD were analyzed using the Wilcoxon signed rank test. The Wilcoxon rank-sum test was used to analyze QTc between the different groups. Results A total of 487 patients with ICM were included in this analysis: 56 sotalol patients (avg. age: 67±10 years, 50% male), 100 amiodarone patients (avg. age: 72±10 years, 50% male), 52 dofetilide patients (avg. age: 66±12 years, 50% male), and 248 control patients (avg. age: 70±13 years, 50% male). Sotalol, amiodarone, and dofetilide patients showed significant increase in QTc 1-7 days post AAD compared to 1-7 days pre AAD (figure 1). For sotalol (post AAD median[IQR] = 426.5[408.8,446.6] msec, pre AAD median[IQR] = 415.9[391,437.3] msec, p=0.004); Dofetilide (post median[IQR] = 432.4[413,458.7] msec; pre median[IQR] = 409.3[397.5,435.2] msec, p&amp;lt;0.001); Amiodarone (post median[IQR] = 426.3[400.9,449] msec, pre median[IQR] = 415.5[397.5,435.2] msec, p = 0.037). The QTc 1-7 days pre AAD compared to 1-7 days post AAD for control patients was insignificant p&amp;gt;0.05. The median change in QTc 1-7 days post AAD from 1-7 days pre AAD was 13.0, 7.5, 21.25 and -1.0 msec for sotalol, amiodarone, dofetilide, and control patients respectively (figure 2). Dofetilide patients had a significantly higher change in QTc compared to the amiodarone patients, p=0.018. QTc for all three AAD groups when compared to control group post AAD were significantly higher, p&amp;lt;0.001. QTc was also higher pre AAD compared to control, p=0.003, p&amp;lt;0.001, p=0.009 for sotalol, amiodarone, and dofetilide respectively. Conclusions There is a significant increase in QTc detected by ICM in patients receiving index drug loading of Sotalol, amiodarone and dofetilide. QTc monitoring via ICM can detect these dynamic changes in QTc allowing for early intervention.Figure 1Figure 2

Research on Jiangxi Cultural Wisdom Tourism Industry Strategy Based on OHDSI General Data Model and Digital Art Empowerment

With the rapid development and wide application of a new round of information technologies, such as the global Internet of Things, the new generation of mobile Internet, cloud computing, etc., informatization has opened a new stage of smart tourism (ST), and ST has become the main trend of tourism development in China at present and for a long time to come. Tourism industry has become one of the important pillar industries in Jiangxi, and it plays an increasingly prominent role in stimulating and supporting economic and social development. Wisdom tourism is an important means to improve tourism service, tourism management and tourism marketing, and an important starting point to promote the upgrading of tourism industry in Jiangxi Province, and to achieve carry-over and leapfrog development and overtaking at corners. This paper is constantly improving and updating on the basis of data warehouse integration mechanism. Under the data warehouse mechanism, through the effective integration of spatial data, virtual reality data and dynamic data, a comprehensive data warehouse oriented to tourism promotion, scenic spot management and department management is formed. It aims to promote the rapid implementation of ST, bring tourists a comfortable tourism experience, save costs and improve efficiency for tourism service providers.

Iterative Development of a Clinical Decision Support Tool to Enhance Naloxone Co-Prescribing.

Background Opioid overdoses have contributed significantly to mortality in the United States. Despite long-standing recommendations from the Centers for Disease Control and Prevention to co-prescribe naloxone for patients receiving opioids who are at high risk of overdose, compliance with these guidelines has remained low. Objectives The objective of this study was to develop and evaluate a hospital-wide electronic health record (EHR)-based clinical decision support (CDS) tool designed to promote naloxone co-prescription for high-risk opioids. Methods We employed an iterative approach to develop a point-of-order, interruptive EHR alert as the primary intervention and assessed naloxone prescription rates, EHR efficiency metrics, and barriers to adoption. Data was obtained from our EHR's clinical data warehouse and analyzed using statistical process control and Chi-square analyses to assess statistically significant differences in prescribing rates during the intervention periods. Results The initial implementation phase of the intervention, spanning from April 2019 to May 2022, yielded a nearly 3-fold increase in the proportion of high-risk patients receiving naloxone, rising from 13.4% [95% CI, 12.9% - 13.8%] to 36.4% [95% CI, 35.2% - 37.5%; p = 1 x 10-38]. Enhancements to the CDS design and logic during the subsequent iteration's study period, June 2022 and December 2023, reduced the number of CDS triggers by more than 30-fold while simultaneously driving an additional increase in naloxone receipt to 42.7% [95% CI, 40.6% - 44.8%; p = 2 x 10-5]. The efficiency of the CDS demonstrated marked improvement, with prescribers accepting the naloxone co-prescription recommendation provided by the CDS in 41.1% of the encounters in version two, compared to 6.2% in version one (p = 6 x 10-9). Conclusion This study offers a sustainable and scalable model to address low rates of naloxone co-prescription and may also be used to target other opportunities for improving guideline-concordant prescribing practices.

Association of race and ethnicity with genomic testing at a comprehensive cancer center in North Carolina.

Genomic testing, which analyzes tumor cells or circulating tumor DNA to guide cancer treatment, is increasingly used in precision oncology. However, data on racial and ethnic disparities in its use are limited. This retrospective cohort study included patients diagnosed with stage IV breast, colorectal, non-small cell lung, or prostate cancer in 2014-2019. Patients were identified from an institutional multilevel data warehouse and their data was linked to an institutional genomics database and electronic health records. Multivariable logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for the receipt of tumor-specific genomic testing among non-Hispanic (NH) Black and NH White patients, stratified by cancer type. The linked dataset included 3,461 patients, 44.5% of whom received genomic testing. NH Black patients with prostate cancer were less likely to use genomic testing vs. NH White patients across the study period (fully adjusted OR: 0.55, 95% CI: 0.32, 0.95), as were NH Black patients diagnosed with breast cancer in 2014-2016 (OR adjusted for patient-level covariates: 0.34, 95% CI: 0.11, 0.99). No differences in testing were observed between NH Black and NH White patients with colorectal or non-small cell lung cancer, or breast cancer diagnosed in 2017-2019. Lower documented use of tumor-specific genomic testing among NH Black patients with prostate cancer across the study period underscores the need to monitor trends and address disparities in the use of emerging precision oncology technologies. Future work targeting multilevel barriers to genomic testing is a crucial next step forward.

Research on AIS Data Analysis and Its Integration into Discipline-specific Teaching

The processing of ship AIS (Automatic Identification System) data plays a crucial role in the digital transformation of the shipping industry. From data collection to data storage, data analysis, and data visualization, AIS data processing technology provides robust support for the shipping sector. In terms of data collection, the platform utilizes distributed message queues to aggregate AIS data. For data storage, the platform efficiently manages and stores massive amounts of AIS data using data warehouses. When it comes to data analysis, the platform mines and analyzes AIS data through Lambda, extracting valuable information such as ship trajectories and sailing speeds, thereby supporting areas such as ship management, navigation safety, and marine research. In data visualization, the platform intuitively displays ship navigation status through charts and maps, providing strong support for decision-making. As an essential component of big data applications, AIS data analysis technology is playing an increasingly important role in the shipping industry. The big data technology major of Jiangsu Maritime Institute integrates big data applications like AIS data analysis into its curriculum, taking the upgrading needs of the shipping industry as an entry point. It actively plans the cultivation of composite big data skilled talents under the backdrop of smart shipping, comprehensively serving the upgrading of the shipping industry and promoting the transformation from a major shipping country to a shipping powerhouse.

Development of an Intelligent Coal Production and Operation Platform Based on a Real-Time Data Warehouse and AI Model

Smart mining solutions currently suffer from inadequate big data support and insufficient AI applications. The main reason for these limitations is the absence of a comprehensive industrial internet cloud platform tailored for the coal industry, which restricts resource integration. This paper presents the development of an innovative platform designed to enhance safety, operational efficiency, and automation in fully mechanized coal mining in China. This platform integrates cloud edge computing, real-time data processing, and AI-driven analytics to improve decision-making and maintenance strategies. Several AI models have been developed for the proactive maintenance of comprehensive mining face equipment, including early warnings for periodic weighting and the detection of common faults such as those in the shearer, hydraulic support, and conveyor. The platform leverages large-scale knowledge graph models and Graph Retrieval-Augmented Generation (GraphRAG) technology to build structured knowledge graphs. This facilitates intelligent Q&amp;A capabilities and precise fault diagnosis, thereby enhancing system responsiveness and improving the accuracy of fault resolution. The practical process of implementing such a platform primarily based on open-source components is summarized in this paper.

Overall Survival in Real-World Patients with Unresectable Hepatocellular Carcinoma Receiving Atezolizumab Plus Bevacizumab Versus Sorafenib or Lenvatinib as First-Line Therapy: Findings from the National Veterans Health Administration Database.

Background/Objectives: This study evaluated comparative overall survival (OS) of United States veterans with unresectable hepatocellular carcinoma (uHCC) receiving first-line (1L) atezolizumab plus bevacizumab vs. sorafenib or lenvatinib, overall and across racial and ethnic groups. Methods: In this retrospective study, patients with uHCC who initiated atezolizumab plus bevacizumab (post-2020) or sorafenib or lenvatinib (post-2018) were identified from the Veterans Health Administration National Corporate Data Warehouse (1 January 2017-31 December 2022). Patient characteristics were evaluated in the year prior to 1L treatment initiation. Kaplan-Meier and multivariable Cox regression methods were used to compare OS starting from treatment between cohorts, both overall and by race and ethnicity. Results: Among the 1874 patients included, 405 (21.6%) received 1L atezolizumab plus bevacizumab, 1016 (54.2%) received sorafenib, and 453 (24.2%) received lenvatinib, with a median follow-up time of 8.5, 7.6, and 8.2 months, respectively. Overall, patients receiving atezolizumab plus bevacizumab had longer unadjusted median OS (12.8 [95% CI: 10.6, 17.1] months) than patients receiving sorafenib (8.0 [7.1, 8.6] months) or lenvatinib (9.5 [7.8, 11.4] months; both log-rank p < 0.001). After adjustment, atezolizumab plus bevacizumab was associated with a reduced risk of death by 30% vs. sorafenib (adjusted HR: 0.70 [95% CI: 0.60, 0.82]) and by 26% vs. lenvatinib (0.74 [0.62, 0.88]; both p < 0.001). OS trends in the White, Black, and Hispanic patient cohorts were consistent with that of the overall population. Conclusions: Atezolizumab plus bevacizumab was associated with improved survival outcomes compared with sorafenib and lenvatinib in patients with uHCC, both overall and across racial and ethnic subgroups.

P10.06.A AGENT-BASED MODELING FOR PATIENT-CENTERED CLINICAL DECISION SUPPORT IN NEURO-ONCOLOGY

Abstract BACKGROUND Complex clinical decision-making in neuro-oncology is a multifaceted process involving numerous specialties influenced by various objective factors and patient preferences. Modeling these multidisciplinary neuro-oncology discussions presents significant challenges, particularly given the multimodal nature of the data. Recent advancements in large language models (LLMs) have enabled the development of LLM “agents”. These agents pave the way for multi-agent systems capable of simulating interacting components, thus capturing the nuances of interdisciplinary clinical discussions. MATERIAL AND METHODS Utilizing GPT-4, we developed conversational agents with an agent-based modeling (ABM) approach; each agent was assigned a specific specialty. We simulated their interactions in various clinical decision-making scenarios. Lastly, we simulated 30 patient-agents and included them in the ABM system. We applied this ABM approach to a dataset of 16 clinical vignettes extracted from the UCSF Clinical Data Warehouse of primary adult glioma patients who underwent tumor board (TB) reviews. The TB decision was used as ground truth for quantitative evaluation. Seven independent clinicians evaluated the final decisions. RESULTS Our quantitative assessment showed that, when blinded to whether ABM or TB generated responses, evaluators preferred ABM recommendations 47.8% of the time, followed by TB recommendations (38%). The highest recommendation concordance was between TB and ABM (50.6%). The ABM was deemed clinically accurate in 76.8% of cases and provided clinically appropriate reasoning in 73.2%. Our qualitative assessment showed that including the patient agent significantly influenced the course of discussions, notably when the patient agents participated in the system by stating their healthcare goals. The most recommended individual treatment was standard-of-care chemotherapy, followed by targeted therapies, clinical trials, immunotherapy, and palliative care. The least frequently recommended management options were radiation, surgery, consultation with other specialties, and devices. CONCLUSION This study underscores the potential of ABM in neuro-oncology, particularly in simulating complex, real-world scenarios. ABM not only empowers patients but also offers clinicians a novel tool for understanding and navigating the multifaceted landscape of patient preferences and treatment options. It provides a platform for clinicians to visualize the potential impacts of various decisions, enhancing their ability to provide personalized, effective care. We envision future work where ABM systems are patient-centric, enabling patients to explore the potential impacts of their individual characteristics on clinical management options.

Developing a Competitive Edge: Building an Effective Portfolio in Snowflake and Teradata Data Engineering

This article examines the critical components of an effective portfolio for data engineering professionals specializing in Snowflake and Teradata platforms. As the data landscape evolves, the ability to showcase practical skills alongside theoretical knowledge has become paramount for career advancement. Through an analysis of industry trends and expert interviews, we identify key elements that contribute to a compelling portfolio, including project showcases demonstrating complex problem-solving, relevant certifications, and evidence of continuous learning. The article highlights the importance of balancing technical proficiency in SQL, data warehousing, and cloud technologies with strong documentation and communication skills. Our findings suggest that a well-curated portfolio not only validates a candidate's expertise but also provides tangible evidence of their ability to navigate the challenges of modern data ecosystems. This article provides a framework for data engineers to construct portfolios that effectively demonstrate their capabilities, potentially increasing their competitiveness in the job market and facilitating career growth in the rapidly evolving field of data engineering.

The Growing Trend of Cloud-Based Data Integration and Warehousing

This article examines the growing trend of cloud-based data integration and warehousing solutions in response to the exponential growth of data generation and the need for scalable, flexible analytics capabilities. It explores the key drivers behind cloud adoption, including data volume increases, demand for real-time insights, cost efficiencies, and advanced analytics requirements. The benefits of cloud solutions are discussed, such as scalability, cost-effectiveness, agility, and enhanced analytical capabilities. The article also addresses challenges like data security, vendor lock-in, and performance issues, providing best practices for successful implementation. Emerging trends in automated ETL processes, machine learning integration, and multi-cloud approaches are highlighted, showcasing the evolving landscape of cloud-based data management. Through analysis of market projections, case studies, and industry statistics, this comprehensive overview equips organizations with the knowledge to leverage cloud technologies to transform their data management capabilities and gain competitive advantages in the digital age.

Anomaly Detection in Healthcare Data Warehouses using ML

Anomaly Detection in Healthcare Data Warehouses using ML

Patient outcomes following buprenorphine treatment for opioid use disorder: A retrospective analysis of the influence of patient- and prescriber-level characteristics in Massachusetts, USA.

Opioid use disorder (OUD) is treatable with buprenorphine/naloxone (buprenorphine), but many patients discontinue treatment prematurely. The aim of this study was to assess the influence of patient- and prescriber-level characteristics relative to several patient outcomes following the initiation of buprenorphine treatment for OUD. This was a retrospective observational investigation. We used the Public Health Data Warehouse from the Massachusetts Department of Public Health to construct a sample of patients who initiated buprenorphine treatment between 2015 and 2019. We attributed each patient to a prescriber based on information from prescription claims. We used multilevel models to assess the influence of patient- and prescriber-level characteristics on each outcome. Massachusetts, USA. The study cohort comprised 37 955 unique patients and 2146 prescribers. Among patients, 64.6% were male, 52.6% were under the age of 35 and 82.2% were White, non-Hispanic. For insurance coverage, 72.1% had Medicaid. The outcome measures were poor medication continuity, treatment discontinuation and opioid overdose, all assessed within a 12-month follow-up period that began with a focal prescription for buprenorphine. Each patient had a single follow-up period. Poor medication continuity was defined as medication gaps totaling more than 7days during the initial 180 days of buprenorphine treatment and treatment discontinuation was defined as having a medication gap for 2 consecutive months within the 12-month follow-up period. The patient-level rates for poor medication continuity, treatment discontinuation and opioid overdose were 59.7% [95% confidence interval (CI) = 59.2-60.2], 57.4% (95% CI = 56.9-57.9) and 10.3% (95% CI = 10.0-10.6), respectively, with 1.1% (95% CI = 1.0-1.2) experiencing a fatal opioid overdose. At the patient level, after adjustment for covariates, adverse outcomes were associated with race/ethnicity as both Black, non-Hispanic and Hispanic patients had worse outcomes than did White, non-Hispanic patients (Black, non-Hispanic -- poor continuity: 1.50, 95% CI = 1.34-1.68; discontinuation: 1.44, 95% CI = 1.30-1.60; Hispanic -- poor continuity: 1.21, 95% CI = 1.12-1.31; discontinuation: 1.38, 95% CI = 1.28-1.48). Patients with insurance coverage through Medicaid also had worse outcomes than those with commercial insurance (poor continuity: 1.18, 95% CI = 1.11-1.26; discontinuation: 1.09, 95% CI = 1.03-1.16; overdose: 1.98, 95% CI = 1.75-2.23). Pre-treatment mental health conditions and other types of chronic illness were also associated with worse outcomes (History of mental health conditions -- poor continuity: 1.11, 95% CI = 1.06-1.17; discontinuation: 1.05, CI = 1.01-1.10; overdose: 1.47, 95% CI = 1.36-1.60; Chronic health conditions -- poor continuity: 1.15, 95% CI = 1.05-1.27; discontinuation: 1.15, 95% CI = 1.05-1.26; overdose: 1.83, 95% CI = 1.60-2.10; History of substance use disorder other than for opioids -- poor continuity: 1.54, 95% CI = 1.46-1.62; discontinuation: 1.54, 95% CI = 1.47-1.62; overdose: 1.93, 95% CI = 1.80-2.07). At the prescriber level, after adjustments for covariates, adverse outcomes were associated with clinical training, as primary care physicians had higher rates of adverse outcomes than psychiatrists (poor continuity: 1.12, 95% CI = 1.02-1.23; discontinuation: 1.04, 95% CI = 1.01-1.09). A larger prescriber panel size, based on number of patients being prescribed buprenorphine, was also associated with higher rates of adverse outcomes (poor continuity: 1.36, 95% CI = 1.27-1.46; discontinuation: 1.21, 95% CI = 1.14-1.28; overdose: 1.10, 95% CI = 1.01-1.19). Between 9% and 15% of the variation among patients for the outcomes was accounted for at the prescriber level. Patient- and prescriber-level characteristics appear to be associated with patient outcomes following buprenorphine treatment for opioid use disorder. In particular, patients' race/ethnicity and insurance coverage appear to be associated with substantial disparities in outcomes, and prescriber characteristics appear to be most closely associated with medication continuity during early treatment.

Polycystic ovary syndrome underdiagnosis patterns by individual-level and spatial social vulnerability measures.

To use electronic health records (EHR) data at Boston Medical Center (BMC) to identify individual-level and spatial predictors of missed diagnosis, among those who meet diagnostic criteria for PCOS. The BMC Clinical Data Warehouse was used to source patients who presented between October 1, 2003 and September 30, 2015 for any of the following: androgen blood tests, hirsutism, evaluation of menstrual regularity, pelvic ultrasound for any reason, or PCOS. Algorithm PCOS cases were identified as those with International Classification of disease (ICD) codes for irregular menstruation and either an ICD code for hirsutism, elevated testosterone lab, or polycystic ovarian morphology as identified using natural language processing on pelvic ultrasounds. Logistic regression models were used to estimate odds ratios (ORs) of missed PCOS diagnosis by age, race/ethnicity, education, primary language, body mass index (BMI), insurance type and social vulnerability index (SVI) score. In the 2003-2015 BMC-EHR PCOS at-risk cohort (n=23,786), there were 1,199 physician-diagnosed PCOS cases and 730 algorithm PCOS cases. In logistic regression models controlling for age, year, education, and SVI scores, Black/African American patients were more likely to have missed a PCOS diagnosis (OR = 1.69 [95% CI, 1.28, 2.24]) compared to non-Hispanic White patients, and relying on Medicaid or charity for insurance was associated with an increased odds of missed diagnosis when compared to private insurance (OR = 1.90 [95% CI, 1.47, 2.46], OR = 1.90 [95% CI, 1.41, 2.56], respectively). Higher SVI scores were associated with increased odds of missed diagnosis in univariate models. We observed individual-level and spatial disparities within the PCOS diagnosis. Further research should explore drivers of disparities for earlier intervention.