HOME
QUICK SEARCH
AUTHORS
KEYWORDS
HELP
BUGSPHUSE US 2025 Papers - Index
222 papers. Source: PHUSE (Pharmaceutical Users Software Exchange) US Connect conference proceedings.
| # | Title× | Authors× | Abstract× | Keywords× | Extras | Pages |
|---|---|---|---|---|---|---|
| 1 | "Will AI Render Programmers Obsolete?" | Kishore Pothuri, Ajay Gunda |
Learning and writing code for any new programming language is a crucial yet daunting task characterized by its complexity, demanding significant time and effort from programmers. But with the recent AI evolution, it has become easy to understand and acquire code, at the same time it can be challenging to shuffle between AI and coding platforms. Having AI integrated with the coding platforms can make it more convenient and efficient. R can be more flexible for AI integration as it is an open-source tool | machine learning, deep learning | source | 5 |
| 2 | 4 Steps to Revolutionise Clinical Data Processes: Harnessing Cryptographic Technology for Seamless Interoperability | Alistair Dootson | clinical data, clinical data management | source presentation source |
3 | |
| 3 | A Customized Algorithm to Evaluate the Quality and Reliability of Real-World Lab Data | Lihai Song | The FDA has recently released guidance on the use of real-world data (RWD), emphasizing the importance of ensuring the relevance and reliability of data through governance and established processes. In response to this guidance, we have developed a customized algorithm to assess the quality of lab data in RWD. This algorithm, based on the OMOP Common Data Model (CDM), employs predefined criteria to evaluate conformance, completeness, and plausibility. The resulting assessment report, available as an R Shiny dashboard and a detailed algorithmic spreadsheet, enables researchers to determine the reliability and feasibility of lab data within RWD databases. Our innovative approach, with its well-documented process and user-friendly implementation, provides a practical means of evaluating the usability of RWD. Additionally, it serves as a valuable example of a complaint process for the effective use of RWD in regulatory submissions | proc sql, data cleaning, data audit, laboratory data, clinical research, sample | source presentation source |
10 |
| 4 | A Gen AI Agent for Statistical Output Review | Peng Zhang, Lizhong Liu |
clinical trial, clinical data, data cleaning | source presentation source |
6 | |
| 5 | A Guided User Experience for Automated Mapping | Nathan Johnson | The hype and promise of fully automated mapping have run counter to the reality of complex, messy source data. In this session, we'll present an approach that aims to bridge that gap - a guided user experience for automated mapping. Rather than being a black box turnkey solution, here the focus is on having a human in the loop expert involved in the most complex aspects, while easing the process for the user with an AI-guided experience. Most attempts at automated SDTM solutions have met the 80/20 rule, capable of automating 80% of mappings. This session will address and consider the challenges and limitations faced by these solutions and present an approach that closes in on the 20% so that we can arrive at a fully automated system that keeps the human in the loop. INTRODUCTION: AUTOMATED MAPPING AND THE NEED FOR THE HUMAN AT THE HELM Let's take a moment to think about the USS Enterprise from Star Trek | sdtm programming, data transformation, audit trail, clinical trial, quality control, clinical data, data cleaning, clinical research, raw data, cdisc | source presentation source |
9 |
| 6 | A Journey Through an Innovative Approach to Migration of Legacy System to New Multilingual Statistical Computing Environment (SCE) | Kajal Tahiliani | Clinical trial data analysis is continuously evolving and the need to expediate data analysis has led many companies to move to a new Statistical Computing Environment (SCE) with built-in capabilities. Migration from a legacy SCE to a new SCE is a very important process to ensure a smooth transition and no loss of data. This paper delves into an agile migration process, highlighting some key factors such as the importance of a wellprepared migration team, the establishment of comprehensive processes and training, an awareness of people's mindset, a close collaboration with technical teams and robust stakeholder management. This paper also discusses the challenges faced during migration, such as handling various study scenarios, migrating through critical timelines, handling restricted data and maintaining an ever-growing inventory | data integrity, access control, data science, data cleaning | source presentation source |
9 |
| 7 | A New Approach to Visualize Adverse Events | Eva Kelty | The life sciences sector is witnessing a paradigm shift driven by emerging technologies that enhance data management and reporting. Among these, advanced data visualization, graph databases, and metadata-driven approaches are at the forefront, offering innovative solutions for handling complex datasets. Integrating adverse event data from various sources and standardizing it with established medical terminologies improves data accuracy. Graph databases are pivotal in this process, enabling efficient modeling and visualization of complex relationships between data points. They enhance decision-making through advanced relationship mapping, contributing to better patient safety and regulatory compliance. These methods boost operational efficiency and the overall effectiveness of clinical trials. This presentation will explore the advantages of graph databases, emphasizing their role in modernizing data management and visualization in the life sciences sector | adverse events, clinical trial, life sciences | source presentation source |
3 |
| 8 | A Novel Pipeline for Generating Realistic Synthetic CDISC ADaM Datasets Using Large Language Models and Knowledge Graphs | Chao Su, Jaime Yan |
This paper presents a systematic pipeline for generating realistic synthetic Clinical Data Interchange Standards Consortium (CDISC) Analysis Data Model (ADaM) datasets. We identify and address two fundamental challenges in direct JSON schema-based generation: the lack of domain-specific knowledge (e.g., laboratory test ranges, MedDRA codes) and insufficient capture of ADaM dataset structures and variable relationships. Our methodology introduces a two-pronged approach to schema enhancement: (1) knowledge enrichment through both a structured knowledge query system leveraging well-organized clinical trial documentation and Large Language Model (LLM) assistance for generating realistic value ranges and distributions, and (2) structure optimization using rule- based LLM prompts to reorganize schemas according to specific ADaM data structures (ADSL, BDS, OCCDS) | adam dataset, clinical trial, data cleaning, clinical data, cdisc | source presentation source |
15 |
| 9 | Accelerating Data Integration: Utilizing Python Scripts and Semantic Search for Variable Mapping in Multiple Studies | Xiaopeng Li | sdtm domain, response variable, data integration, statistical programmer, label statement, cdisc | source presentation source |
6 | |
| 10 | Accelerating Patient Safety with Knowledge Graphs: Lessons from a Successful Implementation | Partha Bhattacharjee, Kirti Aggarwal |
Three challenges stymie the ability to maximize value generation from patient safety systems. First, not all relevant data such as in-house safety reporting applications, drug labels, industry taxonomies, clinical trials, and real-world data are integrated . This ecosystem of data siloes spawns inefficient analysis and reporting processes. Second, well - compensated Subject Matter Experts (SMEs) spend more time data wrangling than analyzing it for scientific insights. Third, data engineered in such information systems is suboptimal input to Large Language Models (LLMs) due to structural constraints such as the inability to model complex relationships, semantically represent data, integrate semi- and unstructured data, and explicitly state relationships between entities. Knowledge graphs are an innovative approach for semantically integrating heterogeneous data for analysis, reporting, and input into LLMs | unstructured data, factor analysis, adverse events, clinical trial, data cleaning, medical history, whodrug, machine learning | source presentation source |
10 |
| 11 | Accelerating PKPD Analysis with Automated DDI Variable Creation: An SAS-Based Approach | Sunil Mummidisetty, Harish Ganesan |
Creating time-varying drug-drug interaction (DDI) variables is essential but often time-consuming in PKPD dataset construction. Our SAS-based standardized script automates this process, leveraging a continuously updated internal database of drug lists and potencies, curated by our DIDB clinical experts. This has led to substantial efficiency and quality gains | exposure, feature engineering, sample, concomitant medications, missing value | source presentation source |
7 |
| 12 | Accelerating Shiny Application Development and Validation with Rhino | André Veríssimo, Vedha Viyash |
data cleaning, role-based access control | source presentation source |
5 | |
| 13 | Actual Intelligence and Artificial Intelligence Joining Forces to Perform Statistical Analysis, Turning AI Hype into Reality | Songgu Xie, Toshio Kimura |
AI "could" automate analysis … "could" transform the pharmaceutical industry … "could", "could", "could" … but will it really? We will demonstrate a proof of concept showing AI/LLM performing statistical analysis and provide technical insights into this achievement. The solution parallels the usual workflow followed by programmers: 1) Specification: Biostatisticians provides specifications 2) Clarification: Programmer asks clarification questions 3) Execution: Programmer executes using gathered information Programmers relate analysis concepts communicated through natural language to programming procedures, datasets and variables. What if we operationalize this using AI/LLM? 1) Specification: How will biostatistician provide specifications to AI/LLM? 2) Clarification: How will AI/LLM know what questions to ask? 3) Execution: How will AI/LLM execute the analysis? This PoC integrates AI/LLM with Python, Databases and SAS/R using APIs and JSON |
data science, statistical programmer, pharmaceutical industry, artificial intelligence, clinical trial, macro definition, clinical data | source presentation source |
11 |
| 14 | Adapting Leadership in Statistical Programming: Navigating Technological Advancements and Evolving Roles in the Pharmaceutical Industry | Jaya Mukkamala | Technological advancements are transforming the pharmaceutical industry, particularly in clinical programming. Emerging technologies are reshaping roles traditionally centered on SAS programming, requiring professionals to continuously update their skills and integrate new tools into their workflows. To maximize the benefits of these innovations, fostering teamwork and a culture of continuous learning is essential. By staying informed and balancing technical expertise with leadership abilities, clinical programming professionals can successfully navigate this evolving landscape. This paper explores strategies for adapting to these changes and fostering professional growth in an era of rapid technological evolution | clinical programming, clinical research, pharmaceutical industry | source presentation source |
2 |
| 15 | Advanced Clinical Data Visualisation Using Violin Plots in R | Mrityunjay Kumar, Ephicacy Lifescience |
Effective visualisation of clinical data is crucial for insightful analysis and decision -making. The violin plot, a combination of box plot and density plot, offers comprehensive visualisation by displaying data distribution and variability in a single, cohesive format. This paper presents the application of violin plots using R, showcasing how to visualise clinical trial data, emphasizing its utility in highlighting data nuances , often overlooked by traditional methods. Attendees will learn the process of creating and customizing violin plots with ggplot2, using real -world examples to highlight how violin plots can effectively communicate complex data patterns, such as patient response variability, subgroup comparisons, and treatment effects. By leveraging violin plots, programmers or statisticians can reveal data nuances, such as skewness , multimodal distributions and outliers leading to more informed and impactful data-driven decisions | box plot, histogram, density plot | source presentation source |
14 |
| 16 | Advancing Safety Evaluation in Oncology Trials Using Estimands | Reshma Radhakrishnan, Lakshmi Mantha |
In Oncology trials, ensuring patient safety is as crucial as evaluating treatment efficacy. Traditional safety assessment methods can introduce biases, potentially misleading the drug's safety profile and risk -benefit ratio. This paper investigates the app lication of estimands to safety data to address these challenges. We focus on the Exposure - Adjusted Incidence Rate (EAIR), a methodology that provides more accurate comparisons between treatment groups with varying exposure durations. Integrating EAIR with the estimand framework enhances the precision and reliability of safety evaluations. This paper presents an implementation strategy, addresses key challenges, and highlights the framework's potential to enhance safety data interpretation in oncology trials. By comparing EAIRs, we can determine the impact of intercurrent events on incidence rates. Significant differences between these rates may highlight the influence of intercurrent events on the study outcomes | intercurrent event, while-on-treatment, adverse events, treatment policy strategy, exposure-adjusted incidence rate, clinical trial, population, analysis set, treatment effect, trial elements, principal stratum, data science | source presentation source |
8 |
| 17 | Advancing TFL Automation: R-Based Meta-Programming Powered by ARS | Malan Bosman | In the industry's steady move towards TFL automation, there is still a considerable gap when it comes to automating Analysis Results. With the recent development of Analysis Results Standard (ARS, launched by CDISC in April 2024), the opportunity of building TFL automation tools on industry standards was introduced. siera is an R package that ingests ARS metadata and generates fully functional R code - code that can be run as-is to produce the Analysis Results Dataset (ARD). The ARD is a single dataset that contains all the results required in the final TFLs described in the ARS. Far from being a "black box" producing the ARD, this R package instead provides the code, making it inspectable and changeable. It brings the user one step closer to automating TFLs and ultimately plays an important role in end-to-end TFL automation - all while utilizing industry standards (ARS) and Open-Source technology | adam dataset, analysis set, excel, cdisc | source presentation source |
7 |
| 18 | AI Ethical Considerations in Drug Development and Regulatory Review | Tara Chavda, Qingying Ally Lu |
clinical research, clinical trial, regulatory submission | source presentation source |
5 | |
| 19 | AI-Infused Leadership: Empowering C-Suite Decisions with Code-Driven Insights | Ravindra Shukla, Aman Bahl |
As the digital landscape evolves, C-level executives must integrate strategic vision with technological expertise, particularly in AI and automation. This paper explores actionable strategies for healthcare and pharmaceutical leaders to leverage AI for enhanced decision-making, innovation, and operational efficiency. By understanding the foundations of AI and identifying shortcomings in both current AI implementations and other technology systems, executives can refine and optimize their technological investments to better align with organizational goals, maintain a competitive edge, and drive industry advancements. Decision-making through AI will empower C-level executives by providing predictive insights into market trends, financial forecasting, and operational efficiencies. Some of the areas that can be prioritized include the Chief Strategy Officer identifying high-potential therapeutic areas, biomarkers, and personalized medicine | data cleaning, artificial intelligence | source presentation source |
7 |
| 20 | An End-to-End Approach to Fine-Tuning Small LLMs for Generating Admiral R Code in Statistical Programming | Tingting Tian, Jaime Yan |
Tingting Tian, Merck & Co., Inc., Rahway, NJ, USA Automated Admiral R Code Generation 1.2 Proposed Hybrid Approach Toaddressthesechallenges,thispaperpresentsa hybridapproach thatenablestheautomaticgenerationof Admiral-based ADaM datasetswhile offering two deployment options: Cloud-BasedWorkflow (foruserswithoutprivacyconcerns): Weintroducea pipelinethatconverts EXCEL-based ADaM specificationsinto a structuredJSON format. This JSON representation is then used within aknowledge graph query systemand a structuredprompting frameworkto automatically batch-generate Admiral code using cloud-based LLMs. Local Fine-Tuned LLM Workflow(for users requiring data security): We fine-tuneLLaMA 3.1 8Busing Low-Rank Adaptation (LoRA)and domain-specific knowledge to generate Admiral R code locally. While not as powerful as state-of-the-art models likeGPT-4o, our fine-tuned model achievesstrong performancein generating ADaM-compliant code without exposing any data externally |
data cleaning, adam dataset, clinical programming, cdisc | source presentation source |
15 |
| 21 | An Integrated Approach to Protocol Deviation Management | Jason Konn | Clinical trials depend on protocols ensuring scientific validity, ethical conduct, and regulatory compliance. As such, processes to address protocol deviations (PD) are critical for data integrity and patient safety. However, survey results from sponsors, sites and institutional review boards (IRBs) confirm that existing protocol deviation management processes are complex, varied, and broadly impacted by a wide range of definitions for ID and classification. Learn how a modern approach to this process that leverages data integration and analytics can enhance and add efficiency to trial oversight by offering the ability to ingest, standardize and manage protocol deviation data. By bringing data together from various incoming sources and systems into one interface, stakeholders can manage deviation records and enforce standards defined in the study Protocol Deviation Assessment Plan (PDAP) | protocol deviations, clinical trial, data integration | source poster source |
3 |
| 22 | An Intuitive Way to Derive the Last Known Alive Date in R | Marckenley Mercie, Diana Tai |
Accurate derivation of the Last Known Alive Date (LSTALVDT) is critical for the creation of CDISC-compliant ADaM datasets, as it directly influences key efficacy endpoints such as progression-free survival (PFS), overall survival (OS), and long-term follow-up metrics in clinical trials. Traditionally, SAS has been the predominant tool for generating analysis datasets. On the other hand, there is a growing trend in the industry towards adopting open- source solutions. This paper introduces a comprehensive R-based methodology leveraging tidyverse packages to derive LSTALVDT. Furthermore, we present a Shiny application designed to streamline this derivation process, thereby enhancing the efficiency and accuracy of dataset production and validation | sample, sdtm domain, date variable, data cleaning, imputation, adverse events, clinical data, clinical trial, overall survival, missing value, adam dataset, cdisc | source presentation source |
12 |
| 23 | API Interactivity with SAS: A Focus on JSON | Mary Dolegowski, Pritesh Desai |
This paper will showcase how SAS can be used to interact dynamically with a RESTful API to help facilitate better interoperability between a user's SAS environment and other endpoints. There are three primary components users should be aware of when interacting with an API - how to ingest JSON, how to export JSON, and how to establish connections to the respective API. We will cover all three topics and provide options for each. While this will not be an exhaustive dive into APIs, we hope this will provide a starting point for SAS users to increase their interoperability options | proc step, proc http | source poster source |
4 |
| 24 | Applications of Machine Learning and Artificial Intelligence in Real-World Data in Personalised Medicine for Non-Small Cell Lung Cancer Patients | Samiul Haque | In this work, we present a platform approach for analyzing Real World Data (RWD) that provided faster insight for personalized medicine for Non-Small Cell Lung Cancer (NSCLC) patients. NSCLC is the leading cause of cancer death. Personalized medicine is a viable option to improve patient outcomes when patient journeys are illustrated using Machine Learning (ML) and Real World Data (RWD).Chronic Obstructive Pulmonary Disease (COPD) persists during cancer treatment and requires interventions negating the effectiveness of the treatment. Patient journeys for 1.2M NSCLC (ICD-10 C34.*) patients from 2019-2020 using Symphony Claims (ICON Plc) were established and associations were examined using ML on SAS Viya 4.0 (SAS Institute, Inc.)Over 39% of patients had Metastatic Cancer (MC). The factors most related to MC were Dehydration, Weight Loss, Coagulopathy and Liver Disease, respectively. Over 30% of patients also had COPD and experienced an impact on their care | machine learning, likelihood function, gradient boosting, logistic regression, proc hpforest, correlation, logistic distribution, regression analysis | source presentation source |
8 |
| 25 | Applied Strategy: The Human Touch to Success | Shereen Anisa Khwajazada | data science, machine learning | source presentation source |
3 | |
| 26 | Automating Preliminary CDISC SDTM Compliance Checks with a Customised SAS Macro | Wei Qian, Danny (Shih-Che) Hsu |
Validation software, such as Pinnacle 21, exists to verify CDISC standard compliance which is critical in the preparation of regulatory submissions; however, they require uploading datasets and files to an online platform to execute the checks, which adds effort and time. Some organizations further perform these checks right before the submission, causing a delay in the identification of any compliance issues and sometimes in a re-run of SDTM, ADaM, or TLFs. It would be efficient and beneficial to have a customized SAS macro that automates preliminary compliance checks and provides a real-time report after each time SDTM/ADaM datasets are generated. CDISC compliance issues would then be detected in the early stage before creating ADaM and TLFs. This process uses existing software (such as SAS) to improve the efficiency of the data validation process, leading to a superior CDISC compliance | data cleaning, adam dataset, cdisc | source poster source |
3 |
| 27 | Automating SAS and R Code Interpretation and Debugging: A Practical Pipeline for Statistical Programmers | Tingting Tian, Jaime Yan |
This paper introduces a practical pipeline for automating the interpretation and debugging of SAS and R code, specifically designed for statistical programmers in the pharmaceutical and healthcare industries. The system leverages Large Language Models (LLMs), such as Claude or GPT-4, to translate user queries into executable code, run the code within a Jupyter kernel environment, and automatically debug any resulting errors. The pipeline includes a robust, iterative error-handling mechanism that utilizes the LLM to refine and correct the code until accurate results are achieved. A Gradio-based user interface provides a clear and interactive display of both the generated code and its execution output. This approach significantly improves productivity and accessibility in data analysis by automating code generation, execution, and debugging tasks | clinical programming, exception handling, data science | source presentation source |
13 |
| 28 | Automating SDTM Annotated CRF (aCRF) Creation | David Postma, Christine Rossin |
This paper will cover details for the "SDTM aCRF Annotator Tool" used at Pfizer to create submission-ready SDTM annotated CRFs (aCRF), leveraging study documents from data management (DM). Within the tool, an Annotated Study Book (ASB) containing database metadata is used to identify the exact location for the SDTM annotations based on the form and field reference identifiers. A metadata repository (MDR) extract file, which is ultimately uploaded into the tool, links the database field to SDTM variables as annotations. The ASB is first annotated with the SDTM variables, then the Blank CRF is annotated with a mirror of the ASB annotations. Finally, the tool uses study visit metadata stored in a spreadsheet from DM outlining the Schedule of Assessments, to create the table of contents (TOC) and bookmarks for the final aCRF. This tool has proven to take the laborious and time-consuming process of annotating the Blank CRF and make it a quicker and more efficient experience | annotated crf, crf, cdisc | source presentation source |
6 |
| 29 | Automating your safety data reconciliation: An approach towards reliable and efficient SAE reconciliation process | Ritu Dogra | Serious Adverse Event (SAE) reconciliation is a critical step in the management of clinical trials. It involves comparing the safety and clinical database to identify, resolve, and document any discrepancies. Poor quality of the reconciliation may result in erroneous information being reported to health authorities, potentially jeopardizing the approval of new drugs/devices. To mitigate these risks, SGS implemented an automated process that enhances the accuracy and efficiency of SAE reconciliation. This paper will guide you through our automated SAE reconciliation process. It begins with the upload of safety database extract into our internal database. Simultaneously, the AEREFIDs from the safety database are added into the SDTM database, linking both databases along with other variables, e.g. AEDECOD and USUBJID. Afterwards, automatic checks run comparing both databases and listing all discrepancies for review by the Clinical Data Manager | clinical database, excel, adverse events, sdtm domain, cdisc | source presentation source |
11 |
| 30 | Big Data Analytics: Unlocking Insights from Vast Clinical Datasets | Bill Qubeck, Kala Shivalingaiah |
Big Data analytics is unlocking unprecedented insights from the vast amounts of data generated in clinical trials, enabling more informed decision-making throughout the drug development process. Advanced analytical techniques allow researchers to identify patterns, correlations, and trends that were previously hidden in large, complex datasets. This abstract highlights the tools and methodologies that are driving the integration of Big Data into clinical research, leading to more precise and targeted therapeutic interventions. The ability to harness and analyze Big Data effectively is becoming a critical competitive advantage in pharmaceutical R&D | clinical research, clinical trial, machine learning, pharmaceutical industry | source presentation source |
16 |
| 31 | Biomedical Concepts: What Are They and Why Should I Care? | Stuart Malcolm | The CDISC initiative to develop standard Biomedical Concepts (BC) that link all foundation standards offers exciting possibilities - but what exactly is a biomedical concept? What problems does it solve? How and why should we use them? The CDISC Biomedical Concept initiative started in 2022 with an aim to create biomedical concepts and integrate them into all the Foundation Standards (CDASH, SDTM, ADaM), and during 2024 the initiative has completed its first public review of BCs with SDTM dataset specializations [1]. This presentation aims to offer a "Getting Started" guide to CDISC Biomedical Concepts initiative, introducing the concepts, walking through practical examples, and discussing possible use-cases and possible "quick-win" implementation scenarios | sdtm domain, clinical trial, unique key, clinical research, data governance, raw data, cdisc | source presentation source |
7 |
| 32 | Bridging the Gap: Creating and Debugging User-Defined R Functions from an SAS Programmer's Perspective | Hrideep Antony | As the field of data science continues to evolve, many SAS programmers are transitioning to R, attracted by its versatility, open-source advantages, and growing prominence in the pharmaceutical industry. However, this shift from SAS's procedural programming approach to R's functional paradigm introduces unique challenges, particularly in the creation and debugging of user-defined functions. While R functions and SAS macros both serve as essential tools for modularizing and automating tasks within their respective programming environments, they differ substantially in their structure, execution, and underlying concepts | macro definition, sas programmer, macro variable | source presentation source |
9 |
| 33 | Building High-Performing Biometrics Teams: The SELF-REINFORCING LOOP and Leadership in Clinical Trials | Arman Melkonyan | In the evolving field of biometrics teams within pharma, effective leadership is crucial for driving innovation, maintaining compliance, and achieving outcomes. This article explores essential leadership practises that elevate biometric teams into high-performing units capable of overcoming complex challenges. Key leadership skills include technical expertise, clear communication, and strong project management. A significant focus is on the SELF- REINFORCING LOOP, which integrates management and training by interviewing trainees, providing managers with feedback to address skill gaps, assigning relevant tasks, and conducting follow-up interviews to assess progress. This continuous loop fosters growth, development, and the cultivation of highly skilled and industry-ready biometric professionals | clinical trial, clinical programming | source presentation source |
6 |
| 34 | Building Rome in a Few Months: How Open-Source Technologies Empower Small Pharma | Chendi Liao, Kasra Yousefi |
clinical data, clinical trial, clinical research, data science | source presentation source |
5 | |
| 35 | Can AI Help Overcome the Validation Burden for Open Source? | Marc Jantke | data cleaning, validation documentation, pharmaceutical industry, clinical research, training set | source presentation source |
8 | |
| 36 | Can AI/ML Drive Data Analysis and Data Mapping in Clinical Trials? | Philip Ross, Revvity Signals |
How can we improve clinical trial data mapping and analysis using AI and machine learning? We share our experiences in data automapping using machine learning to recommend data mapping so that clinical team members can focus on resolution of unresolved data mapping issues. Automated data loading, mapping and transformation provide daily refreshes of emerging standardized data useful for accelerated data monitoring, data cleaning and clinical review. We also explore the use of Generative AI for analysing clinical trial data. AI can help to identify patterns and generate insights from large datasets; however, its use in the regulated clinical trial environment requires careful consideration. The technology may assist in data interpretation and hypothesis generation, but caution is needed to ensure accuracy and adherence to standards. We share our learnings and discuss our ongoing efforts to use AI/ML to improve data understanding and drive clinical strategy and execution | clinical trial, clinical data, adverse events, data mapping, data cleaning | source presentation source |
11 |
| 37 | CDISC 360i: Implementing the CDISC Roadmap Connecting Standards from Design to Analysis | Chris Decker, Peter Van Reusel |
clinical research, change from baseline, cdisc | source presentation source |
4 | |
| 38 | CDISC Open Rules: A Deep Dive into Custom Rule Creation | Roman Radelicki | The CDISC Open Rules project encourages the use of its open-source software to test study data for conformance to various standards by using the same set of unambiguous rules governed by CDISC. Next to these rules, establishing an industry-standard method for rule creation offers a promising prospect. Emphasizing the versatile application of Conformance Rules not just as submission standards, but as essential tools throughout the clinical data lifecycle, facilitating communication and collaboration among stakeholders. This paper will explore creating custom rules using the Conformance Rule Authoring tool. Our journey/challenges of implementing the authoring tool in our cloud environment and running custom rules with the CDISC Open Rules Engine (CORE) , will be shared. We'll further explore several use cases, including data listings, validation of non-SDTM clinical data with a custom define.xml, e.g. external vendor data | data cleaning, clinical data, excel, data validation rules, clinical research, proc print, cdisc | source presentation source |
12 |
| 39 | Challenges of Genomic Data in Clinical Trials | Hong Zheng, Lingsong Yun |
sample, clinical trial, quality control, clinical research, raw data, clinical data, proc freq | source presentation source |
8 | |
| 40 | Changes to the Job of a Statistical Programmer in the Pharmaceutical Industry | James J. Kim | This paper discusses the evolving roles of statistical programmers in the pharmaceutical industry compared to traditional computer science programmers. The distinction was clear when the main tool of statistical programmers has been proprietary software for the last 30 years. But as the adoption of open-source tools like R is gaining momentum recently, the distinction between them has become unclear and created confusion on what their main responsibilities should be in their day-to-day activities. While there is no correct answer and their roles can certainly be overlapped to a certain extent, understanding the differences will help to plan programming capabilities efficiently, especially for smaller-sized companies with limited resources | data science, data cleaning, quality control, regulatory submission, statistical programmer, life sciences | source presentation source |
8 |
| 41 | Clearing Up Confusion with Reference Dates | Swathi Kotla, Dana Booth |
RFSTDTC, RFXSTDTC, RFCSTDTC, RFICDTC (and RFENDTC, RFXENDTC, RFCENDTC, RFPENDTC) are all different reference dates available in the DM and upcoming DC demographics domains. But what do they mean? When should you use them? And why are there so many? | per protocol, washout period | source poster source |
5 |
| 42 | Clinical Trial Reporting Automation Based on Knowledge Graph | Jaya Simha Inampudi | Automating the clinical trial reporting process is a complex challenge that requires harmonizing diverse data sources and analytical requirements. This abstract presents an innovative approach leveraging a knowledge graph to enable end-to-end reporting automation, including the generation of standards-compliant specifications and data. At the core of this framework is a comprehensive knowledge graph that links therapeutic area endpoints to specific analyses, which are then mapped to ADaM datasets and their underlying SDTM and raw data sources. This connected information backbone allows the system to dynamically generate not only the necessary study components, such as tables, figures, and listings, but also the standards-compliant SDTM and ADaM data specifications. An AI/ML-driven automation engine leverages the knowledge graph to streamline the reporting workflow, from data transformation to output generation | clinical trial, data science, cdisc | source poster source |
5 |
| 43 | clinify: Bridging the Gap in Regulatory-Compliant TFL Generation with R | Michael Stackhouse | tables, label statement | source presentation source |
10 | |
| 44 | Collaborating on Standards: An Approach to Harmonizing Vaccine Regulatory Submissions | Fred Bermont, Médéric Celle, Estella Sani, Erin Gnas, Jennifer Mastri, Leslie Morgan, Naila Musani, Ramana Setty, Sandra VanPelt Nguyen, Yun Zhang |
In a significant collaboration, seven leading vaccine organizations -AstraZeneca, GlaxoSmithKline, Johnson & Johnson, Merck, Moderna, Pfizer, and Sanofi -have formed the Vaccines Industry Standards Group (VISG). Over the past two years, this initiative has focused on harmonizing interpretations of regulatory submission guidance and CDISC data standards. T he group recognizes that aligning on processes-such as participant diary data collection and submission of reactogenicity and efficacy data - accelerates time to market and benefits global health. This unified approach enables proactive alignment on standards in preparation for engagement with health authorities and standards organizations, in addition to promoting clarity and consistency in submission standards | data standard, regulatory submission, cdisc | source presentation source |
8 |
| 45 | Comparing Confidence Intervals for a Difference in Adjusted Predicted Probabilities | Cralen Davis, Trevano Dean |
As a statistical method, calculating adjusted predicted probabilities (APPs) is a common and straight-forward analysis when modeling a binary outcome. The procedure, PROC LOGISTIC (PL) within the analysis software package, SAS (Cary, NC) is one of the most used approaches to accomplish this task. However, when a goal is to compute a valid confidence interval (CI) of the difference between mean APPs for 2 groups, the analysis is less straight-forward; this is evident by the fact that a CI for the difference of mean APPs is not directly provided within SAS. For a given dataset with 2 groups and a binary outcome, we used the estimates provided by PL and computed the difference between mean APPs for the 2 groups; computed a 95% CI under asymptotic normal approximation (ANA); and compared it to 95% CIs constructed using bootstrap sampling |
logistic distribution, bootstrap, sample, confidence interval, logistic regression, p-value, proc logistic, proc means, proc step, regression analysis, response variable | source presentation source |
5 |
| 46 | Considerations for the Submission of RWD using CDISC with Insights from HL7 FHIR and OMOP | Ingeborg Holt, Sarah Ferko, Jeffrey M. Abolafia |
There is a rapid increase in the use of real-world data (RWD) to support marketing applications for new medical products. However, submitting RWD poses numerous challenges for regulatory reviewers in their analysis of this data. In our five previous papers, we outlined these challenges and presented several high-level solutions. This current paper builds on our previous work, taking a deep dive into an example CDISC SDTM submission and highlighting updates needed to the SDTM to better accommodate RWD. In this paper, we examine Adverse Events (AE) and Medical History (MH) data. Considering both current regulatory requirements and how other data models, designed for RWD, such as HL7 FHIR, OMOP, and PCORnet, handle the elements that would be considered for these two domains. We propose new variables, update existing variables core (cardinality) designations and business/validation rules that are needed to support the use of RWD to accurately represent AE and MH data | adverse events, raw data, insurance claims, clinical research, clinical data, clinical trial, data cleaning, medical history, verbatim term, data standard, vital signs, cdisc | source presentation source |
15 |
| 47 | Data extraction and Graphical visualization for quick decision making | Mukul K Mittal | Summary tables from clinical trials offer a succinct overview of key findings, including participants demographics, treatment effects, and safety outcomes. While these tables provide essential insights, visualization of the data significantly enhances comp rehension and communication. However, publications and older studies often present summarized data without raw data access, complicating accurate trend analysis. To address this, we developed R - Shiny applications, Data ExtractoR to extract data from summary reports in formats such as PDF/RTF/DOCX and Graph DesigneR for visualization. Our solution incorporates a Python environment with the Camelot package for PDF table extraction, docxtractr for word, and striprtf package for RTF files. Users can select and process multiple tables, and the data undergoes automated cleaning operations through custom functions | extract, proc transpose | source presentation source |
6 |
| 48 | Data Quality in Pooled Studies: Overcoming Challenges and Ensuring Accuracy | Monika Ludwinska | Pooling data from multiple studies to accomplish an extensive knowledge of a treatment and disease or to conduct a robust analysis of drug's safety and efficacy is a commonly recognized strategy in the pharmaceutical industry. This method requires integrating data from studies, potentially having similar methodologies. Collecting data and subsequently examining it introduce unique challenges to be effectively considered. These obstacles involve maintaining consistency, compliance with CDISC standards and managing heterogeneity across studies. The integration of such diverse datasets requires meticulous planning and coordination to ensure compatibility and accuracy. The quality of the pooled data depends on the individual studies, as consolidating can mask significant differences. This paper pro vides a guidance on planning a pooled database from studies with similar structure with overview, typical pitfalls and key considerations to develop a highly efficient pool | data cleaning, adam dataset, proc step, numeric variable, proc freq, character variable, data integrity, pharmaceutical industry, excel, cdisc | source presentation source |
11 |
| 49 | Data Transformation: Achieving Compliance with Irregular Data | Reyes Eduardo Elizalde Duarte | External data generated on clinical trials needs to be collected and submitted following Clinical Data Interchange Standards Consortium (CDISC) standards to ensure compliance with regulatory authorities. Early Phase studies face a particular issue due to the short duration of trial execution, so meeting CDISC standards becomes increasingly challenging. The high number of data types, vendors, and external agreements hinders the standardization of data. Data Management works to ensure all data is transformed to meet CDISC standards and ensure SDTM programmers and biostatisticians can perform their needed tasks. It becomes imperative, for early phase trials to plan, collect, organize, and analyze data as efficiently and correctly as possible. This paper will go through the various challenges, steps, strategies, and insights to understand CDISC requirements, ensuring any delays are prevented and data meets all regulatory requirements | data transformation, clinical trial, clinical database, sdtm programmer, normalization, data governance, sample, data science, cdisc | source presentation source |
6 |
| 50 | Demystifying the Handling and Mapping of External Data to CDISC Standards in Clinical Trials | Murali Kanakenahalli, Vamsi Kandimalla |
As the complexity of clinical trials has grown over the years, the dependency on different kinds of external data has increased tremendously. Different external data types like pharmacokinetic, immunogenicity, biomarker, genomics, cell phenotype, biospecimen, microbiology findings, etc., have been used to support the primary, secondary end points of a study and exploratory analysis. It's a substantial task for statistical programming teams to understand, analyze and maintain external data. A great deal of resources is required to develop an infrastructure to process the external data along with using CDISC implementation to map to SDTM standards. We will discuss the various types of external data. Use data transfer plan with labs to receive clean data transfers. Understand external data and use CDISC to map to SDTM. Reconcile external data and handle some of the data reconciliation challenges. Also, discuss the documentation which can be reference in understanding and mapping to | sample, clinical trial, quality control, raw data, clinical programming, data science, cdisc | source presentation source |
12 |
| 51 | Differential Privacy (DP) Approach for Sharing Survival Analysis Data | Elena Valkanova, Bulat Fatykhov |
Privacy-preserving methods in clinical trials have become crucial for promoting transparency in medical research. Many of the existing privacy methods do not provide provable privacy guarantees. Both clinical trial data sharing and privacy are stipulated requirements by the National Institutes of Health (NIH) and the International Committee of Medical Journal Editors (ICMJE). The objective of this paper is to identify a range of values for the privacy parameter before sharing clinical trial data. We apply a differential privacy (DP) method to both simulated data and real data from a clinical trial considering only time-to-event related endpoints. In addition, we compare the magnitude of the trade-off between the data utility and the privacy budget parameter (noise) ranging from 0.01 to 10 using R | survival analysis, clinical trial, median survival, clinical data, interquartile range, sample size, sample | source poster source |
11 |
| 52 | Digital Health Technologies: Evolution, Insights and Future Directions | Vijay Pasapula | Presentation | data cleaning, regulatory submission, quality control, audit trail, clinical data, clinical research, clinical trial | source | 28 |
| 53 | Dismantling the Black Box: Why We Need to Think Bigger About AI Integration | Christina Dinger, Joe Lebers |
AI is transforming the clinical development value chain by generating insights, reducing silos, and automating clinical data transformation and review tasks. However, when AI models are deployed without seamless integration, traceability, and explainability, the value they bring is often lost. The key to effective AI is when it functions as an intuitive, seamless extension of existing workflows. This paper explores how AI can become a "new stakeholder" in the clinical trial process. It will have its own vested interest, priority, relationships, trust, and value, but will contribute to the collective team, not be isolated in a black box. We'll demonstrate how ThoughtSphere's platform enables a collaborative relationship between humans and machines, delivering advanced data aggregation and transformation, real-time analytics, and automated data check generation | clinical trial, clinical research, type i error, clinical data, data ingestion, training set, data cleaning | source presentation source |
7 |
| 54 | Double Trouble: Essential Insights from Successful Dual sBLA Submissions for a Drug-Device Combo | Elma Ferdousy, Aldrich Salva |
This paper explores the insights derived from the unique challenges of two concurrent supplemental Biologics License Application (sBLA) submissions for a drug-device combination product in the midst of improvements due to an identified issue.This submission involved two distinct Phase III studies with multiple datacuts, necessitated by shifting filing timelines due to device remediation efforts. It also included the integration of pooled safety data from three additional studies to support the Integrated Summary of Safety (ISS).These elements led to the creation of numerous data packages, presenting challenges to the team in consistency, data quality, and conformance. Key takeaways for the team in overcoming these challenges included seamless collaboration among various teams, rigorous implementation of data conformance tools, and thorough documentation across the various data packages. Despite the complexities, the eSubmission was successfully completed on time | data cleaning, adam dataset, cdisc | source presentation source |
7 |
| 55 | Dynamic Survival Model for Monitoring Resistance Mutations in Targeted Therapy: Advancing Precision Medicine in Oncology | Hong Zheng | Targeted therapies have revolutionized cancer treatment, offering patients more personalized and effective options. However, the development of resistance mutations poses significant challenges to long-term treatment efficacy. Recently, global health agencies, including the U.S. Food and Drug Administration (FDA) and the European Medicines Agency (EMA), have begun to request resistance mutation analysis for targeted therapies in certain new drug applications. The dynamic nature of cancer presents ongoing challenges, particularly the emergence of acquired resistance mutations following the initial treatment response. In addressing this critical issue, a time-varying Cox proportional hazards model was developed to evaluate and monitor genetic resistance mutations in patients undergoing targeted therapy. This model accounts for longitudinal changes in mutation status and their association with time-to-event endpoints | time-dependent covariate, progression-free survival, proc phreg, clinical research | source presentation source |
5 |
| 56 | E2E R Submission from a Roche Oncology Study | Jingyuan Chen, Hinal Patel |
In the rapidly evolving landscape of pharmaceutical development, Roche embraces the use of open-source R in the regulatory submission process. In this paper we will share a pioneer experience of end-t o-end R submission for a new drug application to the FDA, EMA and the NMPA. We will unfold the journey of overcoming the initial challenges to achieving the filing goal and emphasise the practicalities of preparing clinical trial data in R leading to a comprehensive eSubmission package. We wil l also share our strategic communication and collaboration with regulatory agencies, highlighting how R- generated data and documentation were effectively presented and validated within the submission framework | data cleaning, clinical trial, data science, adam dataset, regulatory submission, cdisc | source presentation source |
6 |
| 57 | Effective Adverse Event Management | Eva Kelty, Catharina Dahlbo |
The International Council for Harmonisation (ICH) introduced the Development Safety Update Report (DSUR) to streamline safety reporting during the development phase of a drug. Adverse events recording and reporting is an important part of a clinical trial that is very time consuming. It is about understanding relationships between adverse events and the drug to ensure the safety of trial participants. To optimize the management of adverse events, choosing a graph database-based interface is an effective approach. The interface enables the smooth integration of various data sources, intuitive visualization of complex relationships, and flexible querying options. By improving the accuracy and efficiency of adverse event data management, this tool contributes to better decision-making and strengthens patient safety | adverse events, relational database, clinical research, clinical trial | source poster source |
6 |
| 58 | Efficient Positioning of QTL and Secondary Limit Thresholds in Clinical Trial Risk-Based Monitoring | Vladimir Shnaydman | The paper discusses a novel methodology for efficiently positioning Quality Tolerance Limits (QTL) and Secondary Limit (SL) thresholds in clinical trials, focusing on risk-based monitoring (RBM) and risk-based quality management (RBQM) - detailed methodology presented in [1,2] | clinical trial, operational risk, likelihood function | source presentation source |
12 |
| 59 | Efficient TFL Review within Statistical Programming - Using R and SAS | Siddharth Kumar | In clinical data analysis, while all outputs undergo a process of independent QC involving double programming, certain issues in RTF (Rich Text Format) outputs may not be detected by PROC COMPARE, resulting in unaddressed quality issues that are later identified during the cross -functional review. To instill confidence in the TLF (Tables, Listings, and Figures) package, a thorough review within the statistical programming team is essential. Manually reviewing over ~300 RTF outputs in a submission package is time -consuming and prone to errors. This presentation addresses these challenges by proposing an automated solution using R and various R packages to streamline the review process |
ods rtf, data cleaning, proc step, clinical trial, clinical programming, proc report, dataset, clinical data, proc compare, adverse events, data science, independent qc, regulatory submission, secondary endpoint, excel | source presentation source |
8 |
| 60 | Effortless SDTM Upversioning: Leveraging Open-Source R for Efficient Data Standardisation | Ramya Gangadi | CDISC has always recognized the importance of developing data standards to streamline the electronic submission process, benefiting the pharmaceutical industry. As part of this effort, the CDISC -SDS team will be regularly updating the implementation guide to introduce new domains and implementation rules. For any new studies submitted to regulatory agencies in 2023 and beyond, it is recommended to use SDTM -IG 3.3. Given the substantial updates, understanding these changes poses a significant challenge for the Pharma/Biotech industry. To address this, we have developed a post -processing approach using R open -source technology. By leveraging a difference file from the CDISC Library, this approach allows for the seamless updating of SDTM datasets from version 3.2 to 3.3, or from 3.3 to 3.4, without affecting the standard macros and programs of an existing sponsor library | data cleaning, sdtm domain, excel, cdisc | source presentation source |
7 |
| 61 | Elevate Your Game: Leveling Up SDTM Validation with the Magic of Data Managers | Julie Ann Hood | Selecting players armed with unique abilities to collaborate with is the key to crafting an unbeatable strategy while on the field, navigating a quest, or in the office. In the realm of standardized clinical trial data, not only can this break down departmental silos, but it can also enhance the quality of study data, leading to the availability of more effective, efficient treatments earlier. Historically, due to the timing and content of SDTM validation, many organizations enlisted SDTM programmers with unraveling issues in these reports. However, there are some rules that require tracing SDTM data back to raw data or data collection to optimize decision-making. Seamlessly integrating data managers into this process can help slash time to resolution of these types of issues as well as boost overall data quality. This poster will provide important considerations and guidance on weaving data managers into the SDTM validation process | data cleaning, raw data, adverse events, sdtm programmer, data governance, data audit, data validation rules, clinical trial, cdisc | source poster source |
4 |
| 62 | Enable and Automate: A Technical Roadmap Defining CDISC's Path to End-to-End Automation | Samuel Hume | Enable and Automate is an important pillar of CDISC's new strategy that provides the vision and roadmap to enable standards-driven automation across the clinical research data lifecycle-from study design to analysis results. This paper provides a high-level synopsis of the technical roadmap, describing how the CDISC community will implement and showcase end-to-end automation. Key initiatives include significant standards updates, open-source software tools, and test datasets that will be used to demonstrate the research data pipeline automation. One of the major deliverables is a fully pre-configured example study, with all study components represented as machine-readable metadata, from design through submission. The roadmap outlines a series of epics that we will develop and implement over the first phase of the 360i program | phase i, raw data, data pipeline, sdtm domain, phase ii, cdisc | source presentation source |
8 |
| 63 | End-to-End Therapeutic Area Data Standards Development and Governance - A Case Study Involving CRS and ICANS | Yaling Teng | The development of therapeutic area data standards in the pharmaceutical industry requires ongoing enhancement, even in disease areas with existing regulatory guidelines. This paper addresses the agile expansion of company data standards, using Cytokine Release Syndrome (CRS) and Immune Effector Cell-Associated Neurotoxicity Syndrome (ICANS) as examples. These toxicities are closely monitored both pre- and post- regulatory approval for cancer immunotherapies. By developing end-to-end data standards for CRS and ICANS and implementing them in applicable clinical trials, companies can more effectively optimize safety profile of new treatments, enhance data quality and interoperability, and ensure regulatory readiness. The proposed framework aims to guide the development of comprehensive data standards for these toxicities, offering insights for advancing oncology data standards | data standard, adverse events, adam dataset, treatment period, vital signs, clinical trial, ctcae grade, cdisc | source presentation source |
7 |
| 64 | Enhancing Clinical Trial Data Queries with LLMs and Neo4j: A Flexible Framework for ADaM Dataset Management | Jaime Yan | CDISC ADaM standards are essential for clinical trial data management, ensuring regulatory alignment and data consistency. While traditional SQL-based querying methods remain prevalent, graph databases offer unique advantages for handling interconnected clinical trial data. This study presents an innovative framework that leverages Neo4j graph database for ADaM datasets, using Large Language Models (LLMs) to generate Cypher queries while benchmarking against conventional SQL approaches. We developed a reusable graph schema specifically designed for ADaM datasets, enabling efficient storage and retrieval of clinical trial data within Neo4j. The framework allows users to input natural language questions, which are processed through a graph similarity search to identify relevant schema elements. These elements serve as input to the LLM, which primarily generates Cypher queries while maintaining SQL compatibility for comparative analysis | proc sql, clinical trial, adverse events, clinical data, serious adverse event, data science, adam dataset, data governance, f1 score, best response, excel, system organ class, concomitant medications, foreign key, key, relational database, clinical trial protocol, data cleaning, data integrity, neural network, clinical research, standard deviation, clinical database, adsl, progression-free survival, cdisc | source presentation source |
28 |
| 65 | Enhancing Data Accuracy and Collaboration Through Efficient Data Reconciliation Using SAS in Clinical Trials | Piyush Makkapati, Swathi Nattam |
Reconciling data between Electronic Data Capture (EDC) systems and external vendor databases is vital for ensuring the accuracy and integrity of clinical trial data. Effective, timely reconciliation mitigates discrepancies that could compromise study outcomes. While reconciliation is important throughout the trial, its significance is heightened as database locks (DBLs) approach, where preventing costly delays and data integrity issues post-DBL is critical. This paper introduces a streamlined reconciliation approach, utilizing key variables within SAS to integrate data from multiple sources. Tailored discrepancy reports and specifications, developed in accordance with Vendor Data Transfer Agreements (DTAs), facilitate collaboration between Data Management (DM) and programming teams | clinical trial, database lock, adverse events, dataset, sample, data governance, data reconciliation, data cleaning | source presentation source |
8 |
| 66 | Enhancing Data Science with Multilingual Capabilities: A Comprehensive GxP-Compliant Solution | Eli Miller, Casey Higgins |
Discover the future of data sciences with our multilingual platform, offering centralized data storage that is fully GxP- compliant out of the box. This innovative solution automates the addition of user and security policies, ensuring seamless and secure access to data. Our platform features OpenVal, a validated R package product, enhancing data integrity and compliance. Additionally, it includes open-sourced SAS macros and ready-to-use Shiny templates for efficient data review. With a robust ecosystem of partners for data storage, integrations, and cloud services, our platform provides a versatile and scalable solution for data sciences. Join us to explore how this comprehensive toolset can streamline your processes, improve data quality, and ensure regulatory compliance | data integrity, cloud storage | source presentation source |
3 |
| 67 | Enhancing Pharmaceutical Quality Assurance Through Large Language Models: A Novel Approach to Intelligent Regulatory Monitoring | Jake Gagnon, Hangyu (Cedric) Liu |
This study introduces an innovative Generative AI framework to enhance regulatory compliance processes in the pharmaceutical industry. With the continuous evolution of FDA and EMA regulations, pharmaceutical companies often face the challenge of managing manual, error-prone workflows for updating internal documents, such as Standard Operating Procedures (SOPs). We propose a two-step AI-driven solution to streamline regulatory update management. The first step leverages semantic search to identify SOPs potentially impacted by new regulatory updates. The second step utilizes a Large Language Model framework, powered by GPT-4o, to perform detailed comparisons between the new regulations and identified SOPs, confirming their impact and pinpointing specific sections requiring updates. This automated approach aims to reduce manual effort, minimize the risk of oversight, and ensure timely alignment with regulatory changes | pharmaceutical industry, likelihood function | source presentation source |
11 |
| 68 | Enhancing the (ASAP) Aggregate Safety Assessment Planning Framework with an Interactive R Markdown Application: Seeing is believing! | Debopriya Ghosh | When developing a compound for multiple diseases indications across various therapeutic areas, planning for ongoing integrated safety review and aggregate analyses can be challenging. Maintaining uniformity across studies for definitions and analyses is imperative. Unblinded safety review for ongoing studies are performed by compound level Data Monitoring Committee, and post study completion, the aggregate safety analyses are carried out using a comprehensive Integrated Safety Assessment Plan. Enhancing the ongoing aggregate safety review during clinical development still pose significant challenges. Since the introduction of ASAP1, organizations have been attempting to implement the process into their standard safety review and thereby develop tools to support the ongoing safety data review | adverse events, clinical research, data review, data science | source presentation source |
7 |
| 69 | Ensuring Data Integrity: A Three Step Approach to Clinical Dashboard Validation | River Meckstroth | In an era where clinical decisions increasingly rely on automated visualizations, ensuring the accuracy of data within PowerBI dashboards is crucial. This presentation explores an innovative three-step validation process that addresses the challenges of data ingestion, model integrity, and visualization accuracy in clinical dashboards. By employing rival programming techniques and rigorous quality control checks, we mitigate risks of data misrepresentation and uphold the ethical use of clinical data. The automation of this process will improve clinical decision-making timeframes and reinforce the reliability of the decisions. Attendees will gain valuable insights into a systematic approach that not only safeguards data integrity but also sets the stage for industry-wide best practices in dashboard validation | data cleaning, data integrity, data ingestion, clinical data, raw data, pharmaceutical industry, quality assurance, quality control | source presentation source |
7 |
| 70 | eProtocol and the Promise of Automation | Bill Qubeck, Suhas Giriraj |
The eProtocol initiative represents a significant advancement in the automation of clinical trial processes, facilitating streamlined protocol design, execution, and monitoring. By digitizing traditionally manual tasks, eProtocol enhances efficiency, reduces human error, and accelerates trial timelines. This abstract explores how eProtocol is transforming the landscape of clinical research by offering a more integrated and automated approach to protocol management. The potential benefits for sponsors, sites, and regulators alike underscore its promise as a cornerstone of modern clinical trials | clinical trial, clinical trial protocol, clinical research, regulatory submission, clinical database, data integrity, protocol amendment, proc power and sample size, data science, sample, data cleaning, sample size, randomized controlled trial, artificial intelligence, data governance, clinical data, edc, metadata management, secondary endpoint | source presentation source |
17 |
| 71 | Evaluation of Open-Source Technologies in Clinical Trial Reporting | Jaimin Patel | Open-source technologies are transforming clinical trial reporting by offering cost-effective and flexible solutions for data analysis and visualization. In clinical trial reporting, R and Python are pivotal for data analysis and visualization. R excels with its specialized packages like ggplot2 for creating high-quality statistical graphics and dplyr for data manipulation, making it ideal for rigorous statistical analysis. Python, on the other hand, offers flexibility through libraries such as pandas for data handling and matplotlib for plotting, and integrates well with other data science tools. Both languages support the generation of accurate and reproducible reports, facilitating the clear communication of clinical trial results. Their complementary strengths make them valuable in ensuring comprehensive and transparent research documentation | clinical trial, data science, clinical data, machine learning, hypothesis testing, survival analysis, box plot, regulatory submission, excel, adverse events, outlier, regression analysis | source poster source |
7 |
| 72 | Evolution of Genomic Data Mapping from PF to GF in SDTM v3.4: Concepts, Challenges and Controlled Terminology with Genetic Variation Examples | Rama Kudaravalli | The Genomics Findings (GF) domain in SDTM v3.4 was introduced to address the need for standardized representation of genomic data in clinical research. It is designed to handle various types of genomic data, including genomic summary measures, and replaces the PF domain from the provisional SDTM Implementation Guide for Pharmacogenomics and Pharmacogenetics (SDTMIG -PGx). This evolution from PF to GF reflects the need for more comprehensive and standardized representation of genomic data in clinical trial data standards. One of the key challenges is to ensure that the domain can accommodate a wide range of genomic findings while maintaining consistency and standardization. Controlled terminology plays a crucial role in the GF domain, providing a standardized vocabulary for describing genomic findings | controlled terminology, clinical trial, sample, cdisc | source presentation source |
7 |
| 73 | Expediting the Regulatory Submission Process Through Automated Tools | Manish Deole, Gautham Selvaraj |
In the biopharmaceutical industry, expediting regulatory submissions is crucial for timely access to life-saving medications. As a statistical programming team, our role involves accelerating the drug approval process by meticulously preparing (eCTD) packages. Hence, to expedite the statistical review and programming process of Mapping SDTM, Deriving ADaM and TLF generation. Hence, in this presentation Cytel team presents the process and benefits obtained by following the metadata driven approach: from mapping to report, which enhances the efficiency in attaining the results and generating the submission packages promptly by reducing manual interventions | mock shell, regulatory submission, cdisc | source poster source |
4 |
| 74 | FAIR Data, Clear Insights: Enhanced PRO Analysis | Fangming Liao | Roche's application of FAIR (Findable, Accessible, Interoperable, and Reusable) data principles facilitated secondary data usage and big data analytics, including the creation of indication-specific data marts under the EDIS (Enhanced Data and Insights Sharing) initiative. This project utilizes the data mart created for Non-Hodgkin Lymphoma to analyze patient-reported outcomes (PRO), demonstrating the prognostic and psychometric value of key PRO scales. We developed a new risk index that incorporates the NHL International Prognostic Index (IPI) and identified subgroups of patients with advanced DLBCL and high unmet medical needs. Currently we are analyzing all studies in the EDIS NHL data mart to inform future Target Product Profiles (TPPs) with PRO endpoints using statistical evaluations of PRO scales through change from baseline and time to deterioration analyses. This work highlights how FAIR data and robust statistical analysis drive innovation and improve patient care | change from baseline, data mart | source presentation source |
2 |
| 75 | Fit for Purpose: How to Prepare a Package for Efficient Data Monitoring Committee (DMC) Data Review | Ke Xiao, Hong Wang |
When it comes to prepare deliverables for Data Monitoring Committee (DMC) data review, there are a few common misconceptions. One of the biggest pitfalls is a DMC package should be as comprehensive as possible. This results in a larger volume package with less valuable information, which, consequently, tends to miss more important signals. In addition, a DMC package is often considered no different than other deliverables, for instance, interim analysis report, final clinical trial report, etc. However, this might not always be the case. Deliverables that do not fit DMC purpose would not facilitate the committee to provide timely and effective recommendations. Lastly, for ongoing studies, the data used to support DMC review might not be perfectly clean as expected. Taking all these into consideration, in this paper, the authors will share examples of those pitfalls, lessons learned from their experience, provide recommendations, potential solutions, and open for further discussion | data review, interim analysis | source poster source |
4 |
| 76 | From Application to Approval: Exploring the FDA's Review Process | Kiran Bonda | data cleaning, clinical research, regulatory submission, unstructured data | source presentation source |
4 | |
| 77 | From Data Manager to Data Scientist: Navigating the Future of Data Management in Clinical Trials | Debra Anderson | The role of Clinical Data Management (CDM) is changing to meet new challenges such as decentralization of clinical trials, new data sources, real-world data, and evolving data management environments. Clinical Data Science (CDS) represents the adoption of new technologies, new processes and new data analytics skills while also continuing to maintain a deep understanding of the clinical trial domain, good clinical data practices, and other core CDM competencies which are essential to support efficient decision-making throughout the life cycle of clinical research. This paper explores the professional development challenge in transforming from CDM to CDS in terms of foundational knowledge, mindset, technical, and soft skills | clinical trial, data science, machine learning, clinical research, data review, clinical data | source presentation source |
6 |
| 78 | From Fragmentation to Fusion: A Journey through Mergers and Acquisitions Data Integration Challenges | Xiao Huang | In today's dynamic pharmaceutical landscape, mergers and acquisitions (M&A) have become integral strategies for companies aiming to enhance their market presence, expand their pipelines, and drive innovation (McKinsey & Company, 2023). However, the seamless integration of data from acquired entities remains a significant challenge. This paper examines the multifaceted challenges that a statistical programming (SP) team may encounter during data integration following pharmaceutical acquisitions. Key topics explored include understanding acquired data, maintaining data quality, and harmonizing variations in company data management practices while ensuring compliance with both corporate and regulatory standards | data migration, data integration, clinical research, data science, data cleaning | source presentation source |
3 |
| 79 | From Frozen to Fluid: How the {slushy} R Package Meets the Unique Needs of Study Teams | Becca Krouse | In the pharma industry, stable and secure R environments are crucial for our work. The ideal solution has often been the use of locked down, "frozen" environments, which are carefully curated and tested by tech teams. However, the fast-paced nature of R development means these environments can become outdated quickly, preventing teams from utilizing the latest advances. Enter {slushy}: a team-friendly R package powered by {renv} and Posit Package Manager. {slushy} was designed to quickly mimic a future controlled environment, with the easy ability to time travel between CRAN snapshots. However {slushy} offers more than just a glimpse into the future. It caters to the unique needs of studies by empowering teams to define and assess the quality of their own custom R environments. Attendees will discover how {slushy} exceeded our expectations by proving its value in a variety of scenarios at GSK, in turn bolstering our R adoption efforts | data cleaning, pharmaceutical industry, r package | source presentation source |
3 |
| 80 | From Manual to Machine: Leveraging Best-of-Breed Approaches in Clinical Data Management | Joseph Lebers | With the explosion of AI technology, CDM solutions have flooded the market, offering novel AI approaches for conducting data reviews and shortening delivery timelines. Many of these "black box" solutions are experimental at best and require massive change management to implement successfully. This puts drug developers in a tough position-deliver cutting-edge innovation, but with a high risk of failure. AI-tech approaches that blend ML and AI capabilities into established and effective human -driven CDM processes offer great promise. By seamlessly marrying Human and AI capital, CDM teams gain efficiency, reduce the change management burden, and advance clinical research with minimal risk. Whether you're evaluating a vendor solution, or devising a change management strategy, join us as we discuss technology adoption that takes a best-of-breed approach | clinical trial, business intelligence, clinical data, recommendation engine, machine learning, data science, key performance indicator, clinical research, data cleaning | source presentation source |
10 |
| 81 | From RAGS to Riches: Creating a CDISC Standard Customer Service Agent Using R and Generative AI | Daniel Woodie | clinical trial | source presentation source |
16 | |
| 82 | From Roadblocks to Roadmaps: Overcoming Challenges in Registry Data Reporting | Sivasankar Konda, Neha Srivastava |
Year after year, Real World Evidence (RWE) remains a hot topic in the world of data analysis. Despite advancement of guidance documents on RWE from regulatory bodies, the RWE programming community relies on the experience from precedents and pilots for best practices. Registries are emerging as a new gold standard for real world research; however, real-world data presents significant challenges in reporting and analysis. Registry programmers often work in non-standardized settings with diverse datasets. Global registries bring humongous unexpected missing data making data completeness checks essential before proceeding with any analysis. Rare disease registries bring large amount of genetic testing data which is still a niche area for majority of the programmers. Facing these challenges with no survival guide in hand feels like driving a 1976 Impala on 2025 roads | missing value, adam dataset, completeness, raw data, cdisc | source presentation source |
4 |
| 83 | From Specification to CSR -- Toward a Unified Analysis and Review Platform | Steve Ross | Over years filled with increasingly complex clinical trials, 'bigger data', and increased regulatory burden, the requirements associated with TLF development and review have similarly ballooned into a collection of disconnected, redundant, and inefficient processes. The working space between TLF shell design and final displays is a distinct group of highly specific tasks requiring a high level of communication and collaboration to achieve a high-quality and error-free product. The current 'Gold Standard' process -- double programming plus extensive manual review, is ripe for reimagination. In today's IT landscape, unified SaaS platforms enhance efficiency, reduce manual effort, and streamline processes | clinical trial, independent qc, data cleaning | source presentation source |
7 |
| 84 | From Topline to Triumph: Visualising the Pathways to Regulatory Approval | Jasperlynn Kao | Achieving positive topline results in a clinical trial marks a critical milestone in the drug development process, yet it is far from the end of the submission journey. Instead, it often signals the beginning of an intricate and time-sensitive submission process aimed at securing regulatory approval. This regulatory defense stage demands rigorous collaboration, meticulous planning, and adaptability to meet the expectations of regulatory agencies. Three different timeline scenarios will be presented starting with positive topline results until the regulatory agency's final decisions to approve the compound. We will compare the milestone dates as well as ad hoc or unexpected requests that may occur from the topline onward and present them visually on a time scale | clinical trial, clinical research | source poster source |
5 |
| 85 | Future-Proofing Clinical Data Standards: Preparing for AI and Machine Learning | Shivoy Pandita | As Artificial Intelligence (AI) and Machine Learning (ML) become increasingly available to clinical research, future-proofing clinical data standards is essential to maximize their potential. This paper highlights some of the possible ways to modifying current data standard governance models, promoting a forward-looking strategy that embraces technological advancements and ensures compliance with regulatory requirements. The Clinical Data Interchange Standards Consortium (CDISC) provides structured data and metadata standards essential for regulatory compliance and interoperability in clinical research. Traditionally, CDISC rules are defined in static formats such as EXCEL, limiting their integration with modern clinical systems and AI applications |
data cleaning, data validation rules, data standard, excel, clinical data, clinical research, clinical trial, machine learning, cdisc | source presentation source |
8 |
| 86 | Gen AI in People Leadership: Unlocking Engagement and Inclusivity | Rinki Jajoo, Shaveta Bansal |
resource allocation, artificial intelligence | source presentation source |
8 | |
| 87 | Gen AI-Powered Trial Performance Co-Pilot to Accelerate Clinical Trials | Venkata Sureshbabu Bondu, Mohit Sehgal |
clinical trial, clinical research, clinical data | source presentation source |
4 | |
| 88 | Gen-AI for Statistical Reporting and Clinical Study Report (CSR) Writing | Leveraging CDISC | Imagine a future where writing Clinical Study Reports (CSRs) is faster, more accurate, and nearly effortless - this is the promise of our Gen-AI-driven model leveraging CDISC Analysis Results Standard (ARS). In this paper, we present a model fine-tuned on ARS and Analysis Results Data (ARD) metadata that streamlines the interpretation of analysis results and generates narrative summaries for the CSR's safety and efficacy sections. By processing pre-structured ARD datasets, the model identifies key findings and produces detailed summaries, significantly reducing manual effort. This approach enhances efficiency, saving time for biostatisticians, clinicians, and medical writers, while ensuring accuracy and consistency across Tables, Figures and Listings (TFL) outputs and CSR text. We also discuss the pros and cons of building this solution, including scalability, regulatory alignment, and data privacy challenges | clinical trial, clinical data, proof of concept, data science, data cleaning, adam dataset, regulatory submission, csr, privacy, subject-level data, adverse events, cdisc | source presentation source |
10 |
| 89 | Generate Clinical Study Reports (CSRs) Using {quarto} and {shiny} | Peng Zhang, Jimmy Chen |
In the pharmaceutical industry, TFLs can be generated through programming to support final clinical study report. However, by referring to the TFL results, medical writers need to manually transcribe or copy/paste the outputs into CSR. Such an entire process can consume several months depending on the complexity of the study and regulatory requirements. To overcome such difficulty, we would suggest two alternatives by using {quarto} or {shiny}, which is inclusive to derive description, narratives, inline tables, figures, etc., to derive CSR, through template or interactively. In this paper, we aim to address the availability of open-source solutions that can help the clinical reporting process. Similar use cases and examples will be provided as discussion | data cleaning, csr, clinical trial | source presentation source |
9 |
| 90 | GPT-Powered Data Insights: Advancing Clinical Data Analytics with Large Language Models (LLMs) | Melanie Hullings | clinical data, clinical research, data science, data cleaning, clinical trial, adverse events | source presentation source |
5 | |
| 91 | Harmonising Heterogeneity: Orchestrating Innovation Through Interdisciplinary Collaboration and Strategic Alignment | Dorothy Dlamini, Srinivas Veeragoni |
clinical research, best estimate liability | source presentation source |
5 | |
| 92 | Harnessing Analysis Results Datasets (ARDs) for Clinical Reporting in R: Our First ARD-Based Filing Experience | Daniel Sjoberg, Nolan Steed |
We present a pioneering application of the Analysis Results Dataset (ARD) framework for preparing a clinical trial readout. By leveraging open-source | clinical trial, data science, csr | source | 6 |
| 93 | Harnessing the Power of R Shiny in a GxP-Compliant and Validated Manner for Clinical Trial Use | Sanyogeeta Andhale, Shilpa Sood |
R Shiny is being increasingly used for interactive data visualization to create intuitive and powerful dashboards. We will see how cross-functional teams can communicate in real-time and users can dynamically design visualizations based on input changes an d underlying data updates. Additionally, for clinical trials data insights these R Shiny apps are developed and published on a GxP compliant platform that supports version control, audit trails, and tracking of datasets and application changes for maintaining data integrity and analytical reproducibility. Access control mechanisms ensure data visualization security by assigning role-based permissions within repositories, safeguarding sensitive information while promoting efficient data -driven decision -making processes. From simple visualizations to complex interactive dashboards, this flexible approach accommodates diverse data presentation needs | clinical trial, data visualization, medical history, adverse events, concomitant medications, clinical data, regulatory submission, 21 cfr part 11, clinical research, data science, audit trail, access control, protocol deviations | source presentation source |
13 |
| 94 | Health Authority Interactions and Communication During the Submission Process | Ritesh Dhimmar, Raman Bassi |
clinical research, data science | source presentation source |
5 | |
| 95 | Health Equity Explorer: Technical Deep Dive into an R Shiny Application for Revealing Structural Inequity and Systemic Racism in Healthcare | Pavel Demin, William G. Adams |
machine learning, proc sql, data cleaning, sql server | source presentation source |
7 | |
| 96 | How Unique is it for the Subject Identifier in a Study To Have Multiple Screenings and Enrolments? | Anahita Vyas, Rama Kudaravalli |
In clinical studies, the process of mapping Subject Identifiers (SUBJID) and Unique Subject Identifiers (USUBJID) within Study Data Tabulation Model (SDTM) datasets poses significant challenges, particularly when subjects undergo multiple screenings across one or more clinical tri als. These challenges are primarily related to maintaining data integrity, ensuring accurate subject tracking, and adhering to the regulatory standards established by authorities such as the FDA. By addressing these challenges through systematic processes and robust data management practices, sponsors can significantly enhance the integrity of their clinical data submissions and improve compliance with regulatory standards. When subjects participate in multiple studies or are re-enrolled within the same study, the USUBJID is designed to uniquely identify a subject across all studies and submissions, preventing data duplication and ensuring consistency | clinical trial, data governance, data integrity, sdtm domain, crf, randomization, clinical research, data science, data cleaning, adverse events, cdisc | source presentation source |
8 |
| 97 | Implementation of a risk-governed, reproducible, and collaborative R workflow for Real World Evidence Projects | Ting Shi | The significance of Real-World Evidence studies has grown considerably, particularly in their implications for healthcare decision-making and regulatory submissions. Within this context, the reliability of formal Analysis and Reporting deliverables is critical, with a strong emphasis on ensuring traceability, reproducibility, and quality. Nonetheless, the open-source nature of R packages presents inherent challenges, as their development by a diverse community of contributors can lead to variability in validation and overall quality. This inconsistency raises critical concerns about the suitability of these tools in regulated environments, where adherence to stringent quality standards is non-negotiable. The U.S. Food and Drug Administration (FDA) underscores that any computer software employed for data management and statistical analysis must demonstrate reliability and have appropriate documentation of software testing procedures | data cleaning, regulatory submission | source poster source |
8 |
| 98 | Improvement on Sankey Diagram SAS Macro and Implementation | Alvin Mathew | Sankey diagrams have garnered attention and application in the pharmaceutical industry as a tool in visualizing patient flow in longitudinal data. Generation of Sankey diagrams through a SAS macro was published in 2015. A modification allowing incomplete data was made in 2016. While the existing macro shows the overall patient flow, the important information of how many patients out-flow to and in-flow from categories (e.g., severity) across visits is missing. We made three major improvements to the existing macro: 1) provide the option to display the number and percentage of patients out-flowing to and/or in-flowing from categories across visits, 2) add color customization of categories, and 3) generate a summary table with detailed information pairing with the Sankey plot. These modifications improve the existing SAS macro not only in visualizing the overall patient flow but also providing insight into important details that medical researchers seek | pharmaceutical industry, macro definition | source presentation source |
8 |
| 99 | Innovative Cloud Solution for Advanced Image Analysis | Gopal Joshi, Sanjay Koshatwar |
The rapid growth of digital imaging in diagnosis of health care has resulted in a dramatic increase in the volume and complexity of image data that is generated by various radiological instruments and microscopic devices [1]. Traditional methods of storage and processing are becoming inadequate to manage the scale and speed of this data [2] Cloud computing, with its scalability, flexibility, and cost-effectiveness, presents a promising solution to these challenges. This paper examines the innovative uses of cloud computing for hosting and managing imaging data. We explore how cloud-based object storage services can efficiently handle large-scale image datasets, enabling effective storage and retrieval. Additionally, we address the role of data transfer optimization techniques in reducing latency and maximizing throughput | cloud storage, amazon s3, access control, machine learning, data governance | source presentation source |
9 |
| 100 | Innovative Data Cleaning to Prevent Unblinding Errors in Double-Blind Trials | Ling Zhang, Nan Zhou |
clinical trial, clinical database, data integrity, data governance, data cleaning, clinical data | source presentation source |
4 | |
| 101 | Innovative Storage of SDTM Mappings as JSON: Boosting Efficiency and Data Consistency | Rohan Parmar, Edgar Manukyan |
sdtm programming, proc sql, sdtm domain, data integrity, clinical data, data governance, data cleaning, clinical trial, cdisc | source presentation source |
5 | |
| 102 | Integrating Gen AI with an Open-Source Toolkit for Clinical Insight Generation to Boost Data Science Efficiency | Vincent Shen | data science, production run | source presentation source |
3 | |
| 103 | Integration of AI with R Studio: Advantages and Disadvantages | Mahesh Minnakanti, Karthikrao Perala |
Learning and writing code for any new programming language is a crucial yet daunting task characterized by its complexity, demanding significant time and effort from programmers. But with the recent AI evolution, it has become easy to understand and acquire code, at the same time it can be challenging to shuffle between AI and coding platforms. Having AI integrated with the coding platforms can make it more convenient and efficient. R can be more flexible for AI integration as it is an open-source tool. In this paper we will show ways to integrate AI with R studio, its usage for quick learning and efficiency, along with its advantages and disadvantages | data visualization, data science | source presentation source |
8 |
| 104 | Intelligent Automation in Anonymisation for CTIS Submission | Pinku Hooda, Hiteshkumar Kanaiyalal Raval |
The automation of redaction processes across languages presents a formidable challenge due to linguistic variability, cultural nuances, translation accuracy, and character encoding differences. This paper define s the current process of redaction of other language documents for CTIS submission, challenges with current process, the complexities involved in automating redaction using redacted English documents as a reference point and proposes the best solution to automate the process fully using technology. Challenges such as language variability and cultural nuances require sophisticated Natural Language Processing (NLP) algorithms capable of understanding the grammatical structures and cultural sensitivities of different languages. Machin e translation systems play a crucial role in ensuring accurate translation of redacted English documents while preserving context and meaning | machine learning, clinical trial, privacy, data protection | source presentation source |
11 |
| 105 | Interactive Dashboards - Empowering Clinicians to Generate Reports Instantly | Inderjeet Arora | adverse events, clinical trial, predictive analytics, machine learning, vital signs, clinical research, clinical data, excel | source presentation source |
8 | |
| 106 | Interactive Visualisation of Pooled Clinical Trial Safety Data | Sara Bodach, Ling Zhang |
adverse events, clinical trial, data review, adam dataset, regulatory submission, cdisc | source presentation source |
7 | |
| 107 | Introducing TFL Viewer: A Centralised Platform for TFL Review, Collaboration and Approval | Bhavin Busa | Why go through hundreds of RTF and PDF files when you can manage, review, and approve tables, figures, and listings (TFLs) on a unified web-based platform? In this paper, we will introduce TFL Viewer, highlighting features like version control, easy navigation, and comment tracking to simplify the review process. Currently, feedback from reviewers is often disorganized, with comments added to individual Word documents or manually collated into EXCEL, making it difficult to track and incorporate feedback. TFL Viewer centralizes this process, allowing teams to collaborate more effectively by consolidating comments in one place and streamlining review cycles. The tool also allows users to publish selected TFLs into a single report for submission to the Clinical Study Report (CSR), format TFLs for in-text CSR integration, and more. Add-on functionality includes AI-powered features to further streamline the generation of statistical reports and sections of the CSR |
clinical trial, data integrity | source presentation source |
5 |
| 108 | Key Metrics for Success: Measuring Project Performance and Impact in the Face of Complex Clinical Trials | Kavya Yachamaneni, Hayley MacKenzie |
Clinical trials are a long and arduous process, requiring meticulous planning and execution to ensure success. Their timelines are deliberately long to ensure long-term impacts are identified and mitigated. However, with patients eagerly awaiting positive outcomes, it is vital that the trials progress with highest compliance, efficiency, and precision through such lengthy timelines. Clinical Project Management acts as a pillar to these principles, ensuring that clinical trials progress toward positive patient outcomes. Clinical Project Management teams approach each project through four critical lenses: governance, operational, quality, and financial. They develop robust processes to operate each approach and employ unambiguous key metrics. This comprehensive approach allows teams to maintain stringent oversight, optimize operational workflows, uphold high quality standards, and manage resources efficiently, all of which are crucial for the successful execution of clinical trials | clinical trial, resource allocation, data integrity, protocol deviations | source presentation source |
9 |
| 109 | Kraken: A Shiny Application for Visualising SDTM Data in Early-Phase Clinical Trials | Rutu Patel | Kraken is a cutting-edge Shiny application designed to revolutionize how clinicians and statisticians visualize SDTM data in early-phase portfolio studies. The application provides a user-friendly interface for generating summary tables and graphs from SDTM data, that help clinicians explore and understand investigational drug performance and safety profiles with the click of a button. This capability allows for faster decision-making and supports the innovative exploration of trial data. Built on R's Shiny framework, kraken's visualizations are clickable, customizable, downloadable, and come with supporting listings for each analysis. As an essential tool for early-phase studies, Kraken promotes efficiency and progress in pharmaceutical research, empowering researchers to confidently make data-driven decisions. This paper explores the capabilities, methodologies, and challenges in the context of early-phase clinical research | clinical data, clinical trial, disposition table, ods rtf, swimmer plot, excel, data science, sdtm domain, clinical research, cdisc | source presentation source |
12 |
| 110 | Leadership Presentations: Adapting to the Future: How CROs are Evolving to Meet the Shifting Landscape of Programming and Analysis Generation | Lavanya Peddibhotla | Presentation | clinical data, clinical research | source | 10 |
| 111 | Leadership Presentations: Future Leadership Requires Being Uncomfortable with Change | Chris Decker | Presentation | source | 12 | |
| 112 | Leadership Presentations: How and Why, Statistical Programming Leaders Need to Plan for Next Generation Statistical Programming | Francis Kendall | Presentation | data science, clinical programming, statistical programmer, clinical research, machine learning, data wrangling, data governance | source | 29 |
| 113 | Leadership Presentations: Turning Disruption into Opportunity - Guiding Teams Through AI-Powered Transformation | Li Zheng | Presentation | clinical programming, data science, clinical trial, clinical research, statistical programmer, data cleaning, outlier | source | 20 |
| 114 | Learning by Doing: Empowering Professional Growth Through Collaborative Python Automation | Azadeh Gilanpour, Vanessa Chavez |
The Nonclinical Topics (NT) working group is increasingly recognizing the value of data science using the Standard for Exchange of Nonclinical Data (SEND) model. As organizations adopt data science initiatives, SEND experts may lack the programming skills nece ssary to meet industry needs. This paper provides a framework for nonclinical data managers to obtain programming skills and develop a relevant Python programming program within their organization. Participants will engage in a SEND -relevant learning proje ct, fostering collaboration and creating a supportive community among colleagues with varying skill levels. By the end, participants will have not only developed valuable Python skills but also critical competencies such as time management, teamwork, and l eadership to contribute to an innovative work environment. The benefits, challenges, and the overall results of this project from the first training cohort are discussed | data science, data visualization | source presentation source |
11 |
| 115 | Learning to skate, pedal, and drive - Change leadership in the world of Agile delivery | Eileen Ching | Many organizations choose to adopt Agile framework to implement new systems and tools because it fosters innovation, enhances collaboration, and allows for faster delivery of high-quality products. However, not only do project teams face challenges in making the switch from the traditional Waterfall method to Agile, but it also takes a fundamental mindset shift for users to anticipate, prepare for, and adopt iterative deliveries of products along with process changes that may require ongoing learning of new skills. The presentation delves into the strategies and methodologies that facilitate user acceptance in an Agile environment. By leveraging continuous feedback loops, promoting a culture of adaptability, and ensuring transparent communication, organizations can foster a sense of ownership and engagement among users | key performance indicator, adam dataset | source presentation source |
9 |
| 116 | Learnings from a Multistakeholder Estimands Project: A White Paper Journey | Matt Baldwin, Banoo Madhanagopal |
Presentation | treatment effect | source | 18 |
| 117 | Let the Machine Do the Coding: A SAS-Based Framework for Automating TFL Generation with ARS Metadata | Julie Wachara | In today's fast-paced clinical trial environment, efficiency and adaptability are key to meeting programming timelines. This paper presents an approach to automating study programming tasks by leveraging TFL shells and CDISC ARS metadata. By automating much of the programming, this framework reduces manual effort and ensures consistency across deliverables. This SAS-based solution generates code templates for all required TFLs, pre-populating key sections with accurate, study-specific information. Its metadata-driven automation accommodates frequent TFL additions and modifications, offering flexibility during early study stages. For project managers and lead programmers, this approach optimizes oversight, boosts productivity, and positions automation at the core of the programming process. Let the machine do the programming for you | mock shell, macro variable, macro definition, cdisc | source presentation source |
9 |
| 118 | Let's Talk Future: Automated Solution for Standard-Compliant Data, Mock Shells and TFL Generation Using Open-Source Platforms with AI | Inderjeet Arora, Abhijeet Arora |
The traditional approach to standard-compliant data transformation, such as CDISC-SDTM and ADaM, as well as generating mock shells and Tables, Figures, and Listings (TFLs), heavily relies on manual, rule-based processes. This method is time-consuming, prone to errors, and demands specialized technical expertise. In contrast, automated solutions using open-source platforms like R and Python streamline the generation of standard data, mock shells, and TFLs. The automated process offers consistency, faster production, and reproducibility across studies, including an audit trail that enhances efficiency. Additionally, by incorporating artificial intelligence (AI) with a Human-in-the- Loop (HIL) approach, data transformation processes and TFL generation are further optimized, providing a more efficient, automated solution | mock shell, clinical trial, data transformation, clinical data, data cleaning, raw data, sdtm domain, clinical programming, audit trail, sdtm programming, data science, cdisc | source presentation source |
8 |
| 119 | Leveraging HASH Technologies for Enhanced Clinical Trial Programming Efficiency | Dmitrii Fedorychev, Raviteja Nannapaneni |
In the ever-evolving landscape of clinical trial programming, the integration of HASH technologies emerges as a pivotal solution. This presentation explores the potential of integrating HASH technologies into clinical trial programming workflows, to expedite processes and alleviate server load. Particularly beneficial for studies with extensive databases and multifaceted integration projects spanning a multitude of trials, HASH technologies offer a promising avenue for optimizing performance. Utilizing the capabilities of HASH opens up avenues for streamlining tasks and optimizing efficiency in clinical trial programming. With this presentation we will delve into the diverse array of tools available to capitalize on HASH technologies, offering insights into how its adoption can revolutionize efficiency and scalability within the clinical trial ecosystem | hash table, clinical trial, proc step | source presentation source |
5 |
| 120 | Leveraging NLP for Clinical Trial Efficiency: Development of a Protocol-Aware AI Chatbot | Seamus Gallivan, Sairam Tangirala |
Jazz Pharmaceutical PLC's technological transformation is modernizing its operations by integrating AI/ML tools to streamline clinical trial workflows. Its Integrated Data Analytics and Statistical Programming (IDASP) division has developed a proof -of-concept (PoC) AI -powered chat agent to improve clinical protocol document management by enabling natural language interactions. This chat agent leverages Retrieval-Augmented Generation (RAG) techniques to enhance information retrieval while mitigating LLM hallucinations, ensuring accuracy, compliance, and efficiency. This GenAI tool was tested for its performance using a structured questions-based assessment tool, for evaluating its effectiveness in providing contextually accurate responses. The PoC demonstrates the potential for AI-driven solutions in the pharmaceutical industry while highlighting the need for continuous refinement, regulatory oversight, and human- in-the-loop mechanisms | clinical trial, data science | source presentation source |
8 |
| 121 | Leveraging Vector Databases to Enhance AI Applications in Pharmaceutical Research | Angu Sundaresh Krishna Kumar, Ramanathan Srinivasan |
Vector databases are emerging as a crucial component in advancing AI and machine learning applications, particularly in the pharmaceutical industry. This presentation explores how vector databases can significantly improve the efficiency and accuracy of large language models. We'll examine a case study where a vector database was implemented to enhance the search and retrieval of complex clinical trial data, leading to accelerated clinical trials. The talk will address the technical challenges of integrating vector databases with existing | clinical trial, clinical data, stage 2, unstructured data | source presentation source |
11 |
| 122 | Life Cycle for the Management of ODS Regulatory Referenced Deliverables | Dipen Kachhiapatel | Poster | source | 1 | |
| 123 | LLMs in Clinical Trials: Overcoming Challenges Related to Patient Confidentiality and Intellectual Property | Amit Javkhedkar, Samip Shah |
In the rapidly evolving landscape of clinical trials, Large Language Models (LLMs) will play a pivotal role in radically changing the way clinical trial data is collected, transformed, analyzed and submitted to regulatory bodies. These Large Language Models can be designed to understand, interpret, and generate human-like text, thereby simplifying and speeding up the process leading up to submissions. Although LLMs offer promising solutions and show exciting prospects, there are some inherent challenges. Key among those include setting up sophisticated technical infrastructure for LLMs and ensuring data privacy, particularly when training with Personal Health Information (PHI) and organization-specific standards. This paper discusses how to potentially apply LLMs to visionary use cases that will empower the industry and possibly revolutionize the various aspects of the drug discovery process | clinical trial, clinical data, statistical programmer, clinical programming, data science, privacy, mock shell, training set, clinical research, excel, access control | source presentation source |
7 |
| 124 | LLMs in Drug Development - The Future is Closer Than We | Think Chris McSpiritt | clinical research, clinical trial, neural network, life sciences, csr, clinical trial protocol, data cleaning, adverse events, deep learning, pharmaceutical industry | source presentation source |
9 | |
| 125 | Machine Learning-Enabled Personalised Medicine for Rapidly Evolving Infectious Diseases | Jonathon Schisler, Scott McClain |
Patients hospitalized with infection-related respiratory symptoms urgently require personalized therapies. Standardized clinical practices often result in varying effectiveness across diverse populations, whereas personalized medicine offers a tailored approach to address heterogeneous diseases. Despite extensive molecular and clinical data availability, a significant obstacle in precision medicine is the lack of real-time analytical tools to predict individualized treatment efficacy for specific cohorts. Our study focuses on patients admitted to intensive care units for respiratory failure treatment during the initial three years of the COVID-19 pandemic. Previously, we identified distinct molecular endotypes by analyzing omics and cytokine data in deep lung tissue and peripheral blood. In this presentation, we leveraged machine learning methodologies to integrate electronic health records (EHR) data with molecular insights, enhancing predictive capabilities | machine learning, sample | source presentation source |
8 |
| 126 | Management of Regulatory Submissions Using eCTD Sequences and Life Cycle Operators: Lessons from Real-Time Oncology Review (RTOR) and Regular Submissions | Madhusudhan Reddy Papasani, Elisabeth Pyle, Mary N. Varughese |
The eCTD format has become the standard for regulatory submissions by the FDA, resulting in a streamlined and efficient review process. This paper aims to provide practical examples of effectively utilizing eCTD sequences and lifecycle operators in two scenarios: Real -Time Oncology Review (RTOR) submissions, including rolling submissions, and regular submissions requiring updates. The eCTD structure organizes submission content into five modules (m1 - m5), with leaf elements representing individual files or documents within these modules. Each submission is assigned a unique sequence number at a specific time point , which encompasses a collection of modules and leaf elements. Throughout the submission lifecycle, l ifecycle operat ors, like 'new,' 'replace,' 'delete,' and 'append' facilitate the management of file status and updates | adam dataset, reviewer guide, cdisc | source presentation source |
8 |
| 127 | Mapping OMAP Data to SDTM - Use Case for RWD in CAR-T Registry Studies | Kai Deng, Joe Xi |
This paper shares our experience in mapping RWD (Real World Data) in OMOP (Observational Medical Outcomes Partnership) to CDISC SDTM-like model. Below perspectives are included: How our OMOP data (CAR-T specific) are organized How we convert and remap these OMOP data into SDTM-like domains * Decoding, traceability and data masking issues At the end, we summarize the advantage of using OMOP as potential RWD collection model for pharmaceutical industry | sample, pharmaceutical industry, raw data, adverse events, data standard, cdisc | source presentation source |
5 |
| 128 | Measure Total Steroid Dose Using Real-World Data | Yi-Chien Lee | Real-world data provides valuable information for outcomes research. However, the data can be messy and there is no source document to verify or correct the data points. This poster illustrates how the real-world data is cleaned and handled. The goal of the analyses is to measure total steroid dose for each patient over a 90-day period. Days of supply, Quantity, and Strength are used to calculate the daily dose for each drug record. Data cleaning was done to handle missing or outliers. More importantly, it's critical to handle the overlapped prescriptions as they can be used sequentially or concurrently. Input from clinical experts such as pharmacists are valuable to make assumptions and decisions. Multiple overlapping scenarios, such as (1) same drug, same daily dose, same route vs. (2) different medication or same medication but with different daily dose, and the corresponding actions are detailed in this poster | data validation rules, insurance claims | source poster source |
2 |
| 129 | Metaprogramming in R | Sudhir Kedare, Rajprakash Chennamaneni |
Metaprogramming techniques facilitate the generation, modification, and reuse of programming syntax through programmatic statements. Although metaprogramming has its roots in early languages like Lisp, its application in R offers robust capabilities for code manipulation. This paper explores R's distinct metaprogramming features, focusing on fundamental concepts such as defusing and injecting expressions. The analysis demonstrates how these R- specific metaprogramming approaches enhance programming efficiency and code reusability. Through a practical implementation involving dynamic table generation, this paper illustrates how R's metaprogramming capabilities can solve complex programming challenges in data analysis and reporting | proc means, clinical trial, data science | source presentation source |
5 |
| 130 | More Than a p-Value: Understanding the Differences Between Clinical and Statistical Significance | Jasperlynn Kao, Elnaz Ghadimi |
It is possible for a result to be statistically significant but not clinically significant. Conversely, a result can be clinically significant but not statistically significant. Both statistical and clinical significance are crucial for a comprehensive understanding of research findings. Statistical significance helps determine whether an effect exists, while clinical significance evaluates whether the effect is meaningful in a practical sense. To illustrate these concepts, we will explore two real-world examples: one demonstrating a clinically significant finding and another highlighting a statistically significant result, to better understand how these measures impact research and practice | clinical significance, treatment effect, confidence interval, effect size, null hypothesis, sample size, sample | source presentation source |
4 |
| 131 | Multiple Imputation in Composite Endpoints - It's Not That Hard! | Ashley Mao | Missing data in clinical trials is a common issue that can potentially introduce bias and imprecisions in results interpretation. Various approaches exist to address missing data, among which Multiple Imputation (MI) has become the preferred method due to its ability to provide more reliable and valid statistical inferences under realistic assumptions of the missingness mechanism. MI takes into account the uncertainty associated with missing data by creating multiple filled-in datasets, thus yielding more precise estimates of standard errors and confidence intervals. While MI is effective in missing data imputation for continuous or binary endpoints, its application in composite endpoints is not always straightforward. Careful consideration is required in model and variable selection, pooling strategy and result interpretation. In this presentation, we will illustrate in a case study how we apply MI techniques to handle the missing data in composite endpoint | imputation, missing value, multiple imputation, proc calis, inferential statistics, treatment arm, regression analysis | source presentation source |
6 |
| 132 | Natural Language to Code: AI-Powered Automation for CDISC-Compliant Datasets | Vina Ro | This paper presents an innovative approach to automating the creation of Clinical Data Interchange Standards Consortium (CDISC)-compliant datasets, which traditionally demand extensive manual effort and advanced programming skills. By leveraging advanced generative AI technologies, this methodology transforms natural language coding specifications for SDTM and ADaM datasets into R code. The process employs a two -step workflow: First, a Retrieval-Augmented Generation (RAG) framework enhances the precision and logic of the coding specifications using a Large Language Model (LLM) and prompt engineering. Second, the LLM generates highly accurate R code from the enhanced specifications. Testing on real -world studies demonstrated 88% accuracy in variable generation, with most errors stemming from data or human factors, not the AI system | sdtm domain, clinical trial, cdisc | source presentation source |
5 |
| 133 | Navigating the Complexities of AESI: A Centralised Approach to Molecule-Level Management | Tian Fang, Xiaojing Zhu |
Adverse Event of Special Interest (AESI) is an event, whether serious or non-serious, of scientific and medical concern specific to the sponsor's product or program, for which continuous monitoring and prompt communication by the investigator to the sponsor can be appropriate (International Council for Harmonisation [ICH], 2022). During the active development of an investigational medical product in clinical trials, the report of AESI supports the safety assessment and fulfills health authority requirements in regulatory documents. The search strategy of AESI changes over time with accumulated knowledge from evolving data. The implementation of AESI and its version control become challenging particularly in a large project across multiple indications with many ongoing trials. Support from cross-functional teams in a single study as well as the development team at indication or molecule level further adds complexity to the alignment and maintenance of AESI definitions | preferred term, data science | source presentation source |
7 |
| 134 | Navigating the Transition to a New Statistical Computing Environment: Streamlining Compliance and Operational Excellence | Shilpa Sood, Kala Shivalingaiah |
The introduction of a new, GxP-compliant environment for statistical programming and data sciences marks a significant advancement in efficiency and compliance within organizations. This paper examines how well-defined use cases, aligned with customer SOPs, can facilitate a smooth transition to the new system. Emphasizing a short learning curve, streamlined training, and accelerated adoption, we discuss strategies to ensure that teams quickly embrace the new system, achieving both operational excellence and regulatory compliance | data cleaning, clinical programming, 21 cfr part 11, audit trail, data integrity, access control, data visualization, data migration, data wrangling | source poster source |
11 |
| 135 | Neurodiverse Teams: Fostering Innovation and Inclusion | Laura Needleman | source presentation source |
8 | ||
| 136 | Next-Generation Deliverable Tracker | Michael Byrne, Stuart Malcolm |
Deliverable Trackers are a critical tool in biometrics workflow and are used for everything from project planning, resource management, through to QC and validation tracking, and as a quality audit record. The standard approach is to use Microsoft EXCEL- this offers a well understood user interface; however it relies on manual entry that is often out of date, and data is duplicated in multiple places with limited ability to link and synchronize data from other systems. This presentation demonstrates a new approach to implementing a deliverable tracker that uses an interactive user- interface (R-Shiny) backed with a database populated by harvesting metrics directly from the SCE and from other systems via API calls |
clinical trial, proc freq | source presentation source |
5 |
| 137 | OpenRBQM Extended: Endpoint Monitoring, Root Cause Analysis and Workflows | Spencer Childress, Jeremy Wildfire |
OpenRBQM is a PHUSE-sponsored collaborative effort to create a toolbox of open-source R packages focused on risk-based quality management (RBQM). In the past year the OpenRBQM framework has been refactored and extended to support many new use cases, including endpoint monitoring and root-cause analysis. This presentation describes the process of creating new tools that are compatible with OpenRBQM and introduces several new R packages using this framework | adverse events, clinical trial | source presentation source |
4 |
| 138 | Optimising AI Software for Automated Protection of Personal and Commercially Sensitive Data | Honz Slipka | Ensuring clinical data transparency while protecting sensitive patient information has become a major challenge due to the vast volume and variability of data in pharmaceutical research. Traditional anonymization methods often lack consistency and efficiency, increasing the risk of data breaches and regulatory non-compliance. To address these issues, integrating artificial intelligence (AI) and machine learning (ML) offers a transformative solution by automating anonymization processes, improving accuracy, and reducing human error. The SMART approach developed by Certara, along with the adoption of the hybrid methodology, promotes proactive data protection by embedding anonymization strategies early in drug development and leveraging AI automation. By doing this, the pharmaceutical industry can achieve secure data sharing while safeguarding patient privacy, ensuring compliance, and mitigating financial and reputational risks | medical history, adverse events, clinical data, concomitant medications, data protection, pharmaceutical industry, regulatory submission, clinical research | source poster source |
7 |
| 139 | Optimising Collaboration and Quality Control in Outsourced Statistical Programming | Mukesh Patel, Nilesh Patel |
clinical programming, clinical trial, quality control, key performance indicator, regulatory submission, data cleaning | source presentation source |
7 | |
| 140 | Optimising Leadership Strategies for Global Teams Through Time Zone-Based Task Assignment | Tigran Alikhanyan, Erik Torosyan |
clinical trial, data science | source presentation source |
7 | |
| 141 | Optimising Regulatory Submissions: The Role of eCTD, Real-World Evidence and AI | Kishore Pothuri | Recent developments in regulatory submission components have introduced new strategies and technologies aimed at enhancing the efficiency and quality of submissions. The FDA and EMA have placed greater emphasis on the use of electronic Common Technical Documents (eCTD) with updated modules for more streamlined submission processes. This includes the integration of real-world evidence (RWE) and patient-reported outcomes (PROs) to support the regulatory approval of new drugs and therapies. Additionally, advancements in AI and machine learning are being leveraged to automate data processing and analysis, ensuring more accurate and timely submissions. The use of adaptive designs and novel statistical methodologies is also becoming more prominent, reflecting the industry's shift towards more flexible and patient-centric approaches in clinical trials | regulatory submission, adaptive design | source presentation source |
5 |
| 142 | Optimising SDTM and ADaM: Remediation Techniques for Data Accuracy | Kishore Pothuri | Remediation of SDTM (Study Data Tabulation Model) and ADaM (Analysis Data Model) datasets using legacy studies is a critical process in ensuring regulatory compliance and data consistency in clinical trials. Legacy studies often predate current CDISC (Clinical Data Interchange Standards Consortium) standards, necessitating the transformation and remediation of historical data to align with modern SDTM and ADaM formats. This process involves mapping legacy data structures to CDISC-compliant formats, addressing any discrepancies, and ensuring that the datasets are complete, accurate, and ready for submission. The remediation process enhances the reusability of legacy data, supports integrated analyses across multiple studies, and facilitates more efficient regulatory reviews. However, it presents challenges, including the need for deep domain expertise, extensive quality checks, and careful handling of data nuances | data cleaning, adam dataset, data science, controlled terminology, regulatory submission, sdtm domain, clinical trial, liquidity gap, adverse events, cdisc | source presentation source |
10 |
| 143 | Optimizing Development and Regulatory Submission of eSUB Package for Exposure-Response and PK/PD Analysis | Saumilkumar Tripathi, Ephicacy Lifescience |
Exposure-Response (E-R) and PK/PD models and simulations are powerful tools utilized throughout the drug lifecycle, from early research and development to post -marketing surveillance. These models play an important role for dose optimization, assessing the relationship between drug exposure and therapeutic response, as well as adverse events. E-R and PKPD analysis are integral to regulatory submissions, providing critical evidence that supports the safety, efficacy and effects of various population and co-variates on drug. Creating electronic submission(eSUB) package for this analysis is slightly different from the analysis data package we prepare for CSR submission and requires meticulous planning, robust data management, strict adherence to regulatory standards and efficient collaboration between programmer and pharmacometrician | adam dataset, regulatory submission, in=, population pk, clinical research, key, pd, clinical trial, exposure, bds, data cleaning, data science, cdisc | source presentation source |
10 |
| 144 | PHUSE in Japan: Past, Present and Future | Koichi Yamaguchi | Poster | clinical trial, data cleaning | source | 1 |
| 145 | PHUSE Working Groups 2024 Highlights | Chris Price | Poster | source | 1 | |
| 146 | Pivoting to AI: Turning Career Transition into Innovation | Rick Stephens, Good Gorilla |
After a recent layoff due to a reorganization, I embraced the opportunity to pivot to a field I am passionate about - Artificial Intelligence. This paper highlights several AI-driven projects completed during this transition, showcasing the potential for AI to enhance processes and increase value. Featured projects include: AI-Powered Document Search Chatbot: Leveraging Large Language Models (LLMs) to improve search functionality and document interaction across extensive repositories. Meeting Minute Classifier: Achieving 98% accuracy in identifying potential audit issues through an AI- based classification model. AI in SharePoint Design: Enabling non-technical users to contribute through LLM integration, democratizing content creation. Gamifying Training: Utilizing AI to enhance engagement and retention in training sessions through gamified elements. * AI-Driven Meeting Summaries: Automating the generation of concise and actionable meeting summaries | data science, machine learning | source poster source |
5 |
| 147 | Planning Ahead: A Novel Approach to Creating Synthetic Immunogenicity Data | Ruilan Wei | For the NDA submission, immunogenicity analyses, including ADA and NAb, are required. The testing follows a tiered approach: an initial ADA screening test is conducted, and if positive, it is followed by a confirmatory test. If the confirmatory test is also positive, titer and NAb tests are performed. The immunogenicity endpoints encompass ADA positivity, treatment-unaffected, treatment-emergent, treatment- induced, treatment-boosted, persistent ADA, transient ADA, titer, and NAb. In a blinded clinical trial, the sensitivity of ADA and NAb results prevents access until unblinding, creating delays in pre-programming activities. Often, the test files provided by vendors fall short in accurately capturing the tiered approach necessary to derive the required parameters | sample, clinical trial | source presentation source |
5 |
| 148 | Posit Team: Metadata-Driven Risk Assessment and Surfacing Validated R Packages | Philip Bowsher, Michael Mayer |
Many companies now make increasing use of open-source languages like R and the Posit Team software suite for late-stage clinical development. The Computerized Systems where these late-stage workloads run-such as Statistical Compute Environments (SCE)-are governed by GxP principles that put very strict controls on their use, and as a consequence there is now heightened scrutiny on the software testing and validation of R and R packages. In this paper we outline a technical approach to conducting risk-based validation of infrastructure that includes R and a library of R packages using some of the unique features of Posit Team | data cleaning, clinical research, insurance risk score | source presentation source |
5 |
| 149 | Positron: Overview for Pharma | Philip Bowsher, Cole Arendt |
data science, clinical programming | source presentation source |
18 | |
| 150 | Practical Evaluation of AutoML Platforms for Clinical Data Science | Shubhranshu Dutta | The objective of this paper is to evaluate and analyze AutoML platforms. Platforms available in the industry were analyzed by domain-specific use, data support, automation level, model support, scalability and integration, performance, user interface, ease o f customization, and ease of explainability. These kinds of cross-comparison results help customers choose the right and fit-for-purpose machine learning platforms by use. The paper/poster also emphasizes how new machine learning programmers, both in industry an d academia, can benefit and assess the platforms they are using and suggest improvements and enhancements | machine learning, clinical data, data science | source poster source |
4 |
| 151 | Predictive Analytics for Adverse Event Underreporting | Steven A. Gilbert | adverse events, outlier, regression analysis, histogram, scatter plot, binomial distribution, machine learning, negative binomial regression, normal distribution, poisson distribution, box plot, confidence interval | source presentation source |
10 | |
| 152 | Python's Future in Clinical Trials: Innovations and Collaborative Advancements | Shweta Shukla, Sohan Lal |
In the Field of clinical trials, effective collaboration among study teams is crucial for ensuring the success and integrity of research undertakings. This abstract deep dive into the pivotal role of Python programming language in study team collaboration within clinical trials, particularly through its power in handling PDF files and leveraging AL/ML techniques and python has the ideal package to update Word documents with the data automatically. Python, known for its versatility and robustness, offers plenty of tools and libraries tailored to the needs of clinical researchers. Leveraging Python's capabilities, study teams can streamline data analysis visualization, uncover the data hidden patterns, predict outcomes, optimize the clinical trial protocols, and improve the communication processes among study teams | clinical trial, machine learning, artificial intelligence, adverse events, clinical research, extract, clinical data, proc hpforest | source presentation source |
10 |
| 153 | R Consortium's R-Based Test Submission Package for FDA Evaluation: A Milestone and a New Frontier in R-Based Regulatory Submissions | Joel Laxamana | clinical trial, regulatory submission, clinical data, adam dataset, adverse events, cdisc | source poster source |
7 | |
| 154 | R Package Risk Assessment: Balancing Programmatic and Human Input | Chelsea Dickens | As pharmaceutical companies embrace open-source software like R, there are countless considerations to account for. One of the biggest factors in these considerations is risk - how do we quantify risk in using open-source software and how do we mitigate that risk? Some strategies for determining package risk involve using tools like {riskmetric}, which allow for an automated and programmatic risk assessment and the generation of a numeric risk score based on factors including package type, maintenance, community usage, and testing (R Validation Hub, 2024). However, other risk factors, including robustness of documentation and references, maintainer expertise, and even the existence of critical GitHub issues, are often overlooked in these approaches. In this paper, we will discuss our experiences with both programmatic and human inputs during risk assessment to curate and catalogue a subscription-based repository of validated R packages | data cleaning, pharmaceutical industry, regulatory submission, insurance risk score, r package | source presentation source |
4 |
| 155 | Real-World Data - Analysis-Ready Datasets: A Standardised, Specialised and Innovative Data Solution for Efficient Real-World Data Analysis | Nayan Chaudhary, Navdeep Pal |
clinical trial, raw data, data science, machine learning, data cleaning, clinical research, life sciences | source presentation source |
12 | |
| 156 | Real-World Examples of eCOA Compliance and Audit Trail Review Using Advanced Statistical Methodology | Jamie O'Neill | audit trail, clinical trial, adverse events | source presentation source |
5 | |
| 157 | Redefining FDA Submissions: The Future of Interactive R Shiny Dashboards | André Veríssimo, Vedha Viyash |
clinical trial, adam dataset, cdisc | source presentation source |
5 | |
| 158 | Regulatory Challenges and Strategies for AI/ML-Enabled Medical Device (SaMD) Data Submission | Shashikant Kumar, Mrityunjay Kumar |
With the latest advancement in medical device industry and its regulatory requirement, it has become evident that the industry needs to adapt new approaches of data collection and submission. Recent trends show that new approaches of data collection and submission necessitates the increased use of real-world evidence (RWE), digital health technologies, and artificial intelligence (AI) -driven devices. These innovations often lead to data in non - traditional formats, complicating the submission process across Regulatory Authorities (RA). Key challenges include the lack of regulatory frameworks for AI/ML enabled devices (e.g SaMD Software as Medical Device) and harmonization among regulatory authorities for submission. This presentation will discuss effective strategies to handle software v ersion control, including the adoption of standardized data formats (ISO 13485), analytics tools, and employing new AI/ML methods | feature importance, regulatory submission | source presentation source |
12 |
| 159 | Regulatory Submissions - Exciting, Challenging, or Both? | Ganeshchandra Gupta, Randi McFarland |
data science, adverse events | source presentation source |
5 | |
| 160 | Revolutionising Cancer Trials with Adaptive Designs and Bayesian Analysis | Navneet Jha | Cancer trials are undergoing a transformation with Bayesian adaptive designs, allowing real -time modifications based on accumulating data. This research explores the application of Bayesian adaptive designs in cancer trials, focusing on relapsed/refractory Acute Myeloid Leukemia (AML). Using Kaplan -Meier survival analysis, Bayesian Cox Proportional Hazards models, and Bayesian predictive probability, we evaluate overall survival (OS) and progression- free survival (PFS) as key endpoints. The results indicate that while Kaplan -Meier survival curves show modest improvements in OS and PFS for the experimental treatment, Bayesian predictive probability estimates (~99%) strongly support trial continuation. This highlights the potential of Bayesian adaptive designs to optimize trial efficiency by enabling real-time decision-making | predictive probability, adaptive design, overall survival, survival analysis, sample, progression-free survival, clinical trial, sample size, hazard ratio, proc phreg, posterior probability, density plot, treatment effect, scatter plot, likelihood function, bayesian statistics, monte carlo | source presentation source |
11 |
| 161 | Revolutionising Clinical Trials: Gen AI-Powered Biometrics Breakthroughs | Kevin Lee | The pharmaceutical industry is on the brink of a paradigm shift, leveraging Generative AI (e.g., ChatGPT) to transform clinical trial processes. This presentation explores the integration of Gen AI in biometrics, showcasing its potential to optimize workflows and enhance clinical trials process. The presentation will introduce Gen AI and its applications in biometrics, including code generation, document (e.g. SAP, CSR) generation, data analysis, visualization and patient profile. It'll cover Gen AI tools, systems, and processes that are reshaping Clinical Trials. It'll also examine how biometrics teams incorporate Gen AI to be "super biometrics teams," dramatically improving efficiency and effectiveness. Looking ahead, the presentation will consider the future of Gen AI in clinical trials and its impact | clinical trial, phase iii, data cleaning, data science, adverse events, clinical research, pharmaceutical industry, secondary endpoint | source presentation source |
6 |
| 162 | Revolutionizing Clinical Trials with Intelligent Protocol Automation | Zach Martin | Designing clinical trial protocols and associated documents is time-consuming and labor -intensive, requiring ingest, aggregation and analysis of data from multiple sources. Conventional, manual approaches often result in inconsistency, errors, and delays. This hinders trial effectiveness and efficiency. We harness SAS's Natural Language Processing (NLP) capabilities along with Large Language Models (LLMs) to generate draft clinical protocols in a simple, governed and trusted manner. Low -code interfaces facilitate accessible prompt instructions and output specification using desired templates. Our unified AI platform, SAS Viya, enhances output quality by implementing a Retrieval Augmented Generation (RAG) approach which ensures right context, thus mitigating risk of hallucinations | clinical trial, clinical trial protocol, sas viya, life sciences | source presentation source |
5 |
| 163 | Risk Based Quality Management (RBQM) in Clinical Data Management | Marcin Kot, Anna Górecka |
Risk-Based Quality Management (RBQM) is key to ensuring patient safety and data integrity in clinical studies. This presentation, condensed from a comprehensive workshop, highlights the transformation of Clinical Data Management and the role of Centralized Monitoring in real-time data validation. We explore how to define study-specific Key Risk Indicators (KRIs) and Quality Tolerance Limits (QTLs) from protocol analysis and Critical to Quality factors. Attendees gain insights into the utilization of Centralized Monitoring technology for achieving high-quality clinical data. This session is ideal for Clinical Data Managers seeking to enhance their understanding of cross-functional data-relevant collaboration with Centralized Monitoring | clinical trial, likelihood function, clinical research, data visualization, clinical data, data cleaning, outlier | source presentation source |
9 |
| 164 | Risk-Informed Data Strategies to Streamline Trial Complexity | Tiina Erb | clinical trial, clinical research, life sciences | source presentation source |
3 | |
| 165 | Risk.assessr: A Tool for Assessing and Mitigating Risks with Open-Source R Packages in Clinical Trials | Andre Couturier, Edward Gillian |
data cleaning, clinical trial, regulatory submission, r package | source presentation source |
10 | |
| 166 | Role of PK/PD, Biomarker Analysis in the Expansion of Phase 1 Trials | Lavanya Peddibhotla | Traditional phase I drug development involves single ascending dose studies followed by multiple ascending dose studies in healthy volunteers with safety, tolerability and PK as endpoints. However, this conventional approach presents several challenges, including limited generalizability to patient populations, inefficiencies in dose selection, and prolonged development timelines. The reliance on healthy volunteers may not fully capture the pharmacodynamic (PD) and biomarker responses relevant to the target disease, potentially leading to suboptimal dose predictions for later phases. To address these challenges, drug development is shifting toward more adaptive and biomarker-driven Phase I designs to enhance the success of the trials through better identification of responders, improved dose optimization, and earlier efficacy signals leading to better success in later phases of the trials | change from baseline, phase i, dose, percent change from baseline, clinical trial, pharmacodynamic biomarker, pk parameter, biomarker, clinical research, regression analysis, correlation | source presentation source |
7 |
| 167 | SAS Programming approach for adjustment of treatment switching using Two-Stage Method | Mani Mudumba | Overall survival (OS) is an ideal primary end point for the majority of Phase III clinical studies in oncology. Treatment switching happens frequently in oncology randomized controlled trials. Some control participants switch from the control treatment to the experimental treatment or new anti-cancer therapy after progression of disease (PD), which causes a bias in analyses for endpoints measured after treatment switch. In that instance, the treatment effect of the investigational drug may be underestimated The two-stage method is one of the methods generally used for estimating the treatment effect taking treatment switching in to consideration. The development of statistical programming deliverables, which include analysis dataset and outputs, along with our learnings using the "two-stage method" proposed by Latimer et al., will be explained in this paper | overall survival, treatment effect, hazard ratio, adam dataset, proc step, cdisc | source presentation source |
13 |
| 168 | SASBuddy: Enhancing SAS Programming with Large Language Model Integration | Derek P. Morgan, Karma Tarap |
SASBuddy, dubbed "your friendly SAS programming assistant," is a pioneering SAS macro designed to facilitate SAS programming through Large Language Models (LLMs). This tool empowers users, particularly those with limited SAS coding expertise, to generate precise SAS code efficiently by interpreting natural language inputs (Wei et al., 2023). Beyond code generation, SASBuddy now integrates iterative error resolution, session tracking, macro utilization, retrieval-augmented generation (RAG), and an API-driven architecture that separates execution from intelligence. This architecture enables advanced session management and structured debugging while maintaining context across interactions. The system implements react-style code fixing for targeted error resolution and incorporates Search-O1 (Li et al., 2025) with a "Reason-in-Documents" module for efficient knowledge retrieval | sas programming, clinical programming, clinical sas, pharma programming, macro definition | source presentation source |
9 |
| 169 | Scaling up: accelerating self-hosted open-access LLMs to compete with the Big Three (ChatGPT, Claude, Gemini) | Kostiantyn Drach | The rapid development of large language models (LLMs) offers significant advantages in data handling automation. However, industries with strict data privacy requirements, such as clinical research, face unresolved challenges in working with data during LLM training and interaction. Using leading LLMs like ChatGPT, Claude, and Gemini provides access to their immense capabilities but necessitates submitting sensitive data outside the organization. To avoid this, organizations might opt for training self-hosted, open-access LLMs that are typically smaller and hence less powerful. In this paper, we perform an experiment to address this limitation by training several self-hosted open- access models, such as Llama and others, using the Retrieval-Augmented Generation (RAG) approach. These LLMs are then utilized to generate automatic summary reports of findings within the workflow for analyzing clinical data using Topological Data Analysis (TDA) | clinical trial, correlation, clinical data, data science | source presentation source |
12 |
| 170 | SDTM Automation by Using Metadata to Drive R Functions for Data Transformation | Krishna Gopal, Jaya Simha Inampudi |
This paper outlines an approach leveraging metadata and R programming to streamline the raw to SDTM conversion process. By defining dataset specifications, variable mappings, and transformation rules in a centralized metadata repository, the automation framework can dynamically generate the study SDTM Specification and the necessary R code to execute the source-to-SDTM data transformation. This metadata-driven approach ensures consistency, reduces manual effort, and facilitates updates as SDTM standards evolve. Importantly, the generated SDTM specification is user-editable, allowing subject matter experts to review, customize, and maintain the transformation logic as needed. Having digitized the specification, we can automate the generation of R code to transform the source data to SDTM data | clinical data, clinical data management, cdisc | source presentation source |
3 |
| 171 | SDTM Specification Automation Using an ML Approach to Provide Mappings | Benzi Mathew, Jaya Simha Inampudi |
We present an AI-powered clinical data management tool for automating the creation of SDTM specifications. By leveraging past mappings and knowledge within a therapeutic area, along with a deep understanding of CDISC standards, users can dynamically generate customized SDTM specifications for new studies. The AI engine analyzes historical data transformations, mapping patterns, and CDISC guidelines to propose a tailored set of SDTM datasets, variables, and conversion rules. This automation eliminates the tedium of manual effort typically required to design SDTM specifications from scratch, saving time and ensuring consistency across studies. Furthermore, the generated specifications are user-editable, enabling subject matter experts to refine the AI- proposed logic as needed. This collaborative approach combines the efficiency of AI with the flexibility for human oversight, optimizing the SDTM specification development process | clinical data, clinical data management, data governance, data cleaning, clinical trial, sdtm programming, raw data, cdisc | source presentation source |
5 |
| 172 | SDTM UP - Versioning - Do's & Don'ts | Utkarsh Pathak, Ephicacy Lifescience |
Regulatory submissions are required to comply with or endorse current versions of the CDISC standards, which are continuously updated. More standardized data must be incorporated into the new standards version, which will help with its reuse in different submissions and purposes. This article aims to provide insights to organizations who are putting new standards into practice. The authors of this paper share some challenges and best practices they encountered during the process of up -versioning to CDISC standards. We will also highlight about the utility -driven conversion of SDTM 3.2 datasets to higher CDISC standards (i.e. SDTM 3.3/SDTM 3.4). Provide knowledge on how to read and validate these SDTM outputs in accordance with the conformance standards | liquidity gap, create table, data cleaning, proc sql, clinical trial, cdisc | source presentation source |
9 |
| 173 | Selecting GeARS: Analysis Results Standard Implementation Options for Accelerated Analysis and TFL Automation | Bhavin Busa, Richard Marshall |
Published as a new foundational CDISC standard in April 2024, the Analysis Results Standard (ARS) is complex; the model contains multiple classes, attributes and relationships. Most available ARS training focuses on understanding and use of the components of the model that describe analyses and outputs. However, the model contains additional components to support the generation of submission metadata and the automation of analyses and TFL creation. These additional classes are designed to allow different im plementation strategies such as integration with existing macro libraries or fully integrated automation solutions. Understanding how ARS metadata can be interpreted and how the additional classes can be used in different implementation scenarios is not a walk in the park | system organ class, group by, safety population, analysis set, cdisc | source presentation source |
10 |
| 174 | Solving real-world data (RWD) reliability challenges for use in regulatory decision-making | Tasha Nagamine | FDA final guidance on real-world data (RWD) emphasizes the requirement for data reliability when using RWE in regulatory decisions. Unlike clinical trial data, RWD is never collected in any standard/format. As a result, traditional extract-transform-load (ETL) tools and processes relying entirely on the structure (e.g., schema, data model) of data are inadequate for processing RWD, leading to massive data loss, rendering the RWE unreliable for regulatory use. This inadequacy results in bottlenecks, delays, and rejection by the regulator on data reliability grounds. Droice Labs recently discussed its unique approach for RWD reliability with FDA that utilizes both structure and content of RWD for data transformation from various RWD sources-including sites, commercial vendors, registries, and claims-into reliable analysis-ready data as per regulatory requirements | regulatory submission, data standard | source presentation source |
5 |
| 175 | Spot-It: A Shiny Application for Visual QC at Scale | Samiar Ashtiany, Pravin Lakkaraju |
In the pharmaceutical industry, automating repetitive, labour-intensive tasks is crucial for efficiency and accuracy. Spot-It is an innovative application designed to automate visual oversight quality control, addressing the challenge of managing large-scale document comparison. By handling various file types, including RTF, DOC, EXCEL, and PDF, Spot-It converts and organizes documents, ensuring content consistency and precision. This automation significantly reduces manual effort, minimizes errors, and enhances productivity for statisticians and programmers. By streamlining oversight QC processes, Spot-It supports the pharmaceutical industry's need for stringent quality control, aligning with current trends in automation and digital transformation. This paper provides an in-depth look at Spot-It's capabilities, mechanisms, and benefits, highlighting its potential to revolutionise quality control in the pharmaceutical sector |
excel, quality control, ods excel, extract, pharmaceutical industry, ods rtf | source presentation source |
13 |
| 176 | Standardised and Automated Template for Dose Escalation Decision-Making | Jasmina Uzunovic, Aditi Qamra |
Dose escalation (DE) decision making is a routine process in Phase I studies, which involves thorough review of safety data collected at current dose levels, including monitoring of adverse events, vital signs and laboratory data. First DE meetings can occur shortly after study start-up, where ADaM datasets are often not yet available, leading to various adhoc solutions by study teams to present the data. This leads to lack of standardization in structure, content and format of the slides used across and within therapeutic areas. We will present an R-based solution to automatically populate slide decks with standardized text, templates, and outputs to improve efficiency for this decision making process | dose escalation, sdtm domain, adverse events, adam dataset, cdisc | source presentation source |
6 |
| 177 | Staying Out of the Ditch - How Governance Can Provide Your Team Guardrails | Matt Tendler, Chris McSpiritt |
clinical trial, data cleaning, regulatory submission, data integrity, audit trail | source presentation source |
5 | |
| 178 | Strategic Authoring of Clinical Study Reports to Leverage AI for Data Privacy and Public Disclosure | Anaya Rehman | Since the push towards transparency of clinical trials with the establishment of clinical trial registries, sponsors face increasing challenges in ensuring timely regulatory compliance. This increasing demand for public disclosure necessitates a strategic approach to data protection. Personal and company information is often embedded in lengthy clinical documents which makes its accurate capture a laborious process. AI-driven anonymization offers a scalable solution for safeguarding information. However, the AI tools' effectiveness depends upon structured and standardized document authoring. By implementing proactive standardized authoring through Strategic Medical Authoring for Regulatory Transparency (SMART) approach, sponsors can optimize clinical document development to improve automation in anonymization. This approach uses the regulatory understanding of the guidance to devise best practices in medical writing with future disclosure in mind | clinical trial, data protection, clinical data | source presentation source |
6 |
| 179 | Streamlining Machine Learning Modelling in Neuroscience Clinical Trials Through Cross-Functional Collaboration | Jing Dai | Neuroscience clinical trials face common challenges including high placebo responses and high attrition rates. Machine learning (ML) effectively identifies relevant baseline characteristics to inform future trial designs. Unlike traditional statistical methods, ML techniques offer advantages such as handling high dimensionality and performing automated model tuning, leading to more robust and accurate predictive models. This paper focuses on streamlining ML modeling using neuroscience clinical trial data in a cross -functional setting. Key steps include (1) creating a ML-ready data file from ADaM datasets by clinical biostatisticians or programmers, (2) consulting clinicians for baseline feature pre -selection, (3) training, assessing, and fine -tuning ML models by data scientists, and (4) summarizing, visualizing, and interpreting ML results in collaboration with clinicians | clinical trial, baseline characteristics, feature engineering, machine learning, missing value, data cleaning, imputation, regression analysis, biostatistician, data science, feature importance, model training, multiple imputation, clinical research, elastic net, imbalanced data, proc princomp | source presentation source |
5 |
| 180 | Surviving a Sunset: Being Resilient in the Face of Open-Source Disruption | Franklin E. Menius | Let's face it, the world of data science and programming is changing. Open-source technologies like R and Python are gaining momentum across many industries where proprietary institutionalized software solutions once stood unchallenged. The problems and promises of larger decentralized data sources; real world data; machine learning; artificial intelligence; and faster outcome expectations are driving companies to seek out new, innovative solutions in pursuit of migrating market share. The pharmaceutical and life science industry is no different and is diving headlong into adopting open-source as an efficient tool to get therapies to market faster. So, how do you, as a statistical programmer, survive this disruption without being left behind? This paper will focus on the key skills you need to hone today to ensure your programming future without wandering off into the sunset | clinical research, statistical programmer, data science | source presentation source |
5 |
| 181 | Synthetic Data Generation: A Process Paradigm | Sundaresh Sankaran | Organisations face challenges to analytics innovation when confronted with data that's either sensitive, restricted, imbalanced in proportion, or insufficient in volume. In short, a lot can go wrong with data! Synthetic data generation has been acknowledged to be a useful & effective tool to help mitigate these challenges. It provides opportunities to capitalise on trends and patterns suggested by real data, while not compromising privacy and mitigating risk of leakage of private and sensitive informatio n. However, for synthetic data to truly realise value for life sciences organisations, it's important to view the concept of synthetic data generation at two levels - one, as an operation which provides usable, high-quality data, and the other is as a robust and comprehensive process, in which the operation of synthetic data generation plays a vital par t. In this article, we focus on the process paradigm through which we should view synthetic data generation | correlation, machine learning, life sciences, adverse events, data governance, data cleaning | source presentation source |
9 |
| 182 | Synthetic Data Will Save the Day: Fact or Fiction | Sherrine Eid | From their early adoption in the 1980s to their current applications in genomic simulations, drug discovery, and machine learning, synthetic datasets have revolutionized research in fields such as genomics and personalized medicine. Despite their potential, synthetic datasets present several challenges and limitations that must be addressed. Generating realistic synthetic data requires a deep understanding of the underlying biological processes, as well as sophisticated computational methods for synthesis and validation. Moreover, ensuring the quality, diversity, and representativeness of synthetic data poses significant challenges | machine learning, deep learning, clinical trial, data cleaning, neural network, oversampling, correlation | source presentation source |
7 |
| 183 | Table Numbers Presentation Made Easy with Custom Functions | Sunil Kumar Pusarla | SAS software includes a vast array of functions but using them may not always yield the desired solution. Since version 9.2, SAS Institute has provided users with the flexibility to create their own functions, offering a more straightforward approach. Although many production macros exist, some might believe that PROC FCMP and user-defined functions have limited utility. However, combining simple formats or standard functions with PROC FCMP can produce refined and compact solutions for everyday tasks. Our custom user-defined functions are designed to be user-friendly, significantly reducing code complexity. This paper will discuss a few such user-defined functions that facilitate the elegant presentation of numbers in tables |
proc step, proc fcmp, put statement, proc format, sas function | source presentation source |
7 |
| 184 | The Best of Both Worlds: Machine Learning and Human-in-the-Loop Processing in TFl Design, Production and Validation | Steve Ross, Ilan Carmeli |
data cleaning, clinical trial, clinical research | source presentation source |
6 | |
| 185 | The Development of the CDISC eTFL Portal To Support the Adoption and Implementation of the Analysis Result Standard | Bess LeRoy, Bhavin Busa |
adam dataset, clinical trial, regulatory submission, change from baseline, clinical data, system organ class, adverse events, data science, cdisc | source presentation source |
8 | |
| 186 | The Future of Data Surveillance in Clinical Studies: A Comprehensive Approach to Data Monitoring Using an R Shiny Application for Study Health | Jacky Jin, Ruchitbhai Patel |
The Data Surveillance R Shiny Application enhances study health monitoring by integrating dynamic data processing with standardized reporting modules. It is designed to improve operational efficiency and ensure consistency in generating customized data sur veillance reports across multiple studies. The application's data processing layer creates standardized input datasets from various sources, including SDTM, ADaM, and their combinations. Custom - built R functions utilize these standardized datasets to produ ce interactive summary reports and advanced visual analytics, requiring updates only to the data processing layer for new studies. This approach simplifies report generation across multiple studies while maintaining flexibility. With several interactive R Shiny features, the application also supports on -demand custom report generation, aiding in assessing study health and evaluating impacts on key outcomes | adam dataset, clinical research, in=, data wrangling, cdisc | source presentation source |
5 |
| 187 | The Implementation of {teal} Shiny Apps for DMC Activities | Zhen Wu, Peng Zhang |
The integration of R Shiny applications into regulatory submission processes has gained attention through recent pilot studies, highlighting their potential to facilitate efficient and interactive reviews by regulatory authorities such as the FDA. While R Shiny applications have primarily been used for internal reviews within organizations, there is growing interest in leveraging these tools for Data Monitoring Committee (DMC) activities. Interactive dashboards built with R enable customized, study-specific analyses and endpoint evaluations, supporting more dynamic and responsive monitoring. Tools like {teal}¹, {rtables}², and {tern}³ further enhance the creation of portable and downloadable files, ensuring that critical information remains accessible for ongoing reference and decision-making | clinical trial, clinical data, adam dataset, adverse events, data science, concomitant medications, cdisc | source presentation source |
7 |
| 188 | The Olympics of Data Ingestions | Cathryn Herbst | clinical research, protocol deviations, regulatory submission, data lineage, clinical data, data cleaning, clinical trial, gold layer, self-service bi, access control, adverse events, data ingestion, data wrangling, silver layer, outlier | source presentation source |
8 | |
| 189 | The R Test Kitchen: Cooking Up Quality in Clinical Analytics & Reporting | Emily Yates | In recent years, R has gained popularity in the clinical analytics and reporting space. Ensuring the validity and reliability of R-generated outputs is crucial for making informed, high-impact, clinical decisions based on reports, dashboards, and statistical analyses. This paper demonstrates the powerful ability to automate testing in R using packages like {testthat} and {shinytest}. This versatile framework can be applied to any number of clinical R use cases including R packages, RShiny dashboards, or ad hoc R analyses. We will cover how you can practically incorporate testing into your existing R workflows. Additionally, we will learn how adoption of good testing practices can translate into significant time savings and the delivery of high-quality R-based deliverables | clinical data, data cleaning, data science | source presentation source |
6 |
| 190 | The Right Tool for the Job: Enhancing Risk-Based Monitoring and Data Cleaning | Casey Higgins | Integrating advanced risk management and data-cleaning techniques is crucial for clinical trials. Visual analytics, like R Shiny, enable real-time interaction with data, which is ideal for initial exploration and monitoring key metrics such as adverse event rates. Visualizations highlight areas of concern quickly, aiding prompt decision-making. Using a range of programming and statistical techniques, from basic to more advanced, data science tools can detect subtle anomalies not visible through manual inspection, automating the identification of outliers and inconsistencies. It's crucial to combine aggregate metrics with detailed reports to provide a granular view, supporting rigorous inspection, validation, and issue tracking. This thorough process ensures that no detail is overlooked, providing reassurance in the validation process. Combining visual analytics with statistical/clinical programming creates a robust risk management strategy | clinical trial, adverse events, protocol deviations, data cleaning, data governance, trend analysis, clinical data, outlier | source presentation source |
4 |
| 191 | The Role of Quantitative Bias Analysis (QBA) in Ensuring the Reliability of Real-World Evidence (RWE) in Regulatory Submissions | Brian Gillette, Tasha Nagamine |
data cleaning, clinical trial, regulatory submission, sample size, raw data, sample | source presentation source |
6 | |
| 192 | The Show Must Go On: Best Practices for Submitting SDTM Data for Ongoing Studies | Kristin Kelly | Though the CDISC SDTM Implementation Guide provides advice on how to prepare SDTM datasets for completed studies, there is little guidance on what to do when the study is ongoing, leading to varied implementation practices across the industry. At times, it may be difficult for a regulatory reviewer to readily determine that a study is still in progress without looking in the Clinical Study Data Reviewer's Guide (cSDRG). The recent addition of the ONGOSIND (Ongoing Study Indicator) parameter in the FDA Study Data Technical Conformance Guide (sdTCG) allows sponsors to clearly specify within the data whether a study is ongoing. In this paper, some considerations for preparing domains such as Demographics (DM), Disposition (DS), and Trial Summary (TS) for an ongoing study as well as strategies to ensure data transparency across the SDTM submission package will be discussed | data cleaning, trial summary, clinical trial, sdtm domain, cdisc | source presentation source |
8 |
| 193 | The Ultimate Randomisation App | Chen Li, Ivo Marek |
Randomization is the foundation for effectiveness research and is used as a basis for submissions of regulatory dossiers when seeking marketing authorization for new drugs. For this reason, it is very important to select a randomization design that will produce interpretable, valid results. To achieve this goal, the randomization design may be adjusted several times prior to the performance of the trial. It is perfectly obvious that a randomization application needs to restore randomization specifications and must have audit trail functionality because it creates electronic records - the specifications. InnoRand is an R Shiny-based system that is able to maintain randomization specifications and to create randomization lists. The specifications are stored in an on-premises database and can be restored within the user interface. During the storage process the audit trail information is automatically created | randomization, audit trail | source presentation source |
9 |
| 194 | Three Pointers for Effective and Accurate LLM Integration! | Sundaresh Sankaran, Sherrine Eid |
Life Sciences organisations look to integrate Large Language Models (LLMs) in business and analytical processes. We offer guidelines and considerations to ensure LLMs provide you effective, accurate and usable insights that increase your organization's productivity and enriches user experience. Our suggestions cover three broad areas - one - access and integration of external or hosted LLM services with your analytics; two - design considerations for right parameters and instructions; and three - robust approaches based on Retrieval Augmented Generation (RAG), which improve accuracy and increase trust in LLM results. We detail how hybrid approaches combining traditional Machine Learning methods, Natural Language Processing and Decisioning capabilities enhance LLM quality manifold, while ensuring efficient cost of operations | machine learning, data wrangling, analytics platform, life sciences, sas viya | source presentation source |
9 |
| 195 | Tips and Tricks for Using the CMS Platform | Kena Patel | This paper will cover lessons learned as a new user of the Center for Medicare and Medicaid Services (CMS) Virtual Research Data Center (VRDC) platform. Topics covered will include use of R, Python, SQL, and SAS for data analysis in both Databricks and the SAS Enterprise Guide, structure of the CMS data, resources available to programmers through CMS, and tips and tricks learned while working in the CMS VRDC environment | proc sql, calendar year | source poster source |
2 |
| 196 | Topline Considerations for Double-Blind Trials | Veena Palle, Naresh Gopalkrishna |
Topline results, often referred to as "first -look" results of a randomized double-blind controlled clinical trial, provide an early snapshot of clinical trial outcomes. These results are critical for sponsors and stakeholders to make informed decisions quickly. However, producing accurate topline results involves numerous factors that can affect the timeline and scope of delivery. This paper and supporting poster explore the process of producing topline results, the challenges faced by statistical and programming teams, and best practices for optimizing timelines without compromising data accuracy. Several real -world examples and suggestions are provided to help navigate these complexities, ensuring timely and reliable topline reports | clinical trial, csr, database lock, adverse events, data science | source poster source |
5 |
| 197 | Transforming Clinical Trial Risk Assessments: From Checkbox to Quality-Driven Success | Kristin Stallcup, Rachael Geedey |
Poster | clinical trial, phase i | source | 1 |
| 198 | Transforming Clinical Trials with AI-Driven Protocol Optimization and Next-Gen Statistical Programming | Ravindra Shukla | Generative AI holds the potential to revolutionize clinical trials by significantly reducing their duration, optimizing costs, and speeding up the drug-to-market process. By automating code generation, optimizing data cleaning, and efficiently managing complex datasets, AI enhances trial productivity and reliability while reducing errors. This allows researchers to focus on critical aspects such as protocol development and data analysis. This paper explores the current challenges in statistical programming and the transformative potential of generative AI in addressing these issues. Key areas include automated code generation, intelligent data preparation, and optimized, reproducible analysis. Additionally, it highlights practical examples of AI applications in statistical programming and other AI-driven innovations in clinical trials, such as accelerated trial design, AI-driven recruitment, monitoring, and data optimization | clinical trial, clinical programming, clinical research, machine learning, outlier, data wrangling, data cleaning, adam dataset, data science, cdisc | source presentation source |
8 |
| 199 | Transforming Clinical Trials: Leveraging Bayesian Statistics in Innovative Trial Design, Analysis, Applications and Navigating Challenges | Jaskaran Singh Saini | The accumulation of clinical drug development data over time will facilitate the use Bayesian Statistics in clinical trials design, by allowing the incorporation of prior knowledge to improve the analysis of trial data. Bayesian Statistics will also replace the usage paradigm over the routinely used frequentist statistical methods by pharmaceutical industry and its global regulators in the near future. This paper aims to explore the transformative power of Bayesian methods, highlighting their applications in Adaptive designs, incorporation of Historical data, Survival Analysis, Handling missing data and its application with Machine Learning. The proposal highlights the value of Bayesian methods in drug development, discuss Regulatory perspectives and barriers to its application and also emphasize the need for collaboration in its effective implementation as a powerful tool | clinical trial, missing value, prior probability, survival analysis, machine learning, sample size, imputation, adaptive design, bayesian statistics, sample, clinical research, pharmaceutical industry, sensitivity analysis, confidence interval, posterior probability, multiple imputation, treatment effect, neural network, data science | source presentation source |
11 |
| 200 | Transforming Pharma Master Data Management: Integrating Neo4j Knowledge Graphs with GraphRAG and Generative AI for Enhanced Data Querying and Analytics | Sanjeev Kumar, Sanjay Koshatwar |
Master Data Management (MDM) is critical in the pharmaceutical industry, ensuring data accuracy, consistency, and accessibility-essential for informed decision-making and regulatory compliance. This research explores the integration of Neo4j, a graph database, with GraphRAG and Generative AI to enhance MDM and data querying capabilities. Neo4j's Knowledge Graphs (KGs) structure pharmaceutical data into interconnected entities, such as drugs, patients, and clinical trials, capturing complex relationships that traditional databases struggle to represent. GraphRAG (Graph- based Retrieval-Augmented Generation) enhances this by leveraging knowledge graphs and vector embeddings to retrieve relevant contextual information, improving the accuracy and relevance of AI -generated responses. Vectors, representing semantic relationships in multi-dimensional space, further refine query processing by enabling similarity - based searches and context-aware retrieval | clinical trial, relational database, pharmaceutical industry, accuracy, data cleaning, duplicate detection | source presentation source |
7 |
| 201 | Turbocharge Your Trials: Transform CRF Creation and the CDISC standardization process with Xbiom's Smart Metadata Repository and AI/ML Integrations | Harinarayanan Gopichandran | data cleaning, clinical trial, metadata management, regulatory submission, data standardization, audit trail, clinical data, data governance, edc, crf, clinical data management, access control, adam dataset, data science, data validation rules, cdisc | source presentation source |
8 | |
| 202 | Unleashing the Power of {teal}: New Features and Cross-Industry Collaboration | Nina Qi, Dony Unardi |
data science, clinical data, proof of concept, data transformation, clinical trial, pharmaceutical industry | source presentation source |
6 | |
| 203 | Unlocking the Secrets of Health: A Unified Approach to Biomarker and Clinical Data Integration | Pavan Kumar Tatikonda | In pharmaceutical research, biomarker data analyses are often exploratory and used to better understand the study drug and inform future trials. Integrating biomarker data and clinical data is necessary to perform meaningful analyses. Biomarker data such as genetic variations, protein levels, and metabolic profiles can be high-dimensional and complex. This data is often not CDISC compliant. This paper explains a standardized integration process utilizing ADaM datasets and raw biomarker data. We produced a comprehensive, integrated dataset by creating a unified template and applying rigorous validation, enabling robust exploratory analyses. Our findings demonstrate the feasibility and potential of this approach to enhance clinical insights and therapeutic strategies. This work paves the way for more efficient and accurate biomarker and clinical data integration | change from baseline, amazon s3, clinical data, clinical trial, sample, adam dataset, clinical research, data science, data cleaning, cdisc | source presentation source |
8 |
| 204 | Unravelling the Complexity: Efficacy Datasets in Oncology Basket Trials Across Multiple Diseases | Venkatesh Burla, Manasa Ravi |
Oncology basket trails are designed to assess the efficacy of treatments across various cohorts. The ADaM standards established by CDISC provide a framework to standardize the structure and format of analysis datasets in clinical trials, ensuring consistency and facilitating data exchange among researchers and regulatory authorities. This paper aims to outline the development of ADaM datasets within the context of a basket trial, focusing on a cohort with Von Hippel-Lindau (VHL) disease, where subjects exhibit multiple diseases, including Renal Cell Carcinoma (RCC), Pheochromocytoma and Paraganglioma (PPGL), pancreatic Neuroendocrine Tumor (pNET), and Central Nervous System (CNS) Hemangioblastoma. Key endpoints include Progression-Free Survival (PFS), Overall Survival (OS), Best Overall Response, and Tumor Percentage Change from baseline | adam dataset, progression-free survival, duration of response, cdisc | source presentation source |
7 |
| 205 | Using Custom GPT to Write Edit Checks | Andrew Burd | Study builders across the industry must master various EDC systems, each with its own programming language for edit checks-Medidata uses C#, Viedoc uses JavaScript, and Veeva and REDCap rely on formulas similar to EXCEL. Often, the focus on perfecting the eCRFs and Visit Schedule leaves automated edit checks, the first line of defense against bad data, as an afterthought. Robust edit checks ensure data integrity and patient safety while also providing cost savings by reducing time on rework and manual data review. To address this, we developed a custom GPT, an AI model tailored for specific tasks, which leverages the study builder user guide to automatically generate edit checks directly from a study design specification. Automating this process significantly reduces the time and effort required to program edit checks during study start- up and ensures that edit checks are no longer overlooked |
data validation rules, edc, data cleaning, excel | source presentation source |
3 |
| 206 | Using Large Language Models for Consistency Checking of Tables, Listings and Graphs | Yingying Wei | Programmatic data quality checking is essential for accurately summarizing clinical trial results. Automated checks are common at various levels of clinical trial data, including on the raw datasets from the case report forms (CRFs), SDTMs, and ADaMs. Unfortunately automated checking is often difficult for tables, listings, and graphs (TLGs) due to their unstructured nature. Large language models (LLMs) however EXCEL at parsing unstructured text and can potentially automate |
adverse events, clinical trial, trial summary, data cleaning | source presentation source |
7 |
| 207 | Using the USDM and Ensuring Governance and Traceability to unlock the potential of Real-World Data sources for regulatory submissions | Berber Snoeijer | The goal of real-world data (RWD) is to be used as evidence (RWE) in regulatory submissions. The governance and traceability of this RWD are paramount, ensuring integrity, reliability, and compliance for regulatory submission. The use case described in this paper shows how the eligibility criteria of the Lilly LZZT study are populated in the CDISC unified study definitions model (USDM) using syntax template capabilities and how this drives the selection of real- world data for an external control arm. Furthermore, this paper explores challenges and best practices associated with governing RWD throughout its lifecycle, from collection at the source to submission, providing regulators proof of source reliability without regulators needing to access third-party source data. Building upon the ideas of multiple sponsors and regulators through working group collaborations, we developed a method to demonstrate principles and best practices in data governance, quality, and traceability | proof of concept, raw data, data lineage, regulatory submission, sdtm domain, clinical trial, cdisc | source presentation source |
12 |
| 208 | Utility Macro for Ensuring Consistent Updates of CRF Data in SDTM | Naren Surampalli | The source data collected from the Case Report Form (CRF) pages is generally recorded in clinical trial databases. Various organizations may employ different tools to extract this source data and appropriately map it to the corresponding SDTM datasets. Commonly, not all CRF page data is available at the onset of a trial. As the trial advances, additional data is gathered from growing CRF pages. The updated information from the new CRFs must be accurately extracted and synchronized for both the source and SDTM data. This paper will outline a utility macro that incorporates multiple checks to ensure that updates to source data are reliably captured in the SDTM datasets from the initiation of a clinical trial until its completion, thereby reducing the likelihood of overlooking critical data | sdtm domain, raw data, clinical trial, dataset, create table, proc sql, cdisc | source presentation source |
8 |
| 209 | Validated by Verisian: The Future of Quality Control and Validation in Clinical Trial Analyses | Henning Kuich | Outdated validation techniques in clinical trial analyses waste billions while delaying critical treatments for patients. Today's approach, heav- ily reliant on double programming and manual processes, wastes time and resources while paradoxically undermining the very quality it aims to guarantee. In this article, we introduce the Validated by Verisian process. Powered by Full Study Traceability, it enables the required paradigm shift in how clinical trial analyses are validated, enabling quicker, more confident decisions that move medicine forward and bring treatments to patients faster and at lower cost. January 2025 Contents 1 Clinical Trial Validation and QC is Broken 2 2 Validated by Verisian 5 2.1 The Mission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.2 The Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.3 The Process | data cleaning, clinical trial, independent qc, raw data, quality control, data science, excel | source presentation source |
12 |
| 210 | Verifiable Compute to Trust AI in Regulated Workflows | Alistair Dootson, Jonathan Dotan |
data integrity, audit trail | source presentation source |
5 | |
| 211 | Verify: A First-in-Class Review and Validation Platform for Biometrics in Clinical Research | Achinoam Ravet Perel | Double programming of TLFs is one of many redundancies in the development and review of analysis outputs in clinical research. Overlapping review responsibilities begin with S&P and end with medical writing; iterative table review for each deliverable in each study; multiple disparate communication channels; each duplicative by design. The maintenance and perpetuation of each of these inefficiencies has direct time and financial costs, and indirect costs (burnout, turnover, boredom with repetitive tasks), which detract from a bottom line. Verify provides a first-in-class, standalone suite of AI-enabled verification and validation checks and a process map designed to minimize silos and redundancies. Leveraging novel technologies that optimize the strengths of human and machine, Verify provides a platform to streamline TLF development, validation, and review cycles, enabling significant reduction in validation timelines | data cleaning, clinical research, independent qc | source presentation source |
3 |
| 212 | Visualising Longitudinal Patients' Responses in a Clinical Study: A Unified Graphical Approach | Avani Reddy Alla, Sunil Kumar Pusarla |
In clinical trials, analyzing patient responses over time is essential for evaluating treatment outcomes and efficacy. This poster demonstrates a unified method for visualizing longitudinal patient data in a single, comprehensive graph using SG PLOT Procedure in SAS. It outlines how all the patients' responses can be correlated with treatment administrations on a unified timeline by overlapping multiple scatter plots in a single output. Our results highlight the versatility of SAS' graphic capabilities and showcases how it is a valuable tool for researchers in clinical trials. This graph also aids in effective communication of all vital study findings in a snapshot |
scatter plot, series plot, outlier, clinical trial, proc sgplot | source poster source |
6 |
| 213 | Wearable Technology: Enhancing Patient Monitoring and Real-World Evidence | Bill Qubeck, Sridhar Vijendra |
Wearable technology is becoming a pivotal tool in the real-time collection of patient data and the remote monitoring of clinical trial participants. These devices provide continuous, non-invasive insights into patient health metrics, enabling more precise and timely interventions. This abstract analyzes the growing role of wearables in enhancing patient engagement and compliance, as well as improving the accuracy of trial outcomes and ethical considerations. As the adoption of wearable technology expands, its impact on patient-centric clinical research is set to grow, offering new opportunities for more personalized medicine | clinical trial, data cleaning, clinical research, data integrity, data wrangling, 21 cfr part 11, data pipeline, machine learning, audit trail, real-time processing, regulatory submission, clinical data, access control, adverse events, clinical database, accuracy, data ingestion, data governance, raw data, proc freq, data science | source presentation source |
15 |
| 214 | What R We Counting? The Quest to Quantify R Usage in Biostatistics Organizations | Benjamin Arancibia | clinical trial, rest api | source presentation source |
3 | |
| 215 | What's In a Message? Enhancing Error and Warning Messages for the Admiral R Package for ADaM Datasets | Benjamin Straub, Daniel Sjoberg |
Effective software projects often prioritize delivering clear and informative messages to users when issues arise. Incorporating a touch of humor, as illustrated in the image at the end of this section, can enhance user experience by enabling quicker self-diagnosis and resolution. Users being albe to quickly self-diagnosis issues they are having with your software can provide a more positive experience for the user. The admiral R package, which seeks to help users of R to build ADaM datasets, is keen on this type of messaging. In this paper, we will walk you through several simple examples looking at how to improve messaging from R using base R and the cli and glue R packages. We will then look at admiral for additional examples where the team improved the messaging of the function using the cli and glue packages. Finally, we will conclude with a brief collection of utlities that can aid you in your quest for improving your messaging | error message, adam dataset | source presentation source |
8 |
| 216 | What's Up with icSDRG? Navigating Through an Example | Srinivas Kovvuri, Bidhya Basnet |
Poster: cSDRG vs icSDRG - What's new? Disclaimer: All views expressed in this paper are those of the authors and are not necessarily those of PHUSE, CDISC, regulatory agencies, or the author's companies: Roche-Genentech, ADC Therapeutics, Ephicacy Consulting Services, Fortrea, and Lilly | sdtm domain, cdisc | source | 1 |
| 217 | When and Why You Should Deviate from CDASH Standards | Agnes Nkenglefac Akawung, Trevano Dean |
The Clinical Data Acquisition Standards Harmonization (CDASH) standardizes data collection to ensure traceability, reduce errors, and facilitate efficient transformation into Study Data Tabulation Model (SDTM) datasets and subsequent Tables, Listings, and Figures (TLFs). However, strict adherence to CDASH may not always be practical. Protocol-specific requirements, operational or financial efficiency often necessitate deviations. This paper explores the benefits of CDASH compliance, including improved data traceability, reduced data cleaning efforts, and streamlined TLF generation. Additionally, it highlights domain-specific examples and discusses scenarios where deviations from CDASH are justified to optimize data collection, analysis, and reporting | controlled terminology, clinical data, cdisc | source poster source |
2 |
| 218 | Why Be Different? Shouldn't Everyone Be Using Agile to Deliver Their Clinical Trials? | Michael S Rimler | clinical trial, clinical data, data science | source presentation source |
5 | |
| 219 | Write Concise ADaM Programs with If-less Programming Techniques | Wei Shao, Lei Zhang |
ADaM is a widely adopted CDISC standard for analysis datasets in clinical trials. Conditional statements, or blocks like IF-THEN-ELSE or SELECT-WHEN are commonly used to derive ADaM variables. However, excessive use of conditional code makes ADaM programs cumbersome, hard to read and modify. This paper examines methods to eliminate conditional blocks through if-less programming techniques. It first covers basic if-less programming approaches in SAS and then focuses on how to develop new mapping macros to generalize them. The way to construct these mapping macros is fully explained along with practical code in ADaM scenarios such as flag creation, partial date imputation, and complex variable derivation. The paper underscores the benefits of if-less programming techniques for improving the readability and reusability of ADaM programs. Additionally, the techniques discussed are versatile enough to be adopted in many other programming situations, extending their use beyond ADaM programs | macro function, adam dataset, character constant, proc fcmp, proc step, macro variable, macro definition, %if-%then-%else, cdisc | source presentation source |
17 |
| 220 | Your SCE is Only as Good as Your Data: The Central Role of Information Governance in Clinical Research | Michelle Chen | clinical research, data governance | source presentation source |
3 | |
| 221 | {sdtm.oak} - SDTM Programming in R | Rammprasad Ganapathy, Shiyu Chen |
{sdtm.oak} is an EDC (Electronic Data Capture systems) and Data Standard agnostic solution that enables the pharmaceutical programming community to develop CDISC SDTM datasets in R. The reusable algorithms concept in {sdtm.oak} provides a framework for modular programming and also can potentially automate SDTM creation based on the standard SDTM spec | sdtm programming, edc, sdtm domain, cdisc | source poster source |
5 |
| 222 | {teal.builder}: Our Journey After One Year | Andre Couturier, Freeman Wang |
clinical data, proc princomp | source presentation source |
4 |
This site serves as a convenient hosting location and does not work as a publisher. All copyrights belong to their respective owners.
The information on this site is provided for general reference purposes only. Whilst every effort is made to ensure that the site is up to date and accurate, no warrant, nor any responsibility or liability for, the accuracy or completeness of the content or for any loss which may arise from reliance on information contained in this site is provided. Links to other sites are provided for your convenience but no responsibility or liability for the content of those sites or of any external site which links to this site is accepted. References to any specific commercial product, process, service, or corporation are also for information only, and do not constitute endorsement, recommendation or favoring.
SAS is a registered trademark of SAS Institute Inc. SAS, Statistical Analysis System, and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other countries. ® indicates USA registration.