HOME
QUICK SEARCH
AUTHORS
KEYWORDS
HELP
BUGSPHUSE US 2024 Papers - Index
187 papers. Source: PHUSE (Pharmaceutical Users Software Exchange) US Connect conference proceedings.
| # | Title× | Authors× | Abstract× | Keywords× | Extras | Pages |
|---|---|---|---|---|---|---|
| 1 | A Data Manager's Experience with AI/ML Adoption and Efficiency | Diane Lacroix | AI buzz continues to include exciting predictions along with questions about job security, regulatory guidance and oversight. But AI intent is here in clinical development, and practical use cases are being implemented today with the potential to deliver much-needed efficiency gains. Successful AI adoption requires focusing on the users of these AI algorithms - humans providing necessary supervision and reducing risk, but also fine-tuning models, actioning decisions and finding new, efficient ways to work smarter using this intelligence. This session will examine the data manager's experience with AI/ML from the practitioner perspective. With a focus on AI-enabled data review, the talk will highlight clinically relevant AI use cases for DM and address how data management uses these outputs to augment their data review and adapt typical workflows | data review, machine learning, data governance, data cleaning | source presentation source |
4 |
| 2 | A Data Visualisation App to Reduce Event-Based Requests | Lixiao Chen, Diana Tai |
R Shiny applications have proven to be an effective framework for data exploration and analytics due to the unique features offered. In medical affairs and market access (MMA), programming activities mainly focus on exploratory analysis, supporting posters, abstracts, manuscripts, publications, and conference presentations. Large quantities of event-based requests (EBR) have become a burden for the MMA biostatistics and programming groups. A comprehensive R Shiny app has been built to help statisticians and clinical teams explore data/perform data analytics to help decide whether a formal EBR is required. This app has proven to be an effective tool that has significantly reduced the number of EBRs needed and greatly impacted the budget that is allocated to support EBRs | subject-level data, create table | source poster source |
9 |
| 3 | A Holistic Approach to Trustworthy Artificial Intelligence (AI) | Tara Chavda, Qingying Ally Lu |
Over the past decade, Artificial Intelligence (AI) has advanced significantly to be able to perform many tasks previously done by humans. The key advancement in the AI technology is that algorithms can recognize patterns in data and build models to recommend decisions without requiring explicit programming. Since the AI models learn behavior from the training data, any patterns (e.g., biases) in the training data will be faithfully reproduced in the model behavior. Due to the statistical nature of the models, they are also likely to make errors due to natural scatter in the data and uncertainties in the chosen models. In addition, AI systems often operate as "black boxes" and may produce unfair outcomes and cause other ethical concerns. Therefore, fostering trustworthy AI practices is not only a responsible and ethical choice but also a strategic imperative for organizations looking to build trust, maintain compliance, and thrive in an increasingly AI-driven landscape | training set, machine learning, artificial intelligence, sample | source presentation source |
6 |
| 4 | A Larger Tent - Integrating Multiple Open-Source Environments in SAS | Sundaresh Sankaran, Samiul Haque |
Presentation | sas studio, proc step, proc iml, dataset, proc python | source | 15 |
| 5 | A Practical Approach to Automating SDTM Using a Metadata-Driven Method that Leverages eCRF Specifications and SDTM Programming Standards | Xiangchen (Bob) Cui, Min Chen |
Automated SDTM generation has several benefits , including efficiency, accuracy, compliance with regulatory requirements, and the speeding up of the data analysis process. However, due to the dissimilarity and varying complexity of different CRFs, SDTM domains, and eSource systems among different studies, development of a tool to automate SDTM has been a challenging task for sponsors, CROs, and EDC service providers. We propose a new approach in automatic generation of SAS code for SDTM. A SAS-based macro is developed based on CRF specifications from an EDC database and SDTM standards. Our approach is user-friendly with high transparency, easily scalable to multiple studies, and especially useful for relatively smaller sponsors and CROs, for there is no requirement to standardize CRF s and raw dataset variables' attributes (which is the best practice but can be too work-intensive) and no required expertise in other computer languages | sdtm programming, sdtm domain, data cleaning, dataset, clinical trial, macro call, macro definition, independent qc, sample, cdisc | source presentation source |
12 |
| 6 | A Primer to Pharma DevOps: Empowering R Package Development with DevOps Workflows and Automation | Benjamin Straub | DevOps practices are increasingly crucial for maintaining high-quality, compliant R packages in Pharma. The PhUSE DevOps Working Group is pioneering a revolutionary initiative to transform Pharma De- vOps with an extensive knowledge base. This project aims to establish a robust knowledge base featuring tailored Continuous Integration/Continuous Deployment (CI/CD) workflows for R packages. These pre- configured workflows will serve as a launching pad for pharma researchers and developers, streamlining project initiation. Automation will free up developers to focus on scientific innovation rather than repet- itive tasks. The knowledge base will meticulously document best practices and guidelines for R package development, promoting efficiency and quality in pharma R&D. Adhering to these best practices ensures work reproducibility and reliability. Our initiative will introduce flexible CI/CD workflows, allowing users to customize pipelines to match their project needs | data cleaning, likelihood function, regression analysis, r package | source presentation source |
5 |
| 7 | A Privacy Methodology Toolkit to Improve Cross-Industry Clinical Data Reuse | Patty Hegarty | Secondary use of clinical trial data can help accelerate the development of medicines through reducing patient burden, generating insights, and enabling evidence-driven decision making. However, differences in the current data privacy methodologies can make the data usability extremely challenging, especially for cross-study analysis. TransCelerate Biopharma has developed a Privacy Methodology, Supporting Materials and toolkit that aim to help companies improve reusability of data and data transparency to researchers on the applied safeguards to protect patient privacy. Despite being initially developed for TransCelerate's Clinical Data Sharing Initiatives, the Privacy Methodology paper was submitted to public commenting in November 2022 - March 2023 to collect industry feedback on potential areas of improvement and determine whether the solution could be useful to other stakeholders and data sharing platforms | clinical trial, privacy, clinical data, proc freq | source presentation source |
7 |
| 8 | A Programmer's Perspective on Navigating the Hurdles of Investigator-Sponsored Research (ISR) Trials | Jyotsna Lakshmikumar | Independent research may be defined within the industry as Investigator Sponsored Research (ISR), Investigator Initiated Research (IIR), Investigator Sponsored Trials (IST) or Investigator Initiated Trials (IIT). The Success of this investigator-sponsored trail holds the promise of a breakthrough in treatment, potentially improving the lives of countless individuals. These "homegrown" studies are unlike our regular Sponsor defined studies and pose many challenges in multiple aspects such as data format, non-standard CRF, data mapping to CDISC standard, coding concomitant/ adverse event data as per dictionary with only the verbatim term, integrated summaries with other studies, analyzing nonstandard data to create meaningful results. This paper will delve deeper into the challenges faced in one such trial and will provide approaches that help mitigate them and prepare for similar trials in the future | concomitant medications, adverse events, dose, phase ii, phase iii, raw data, cdisc | source presentation source |
6 |
| 9 | A Review of Open Source in New Drug Applications: Current Trends | Philip Bowsher | Presentation | source | 24 | |
| 10 | A SAS Tool for Generating Adverse Event Summaries in XML Format Using ADaM Datasets ADSL and ADAE for ClinicalTrial.gov and EudraCT Submissions | Karthika Bhavadas, Shabeeb Kannattikuni |
Presentation | adverse events, serious adverse event, adverse event summary, macro quoting function, dataset, macro variable, excel | source | 23 |
| 11 | A standard data set format in regulatory submissions for clinical trials: automatic creation of headers, headers of headers, and footnotes | Kevin R. Viel | Accuracy, efficiency, reproducibility, and traceability are major goals of the programming of tables and listings, but automa tion and reusability have emerged as important targets. Within this framework, standardization is a foundation to achieve these goals. Busa and LeRoy introduced the Analysis Results Standards model to drive automation, especially with re -arrangement of a table or the reproduction of it with various languages, including open-source tools. Viel described a standardized structure and content of a data set to aid in Validation and automatic creation of a table. The goal of this paper is to expand on the latter, emphasizing headers and footnotes and the ease with arranging them, and to introduce a SAS System macro that automates their creation. Importantly, headers and footnotes contain derived data that are subject to Validation, but are often not included in the programmatic approaches, i.e., the COMPARE procedure in SAS, because the data set omits them |
proc report, data cleaning, macro call, proc freq, macro parameter, data step, footnote statement, clinical trial, order by, macro definition, macro quoting function, adverse events, macro function, proc sql | source presentation source |
12 |
| 12 | A Step-by-Step Programming Approach to Summarize Best Response in Patients with Multiple Myeloma Who Had Positive and Negative | MRD Test | Minimal Residual Disease (MRD) test has been considered the most relevant prognostic factor in Multiple Myeloma (MM). MRD testing is done to see whether cancer treatment is working and what could be further course of action for new or ongoing treatments. This test can find even the smallest amount of cancer cells that may be left in the body after treatment and highly efficient method to detect even one cancer cell among a million normal cells. Summarizing the best response in patients with MRD test results for a given line of chemotherapies often appears challenging for statistical programmers. A stepwise logical approach on how to combine various data sources for a patient may be helpful to generate a detailed summary repor t of the best response with MRD Positive and MRD Negative test results. Patients with negative MRD test results display significantly longer survival as compared to those with positive MRD test results | best response, proc sql | source presentation source |
8 |
| 13 | Accelerating Clinical Trial Submissions with Graph Theory and Log Analysis | Henning Kuich, Tomás Sabat Stofsel |
Presentation | root cause analysis, data cleaning, proc sort, adverse events, baseline, change from baseline, proc step | source | 42 |
| 14 | Addressing Challenges in Structuring CDISC SDTM and ADaM Datasets to Report Adverse Events Spanning Two Treatment Periods | Madhusudhan Reddy Papasani | Reporting treatment-emergent adverse events that range from grades 3 to 5 in complex clinical trials presents significant challenges. These challenges are mainly related to accurately tracking toxicity grades, data entry for adverse events, developing SDTM datasets, and capturing important information such as treatment period dates, treatment emergence, and toxicity grades for reporting purposes in ADaM datasets. Here we present a practical example using a two-period study, where an AE starts in period one and keeps changing toxicity and continues through the second period. The details in the paper include practical examples of SDTM, and ADaM mappings in the context of reporting treatment-emergent adverse events of grades 3-5 in the second period of the study | adverse events, adam dataset, ctcae grade, sdtm domain, cdisc | source presentation source |
4 |
| 15 | Agile, Collaborative, Efficient (ACE): A New Perspective on Data Monitoring Committee Data Review Preparation | Ke Xiao, Hong Wang |
Data Monitoring Committee (DMC) has been part of clinical trials since 1960's. Nowadays, an increasing number of sponsors utilize DMC in various situations to monitor safety data, critical efficacy data, and ensure the integrity of study conduct. Given the different purposes of DMCs, there are diverse approaches to facilitate DMC data review. In this paper, we will discuss certain strategies and processes that Boehringer Ingelheim (BI) implement for DMC. To begin with, we will briefly describe the general purpose of a DMC as well as its roles and responsibilities in BI, followed by an introduction to BI's independent safety/statistical analysis team (iSAT). The focus is to present three models that iSAT, as an independent team within sponsor, apply to efficiently and timely assist DMC data review. We will conclude by sharing experience/lessons learned from supporting DMCs in randomized non-pivotal and/or open label registrational studies to open further discussion | open label, clinical trial, phase ii, data science, data review, statistical programmer | source poster source |
6 |
| 16 | AI Exploration and Innovation for the Clinical Data Scientist | Cole Arendt, Philip Bowsher |
standard deviation, proc calis | source presentation source |
15 | |
| 17 | AI-Enhanced Chatbot for Streamlined Clinical Trial Analysis and Document Management | Jaime Yan | The rapid escalation in data volume demands effective utilization and safeguarding of confidential documents, such as | clinical programming, proc sql, proc step, adam dataset, macro call, cdisc | source presentation source |
7 |
| 18 | AI-Facilitated Monitoring Methods and Techniques to Enhance Participant Safety and Data Integrity | Kris Wenzel, Martin Bauwens |
adverse events, clinical trial, artificial intelligence, concomitant medications, data science | source presentation source |
6 | |
| 19 | An Agile Way of Working with Azure DevOps in Clinical Trials | Rambabu Sura | In an era marked by dynamic transformation, industries are transitioning away from conventional static libraries and driving towards cloud-based environments. The time has changed, and so have we. We at Novo Nordisk, have recognized the need for evolution and modernized our ways of working by harnessing the potential of cloud computing through Azure DevOps (ADO). With the implementation of ADO, internal collaboration and structured discussions have improved. The new agile methodologies and technologies offered through ADO, have facilitated a comprehensive overview of the project's progress along with version controlling. Furthermore, ADO empowers us to address various project issues through dedicated issue logs and share knowledge efficiently via wikis, Kanban Boards, Pictoria l presentations, documentation , and much more. This strategic shift underscores our commitment to staying at the forefront of clinical project management | data integrity, audit trail | source presentation source |
10 |
| 20 | An Efficient Way of Managing the Execution and QC of SAS Programs in Multiple Folders | Wei Qian, Danny Hsu |
A clinical trial could involve multiple statistical programmers working on hundreds of SAS programs in different subfolders (e.g., sdtm, adam, and tlfs). It is therefore critical that a lead programmer manages execution and QC efficiently and effectively. This paper proposes a method of setting up a batch file which first executes all programs in proper sequence and then runs a utility program to generate an overall summary report listing file names and authors for programs with unmatched results or unwanted messages (e.g., "ERROR" and "WARNING") in log files. In this way, with one click, hundreds of SAS programs are executed and QCed and the lead programmer gets immediate insight into the study status by utilizing the QC report and assignment information to notify programmers to correct their programs for any issues. This process would streamline the program execution process while continuously tracking QC status | proc compare, proc step, cdisc | source poster source |
2 |
| 21 | An End User's Perspective of Clinical Dosing Data Used for Population PK/PD Analysis with Non-Linear Mixed Effects Modelling | Rebecca Humphrey, Mitali Gaurav |
Observing how clinical dosing data are used in building population pharmacokinetic/pharmacodynamic (PK/PD) analysis datasets lead to a more comprehensive collection and representation of vital clinical data. A model is only as good as the data! Population PK/PD modeling is an important part of a regulatory submission and data integrity is a crucial component to all successful models. Here, we explore the challenges end-users face when the provided dosing data are incomplete, insufficient, or ambiguous, as well as solutions and best practices to address such challenges. Population PK/PD modeling depends on an accurate time-ordered analysis dataset. For example, when dosing times are not collected, imputations and assumptions are made which lowers the integrity of the model. The representation of dosing start and stop dates and times differ for different routes of administration (for example, intravenous versus oral) | population pk, adam dataset, pd, imputation | source poster source |
2 |
| 22 | Analysis of Same Objective - Different SAS Procedures: How to Choose a SAS Procedure Wisely for a Statistical Analysis? | Pradeep Acharya, Ephicacy Lifescience |
Statistics plays a very important role in any clinical trial from design, conduct, analysis, and reporting in terms of controlling or minimizing biases, confounding factors, and measuring random errors. Statistical analyses are typically based on estimating confidence intervals (CI), hypotheses and drawing conclusions based on observed data. In clinical trial reporting, summary statistics can be obtained using different SAS procedures such as Means, Univariate, Summary, Tabulate etc., Similarly there are different SAS procedures which can be used to do statistical analysis for the same objective. To illustrate this through examples, hypothetical random data has been generated similar to clinical trial scenario using the SAS function RANNOR under the assumption of Normal Distribution. An objective of the study has been assumed, the data was analyzed using different SAS procedures, and the results are compared | ancova, proc step, covariance, response variable, proc glm, proc mixed, proc genmod, predictor variable, residual, change from baseline, regression analysis, clinical trial, normal distribution, data science, simple linear regression, scatter plot, anova, proc reg, p-value, hypothesis testing, normality test, proc means, estimate, numeric variable, correlation, homoscedasticity | source presentation source |
9 |
| 23 | Applying Machine Learning for Predicting Potential Adverse Events from Real-World Data | Samiul Haque | Adverse reactions to drugs are common occurrences, with risks often stemming from factors such as age, weight, sex, and comorbidities. Employing machine learning techniques can shed light on the underlying causes of adverse events. Moreover, predictive models can be trained to evaluate the risk of adverse events for individual patients or patient groups. This work introduces a streamlined workflow for deploying machine learning approaches to assess the likelihood of adverse events for patients. Our approach is based on synthetic data extracted from OHDSI/Eunomia, a widely recognized dataset within the OMOP common data model. We delve into various facets of a machine learning project, encompassing data ingestion, cleaning, partition, m odel training, motel tournament, scoring, and model deployment | machine learning, adverse events, neural network, model training, data cleaning, roc curve | source presentation source |
6 |
| 24 | Approaching End-to-End Analysis Standardisation in Global and Therapeutic Areas Level | Rudra Sheth, Xiaohan Zou |
In the long term, standard ization is expected to reduce costs and increase efficiency. The paper outlines a strategic approach for developing and implementing analysis standards end -to-end to bolster efficiency and consistency within Bristol Myers Squibb (BMS). We have devised a model that establishes a standards organization with representation from all therapeutic areas in the relevant functions and ensures resources are available to develop, maintain, implement, and govern the analysis standards . Members of the standards organization seamlessly collaborate and play a pivotal role in standardization. We crafted a streamlined process that supports and facilitates the use of these standards. We approach standardization at the global level in a way that provides the flexibility for Therapeutic Area (TA) specific customization. Each Therapeutic Area (TA) contributes to the development of global standards | clinical trial, data science | source presentation source |
4 |
| 25 | Are Platform Studies the Need of the Hour for Oncology Trials? | Vidya Muthukumar | In the fast-changing, cost-conscious clinical trials industry, platform trials are slowly becoming the need of the hour, especially for oncology studies. Sponsors are often looking to gain level-playing fields to maintain a cost equilibrium for clinical trials. Platform trials are set up to study diseases in specific therapeutic areas, rather than specific interventions. Trials can continue indefinitely, adding new arms to test new therapies, or discontinue existing ones as soon as it becomes clear the drug is ineffective or harmful, and substitute the control arm for a new standard-of-care. The randomization process can be amended so that new participants are more likely to receive a better performing therapy. The key features of platform trials are their scope and adaptability. They are designed to evaluate multiple therapies. Rather than test drugs one at a time, one could test them all at once; and get results quickly to make informed decisions | clinical trial, randomization, treatment effect | source presentation source |
2 |
| 26 | Auditing an Open-Source Software Framework in a Regulatory Environment: Challenges, Gaps and Recommendations | Uday Preetham Palukuru, Shahrzad Talebian |
With the increasing use of open-source software like R in clinical trial analysis and reporting (A&R), ensuring regulatory compliance has become a critical factor for organizations. While open -source software frameworks offer greater flexibility and potential cost-effectiveness than proprietary systems, auditing them for quality and functionality can be difficult. This paper outlines the audit process for open -source software frameworks, from data collection to findings reporting, and highlights the challen ges the process faces, such as technical complexity, lack of documentation, and transparency issues. We share our experiences of auditing an internal open-source R framework for clinical trials A&R and provide recommendations to help organizations establis h effective audit procedures for open -source software frameworks | clinical trial, data cleaning, clinical data, clinical research | source presentation source |
5 |
| 27 | Automating ADaM Creation in Clinical Trials Part 2 | Rie Ichihashi, Lyubov Sushchenko |
The automation of ADaM creation in clinical trials has become a focal point within the pharmaceutical industry. Building on our presentation at PHUSE US Connect 2023 [1], where we introduced Sanofi's initial approach to ADaM/SDTM automation, we are excited to present the continued evolution of this project. As we advanced in the development of the user interface and tested pseudo-codes in pilot studies, we discovered significant support for the creation of value-level metadata (VLM) pseudo-codes. An unexpected yet valuable outcome of this automation initiative has been the reorganization of internal template programs, derivations, and macros, leading to a more streamlined automation process. In this presentation, we will again provide concrete examples of pseudo-code to illustrate its capacity to generate clear, specific, consistent, and unambiguous derivation rules, facilitating seamless automation. We will explore the end-to-end flow and discuss the degree of automation achieved | clinical trial, lookup, cdisc | source presentation source |
7 |
| 28 | Automating TLG Generation for Clinical Trials Leveraging R/R Shiny Technology and Cross-Therapeutic Area Expertise | Vijay Sharma, Margaret Wishart |
Scalability, consistency, and efficiency in creating Tables, Listings, and Graphs (TLGs) have proven impactful within Full Development studies across Therapeutic Areas (TA). This paper presents a use case for clinical programming to rapidly build an application utilizing R Shiny to automate TLG generation. The foundation of the increased efficiency in quickly generating outputs depends on pre-defined global and TA-level standards. To achieve standardization, a Governance Team was instituted with TA-level experts to define standards and maintain a library of outputs. Leveraging an R-based interactive web tooling to utilize standard output templates on a study-by-study basis with a point-and-click selection mechanism. Introducing R within the industry has created a need for commonly used statistical methods to be defined and preserved for uniformities across studies, requiring experts with cross Therapeutic Area knowledge | clinical trial, data science | source presentation source |
6 |
| 29 | Automation of Exploratory Report Generation using R shiny and R Markdown | Denny Sebastian | In the field of clinical data analysis and reporting, the demand for efficient and automated solutions has grown exponentially. This paper presents a novel approach to automate the generation of exploratory reports using the capabilities of R Shiny and the capability of R Markdown for document production. Traditional report generation often involves manual steps, which can be time-consuming and error prone. By leveraging the integration of R Shiny and R Markdown, we have developed a streamlined workflow that significantly reduces manual intervention and accelerates the production of insightful exploratory reports | data science, clinical research, correlation, exploratory data analysis, scatter plot, clinical data, histogram, box plot, outlier | source poster source |
7 |
| 30 | Beyond Excel: Automating Clinical Project Management with Interactive R Shiny Dashboards | Melanie Hullings, Emily Murphy |
source presentation source |
7 | ||
| 31 | Beyond Outliers: Unsupervised Anomaly Detection in Clinical Trial Data | Helena Yaben, Kian Norouzi |
Reviewing clinical trial data during conduct is crucial for ensuring patient safety, data quality, and regulatory compliance. However, manual inspection for identifying anomalies can be resource-intensive and limited, especially with the exponential growth in data volume seen the recent years. To address this challenge, we have developed a Python pipeline which uses a combination of univariate and multivariate unsupervised machine learning algorithms to automatically identify anomalies in SDTM datasets. By employing feature selection, we limit the number of variables used in the multivariate anomaly detecti on, ensuring the model remains explainable. The multivariate approach detects not only outliers but also abnormal data points that are within the acceptable range but appear unusual when considering other variables. Our package generates a user-friendly report that highlights anomalous subjects, visits, and findings, along with a detailed explanation of the detected pattern | outlier, clinical trial, imputation, missing value, clinical data, response variable, correlation, type i error, feature selection, data validation rules, feature importance, study day, regression analysis, likelihood function, data cleaning, cdisc | source presentation source |
9 |
| 32 | Beyond the Hype: A Pragmatic Approach to Leveraging AI/ML for Data Review | Robert Musterer | There is still a lot of hype around the potential of AI/ML to transform the life sciences industry and while advancements in clinical development have accelerated in recent years, the adoption of innovative technologies has lagged. Hype aside, taking a pragmatic approach to AI/ML enables life sciences companies to adopt new capabilities and accomplish objectives in a more efficient, cost-effective manner. This session will cover how a top 20 Sponsor partnered with eClinical Solutions to leverage the AI/ML capabilities to enhance clinical data review processes for greater efficiency using fewer resources. With these capabilities, high quality data is produced in near real-time to support accelerated submission and collaboration with the FDA. We will discuss the pragmatic approach to adoption including how the AI-enabled data review use cases were selected and evaluated, the validation process for ensuring accuracy of the AI/ML models, and the resulting efficiency gains | data review, clinical trial, clinical data, concomitant medications, data science, adverse events, data cleaning, data governance, data validation rules, life sciences, clinical data management, proc freq, medical history, outlier, training set | source presentation source |
10 |
| 33 | Bioinformatics Tools in Digital Health and Remote Monitoring | Sugumaran Muthuraj, Barath Baskaran |
Digital Health and Remote Monitoring have emerged as transformative approaches in healthcare, offering innovative solutions for personalized patient care, data-driven insights, and enhanced disease management. These have also given rise to technical challenges include interoperability issues among diverse de vices, data security concerns, and the need for robust connectivity in remote settings. In addition, ethical and regulatory challenges involve patient data privacy, informed consent in data sharing, and adherence to evolving regulatory standards. Bioinformatics tools can play a pivotal role in addressing these challenges and optimizing the potential of digital health and remote monitoring. By leveraging advanced algorithms and data analytics techniques, bioinformatics aids in deciphering complex pa tient data collected from diverse sources | clinical research, privacy | source presentation source |
7 |
| 34 | Bridging the Gap Between the pharmaverse and Company Internal Resources | Kangjie Zhang | This paper explores the integration of open source R packages to leverage community-developed tools, bridging the gap between the pharmaverse (with a focus on the {admiral}, {metacore}, and {metatools} packages) and Bayer internal resources to create ADaM dataset. Three repositories are created to address various facets of this integration: Testing: Assess packages using real studies, identify and categorize issues as either company-specific or general. Company-specific issues will be documented internally, general concerns will be shared with the open source community, fostering collaboration and improvement. Knowledge Sharing Website: Document the comparison of internal macro libraries and pharmaverse, track ongoing issues, provide workarounds, templates, and use-cases. * Internal R Package: Develop an internal toolbox of functions tailored to company-specific standards and metadata structure | out=, excel, adam dataset, macro definition, cdisc | source presentation source |
8 |
| 35 | Build an Effective and High-Performance Team | Helen Yang, Richa Kala |
People management is the process of hiring, leading, and developing team members to support the growth trajectory of an individual and build teams to align with the organization's overall mission. This paper will discuss different approaches on how to ramp up new hires, develop optimal skill sets, and manage resources across different time zones to optimize performance. We will share experiences particularly from COVID-19 vaccines trial within the last 3+ years | data science, clinical trial | source presentation source |
4 |
| 36 | Building a Global Workforce Ecosystem: Unlocking the True Potential of Offshore Teams | Bhavesh Patkar | regression analysis, customer satisfaction | source presentation source |
9 | |
| 37 | Case study of verification and validation methods for algorithmic predictive biomarkers | Kartik Shetty | Digital Health Technology (DHT) is an emerging field; known for its ability to make healthcare delivery to patients efficient. One application is the development of predictive biomarkers through the application of algorithms on data from wearable devices, enabling the prediction of health conditions; leveraging this potential can accelerate drug development. However, both data and algorithms need to meet regulatory requirements with regards to consistency, accuracy, traceability, and validity to be part of a clinical study. The focus of this paper is to investigate the verification and validation methods for algorithms that are utilized in the formulation of biomarkers. As a part of PHUSE "Emerging Trends & Technologies" working group's project "DHTs", we present a case study where heart rate variability based predictive biomarker is devised using an algorithmic approach. The objective is to develop a framework to guide the use of DHT in clinical trials for comparable scenarios | data cleaning, logistic regression, proc hpforest, regression analysis, clinical trial, f1 score, logistic distribution, clinical research | source presentation source |
4 |
| 38 | CDISC Conformance Rules and the CORE Engine: Continuing the Road to Adoption | Peter Van Reusel | CDISC Conformance Rules are an integral part of the Foundational Standards and serve as the specific guidance to Industry for the correct implementation of the Standards in clinical studies. The overall goal of the CORE Initiative is to provide a governed set of unambiguous and executable Conformance Rules for each Foundational Standard, and to provide an open-source execution engine for the executable Rules which are available from the CDISC Library. This presentation will begin with a brief review of the CDISC CORE program concept --- what it is and why it is important. Progress and status of development of the CDISC Conformance Rules and the CORE Engine will then be covered. The focus then shifts to uptake and implementation to date by the CDISC community including Pharma and Biotech end-users, software ve ndors offering CORE solutions, and Regulatory Agencies. Finally, the CDISC Conformance Rules governance model and the CORE roadmap will be discussed | data cleaning, clinical trial, cdisc | source presentation source |
7 |
| 39 | Change Management: A futuristic opportunity for SAS Programmers to build R Library | Shashikant Kumar, Ephicacy Lifescience |
Pharmaceutical industries are rapidly evolving especially after Covid -19 pandemic. A race has begun to develop time-bound and cost -effective treatments for safeguarding patient's life. Retaining talents without compromising quality and cost has been challenging in recent time. Often, we hear the talks, that AI based tool such as ChatGPT gaining popularity in healthcare industry, and it may take over SAS programmers in clinical industry which appears to be a myth. Leadership could focus on existing SAS Programmers building expertise on open -source software such as R, Python etc. to convert the library of existing standards programs developed over the years using R, Python etc. This novel approach allows SAS Programmers to build expertise using new technologies in more efficient manner and help pharmaceutical companies save cost on licencing which needs renewal on yearly basis. Discussion on some real time strategies would help leaders or SMEs to adapt the change more effectively | statistical programmer, data science | source presentation source |
6 |
| 40 | ChatGPT and AI/ML-Empowered Websites for All | Jayapandian Nagamalaiyan | In today's fast -paced world of pharmaceuticals and clinical research, the integration of AI/ML has become pivotal in optimizing daily productivity and efficiency. This abstract explores the transformative impact of ChatGPT and AI/ML - powered websites on our daily activities. ChatGPT, a versatile AI language model, provides real -time assistance in formulating queries, generating code, and interpreting complex regulatory guidelines, streamlining communication and decision-making within teams. Furthermore, we delve into the landscape of AI/ML -driven websites, offering a transformative approach to enhancing operational efficiency across industries. ChatGPT collaborates as digital assistant, aiding in formulating complex queries, drafting emails, content generation, turbocharging productivity and saving time | clinical trial, clinical research, adverse events | source | 5 |
| 41 | Choose Your Own Adventure: A Data Story | Jiwan Heo | adverse events, pharmaceutical industry | source presentation source |
10 | |
| 42 | Clinical Project Management - Where Art, Science, and Craft Meet | Hayley MacKenzie | Effective Clinical Project Management demands a high degree of forecasting, planning, organization, coordination, and controlling, especially when overseeing multiple clinical trials. It is a balancing act of, Art, encouraging creativity, Science, with universally accepted principles, and Craft, making connections through experiences. In this dynamic clinical research world, it is essential to efficiently allocate resources, track progress, budget, and timelines, in real- time, and ensure compliance with regulatory standards. This paper highlights the importance of identifying goals, arranging priorities, communicating effectively with the study team members, and ultimately mastering the skill of time management & prioritization, organization, communication, and streamlining processes | excel, clinical trial | source presentation source |
9 |
| 43 | Codeless Queries: Harnessing AI to Translate Clinical Trial Questions into Custom Listings Through Intuitive UI Interactions | Aditya Gadiko | The traditional methods of analyzing clinical trials data, often involving complex coding processes, have long been a bottleneck in deriving insights efficiently. This paper introduces a groundbreaking AI-driven solution that revolutionizes the approach to clinical trials data interpretation. By leveraging a human-in-the-loop approach, our innovative UI permits 80% code completion without direct coding. Subject matter experts can articulate questions in English, which are then translated into UI blocks. These blocks capture the essence of the query, and if machine comprehension falters, experts can iteratively refine the machine's understanding. The final outcome: tailored custom listings generated with precision. This method accelerates data analysis and fosters a more intuitive interaction with data, particularly pertinent for those with a clinical trial statistical programming background | clinical trial, data science, clinical data, medical history, data cleaning, machine learning, concomitant medications, privacy, clinical research | source presentation source |
7 |
| 44 | Collaborative Development of a Clinical Trial Data Anomaly Detection Tool within an Industry Consortium | Pekka Tiikkainen | Identifying anomalies early in on -going clinical trials data is crucial for ensuring the integrity and quality of data for submissions. However, this can be a time -consuming and challenging task, often requiring extensive manual review. To improve this, the IMPALA consortium, a collabor ation of world -leading bio-pharmaceutical companies, is working to cooperatively develop an anomaly detection tool originally developed at Bayer for flagging timeline outliers at study sites and subjects. Our vision is to make the tool a widely accepted be st practice, used by both the industry and the regulatory authorities alike. In an effort to promote transparency and collaboration, the tool will be made publicly available as an open -source R package in early 2024. As the IMPALA members continue to enhan ce the tool, FDA and other authorities are encouraged to join the project. Together, we could enhance clinical trial outcomes, and ultimately, improve patient care | clinical trial, clinical data, time series analysis, outlier, data cleaning, quality assurance | source presentation source |
6 |
| 45 | Comparing Analysis Method Implementations in Software (CAMIS): An Open-Source Repository to Document Differences in Statistical Methodology Software | Soma Sekhar Sriadibhatla, Vikash Jain |
Poster | data cleaning, survival analysis | source | 1 |
| 46 | Comparing the Performance of Propensity Score Matching Algorithms from Multiple Programming Languages | Catherine Briggs | With the increased use of real-world data in clinical and healthcare settings, having a comprehensive comparison of propensity score matching (PSM) algorithms available to researchers is vital. In this paper, we will discuss different avenues for PSM and show the strengths and weaknesses of each using simulated data. Secondly, the paper will demonstrate the effect of the matching algorithms on estimating the causal effect of treatment on an outcome. We will cover statistical performance of the algorithms in matching and estimating casual effects | propensity score, propensity score matching, proc step, treatment arm, treatment effect | source presentation source |
7 |
| 47 | Comparison of Standard Log-Rank Test with Selected Alternatives | Elena Valkanova, Bulat Fatykhov |
In oncology studies with time-to-event outcomes, the hazard and survival functions are two ways of determining survival distribution; comparing the survival distributions using the constant hazards assumption could provide misleading conclusions. The clinical trial designs with two groups under the proportional hazards (PH) assumption is likely to be underpowered to detect treatment effects. When the PH assumption is violated, the likelihood estimator depends on the underlying censoring time distribution. The objective of this paper is comparing selected alternative methods for survival analysis when the assumption for the proportional hazards is questionable. We discussed three selected methods: log-rank (LR), Max-Combo (MC) and restricted mean survival time (RMST) applied to simulation data and real clinical trial data supported with graphical presentations using R. In the final section we provide guidance in choosing an alternative method to the standard log- rank test | rank test, log-rank test, exponential distribution, proportional hazards assumption, sample size, weibull distribution, overall survival, treatment effect, type i error, sample, randomized controlled trial, clinical trial, survival function | source poster source |
18 |
| 48 | Comprehensive Evaluation of Large Language Models (LLMs) Such as ChatGPT in Biostatistics and Statistical Programming | Toshio Kimura, Michael Pannucci |
Generative artificial intelligence using large language models (LLMs) such as ChatGPT is an emerging trend. However, discussions using LLMs in biostatistics and statistical programming have been somewhat limited. This paper provides a comprehensive evaluation of major LLMs (ChatGPT, Bing AI, Google BARD, Anthropic Claude 2) in their utility within biostatistics and statistical programming (SAS and R). We tested major LLMs across several challenges: 1) Conceptual Knowledge, 2) Code Generation, 3) Error Catching/Correcting, 4) Code Explanation, and 5) Programming Language Translation. Within each challenge, we asked easy, medium and advanced difficulty level questions related to 3 topics: Data, Statistical Analysis and Display Generation. After providing the same prompts to each LLM, responses were captured and evaluated. For some prompts, LLMs provided incorrect responses, also known as "hallucinations" | data science, statistical programmer, clinical programming | source presentation source |
9 |
| 49 | Confirmation and Cancellation - Best Strategies for Data Collection and Handling within Solid Tumor Studies Using RECIST 1.1 and iRECIST Guidelines | Sherry Ye, Abraham Yeh |
As immuno-oncology compounds are getting popular in clinical studies, RECIST 1.1 and iRECIST are the primary guidelines to handle tumor response assessments. From data management, what are the best ways of collecting the assessment without making mistakes? From a statistical perspective, what are the easiest ways to process the confirmations if the protocol is asking for confirmed assessments for the Best Overall Response? What can be done if there are better assessments after the iUPD in iRECIST which need to be cancelled out while deriving iPFS? This paper presents three main categories from the end-to-end approach on (1) Is it possible to start these processes earlier while collecting data? (2) What is the best approach to handle the confirmation? And finally (3) What is the best approach to handle the cancellation? We will think through the potential of the collection and metadata-driven data handling along the way | best response, data governance, randomization, clinical trial | source presentation source |
10 |
| 50 | Constructing Multilingual Visualisation Toolsets for Clinical Study Reports and Submission | Jane X. Liao, Bingjun Wang |
With the increasing demand for insightful data visualization deliverables for clinical study reports and submissions, robust and powerful data visualization tools are needed. In this paper, we illustrate development and implementation of a multilingual visualization toolset. The system includes a graphic repository with visualization standards, a visualization gallery, internally developed R packages, and SAS macros. Visualization standards define the figure layout, recommendations, and standard themes to promote operational efficiency and consistency across various clinical trials within the organization. The visualization gallery provides code in both SAS and R for commonly used figures. Additionally, R packages and SAS macros are developed based on our organization standards and branding for frequently requested figures | clinical trial, proc sgplot, data visualization, forest plot, macro definition, csr, sample, proc step | source presentation source |
13 |
| 51 | Data Designer: Empowering data mapping, manipulation and analysis for advanced insights and decision making | Mukul K Mittal | Data analysis and visualization are critical aspects of clinical trials decision making. Data Designer, a R Shiny application offers a comprehensive set of functionalities to load data from different sources (sas, xpt, CSV, xlsx, json etc.) and perform basic and complex column/row transformations. The application allows users to interactively manage outliers, duplicate values, variable types (floats, dates, string), remove invisible characters, etc. The interface makes it easy to remap variables to specified format, create new rows and columns. Data can be transposed based on key variables and multiple datasets can be merged horizontally or concatenated vertically. In summary, Data Designer is a comprehensive data manipulation tool designed to simplify and accelerate data analysis process |
clinical trial, data science, proc transpose, outlier | source presentation source |
6 |
| 52 | Data Handling through CDISC Standards in Master Protocols: A Focus on Rescreened and Re-Enrolled Subjects | Swathi Kotla, Madhusudhan Reddy Papasani |
Master protocols are emerging as a promising design for clinical trials that investigate multiple hypotheses with concurrent sub-studies. These protocols offer efficiency in conducting trials by including multiple populations, incorporating various treatments, or allowing the addition and removal of treatments during the trial. However, implementing these protocols introduces scenarios where subjects are screened or enrolled in multiple sub-studies. This paper provides practical examples of data collection and handling for subjects who re-screen and re-enroll in master protocols, including basket, umbrella, and platform trials. The collected data is transformed to SDTM domains such as DM and DC, followed by data handling for analyses using ADaM datasets for subject-level data in ADSL, ADTP, and laboratory data in ADLB. Overall, the paper offers insights into the data handling for re-screened and re-enrolled subjects in master protocols using CDISC standards | adam dataset, laboratory data, sdtm programming, annotated crf, sdtm domain, cdisc | source presentation source |
7 |
| 53 | Data Handling with SAS Access Engine to Handle all your Data Engineering Needs | Pritesh Desai, Samiul Haque |
Presentation | proc sql, relational database, proc step, sample, dataset, data mart, data warehouse | source | 16 |
| 54 | Data Pipeline - Improving Cycle Time and Shifting Data Closer to Analytics | Jayasree Ivaturi | Clinical trials DM and Analytics processes are facing challenges due to the increasing data diversity and sources. Process of sourcing data, managing quality, review, standardization, and statistical programming for regulatory submission involves multiple siloed systems and extensive manual interventions; risking synergies, data integrity and traceability in critical last mile. Our fully automated end -to-end data pipeline offers timely data access to consumers across the data -to-outcome analysis process with quality, integrity, and end-to-end traceability from source to analysis. These data pipelines bring data from various vendors at a set frequency; verify the quality and adherence of incoming data to standards; and offer oversight of data review observations between the DM and vendors until resolution. Finally, the refreshed data is channeled through the data pipeline to the st atistical programmer on demand for regulatory analytics | data pipeline, clinical trial, statistical programmer, data cleaning, data integrity, regulatory analytics, regulatory submission, clinical programming, clinical data, data governance, data science, raw data | source presentation source |
6 |
| 55 | Data Quality Evaluation of EHR FHIR API Real-World Data for FDA and Pharmaceutical Industry Use for Analysis and AI Implementation | Sarah Ferko, Abhishek Pandey, Matthew Deady, Brian Goodness |
The FDA and pharmaceutical industry's pre-market and/or post-market activities could possibly be improved by using real-world data (RWD) from healthcare organizations' electronic health record (EHR) databases and application of artificial intelligence (AI) tools. In the U.S., this aim has been traditionally hindered by the inability to exchange clinical data between healthcare providers, pharmaceutical companies, regulatory agencies, and public health organizations in an interoperable format. However, the recently enacted 21st Century Cures Final Rule requires all healthcare providers to support seamless and secure access, exchange, and use of EHR data through an interoperable FHIR (Fast Healthcare Interoperability Resources) standard API endpoint | clinical trial, data cleaning, adverse events, unstructured data, clinical data, regulatory submission | source presentation source |
10 |
| 56 | Demystifying Submissions to NMPA China | Hiral Raval, Dyuthi Yellamraju |
Presentation | data validation rules, data cleaning, label statement, data science, data standard, sdtm domain, cdisc | source | 19 |
| 57 | Developing Leadership, Fostering Collaboration and Building Perspective Through External Communities | Harper Forbes | External communities of data science range from local coding clubs of a dozen members to thousands across the globe (i.e., PHUSE). Collaborative communities such as PHUSE span a dynamic pharmaceutical industry and have driven innovation, established new ways of working, addressed the implementation of standards and continuously share ideas with health authorities. Participating and contributing to these communities can help individuals develop professionally and holistically across industry - they provide opportunities to lead and connect with other talented data scientists while expanding their pharma knowledge and perspective. This paper will highlight some of the professional development benefits including growing leadership skills, forming effective collaborations and building industry awareness | pharmaceutical industry, data science | source presentation source |
2 |
| 58 | Digital Protocol Vision ... How Digital Information Can Transform and Automate Our Processes | Bill Qubeck, Chris Decker |
Presentation | clinical trial, csr, clinical database, clinical trial protocol, machine learning, artificial intelligence, clinical data | source | 27 |
| 59 | Digitisation of Specifications - Brave New World for Analysis Definitions | Rebecca Bates, Krishna Gopal |
controlled terminology, data science, adverse events, cdisc | source presentation source |
6 | |
| 60 | Does One Size Fit All? | Bidhya Basnet | Submitting data to the Health Authorities is a critical part of our business to obtain approval for our medicines. Currently, three agencies, namely FDA (USA), PMDA (Japan), and NMPA (China), require clinical trial data for filing. Preparing one set of data packages will not satisfy all agencies ' needs. i.e., Variations DO exist. Hence, understanding the agency-specific requirements could help us to s ubmit quality data packages to facilitate efficient review to get our medicines approved faster and get them to the patients sooner. This poster demonstrates these variations across these agencies with respect to the submission requirements and recommendations. This is based on our interpretation of the data requirements across the agencies. It also highlight s the importance of being aware of these variations to help adequately plan time, resources, and contents when considering filing to multiple agencies in parallel | adam dataset, data cleaning, cdisc | source poster source |
4 |
| 61 | Dynamic CDISC Analysis Results Data with the R Package {cards} | Daniel Sjoberg | clinical trial, rest api, r package | source poster source |
5 | |
| 62 | Dynamic document creation framework for streamlining cSDRG | Umayal Annamalai | cSDRG provides guidance and assistance to the clinical reviewers to review and evaluate the clinical trial data. This critical document for regulatory submission is created manually based on templates and guidelines. This leaves it prone to errors and lack s efficiency in the process. There is a clear need for the adoption of automated and reproducible approaches in regulatory documentation to ensure reduced error rates, transparency, improved document consistency, time savings, and regulatory compliance. Hence, we are proposing a novel approach using R Markdown to streamline the cSDRG creation process by pre -populating the cSDRG based on a dynamic document generation framework. R Markdown allows the integration of metadata details, required data, and narrat ive text within a single document, enabling automated and consistent generation of cSDRGs across different studies and analysis scenarios | clinical data, clinical trial, sample, extract | source presentation source |
7 |
| 63 | Effective Workflow Process: Rapid SDTMs and Compliance Dashboards | Sunil Kumar Gupta, Harinarayan Gopichandra |
Now that the pharmaceutical industry has established a formalized process for SDTM development, do standards and compliance dashboards seamlessly integrate into your workflow with each milestone (IA, DSMB/IDMC, DBL)? The utilization of a meticulously designed product, equipped with traceability and controls, empowers teams to standardize SDTMs, thereby eliminating the need for redundant programming and effectively reducing associated costs. The incorporation of a unified raw data to SDTM mapping layer not only results in cost efficiencies but also expedites the creation of SDTMs that meet stringent quality assurance (QA) standards and are well -prepared for downstream analysis. Proactive management of medical coding, controlled terminology, adherence to CDISC requirements, resolution of clinical data issues, m itigation of paired variable inconsistencies, and oversight of derived variables collectively contribute to a streamlined and highly efficient operational framework | data cleaning, controlled terminology, raw data, sdtm domain, sdtm programming, cdisc | source poster source |
5 |
| 64 | Effortless Table Generation: A SAS Macro for Streamlining Your Workflow | Zhen (Laura) Li | This paper presents a SAS macro tool designed to streamline table generation, a critical task in clinical data analysis. It simplifies the process, allowing users to create tables with diverse templates featuring descriptive statistics, counts, and percentages. Practical examples showcase its capabilities. The tool is user-friendly, requiring users to populate an EXCEL file and execute the macro. Its versatility stands out as it can handle raw data, SDTM data, ADaM data, or any combination thereof. For junior programmers, this macro is a game-changer, enabling quick, high-quality output generation and validation. The paper also covers design principles and programming challenges in the macro's development, offering insights into coding intricacies and best programming practices. In conclusion, it explores potential enhancements and future applications for this SAS macro, emphasizing its role in optimizing data analysis workflows in clinical research and promoting efficiency |
raw data, create table, macro variable, excel, proc means, macro conditional, data cleaning, proc step, out=, macro definition, proc sql, adam dataset, data science, macro loop, cdisc | source presentation source |
12 |
| 65 | Electronic Clinical Outcome Assessment (eCOA) in SDTM: Challenges and Solution | Shivam Kumar | Electronic Clinical Outcome Assessment (eCOA) is a method of capturing outcomes data electronically in clinical trials. eCOA employs technologies such as handheld devices, tablets, or the web to allow trial participants, physicians, and caregivers to directly report information related to healthcare outcomes. Data can be submitted by participants from home and integrated with the trial's electronic data capture (EDC) system to reduce the burden on clinical site coordinators to manually input patient information. Because of the different sources, volume, and complexity of eCOA data, it creates a challenge for Study Data Tabulation Model (SDTM), team to map it into standard domain without losing any information from raw data. In this paper, we will discuss the aftermath of receiving eCOA data, challenges for SDTM team to map this data and the solution. This presentation will also emphasize analysis of raw data along with pre planning to achieve streamlined and efficient SDTM output | raw data, sample, data cleaning, cdisc | source presentation source |
5 |
| 66 | Elevating Brands: Empowering Employees as an Ambassador | Supriya Moghe | In the dynamic era of diverse branding costs and information overload, the challenge for brands is to remain memorable. This article advocates a shift in strategy-harnessing employees as brand ambassadors for effective digital promotion. Employee branding, transforming each team member into a promotional medium, emerges as a strategic solution. Ensuring satisfaction through innovative initiatives and flexible policies becomes crucial for successful employee branding. The link between employee satisfaction and brand promotion is clear-unsatisfied employees won't champion a brand, and satisfied employees are more likely to stay. Exploring avenues like word-of-mouth publicity, talk-shows, and social networking, the article highlights these platforms as enablers for brand advertising. The multiplier effect of advertising efforts is evident, especially in managing attrition | source presentation source |
2 | |
| 67 | Eliminating Duplicative Maintenance of Statistical Programs In Parallel Analyses Using a Single Control File | Kavitha Boinapally | Clinical study programming involves working on multiple deliverables in parallel with different source data. This can be efficiently achieved by using a standard folder structure and single master programming environment set-up file. With use of macro variables and SYSUSERID, the switch between analyses, randomization files, blinded and unblinded (restricted) folders can be seamless without duplicating the programs in various folders. Leveraging SYSUSERID adds additional security. A global macro variable to indicate deliverable type and custom macro variables for each such type help maintain a single set of SDTM, ADaM, and sometimes TLF programs to generate datasets and outputs in different analysis folders with different source data. Archiving all programs and files in a zip file in a snapshot folder ensures easy retrieval and rerunning of deliverables (within that snapshot folder) without disruption to the regular study area. This approach can enhance productivity and data quality | macro variable, randomization, clinical trial, out=, raw data, data cleaning, %if-%then-%else, cdisc | source poster source |
4 |
| 68 | Eliminating SDTM Double Programming by Using a Validation Tool and Dummy Data | Glenn Borgmans, Tine Seeuws |
Validation of data conversion from EDC systems into SDTM compliant SAS datasets remains a hurdle in the database set-up phase. Often SDTM double programming is used to validate the conversion. This validation is crucial to ensure that the structure and content of the database are compliant and that all variables from the (e)CRF are present and mapped correctly. An in-house developed SDTM data validation tool and unique approach guarantee exactly that. SGS eliminates SDTM double programming by performing 100% validation through dummy data from test subjects entered in the training eCRF. Data programmers and data managers work simultaneously, speeding up database creation and validation to have the database ready shortly after FPI. This paper gives insights in our validation approach, environments used, the tool, including live issue tracking and solving feature, automated validation issue report and the ability to perform a QC per SDTM domain | data cleaning, independent qc, clinical trial, sdtm domain, cdisc | source presentation source |
8 |
| 69 | Embracing the Future of Digitisation: Uninterrupted Data Flows from Protocol Design to Electronic Data Capture | William Illis, Diana Ye |
Presentation | clinical research, clinical trial, likelihood function | source | 18 |
| 70 | Empowering Excellence: Strategies for Cultivating a Culture of Bold Decision-Making in Teams | Mark Bynens | In a world that values innovation and agility, the role of individual empowerment within teams has never been more critical. This presentation unveils the transformative techniques and strategies to instill a robust culture of empowerment, where team members confidently own their roles and make decisions with assurance. We will delve deep into communication paradigms, growth mindset fostering, and leadership models that celebrate and champion team member autonomy. By illuminating real-world examples and success stories, attendees will grasp how empowered teams can become a game-changer, driving innovation, creativity, and efficiency. Step into a future where each member is an active decision-maker, propelling the team forward with collective expertise and individual confidence. Embrace the tools and insights that not only build a positive work atmosphere but also set the stage for organizational success in our rapidly evolving business landscape | clinical programming, clinical trial, clinical research, pharmaceutical industry, data science | source | 8 |
| 71 | Empowering Healthcare Through Generative AI | Ravindra Shukla, Aman Bahl |
Generative AI has the potential to empower healthcare in various innovative ways. By leveraging the power of machine learning and deep learning algorithms, generative AI can contribute to improving patient care, disease diagnosis, drug discovery, and personalized medicine. Artificial intelligence (AI) has been transforming health and medicine for several decades, whether through diagnostics, personalized medicine, data analytics or other innovations. The advent of generative AI is now reshaping drug discovery, diagnostics, clinical decisions and delivery of care. Generative Artificial Intelligence (AI) is emerging as a disruptive wave with implications that span across diverse business sectors. As one of the most rapidly growing advanced AI technologies, generative AI offers valuable opportunities ranging from personalised health programs to pivotal healthcare cost reduction strategies | clinical trial, artificial intelligence, clinical research, deep learning | source presentation source |
7 |
| 72 | Enabling the Breakthroughs of Tomorrow: The Data Lakehouse for Clinical Trials | Venu Mallarapu | Clinical trials continue to grow in complexity amid decentralization, digital trials, and ever- increasing surge of data from multiple external sources. Meanwhile, leaders in life sciences are under immense pressure to reduce cycle times, gain oversight, and mitigate issues and delays in the clinical data lifecycle. Modern research requires a scalable clinical data lakehouse architecture. In this session, attendees will learn about the key components of a clinical data lakehouse, why it's the ideal compared to the data lake or data warehouse, and what is different about the data lakehouse in a regulated environment. Find out how the data lakehouse blueprint supports the use of advanced techniques like AI/ML to accelerate data science and secondary use of data, beyond the submission pipeline | clinical data, data lake, data lakehouse, clinical trial, life sciences, unstructured data, clinical research, machine learning, data warehouse, data wrangling, data cleaning, access control, data ingestion, data governance, data science | source presentation source |
12 |
| 73 | Enabling the Interoperable Data Ecosystem with the elluminate Clinical Data Cloud | Nathan Johnson | Presentation | clinical trial, rest api, clinical research, data wrangling, sdtm programming, data science, sample, cdisc | source | 15 |
| 74 | Enhancing PK Data Exploration and Visualization by using a R Shiny Application | Runcheng Li, Hanyang Sun, Yiyuan Zeng, Jiannan Kang |
This paper introduces an R Shiny web application for PK data exploration and visualization. This tool offers a range of powerful features, including rapid data review, checklist-driven validation, dynamic plot generation, and downloadable reports for future use. By leveraging R Shiny's interactive capabilities, the tool streamlines workflows, enables real-time data interaction, provides comprehensive and efficient review, checking, and visualization for PK analysis. The user - friendly interface and dynamic features empower researchers in foster ing informed decision-making and expediting research outcomes without the need of complex programming scripts | data review, correlation, dataset, data cleaning | source presentation source |
7 |
| 75 | EOSS_MMRM V1.0: A Shiny Application for Estimand-Oriented Sample Size and Power Simulation Using MMRM Modelling | Xiaoli Kong, Jing Dai |
The ICH E9 (R1) addendum on estimands and sensitivity analysis in clinical trials has been adopted by many regulatory agencies since 2019. Sample size and power calculation, as a critical component in trial design, need to be aligned with the chosen estimands and intercurrent event (ICE) strategies that reflect trial objectives. A Shiny application (EOSS_MMRM V1.0) has been developed to perform estimand oriented sample size and power simulation for clinical trials that use mixed-effect model for repeated measures (MMRM) for their primary endpoint analysis and handle ICEs using treatment policy or hypothetical strategy. EOSS_MMRM V1.0 allows the specification of expected efficacy at post-ICE follow-up visits as well as missing data pattern and extent | baseline, sample size, sample, missing value, hypothetical strategy, clinical trial, while-on-treatment, imputation, primary endpoint, treatment arm, seed, sensitivity analysis, correlation | source presentation source |
8 |
| 76 | Evaluating Safety Compliance in Clinical Trials by Leveraging Patient Narratives and Deep Learning | Sherrine Eid, Sundaresh Sankaran |
Adverse Events (AEs) during clinical trials require identification, classification, risk quantification and treatment. Information embedded within clinical trial narratives can enhance understanding and insights. Leveraging best-in-class Artificial Intelligence (AI) methods helps us accurately and precisely assess Adverse Events, safety signals and physician compliance from patient narratives. Our analysis ensures v aluable insights and productivity gains from a derived structured table with high accuracy and precision, reflected in time savings of hours versus weeks | clinical trial, deep learning, machine learning, adverse events, sas viya, artificial intelligence | source presentation source |
9 |
| 77 | Expanding Collaboration and Innovation in R&D: Leveraging the Power of Open-Source Tools | Ellis Hughes | Departmental silos and disjointed communication channels often impede effective collaboration, decelerate decision- making processes, and hinder innovation. In this presentation, I explore the transformative potential of open-source tools in fostering seamless communication, empowering end users across teams, and promoting cross-functional collaboration. By delving into the inherent flexibility, scalability, and adaptability of open-source platforms, we will demonstrate how organizations can harness and build upon these tools to automate processes, enhance interdepartmental communication, and facilitate knowledge sharing. Furthermore, we will discuss the role of open source in driving innovation by enabling rapid prototyping, iterative development, and the integration of cutting-edge technologies. Discover how the strategic implementation of open-source tools can dismantle silos and supercharge your organization | pharmaceutical industry, data science | source presentation source |
3 |
| 78 | Exploring the Potential Use of Generative AI in the Regulatory Review Process | Qingying Ally Lu, Kathryn Matto |
Presentation | sample, machine learning, adverse events, artificial intelligence, model explainability, quality control, regulatory submission | source | 19 |
| 79 | Fidelity Assessment of Real-World Data (RWD) as an External Control Arm | Mohamed Sameer, Umayal Annamalai |
Real-world data is being extensively used across different areas of clinical research, including past personalized medicine, patient recruitment, treatment effectiveness, and post -approval safety evaluation. Furthermore, the FDA's commitment to considering real -world evidence in regulatory decision -making emphasizes the emerging role of real-world data in clinical research. An important though niche use of real -world data is to serve as a control arm (CA) in clinical trials, which can be a reliable alternative to randomized control trials. However, real -world data presents a unique set of challenges in its use as an external control arm due to the 3 V's - Volume, Velocity, and Veracity. Hence, despite a multitude of advantages, it is evident significant planning and efforts are needed to leverage the power of real-world data as a control arm | clinical trial, clinical research | source poster source |
5 |
| 80 | Fight the Power - Supporting Study Teams with Niche R Packages for Delivering Statistical Packages | Benjamin Arancibia, Alice Ehmann |
data cleaning, clinical trial, r package | source presentation source |
4 | |
| 81 | From Data Iron to Data Gold | Ashwini Yermal Shanbhogue | In recent years, Real World Data (RWD) has gained importance in the clinical trial space. One of the sources of RWD is claims data. Claims data is usually present in large datasets, which until recently were hard to process and gain any meaningful insights from. However, recent advances in Machine Learning (ML) have made this task easier. In this poster, I will demonstrate how an R package that includes several ML algorithms can be used to make predictions and thus gain insights from a claims data set, the CMS 2008- 2010 Data Entrepreneurs' Synthetic Public Use File (DE - SynPUF) | insurance claims, clinical trial, clinical data, data cleaning | source poster source |
9 |
| 82 | From Missing to Meaningful: Tackling Partial Dates in Clinical Trials | Bhargav Koduru | Handling partial or missing dates is a recurring challenge in clinical trials, often necessitating imputation for key analyses like identifying treatment -emergent events. While many sponsors default to their Statistical Analysis Plan (SAP) for guidance, a universal approach is still lacking. Leveraging the existing guidance laid out in the PHUSE's 'SDTM ADaM Implementation FAQ,' we fine -tuned a more robust data imputation guideline using de-identified clinical data from approximately 50 clinical trials (eligible for Medidata analysis through Medidata's data collaboration program). Our paper will present these optimized guidelines, discussing specific scenarios, critical variables, and common pitfalls. This enhanced summary of scenarios will provide a customizable foundation for date imputation that can be further enhanced to fit your trial needs, bolstering the integrity of clinical trial data for regulatory submissions | imputation, clinical trial, data science, clinical data, cdisc | source poster source |
4 |
| 83 | From stowaway to admiral, a R-iveting tale of a hierarchical composite endpoint's journey from idea to dataset | Nicole Major, Otto Magnusson |
Hierarchical composite endpoints (HCE) are complex endpoints that combine outcomes of distinct types and clinical importance into a single ordinal outcome that prioritizes the clinically most important (e.g., most severe) event of a patient. Analysis of these endpoints employs win statistics and maraca plots for visualization. To streamline HCE analysis, the structure and programming of the ADaM dataset requires evaluation. Developing an analysis dataset adhering to CDISC standards while minimizing programming complexity is crucial due to the novel and intricate nature of HCE analysis. The paper addresses dataset programming, suggesting the use of R's hce and maraca packages for win statistics and visualization. Proposing an end-to-end framework, the paper demonstrates how to create the ADHCE dataset using R's admiral package, contributing to a comprehensive approach for analyzing these endpoints | in=, missing value, adam dataset, sample, study day, data cleaning, cdisc | source presentation source |
16 |
| 84 | Front-Loading Statistical Programming Using Automated Synthetic Data | Faezeh Moshfeghifar, Hemanth Padmakar |
An efficient submission of clinical trial data is crucial to successfully introducing new treatments to patients in need. Novo Nordisk's "Fit4Take-Off" aspiration aims to submit a New Drug Application to the pre-defined four markets within four days of the Last Patient Last Visit (LPLV). To support this aspiration, Novo Nordisk has developed an in -house application called Kassandra, which automatically generates trial -specific synthetic data for clinical trials across therapeutic areas. Kassandra's synthetic data generation capabilities offer a solution to streamline clinical trial setup processes. It helps the Electronic Data Capture (EDC) system check procedures, ensuring a robust and adaptable data entry approach surpassing traditional methods. It also enables the front-loading of statistical programming tasks ahead of the First Patient First Visit (FPFV), allowing statistical programmers to initiate these tasks earlier in the trial conduct phase | clinical trial, edc, sample, clinical data, data integrity, statistical programmer, clinical research | source presentation source |
5 |
| 85 | Future Clinical Data Submission Standards: CDISC, FHIR, OMOP, or Hybrid Model | Jeffrey M. Abolafia, Sarah Ferko, Ingeborg Holt |
There is a rapid increase in the use of real world data (RWD) and observational studies to support marketing applications. However, current submission standards for these non-traditional data sources have not evolved. All study data must be submitted using CDISC standards. While CDISC standards are optimized for randomized clinical trials (RCTs), alternate standards may better be a better fit for other study designs. This presentation examines the future of data submission standards. First, we present an environmental scan of possible standards for submitting RWD and RCT data. Second, we discuss the pluses and minuses of submitting data using CDISC as opposed to alternative standards. Third, we look at the pros and cons of submitting data under a single versus a hybrid approach, where each data type is submitted using the standard for which it is best suited. Finally, we provide short term and long recommendations for submitting all types of study data | clinical data, clinical research, data standard, insurance claims, regulatory submission, data cleaning, adam dataset, clinical trial, data validation rules, cdisc | source presentation source |
10 |
| 86 | Future-Proofing Clinical Data Sciences: The Pivotal Role of Therapeutic Area Proficiency | Sascha Ahrweiler | Understanding Therapeutic Areas (TA) is becoming increasingly critical for clinical data scientists in the pharmaceutical industry, particularly as artificial intelligence continues to revolutionize the field. This paper introduces an innovative AI-driven methodology, utilizing ChatGPT technology, for generating educational articles on TAs. These articles are intended for publication on the PHUSE Education website, serving as a peer-reviewed, continuously updated knowledge repository. Emphasizing the role of TA expertise as a career-defining skill in the next 5-10 years, the paper underscores the impending transition from manual programming to specialized data interpretation and analytics. The initiative aims to foster self-learning and knowledge dissemination, aligning with the PHUSE Education mission of creating proficient practitioners capable of teaching others, thus enhancing the data science community's competence in this critical domain | clinical trial, clinical data, data science, baseline characteristics, data governance, artificial intelligence, clinical research | source presentation source |
12 |
| 87 | Getting More Efficiency from your Instruments Using CDISC | Franklin E. Menius | It is common in clinical programming to reuse programs, macros, and snippets for efficiency in time and processes. Entire libraries of programs are stored to be taken "off the shelf" and plugged into new studies. In the emerging area of electronic Clinical Outcome Assessments (eCOA) and electronic Patient Reported Outcomes (ePRO), this translates into a library of standardized questionnaires that can be dropped into new studies for faster startup and quicker subject enrollment. This presentation will focus on expanding the efficiency of questionnaire libraries by applying CDISC standards to collected data. Utilizing the growing library of CDISC QRS supplements and existing questionnaire libraries, vendors can prepopulate SDTM variables with appropriate code lists and controlled terminology | clinical trial, controlled terminology, cdisc | source presentation source |
5 |
| 88 | Getting Started with Quarto | Kevin Putschko | Clinical reports have their own defined reporting styles, and that is well documented. But what about using Quarto Books or Quarto Documents to easily maintain metadata regarding clinical studies? This could include querying your data for quality control, keeping up-to-date listings of all the variables and variable summaries, or including interactive tables and plots to help researchers explore data for new insights. Quarto is the next generation of dynamic and repeatable reporting with R, blending code, visualizations, tables, and documentation. With the ability to create and reuse custom templates, maintaining the look and feel of clinical tables, figures, and listings has never been easier. In this poster, we will explore the transformative capabilities of Quarto with a brief introduction to the creation and use of custom templates. Page 1 of 10 | clinical trial, tables | source poster source |
10 |
| 89 | Getting Started with the New CDISC Analysis Results Standard | Bess LeRoy, Bhavin Busa, Richard Marshall |
The CDISC Analysis Results Standard (ARS) Team has developed standards to improve and facilitate the automation, reproducibility, reusability, and traceability of analysis results (Tables, Figures and Listings, aka TFL). The team created a logical model to incorporate the elements for both analysis results and associated metadata. This paper provides an overview of the CDISC Analysis Results logical model, including its background and development | adam dataset, adverse events, cdisc | source presentation source |
4 |
| 90 | Good Statistical Monitoring {gsm}: An R Package for Risk-Based Monitoring and Clinical Trial Analytics | Matt Roumaya, Jeremy Wildfire |
Risk-based quality management (RBQM) is a regulatory-recommended approach to manage risk in a clinical trial. While there is a significant body of literature describing RBQM implementation and best practices1,2, there are few publicly available open-source tools for RBQM available. The {gsm} R package provides a robust and open-source solution for evaluating risk signals at the site-level. RBQM monitoring can be a costly endeavor, and the {gsm} package is a fundamental component of a growing framework that allows for efficient, accurate, and timely assessment, visualization, and reporting of differences in quality across sites in a clinical trial | clinical trial, adverse events, r package | source presentation source |
5 |
| 91 | Graphical overview for understanding, planning and execution of statistical programming activities using Neo4j | Madhusudhan Reddy Mule Venkata | The world of statistical programming is changing dynamically to meet the complex requirements of clinical trials. As a statistical programmer, our involvement spans across the horizon of clinical trial conduct: from inputs to protocol creation, to finalization of end of text material in clinical study reports. Along with these, we are also associated with other activities like alignment of clinical studies within a project, resource planning & allocation etc. To get a good hold on the activities and understanding the dependencies, graph theory comes to our rescue with data visualization. Some of the examples applied include , displaying dependencies between SDTM and ADaM, Resource utilization in programming SDTM datasets/ ADaM datasets/ TFL outputs, interconnections between SOPs etc. This paper attempts to address the requirements by introducing graph theory and finding solutions using Neo4j applications | clinical trial, csr, cdisc | source presentation source |
13 |
| 92 | Guidance Beyond the SDTM Implementation Guide | Michael Beers, Blue Bell |
A common misconception among preparers of SDTM data seems to be that it is sufficient to just follow the SDTM Implementation Guide when creating the datasets. The truth is that it is more complicated than that. A preparer of SDTM datasets needs to be aware of all the industry guidance available when preparing for regulatory submission, from CDISC and the regulatory agencies, but also from other organizations as well. This presentation will discuss some of the lesser-known guidance in the industry and why they should be referenced, as well as some of the impacts of not using these documents in the creation of SDTM datasets | sdtm domain, sdtmig, clinical trial, data cleaning, medical history, clinical endpoint, data validation rules, cdisc | source presentation source |
10 |
| 93 | Harness the Data's Potential in a Clinical Trial Process | Eva Kelty, Catharina Dahlbo |
Clinical trials and medical data often have many dimensions, as they seek to represent and explain the complexities of biological processes. Using a graph database for handling complex and unstructured data, like clinical trials, can get flexibility across the entire lifecycle from planning to archiving. A repository containing many trials, built, and modelled for searching in data, can be used in different phases of a clinical trial such as planning, execution, reporting and archiving. Accessing data independent of the trial's outcome could be of interest for many purposes like planning new studies or preparing for discussions with authorities, where comprehensive understanding and explanation of results are vital. The poster will explore the significance of rapid and streamlined data access at various stages of the clinical trial process | clinical trial, relational database, unstructured data, life sciences | source poster source |
4 |
| 94 | Harnessing AI for Health Equity and Inclusivity in Risk-Based Monitoring | Finn Janson | hyperparameter tuning, sample, nominal variable | source presentation source |
12 | |
| 95 | HoW - Your First Submission in R | Kristian Trøjelsgaard, Steffen Falgreen Larsen |
Presentation | source | 100 | |
| 96 | How No-Code Visualisation is Fast Becoming an Asset for Data Engineering and Data Science Teams | Pritesh Desai | data science, exploratory data analysis, clinical trial, logistic regression, sas viya, regression analysis, logistic distribution | source presentation source |
6 | |
| 97 | How to Be a Neuro-Inclusive Leader | Laura Needleman | Neurodiversity is a fast-growing diversity category that we as an industry are just beginning to understand and incorporate into Diversity, Equity, and Inclusion (DE&I) plans. Neuro-inclusion means implementing practices that consider the neurodiversity of your employees. Leaders of our industry have the hard task of keeping up with changes in language and understanding of neurodivergent employees. Topics of this paper include discussing terminology and it's importance, how to support a neurodiverse team, and how to create an environment that brings out the best in everyone. I'll also discuss the benefits that come from being more neuro-inclusive in the workplace | source presentation source |
5 | |
| 98 | How to Monitor SDTM Data Health | Amanda Cross | Currently, the majority of SDTM conversions are completed at predefined timepoints during a study lifecycle, typically "just in time" before a reporting event. At these timepoints, data quality is assessed using methods such as double programming against a standard set of specifications and validation rules. It is the opinion of the authors that these controls are both insufficient and inefficient to ensure oversight of SDTM data health, which can also be considered comprehensive data quality surveillance. We define SDTM data health by four key components: data completeness, mapping correctness, SDTM conformance, and data issue management. Data completeness requires that all source data are transferred to SDTM. Mapping correctness requires that the data are mapped correctly to SDTM. SDTM conformance requires the data to conform with SDTM standards. Finally, data issue management concerns ensuring that all data issues are identified and managed appropriately and consistently | raw data, controlled terminology, data cleaning, completeness, independent qc, data validation rules, cdisc | source presentation source |
8 |
| 99 | ICH M11 Clinical Electronic Structured Harmonised Protocol (CeSHarP) and CDISC: Making the Electronic Protocol a Reality | Peter Van Reusel | Presentation | trial summary, clinical trial, ich, clinical trial protocol, tolerance interval, cdisc | source | 33 |
| 100 | iCSR: A Wormhole to the Interactive Data Exploration Universe | Sudhir Kedare, Steve Wade |
proc means, clinical trial, data science | source presentation source |
7 | |
| 101 | Identifying Data Quality Issues in Clinical Trials Using OpenAI Embeddings | Yash Shah, Ashwin Rai |
clinical trial, clinical data, data cleaning, outlier, proc princomp, clinical research, excel | source presentation source |
5 | |
| 102 | Implementation of the ICH E9(R1) Estimands Framework Using Data Standards | Lori Van Meter, Banoo Madhanagopal |
Poster | source | 1 | |
| 103 | Incorporating Real-World Data in a Low-Interventional Study: Challenges and Considerations | Sandra VanPelt Nguyen | This presentation will share an example of a low interventional study that incorporates real-world data along with conventionally collected clinical data to demonstrate real-world evidence of vaccine efficacy. We will explain the use of tokenization to link RWD from external data sources with the clinical database while maintaining patient privacy, how the data was mapped to SDTM, and challenges and considerations for mapping and analyzing the data | clinical trial, data standard, cdisc | source presentation source |
8 |
| 104 | Ingenious Raw Data Tracking in SDTM Programs: A New Perspective | Bhanojinaidu Duppada, Sumanth Goud Bheemagani |
This paper introduces an innovative approach to streamline the process of tracking raw datasets within the context of Study Data Tabulation Model (SDTM) programs, addressing challenges like the traditional need for complex SAS programming. The SDTM mapping process involves ensuring consistency between annotated Case Report Forms (CRFs) and tracking the raw datasets in SDTM programs demanding significant manual effort and time. However, this paper presents a novel method that expedites the identification of raw datasets within SAS programs through a semi-automated procedure. By harnessing the potential of MS DOS prompt, batch scripts, and Microsoft EXCEL, this approach not only offers a practical solution to organize and present raw dataset information more efficiently but also tackles the complexities of complex SAS programming and study folder restrictions. Embracing our approach introduces novel data cleaning practices and enhanced data analysis in EXCEL |
excel, txt, sas programming, synonymous / closely related terms are grouped on the same line, cdisc | source presentation source |
7 |
| 105 | InnoRand - Innovative Randomisation | Yinfei Wu, Ivo Marek |
Randomization is the foundation for effectiveness research and is used as a basis for submissions of regulatory dossiers in request of marketing authorization for new drugs. In this context, the choice of the randomization design plays an important role. A popular choice is the permuted block design that controls imbalance by making treatment assignments according to a specified allocation ratio at random within blocks. However, the consistent imbalance control increases the risk of selection bias and prediction of upcoming allocations. Better alternatives like the block urn design, the big stick design and others have been developed, and are recently gaining attention. Nevertheless a GxP compliant tool remains hard to be found. We present the conceptual implementation of an R Shiny application which generates randomization lists. InnoRand implements new approaches in addition to the well established randomization methods | randomization, power analysis, sample | source presentation source |
4 |
| 106 | It is Time to Get Answers: Handling Questionnaire Data in Respiratory Studies | Valeriia Oreshko | Questionnaires are one of the most important tools for respiratory clinical studies to evaluate the health and quality of patients' lives. However, a large amount of questionnaire data collected in the form of a patient's diary causes difficulties in preparing it for analysis. In this paper, we provide a summary of our experience in calculating time-to- event (TTE) endpoints based on questionnaires utilized in respiratory studies. Unlike other TTE endpoints, where an event of interest occurs at a single time point, our typical TTE endpoint may require the maintenance of the event for a certain period. This paper provides a method to trace patients' answers to produce meaningful outputs. We also consider common pitfalls of TTE design and its implementation in terms of ADaM Basic Data Structure (BDS). The following problems will be considered: dealing with missing data, common TTE calculation, requirements to baseline values, and the difference between event and censor date description | survival analysis, missing value, imputation, proc sort, proc step, clinical trial, cdisc | source presentation source |
10 |
| 107 | Journey of Reactogenicity Data from EDC to SDTM to Analysis Datasets | Charumathy Sreeraman | Vaccines, the necessity of the recent times, play a major role in the prevention of infectious diseases. Vaccine development is a complex process and is continuously evolving to optimize efficacy and minimize reactogenicity and safety issues. The main safety assessment in vaccine development is reactogenicity analysis, which outlines anticipated short-term reactions to vaccines. In clinical trials, reactogenicity is measured by collecting sets of injection site and systemic signs and symptoms that are solicited from study participants over a specified time after vaccination. These symptoms may include pain, redness, swelling for injected vaccines, and systemic symptoms, such as fever, headache, or rash. The collection of these symptoms from subjects is done through various sources like study site Investigator, participant Diary etc | adam dataset, adverse events, clinical trial, sdtm domain, cdisc | source presentation source |
10 |
| 108 | Key to EU-CTIS Success: Plan, Prepare, Participate | Shawn Lantz, Raina Agarwal |
All new clinical trials within the European Union (EU) must be uploaded to the Clinical Trial Information System (CTIS). This requirement poses a significant challenge for many sponsors , as they must interpret the regulations accurately and act swiftly to meet stringent regulatory timelines. This poster presentation delves into how sponsors can align their resources effectively. It emphasizes upon the importance of gathering a team of knowledgeable subject matter experts and meeting submission criteria. The presentation also highlights how a seasoned contract research organization (CRO) can offer invaluable advice on submission planning, enhancing process efficiency, and liaising with the EMA. This guidance ensures the smooth integration of teams for successful submissions and adeptly manages requests for information throughout the clinical trials' CTIS lifecycle | clinical trial | source poster source |
2 |
| 109 | Large Language Models to Support RBM and Clinical Trial Development | Sebastiaan Höppner, Laura Trotta |
clinical trial, adverse events, medical coding, deep learning, clinical data, neural network | source presentation source |
5 | |
| 110 | Leadership Beyond Borders ... | Nalin Gupta, Shweta Shukla |
source presentation source |
3 | ||
| 111 | Leadership Insights - Let's Learn How to Lead the Leaders | Rachna Mittal | Leadershipisanever-evolvingdomainthatrequiresnotjusttheabilitytoleadateambutalsotoguideotherleaderseffectively.Thispaperdelvesintothenuancesofleadingleaders,examiningthedistinctivetraits,challenges,strategiesandactionableinsightsessentialfornavigatingthiscomplexaspectofleadership | data science, resource allocation | source presentation source |
4 |
| 112 | Leading Teams Through Clinical Data Transformation with Storytelling | Alison Banda, Katrina Rice |
In an increasingly fast-paced and demanding development landscape, leaders are challenged to optimize technology, workflows, and skills to effectively manage trial complexity in less time and with fewer resources. As part of this, organizations must also embrace a new vision and mindset for the future. While exploring and implementing new ways of working, story sharing can illustrate the transformative potential of data science optimization and empower team members to 'think big' so they can develop innovations of their own. This session will explore the instrumental role of storytelling in the strategic toolkit of modern data leadership. Attendees will learn practical strategies for helping teams working with clinical data not only confidently navigate change, but also become champions of change | clinical data, data transformation, clinical trial, data science, clinical research | source presentation source |
3 |
| 113 | Lessons for Leadership from Lincoln | Abraham Pulavarti | Leadership is a privilege, and we owe it to ourselves to learn and develop to be the best leader we can be to our teams and to our company. Leadership styles are varied, and the situations are varied but there are some underlying leadership principles we all can learn from people around us and from great leaders of the past. One such leader is Abraham Lincoln. He often is regarded as the greatest US president who by sheer grit and extra ordinary leadership skills, saved the union and led the US in its most difficult period. In this paper, I summarized the leadership lessons in the business classic, 'Lincoln on Leadership' by Donald T Phillips. Lincoln often used anecdotes to persuade and state his position which reveals his exceptional understanding of people. Lincoln stood for all that was right, honest, and self-evident. He grew up in poverty and had a binding link to the common people. He was innovative at a time when the age of discovery and inventions was just beginning | source presentation source |
5 | |
| 114 | Leveraging APIs in a Multilingual World: Run SAS Jobs from Your R Program | Samiul Haque | A common practice for R programmers in the pharmaceutical industry is to ingest SAS datasets through third-party packages and generate desired tables, listings, and figures using R packages of their choice. However, there are situations where they want to mix and match R and SAS capabilities. For example, one may want to trigger an SDTM automation process written with SAS Macros from R Shiny applications. Another example is producing ggplot2 figures based on outputs from a SAS macro. SAS Viya provides public API endpoints to execute SAS processes, which can be leveraged to integrate R and SAS seamlessly. SAS also has an R package that allows R programmers to leverage SAS in-memory massively parallel processing. In this session, we will discuss how SAS jobs and publicly available SAS REST APIs can be leveraged to operationalize a seamless R and SAS workflow | sas viya, rest api, cdisc | source presentation source |
4 |
| 115 | Leveraging GitHub Actions and Continuous Integration in R Package Qualification | Matt Roumaya, Michael Stackhouse |
The open-source Good Statistical Monitoring, or {gsm}, R package is built to support Risk-Based Quality Management in active clinical trials. To support this GxP use case, a robust qualification framework is included in the package. The {gsm} qualification framework builds upon established best practices by leveraging a set of machine -readable specifications and test cases that are evaluated and summarized in a concise qualification report. This process is highly automated using continuous integration, and leverages existing R tools such as testthat, R Markdown, and pkgdown, as well as custom GitHub Actions that are tied to each package release. This paper offers a detailed exploration of the framework's architecture and processes, spotlighting its pivotal role in enhancing the stability and operational utility of the {gsm} R package | adverse events, clinical trial | source presentation source |
4 |
| 116 | Mapping Real-World Data (RWD) to SDTM Samples for External Control Arm or Registry Studies | Joe Xi, Kevin Yang |
This paper shares our experience in mapping RWD to SDTM from different perspectives including: 1. How to organize safety data into proper domains; 2. How to define the reference date in an ECA study; 3. How to handle complicated baseline information; 4. A quick review of Pn21 compliance report; | clinical trial, sample, clinical data, adverse events, sdtm domain, cdisc | source presentation source |
3 |
| 117 | Mastering the Art of Adaptation: Innovative Strategies for Effective Change Management | Mark Bynens | "Mastering the Art of Adaptation: Innovative Strategies for Effective Change Management" presents a comprehensive approach to successfully navigate organizational transformations in today's fast-paced environment. By analyzing some practical case studies from Janssen R&D, this presentation outlines a framework addressing psychological, organizational, and strategic aspects of change management. The presentation highlights the importance of fostering a growth mindset and creating a psychologically safe environment for employees, with a focus on communication, empathy, and trust to overcome resistance to change. We also discuss the roles of leadership and stakeholder management in change management, emphasizing the characteristics of effective change leaders. Lastly, the abstract introduces innovative change management strategies, including leveraging technology, adopting agile methodologies, and utilizing data-driven decision-making | key performance indicator, clinical trial | source presentation source |
13 |
| 118 | Mining Historical Data and Epidemiological Evidence for Real-World Studies | Shubhranshu Dutta | Real-world evidence studies always have been used to find definitive or conclusive patterns in a variety of aspects of health and disease, such as epidemiology, disease burden, treatment patterns, efficacy, safety, treatment outcomes, long -term outcomes, and patient-reported outcomes such as satisfaction, medication adherence, quality of life, and patient experience. One of the main purposes of Real-W orld Studies is evidence generation for the safety and effectiveness of a drug in larger populations to assess and research if the actual Randomized Controlled Trials could run for a shorter duration based on a robust foundation of evidence. Evidence generation requires a lot of research and analysis to make the right interpretation of the outcome of the experiment | clinical research, feature selection | source poster source |
13 |
| 119 | My Data Are Too Good to Be True! Should I Worry About Fraud or Sloppiness? | Johann Proeve | As clinical trials become more complex, and as more trials are placed at sites by sponsor organizations, it is paramount that the data generated support the ability to evaluate the well-being of trial participants and the reliability of the data. This paper will explore areas where the reliability of the data may come into question, and, through a series of examples, demonstrate how Risk-Based Quality Management (RBQM) methods can be used to provide assurance of the robustness of the data as well as protection of the trial participants | clinical trial, data validation rules, adverse events, raw data, data review, chi-square test, data cleaning, audit trail, edc, vital signs, proc freq, outlier | source presentation source |
9 |
| 120 | Navigating RBQM Change Management: Real-World Strategies for Success | Jamie O'Neill | clinical research, clinical trial | source presentation source |
4 | |
| 121 | Navigating the Safe Integration of AI (ChatGPT) into Clinical Trials Programming Workflow | Mykyta Vysotskyi | Working toward the constant push to shorten the time and effort required to process clinical trial data, the industry is diligently trying to replace the current statistical programming process with artificial intelligence (AI). This is a complex and expensive endeavor that does not seem to give much consideration to the tools currently available that could be helpful in the present moment. This article goes against this trend by discussing immediate, free, and safe methods for easily and effectively bringing AI into the regular work of a statistical programmer. In this presentation, practical ideas for smoothly integrating AI (using ChatGPT) into the everyday routine tasks of clinical trial programmers will be introduced. Statistical programming is an area which continually looks for new ideas within traditional methods | clinical trial, hash table, artificial intelligence, proc sql, clinical data, proc step, clinical research, adam dataset, clinical programming, cdisc | source presentation source |
10 |
| 122 | NEST: How Could We Generate and Deliver Clinical Insights Even Faster? | Vincent Shen | Over the past 7 years, NEST has evolved from a small proof of concept to an industry-recognized and well established project containing numerous R packages for streamlining analyses in clinical reporting and beyond. With large-scale adoption internally and ever-increasing engagement from external users, NEST is well positioned to further influence and revolutionize how we generate insights for clinical research in the pharmaceutical industry. In this paper, we will review our journey so far, share NEST's current capabilities and impact, and then highlight 3 prototypes that we are actively pursuing under the NEST umbrella: Multi-purpose Analysis Result Data, Regulatory Output Review Dashboard, and LLM-based NEST chatbot. Following that, we will discuss our long term vision on clinical insights generation and end the paper with a call on further and closer industry collaboration | clinical trial, csr | source presentation source |
4 |
| 123 | Novel Analytics for Risk-Based Monitoring: Harnessing Natural Language Processing for Query Data to Enhance and Optimise Data Quality | Lawrence Dennison-Hall, Mariia Lapaeva |
Inhealthcareanalytics,unstructuredfree-textoftencontainsvitalinformation,posingchallengesfortraditionalprocessingmethods.Ourinitiativepioneersanovelsolutiontoproactivelydetectdataqualityconcernswithinpreviouslyunanalyzedtextualquerydata. Weleveragetwomachinelearning(ML)frameworks:XGBoostforanalyzingquerytextandresponsesforinconsistenciesandpre-trainedSBERTfordetectingreopenedqueries.Ouraimistopreemptivelyaddressdataqualityconcernsandoptimizeresourceallocationbasedonrisk-basedmonitoring(RBM)principles. Themodelsachieveaccuracyof89.6%and90%respectively(testsets:~10000queries).IntegratedintoRoche'sDataQualityOversightexploratorytoolforover450studies,thesemodelsprovidereal-timeinsightsandofferrisk-basedsolutionsforidentifyingdataqualitychallenges.OurqualitativeuserfeedbackunderscoresthetransformativepotentialofML.Thisultimatelyenablesustoofferdeeperinsightsintopotentialgrowthinthedataqualitybacklog,facilitatingtheproactivemitigationofdataqualityissues | proc hpsplit, machine learning | source presentation source |
14 |
| 124 | Novel Ways to Leverage Your Real-World Data Investment in Integrated Evidence Package Submissions | Sherrine Eid | Traditionally, Real World Data (RWD) has been used in health economics and outcomes research, market access evidence generation and post-marketing commitments. The FDA has released the final guidance, and these data can now serve multiple functions across the product value chain including integrated evidence package submissions. This work covers novel applications of RWD to support activities such as clinical trial optimization, external control arms, integrated evidence package submissions, diversity plans and patient journey to determine unmet need, as well as applications in safety and pharmacovigilance and key considerations in their execution. Leveraging RWD in accordance with the FDA RWE Guidance warrants new and advanced methods in machine learning, deep learning, artificial intelligence and simulation to manage data and generate the insights needed to execute these novel applications. We highlight key considerations in this work | clinical trial, synthetic control | source poster source |
6 |
| 125 | Optimizing Clinical Research: Using AI for Automated Validation of Output Tables against ADaM | Steve Ross | ADaM plays a crucial role in the clinical research process, by providing a structured and consistent framework for organizing and analyzing data, making it easier to generate accurate findings. Validation of the output tables against ADaM is essential to ensure data accuracy and consistency in clinical research. By validating output tables against ADaM, clinical programmers can confidently confirm that the analysis results reflect the core data, thereby enhancing the reliability of findings in both clinical trials and research studies. This validation process is performed primarily manually, using double programming or visual review. In this paper, we show how this manual process can be replaced by automation using AI. We present data elements that can be derived from the output and can be matched to the ADaM standards, using multiple examples of automated validation of generated output against ADaM datasets | data cleaning, adam dataset, clinical trial, independent qc, clinical research, proc step, cdisc | source presentation source |
5 |
| 126 | Organisational Considerations When Replacing or Adding a New Software Language | Brian K. Varney | Replacing or adding a new software language for your organization's users can add a lot of value and productivity to your team(s). There are, however, many factors and considerations to consider avoiding chaos and hidden costs | data cleaning, clinical trial, pharmaceutical industry, macro definition | source presentation source |
4 |
| 127 | Population Pharmacokinetic Analysis Data (ADPPK) Programming in {admiral} and the pharmaverse | Jeffrey Dickinson | Population Pharmacokinetic modeling is an important tool for drug development. The CDISC ADaM Population PK Implementation Guide was released on October 6, 2023. Population PK models generally make use of nonlinear mixed effects models that require numeric variables. The data used in the models will include both dosing and concentration records, relative time variables, and numeric covariate variables. {admiral} is an open- source R package for creating CDISC ADaM data. It can be used effectively to create Population PK analysis data (ADPPK). Additional tools from other Pharmaverse packages such as {metacore}, {metatools} and {xportr} can be used to simplify the workflow. I will discuss some of the challenges of Population Pharmacokinetic analysis data programming and show some of the solutions developed in {admiral} and the Pharmaverse | population pk, time variable, label statement, sample, response variable, adam dataset, cdisc | source poster source |
9 |
| 128 | Posit/RStudio Pharma Updates: Reproducibility and Beyond | Cole Arendt, Philip Bowsher |
statistical programmer, data cleaning | source presentation source |
14 | |
| 129 | Post-Pandemic Trends in Study Designs | Charumathy Sreeraman, Jisha George |
From the conventional method of having fixed study designs to the most recent trends, adaptive design model, the clinical trial industry has come a long way. The adaptive model allows for flexibility and optimization, which in turn speeds up the future procedures, by permitting alterations to subsequent study activities based on interim evaluation. Faster and more effective findings are achieved by using a single Master protocol with numerous sub studies and/or "parent-child" studies. Additionally, using this method enables the sponsor to carry out numerous dose tests and evaluate the outcomes concurrently. This paper will discuss the various data gathering and mapping procedures used by vaccine studies to apply the sub - study concept, including (a) dividing the Master Database into several sub -studies and (b) combining the database of a prior study and handling re-screening scenarios in parent-child studies | clinical trial, raw data, clinical research, sdtm domain, cdisc | source presentation source |
7 |
| 130 | Post-Text (TLGs) to In-Text Statistical Table Automatic Conversion for Clinical Study Reports (CSRs) | Divya Kanniah | We have developed a system that uses the power of AI technologies that swiftly constructs in-text tables from post- text TLGs, eliminating the need for manual construction in MS Word or SAS programming by medical writers and statisticians, respectively. The system accommodates post-text tables in Word ,rtf or PDF formats, allowing users to choose font styles, edit titles and footnotes, and merge pertinent post-text into consolidated in-text tables. Users have the flexibility to select footnotes from merged tables, apply filters to categorical tables based on specified percentages, and display summaries in the resulting in-text format. This system significantly saves time, offering a streamlined solution for medical writers | sas programming, artificial intelligence | source presentation source |
5 |
| 131 | Programming Approaches for Integrated Summary of Efficacy (ISE) Development in Oncology | Narendra Babu Kolisetty, Saigovind Chenna |
In clinical trials, the Electronic Common Technical Document (eCTD) is the standard format for submitting regulatory applications including new drug applications (NDA) and biologics license applications (BLA) to the FDA. An integrated summary of efficacy (ISE) is one of the critical components of an eCTD, providing a comprehensive overview of an investigational product's efficacy across multiple studies. However, integrating efficacy data from late-phase clinical trials into a high-quality ISE data package can be challenging due to differences in study design, data collection, and mapping procedures. Further challenges include differences in SDTM and ADaM implementation guides, inconsistent analysis result derivation, handling of external data, dealing with missing values, Pinnacle21 issues, and managing value truncations | adam dataset, clinical trial, missing value, numeric variable, dataset, cdisc | source presentation source |
7 |
| 132 | QTLs/Acceptable Ranges: How to Set Thresholds | Debra Jendrasek, Marcus Muli |
Presentation | source | 13 | |
| 133 | R Shiny Application to Optimise Review of Multiple Myeloma Efficacy Data Listings | Dinesh Putluri | This paper explores the challenges encountered by clinical teams in Multiple Myeloma studies during manual reviews of SAS-generated listings. Emphasizing the importance of data like biochemical test results, BMA, BMB, bone lesion, and plasmacytomas data, crucial for the adjudication process, we delve into the labor-intensive nature of existing review processes, particularly in larger studies. The paper introduces a transformative solution - an intuitive web application developed using R Shiny. This innovation empowers clinical teams through personalized data scrutiny, comparative analysis, and seamless collaboration, promising heightened efficiency, data integrity, and improved team collaboration | data review, adam dataset | source poster source |
2 |
| 134 | R-Guru Resource Hub for Rapid R Learning | Sunil Kumar Gupta | Are you behind in leveraging Pharmaverse R package releases for SDTMs, ADaMs and TLFs? Do you understand the benefits of learning and applying R instead of only programming in SAS? Do you know how to apply basic R syntax using common and validated R packages by R Studio and the pharma industry? Are you interested in adding R to your programming skill set so that you have multiple methods for data input, management and analysis? Learning R for professional development is a new trend for early adaptor corporations and SAS programmers. Because R is more an object-oriented and syntax direct language, programmers can now write concise code for targeted tasks. This empowers programmers with a competitive advantage for advancing their career | sas programmer, dataset | source presentation source |
8 |
| 135 | Real-Time Data Cleaning -- A Reality: When Automation Connects Digitised, Centralised, Standardised Data, it Transforms into Real-Time Data Cleaning | Vidya Muthukumar | With the constant emphasis and request by sponsors in clinical trials to "clean" data regularly and not wait until close to final database lock, CROs are looking for smart and cost-efficient ways to "clean" data. The exponential increase in different types of data sources along with the rising costs of clinical trials requires bolder, smarter, and leaner solutions to current manual data cleaning processes. When digitized, centralized data flows and standards connect together to provide an integrated, streamlined, and automated data management process, it transforms into a system called Real- Time Data Cleaning (RTDC). Continuous data ingestion and automated cleaning, results in shortened database locks, issue detection and mitigation. The increasing data source types together with the complexities of trial designs require new techniques to meet these challenges efficiently while maintaining high quality, without raising costs. The obvious solution is automation | data governance, clinical trial, database lock, data review | source poster source |
2 |
| 136 | Real-World Application of Natural Language Processing: More Than Just Sentiment Analyses | Scott McClain | Natural language processing, or text analytics, is overlooked as a data science tool more than we appreciate. This is mainly because the use cases and associated functionality offered by NLP are so diverse. A key challenge: how can individuals consistently and rapidly read, extract, sort and catalogue large volumes of text for specific topics with software? And how can it be done without command-line programming? Learn how to utilize NLP through an interface to prioritize keywords, terms, numerical and other data automatically, at a scale beyond what a person could accomplish. Goals for NLP can include finding new associations of key terms. These can lead to training NLP language models to act as repeatable steps that manage, classify and catalogue better than human-in-the-loop only processes. We examine use of text language software to build low code, GUI driven models specific a scientific literature-based proof of concept that can be turned into a repeatable application | data science, machine learning | source presentation source |
7 |
| 137 | Real-World Data - What is Fit for Purpose? | Berber Snoeijer, Parag Shiralkar |
Poster | source | 1 | |
| 138 | Recruiting the Best Talent and Developing them within the Organisation | Naveen Kommuru, Madhusudhan Reddy Papasani |
Statistical programming organizations are facing new challenges with hiring and developing the best talent for their organizations in the pro-remote and hybrid work environments, like all other industries. The employer, as well as the employees, need to understand and work towards making the required changes to attract the best talent in this competitive world. This paper provides different challenges that employees experience, such as presenting their skills and experience in virtual interviews and working in remote positions. We suggest the approaches that they need to take to enhance their skills. Further, we emphasize the importance of continuously expanding their technical knowledge, leadership, and soft skills, which could lead to future career advancement. Alongside this, we encourage employees to consider their long-term growth and development within the industry, rather than the short-term gains and comfort | clinical programming, statistical programmer, sample | source presentation source |
11 |
| 139 | Redefining 'Interactive' Dashboards to Enhance Safety Data Review | Andrew McGowan | The periodic review of safety data is a critical task throughout clinical trials which can be extremely inefficient and make signal detection difficult. Interactive dashboards developed through close collaboration between programmers and study team members provide an optimized method of reviewing accumulated safety data to minimize subject risk. By avoiding the limitations of static listings and visualizations, dynamic displays allow the reviewer many views of the data without being overwhelmed with many different visualizations or listings. By allowing the reviewer to manipulate views of the data, filter data being fed into the display, and mark data to trigger targeted dynamic supporting displays with relevant data, the reviewer can unleash the full power of the increasing volume of clinical trial data | adverse events, clinical trial, standard deviation, line chart, adverse event summary, outlier, data review, box plot | source presentation source |
7 |
| 140 | Revolutionising Informed Consent Form Analysis with Generative AI: Enhancing Efficiency in Drug Development and Data Re-Use | Hangyu (Cedric) Liu | The pharmaceutical industry faces challenges in re-using samples, data, and images for secondary research. Complex guidelines governing data re-use are often buried in Informed Consent Forms (ICFs) that are manually reviewed by subject matter experts (SMEs) on a case-by-case basis. This process is time-consuming and hampers efficiency. We developed a novel Artificial Intelligence (AI)-driven workflow leveraging Large Language Models (LLMs) for computer-assisted analysis of ICFs. This cutting-edge method substantially simplifies our process by using generative AI to summarize and identify allowable data usage promptly and precisely. This adaptable framework is applicable to any company requiring comprehension and analysis of ICFs or data use agreements. Through a domain-specific validation framework, combining manual SME review and quantitative measurements, we achieved 96% accuracy in Biogen study ICFs tested to date | data cleaning, pharmaceutical industry, sample, clinical trial, artificial intelligence, validation set | source presentation source |
11 |
| 141 | Revolutionising SDTM Transformation in Pharma: Introducing {oak} - An Open-Source R Package Sponsored by CDISC COSA | Rammprasad Ganapathy | edc, crf, cdisc | source presentation source |
5 | |
| 142 | Revolutionize Clinical Trial Data Exploration with teal | Chendi Liao, Dony Unardi |
clinical trial, data science | source presentation source |
6 | |
| 143 | Risk-Based Monitoring Using AI/ML Detection of Systematic Bias in Clinical Data | Kostiantyn Drach, Sergey Glushakov |
To effectively plan and execute clinical trials, it is essential to oversee data integrity as part of risk-based monitoring (RBM). Inaccuracy and incompletion in clinical data can be explained by systematic bias in multiple metrics considered for RBM execution. Systematic bias is hard to detect using standard statistical methods. Thus, we propose an approach to reveal this type of inconsistency in data using AI/ML methods together with graph-based machine learning algorithms. To perform the experiment, we use clinical datasets with synthetically generated bias as proof of concept. Our approach relies on pure geometric properties of data combined with AI/ML algorithms, such as neural networks, community detection, etc. The experiment unveils the systematic bias hidden in data with no prior knowledge of its source | clinical trial, regression analysis, neural network, data cleaning, sample, test set, training set, logistic regression, logistic distribution, f1 score, lasso regression, proc hpsplit, feature selection, machine learning, missing value, proc hpforest, standard deviation, clinical data, data science, structured data | source presentation source |
12 |
| 144 | Risk-Based Quality Management: It's Embedded! (Or Is It?) | Michael Walega, Steven A. Gilbert |
Presentation | source | 12 | |
| 145 | SAS ODS Package for Efficient Creation of Zip File for P21 Validation | Venkatesh Burla | Pinnacle 21 validation is a critical step to prepare FDA eSubmission package. It ensures compliance with the regulatory guidelines and validates the accuracy and consistency of clinical trial data. ZIP file is a mandatory input to perform P21 validation. Manually creating a ZIP file leads to inefficiencies and potential errors. Although ZIP compression tools (7- zip and WinZip) are widely used, they are not convenient with in the SAS program. Leveraging the power of ODS, comprehensive solution is developed to automate the process and reduce the time/effort while maintaining data integrity. This paper presents a novel approach utilizing the SAS Output Delivery System (ODS) to streamline the process of creating ZIP files for P21 validation |
macro conditional, data cleaning, adam dataset, output delivery system, clinical data, cdisc | source presentation source |
8 |
| 146 | SAS PROC MI: Your Impeccable Tool for Enhanced Data Analysis Through the Lens of Multiple Imputation | Anush Kurghinyan, Shahrzad Talebian |
This paper investigates the robust capabilities of SAS 's PROC MI in facilitating advanced data analysis through the lens of multiple imputation. Multiple imputation emerges as a crucial statistical technique for addressing the challenges posed by missing data, offering a systematic approach to enhance the reliability and precision of statistical analyses. The paper explores the functionalities and applications of PRO C MI, emphasizing its role in producing more accurate estimates through the generation of multiple datasets with imputed values. The advantages of multiple imputation, particularly in longitudinal studies and clinical trials, are underscored. The study highlights the increasing recognition of multiple imputation as an indispensable tool for researchers across various fields striving to improve the quality of statistical inferences |
missing value, imputation, multiple imputation, data science, proc mi, proc step, inferential statistics, power analysis, sample, estimate, confidence interval, sample size, statistical programmer, type ii error, proc mianalyze, regression analysis | source presentation source |
7 |
| 147 | Scaling ECAs from Projects to the Enterprise: Multicentre, Multisource Infrastructure for External Control Arms | Jussi Leinonen, Mayur Saxena |
raw data, data transformation, data standard, adam dataset, data wrangling, clinical research, cdisc | source presentation source |
4 | |
| 148 | SDTM/ADaM Automation - Challenges and Issues from a Programmer's Perspective While Exploring an Automation Attempt to Address Them | Alistair D-Souza | Programmers' resistance to automation, lack of scalability, flexibility, and additional overhead are some of the reasons for tools failing or being scaled down after launching. We would investigate some of the challenges and points to remember in tool development from a programmer's perspective. This paper will talk about how such factors affect the overall success, effectiveness, and acceptance of the tool. It would also present high-level schematics of an innovative automation tool to address some of these challenges. With machine readable maps it is scalable and flexible due to a unique bottom -up and unit -based integration approach. A tool with which the programmer feels in control, without feeling stuck or stalled due to variations in the input data or unexpected scenarios, and helping the programmer reduce overall effort and time needed for p rogramming. It also talks about the possible additions and efficiencies that can be built later | data step, clinical sas, macro call, sas programming, clinical trial, cdisc | source presentation source |
6 |
| 149 | Software Product Lines Framework: How to Manage End-to-End Metadata-Driven Automation at Sale | Stuart Malcolm | CDISC's ambition of developing and extending clinical standards with "the additional semantics needed to support metadata-driven automation across the end-to-end clinical research data lifecycle" is compelling, and promises benefits of "substantially improved efficiency, consistency, and re-usability across the clinical research data lifecycle." [1] To deliver on this vision means replacing today's manual processes with new processes that support automation, not for a single trial, but for all trials across all phases, treatments, devices, therapeutic areas, etc. The problem this paper explores is how to manage end-to-end automation such that the time and cost of maintenance across all studies does not counter the efficiency gains. In short - how do we scale end-to-end automation? This paper proposes using Software Product Lines (SPL) as a framework for managing automation of clinical trial projects at scale | clinical trial, clinical research, cdisc | source poster source |
5 |
| 150 | Sowing Seeds of Inspiration: Letters to a Pre-Scientist | Julie Ann Hood | A multitude of career possibilities exist today; however, many remain unaware of the range of careers available in Science, Technology, Engineering, and Mathematics (STEM). The Letters to a Pre-Scientist volunteer program aims to bridge that gap by providing children in underprivileged communities the opportunity to explore various STEM careers through pen-pal correspondence with STEM professionals. Throughout the school year, partners exchange themed letters, cultivating personal connections and facilitating idea-sharing. Students gain exposure to a diverse range of STEM careers by discussing their partner's insights and their own thoughts and ideas on these subjects with their classmates. This poster will showcase the Letters to a Pre-Scientist program and share my journey as a novice Pen Pal last year | clinical trial, clinical data | source poster source |
3 |
| 151 | SPACE - GSK Biostatistics continues its journey to a new frontier | Eileen Ching | GSK embarked on a journey to implement a modern fit-for-future statistical computing and data science platform, called SPACE, in early 2020. Since then, we have delivered multiple iterations of our core products (including a new SCE) aimed to accelerate medicine and vaccine development by transforming how we analyze data and share results. In addition to implementing trailblazing solutions, the SPACE team also became trailblazers in adopting Agile framework in a risk-averse environment with ingrained processes. As we continue to build out the platform, support user adoption, enhance the solutions, and tackle change management challenges, we are making progress at extraordinary speed, with a focus on migrating numerous studies from legacy to SPACE very quickly. To achieve success, we also needed to be relentless in applying learnings to continuously improve. We are on our way to new and greater things, and I will share these new chapters of our story | clinical trial, data science | source presentation source |
9 |
| 152 | Speeding Up Submissions - Assessing the Business Value, Feasibility and Risk of New and Emerging Approaches | Jaya Simha Inampudi, Nekzad Shroff |
Presentation | data cleaning, data integration | source | 9 |
| 153 | Spline Analysis to Understand the Non-linear Relationship in Clinical Research | Anwaya Joshi | In clinical research, often we come across a situation where we have several predictor variables to predict the response. There is a possibility that all the predictor variables may not have a linear relationship with response variable. Fitting a higher degree polynomial model to such data may not provide reliable results. The Spline regression analysis allows us to model non-linear relationship of predictor variables by dividing the dataset into bins and fitting a separate model for each bin. Further penalizing this spline model can provide a better fit for the data. In conclusion, the adoption of spline analysis represents a substantial stride towards solving the nuanced non -linear associations present in clinical research | regression analysis, spline, proc reg, polynomial regression, degrees of freedom, clinical research, predictor variable, response variable, scatter plot, proc step | source presentation source |
9 |
| 154 | Standards Library: Getting started - A Medical Device Case Study | Amy Booth | This paper describes a framework for implementing a Clinical Data Standards (CDS) Library which can start small and grow with you as you add more assets to the library. An in-house company standards library is one of the key foundations for achieving efficiency gains through reuse, higher quality, and reduced timelines. However, it is not without challenges - not least maintaining alignment with global standards, mappings global standards to company/franchise standards, version control and change management. This paper walks through a Medical Device case study, and describes the library structure, process flows for SDTM, ADaM and TFL submission deliverables, validation and metrics reporting, lessons learned, and future aims to support automation | sdtm programming, adam dataset, cdisc | source presentation source |
6 |
| 155 | Statistical Monitoring Report: A Path Forward to Interactive Data Analysis and Quantitative Decision-Making | Debopriya Ghosh | Early development clinical trials warrant an agile approach of interactive data analysis for exploratory analyses that take the form of "quick insights" and lead ultimately to the generation of scientific hypotheses. The Statistical Report Tool (START) is a cross-functional collaboration to streamline the process from exploratory analyses to hypotheses generation and early decision making. The framework consists of several fit-for-purpose modules, ranging from data preparation to specific types of analyses, allowing for dynamic previews of safety and preliminary efficacy. The modules are programmed in R and the analyses are presented via a html interface that supports interactivity. In contrast to static outputs, the interactive visualizations and data tables support cross-domain data linking and multi- dimensional data analysis that could be hard to capture using static displays | clinical trial, clinical data, data science, dataset | source presentation source |
3 |
| 156 | Statistical Reporting Metadata Manager: Web Application to Streamline and Centralise Global and Study Metadata Preparation and Management | Faisal Mohammed | SRMM (Statistical Reporting Metadata Manager) is a web-based application built to streamline and centralize Global & Study Metadata (MD) preparation and management. Harnessing the power of graph database technology, SRMM connects metadata end to end from data collection (SDS) to statistical reporting (TLF). The Python backend is connected to the database and handles requests from users who interacts with the application through a Vue.js interface. The User Interface is ergonomic, and enables the user to browse Global Metadata, Study Metadata and inspect connections between different metadata types. The goal of SRMM is to help standardization teams, biostatistician and programmers by saving critical time spent on manual, error-prone work. It empowers them to focus on their area of expertise, thereby accelerating the analysis and reporting of clinical data. In the future, these can be further enriched by connecting to protocol & SAP to drive automation | metadata management, data wrangling | source presentation source |
4 |
| 157 | Streamlining the Integration of SDTM Datasets from Multiple Studies: A Macro Approach | Santosh Reddy Lekkala | In New Drug Application (NDA) or Biologics License Application (BLA) submissions to the FDA, Integrated Summary of Safety (ISS) and Integrated Summary of Efficacy (ISE) play pivotal roles. Building integrated datasets by pooling data from multiple studies is a complex and time-consuming task which is either done at Study Data Tabulation Model (SDTM) or Analysis Data Model (ADaM) stage. This PHUSE paper presents a macro designed to streamline the data integration process, particularly focusing on SDTM datasets. We will address common challenges encountered during dataset integration, such as variables with different lengths, missing data and variable label discrepancies. The macro will facilitate the identification of these inconsistencies and generate an output to help resolve them effectively thereby saving valuable time and effort | excel, sdtm domain, proc step, macro definition, cdisc | source presentation source |
7 |
| 158 | Synergising Large Language Models (LLMs) and Human-in-the-Loop (HITL) for Data Noise Reduction in the RBQM Approach | Mukesh Babu, Vineet Jain |
data cleaning, clinical trial, outlier | source presentation source |
7 | |
| 159 | Teal as an Open-Source Tool Accelerates Clinical Trial Data Research at GSK | Jeffrey Long | GSK is a pioneer in embracing open-source technologies and applications. Teal, a highly engineered and validated tool developed by Roche/Genentech, is aiding the interactive exploration of data with Shiny web applications. Teal apps for RWD, machine learning output and public data provide tools for niche areas to perform exploratory investigation of data that is non-standard and outside the scope of our data visualization tool RAPIDO where the focus is CDISC and Analysis Results Data. Teal is enabling the refinement of patient selection in oncology trials, and supporting the generation of new hypotheses as biomarker and baseline data are interrogated in new ways. GSK's move to a cloud-based platform has enabled custom Docker runtime environments that expedite the installation of {teal} packages and dependencies | clinical trial, adverse events, cdisc | source presentation source |
7 |
| 160 | Teal.builder: A Multifaceted Teal App Builder for Clinical Data Exploration and Reporting | Freeman Wang, Andre Couturier |
clinical data, clinical trial, data science | source presentation source |
5 | |
| 161 | TFL Designer: Automating Tables, Figures and Listings Design and Generation for Analysis Results | Navin Dedhia, Bhavin Busa |
Presentation | adam dataset, macro loop, excel, cdisc | source | 10 |
| 162 | The Data Review Model - Recognising the Benefits of a Single Source of Truth Model for Data Review | Priyank Patel | The Data Review Model (DRM) is J&J Innovative Medicine's proprietary data model which serves a single source of truth conceptualized in -house to meet various stakeholders' needs such as data visualization, data cleaning, risk management and central monitoring, etc., on time. The data model has been intellectualized as the CDISC SDTM+ model, allowing greater flexibility to expand, keeping in mind to produce quality submission artifacts in the future. The DRM model concept, an en abler of various stakeholders' day to day requirements, has invested heavily in an out -of-box framework to increase efficiency and reduce maintenance costs across set processes, tools, and systems to increase productivity, business growth, and sustainabili ty. The article will share a high level framework in data handling space that permits stakeholders to reach data faster, almost in real time | data review, data transformation, surrogate key, data mapping, clinical data, cdisc | source presentation source |
8 |
| 163 | The Implementation of Scrum in Pharmaceutical Data Analytics and Statistical Programming at Jazz Pharmaceuticals | Eliana D'Angelo | Scrum is a popular agile project management framework that is useful for addressing complex environments, such as Pharmaceutical Data Analytics and Statistical Programming. By using Scrum, teams can control risk, minimize investments, and improve predictability. Scrum teams focus on value and learning, rather than relying on opinions, which increases the chances of a good outcome and a high return on investment. In comparison to the traditional waterfall approach, which is commonly used in our industry, Scrum allows teams to constantly adjust and adapt to changes, leading to more efficient and effective project delivery. In this paper, we discuss Scrum's usefulness in complex environments and the benefits and learnings that we have experienced while implementing Scrum across our Data Analytics and Statistical Programming team | clinical programming, clinical research | source presentation source |
6 |
| 164 | The Knowledge Graph - The Power of an Information Network driving Biometric workflows | Nekzad Shroff, Jaya Simha Inampudi |
Clinicaltrialprogrammingrequiresmeticulousmanagementofcode/specificationlineage,dependencies,andtraceabilitybetweencomponentslikeSDTM,ADaM,TLFs,logs,andoutputs. ThispaperexploresKnowledgeGraphdatabasesasatransformativesolutionforthisusecase.BlendingempiricalresearchandacasestudywithSaama'sBiometricsresearchandanalysisinformationnetwork,wedemonstrateKnowledgeGraphs'unparalleledabilitytomaprelationships,lineage,anddependenciesamongSDTM,ADaM,TLFs,outputs,logs,andoutlinefilesastheygetcopied/mutatedduringtheirlifecycleinthesystem.Thisvisualgraph-basedgranularityprovidescomprehensive,intuitivenavigation,reusability, andmetadatamanagementpreviouslyunattainable.KnowledgeGraphsnotonlyenhancetraceabilityandintegrityofcode/specificationlineagebutsignificantlyimprovecomplianceandoperationalefficiency. Inaddition,theavailabilityofthisdigitizedrelationshipmetadataintheKnowledgeGraphisanimportantstepforfutureautomationandstreamliningofmanualstepsinthesubmissionlifecycle | clinical trial, regulatory submission, cdisc | source presentation source |
3 |
| 165 | The Many Use Cases of Standardized Data and Metadata | Sanjiv Ramalingam | Study Data Tabulation Model (SDTM) datasets are predominantly Clinical Study Report (CSR) driven and primarily created to support analysis dataset creation. The data flow is typically unidirectional. The SDTM group has pioneered the creation of SDTM datasets enabling creation of SDTM datasets within days of First-Patient-In (FPI) and automatic refreshes enabling wider use of standardized data across the organization. With this new way of working, the SDTM group is able to serve and enable efficiencies across multiple functions such as Data Management, Data Standards and Governance, Biomarker and Clinical Operations than just Statistical Programming. The use cases for each of these groups have been discussed in this paper | sdtm domain, raw data, cdisc | source presentation source |
4 |
| 166 | The Potential Power of Data | Eva Kelty, Catharina Dahlbo |
Drug development is all about data. All clinical trial activities are geared towards generating data to accept/decline hypotheses, to support claims and to generate new knowledge and ideas. Data needs to be available and understandable for the people consuming it. Utilizing a graph database with an intuitive interface makes it possible to structure data in a way that makes it easily available for non-programmers. Combined with relevant metadata data the usability increases. This presentation will focus on how data can be organized and utilized in clinical trials, e.g., in the planning, execution, reporting and archiving. A repository containing clinical trial data provides opportunities to learn from past experiences, while planning ahead. Access to data as a clinical trial is ongoing could improve quality, speed up the process and support collaboration. During the reporting period data can be easily explored to understand outliers or subgroups | clinical trial, clinical research | source presentation source |
5 |
| 167 | The Role of AI (ML/NLP/Deep Learning) in Handling Sensitive/Confidential Data | Divya Kanniah, Swetha Balaji |
This paper explores the pivotal role of machine learning, deep learning, and natural language processing in the creation of scientific documents, focusing on the challenges of handling unstructured data and ensuring the confidentiality of clinical information. The study investigates algorithms employed for data extraction and text generation using Natural language Processing (NLP) and Natural Language Generation (NLG) | deep learning, logistic distribution, logistic regression, unstructured data, regression analysis | source presentation source |
3 |
| 168 | The Role of FHIR Resources in Testing and Validating an EHR to a Sponsor Data Pipeline | Patrick Genyn, Andy Richardson |
clinical trial, data cleaning | source presentation source |
6 | |
| 169 | The State of Open-Source Technology in Clinical Data Analysis, Reporting and Submissions - A PHUSE Manuscript | Michael S Rimler | A significant amount of time and energy has been invested in recent years exploring the desirability (do we want it?), feasibility (can we do it?), and viability (is it worth it?) of integrating open source solutions into our clinical data pipelines, pipelines which transform source data into clinical study reports and submission data packages. This presentation provides awareness on a PHUSE initiative which aims to comprehensively describe where the pharmaceutical industry is at with respect to embedding open source solutions into all matters of clinical data handling, reporting, and submissions. We will share our vision for a "state of the union"-style manuscript, what we have so far, and (most importantly) how the PHUSE Community can contribute its expertise to further the effort | clinical data, csr | source presentation source |
1 |
| 170 | The TransCelerate/CDISC Digital Data Flow Project: Practical Electronic Study Designs | David Iberson-Hurst | Presentation | clinical data, clinical trial, data standard, pharmaceutical industry, cdisc | source | 28 |
| 171 | Transformative Strategies in People Leadership and Management for Organisational Success | Bhanojinaidu Duppada, Sumanth Goud Bheemagani |
The SDTMIG v3.4 and the SDTM v2.0 were released in July 2022. These versions have been published as HTML documents, rather than the typical PDF's of the past. While the SDTMIG v3.2 which was published in 2013 contained 398 pages and SDTMIG v3.4 would be more than 400 pages if formatted properly and printed to a PDF. Several new domains, new concepts have been added and several domains that had undergone public review in 2014 were subsequently combined with existing domains. The Disease Milestones concept, introduced for the TAUG-Diabetes in 2014, is now included in the latest version of the SDTMIG. A new Section '9' includes Study References with added models for Device Identifiers and Non-host Organism Identifiers. Three versions of the SDTM have been published since the release of v1.4 with the SDTMIG v3.2 | clinical trial, adam dataset, sdtm domain, cdisc | source presentation source |
8 |
| 172 | Transformative Strategies in People Leadership and Management for Organisational Success | Srinivas Atla | In today's rapidly evolving business landscape, effective People Leadership & Management has emerged as a cornerstone for achieving organizational success and sustained growth. This paper delves into the realm of transformative strategies that empower leaders to nurture their teams, foster innovation, a nd drive productivity. Utilizing a wealth of research and captivating real -life examples, our paper guides you through the complexities of successful leadership and management. We focus on the six critical dimensions crucial for success in today's organizations. By blending research with real -world experiences, our goal is to provide practical strategies and actionable insights for driving meaningful change. 1. Adaptive Leadership 2. Emotional Intelligence (EI) 3. Inclusive Team Dynamics 4. Coaching and Mentorship 5. Remote Work Leadership 6. Conflict Resolution | source presentation source |
6 | |
| 173 | Transforming RWD for Regulatory Submissions: How to Use SDTM for RWD | Samuel Hume, Tasha Nagamine |
source presentation source |
5 | ||
| 174 | Transforming Submissions with Saama's Biometrics Research Analysis and Information Network (BRAIN) | Jaya Simha Inampudi, Nekzad Shroff |
Presentation | job scheduling, statistical programmer, clinical programming | source | 6 |
| 175 | Translating Data into Tableau Insights: Bridging the Gap Between Programmers and Medical Physicians | Agnieszka Bel | Tableau dashboards provide valuable insights into clinical trial data. Our extensive library of standard dashboards covers common scenarios across different studies. However, end users often require custom data visualizations tailored to their needs and trial's nuances. The understanding of these requirements and effective communication between diverse study teams, especially medical physicians and programmers, are essential to meet these needs. This presentation will outline our methodology for creating and validating custom Tableau dashboards and addressing potential challenges. The core feature in the process is setup of the demo environment, accessible by the end user. There, we can perform rapid prototyping, and the dashboards can be viewed and validated internally, and by the end user. This allows us to quickly understand if we are going in the right direction | data cleaning, data visualization, clinical trial, adverse events, sdtm domain, clinical data, cdisc | source presentation source |
7 |
| 176 | Unlocking the Future of Tables with the {tfrmt} R Package | Becca Krouse | Clinical trial displays come in many forms; even highly standard displays require modifications. Table customization and subsequent QC drain resources, and creating mock tables separately during analysis planning causes duplicated efforts. With a new industry standard on the horizon, Analysis Results Data (ARD), we worked to rethink the overall reporting process. The {tfrmt} package transforms table creation by leveraging the consistent structure of ARDs. With {tfrmt}, a user defines a table's formatting independently of the data, allowing reuse of code for both mock and final tables. Formatting templates enable automation of standard tables, as well as the ability to quickly refine tables by layering study-specific customizations. These benefits extend beyond R users, as {tfrmt} can generate and consume language-agnostic metadata. Furthermore, the companion Shiny app, {tfrmtbuilder}, simplifies the learning process for newcomers and facilitates customizations for non-programmers | clinical trial, data standard, r package | source presentation source |
8 |
| 177 | Unveiling the Potential of ADNCA via the CDISC ADaM IG: Revolutionising Pharmacokinetic Non-Compartmental Analysis Input Data Standardisation | Saumilkumar Tripathi | In pharmacokinetic data analysis, standardizing input data for non -compartmental analysis (NCA) has historically been challenging due to varying approaches across companies which makes it difficult and time consuming for reviewers and auditors to understand the data flow and lineage. For statistical programmers, conveying these details in the Analysis Data Reviewer's Guide to establish traceability and relationships with SDTM and other PK ADaM datasets such as ADPC/ADPP is also challenging. This paper introduces ADNCA (ADaM NCA input Dataset), utilizing CDISC ADaM IG to convert clinical study PK, Exposure, Exclusions and Disposition data into NCA input data for pharmacokinetic analysis and PK parameters calculation. ADNCA seamlessly integrates with CDISC ADaM IG, simplifying PK data preparation, ensuring data quality, consistency, traceability, and compliance to industry standards. This paper demonstrates methods to automate standard variables derivations using the SAS Macro programs | in=, pk parameter, macro definition, sample, key, statistical programmer, adam dataset, compartmental analysis, sdtm domain, clinical research, cdisc | source presentation source |
10 |
| 178 | Ups, Downs, and FDA Crowns - Lessons Learned from BIMO Submission Adventures! | Sushant Thakare | Presentation | excel, safety population, adverse events, clinical trial, independent qc, primary endpoint | source | 21 |
| 179 | Usage of OMOP-CDM along with CDISC in Clinical Trials: Challenges and Opportunities | Parag Shiralkar | With growing importance of RWD in getting clinical evidence, many companies are finding ways to make the best use of data from real world sources. Claims and healthcare data obtained from EHR systems is a major source of such data. OHDSI organization has developed a Common Data Model (CDM), which can take large observational data from multiple EHR and claims data sources around the world, and read and process such data in a consistent manner. Usage of such data in randomized clinical trials poses challenges as regulatory bodies prefer the submission of the data in CDISC format. How are traceability and provenance issues dealt with in OMOP-CDM? Can such data be easily transformed into SDTM? Would regulatory bodies accept the OMOP CDM as a submission standard? This manuscript provides detailed assessments of efforts taken by industry in usage, and transformation of OMOP CDM along with FHIR to CDISC SDTM | clinical trial, clinical data, pharmaceutical industry, data standard, regulatory submission, insurance claims, clinical research, cdisc | source presentation source |
5 |
| 180 | Using Bundles for R Package Management | Magnus Mengelbier | The broader use of R within Life Sciences from a niche personal statistical analysis tool to an organization wide environment is making R package management a more complex and comprehensive task. As the organization grows and the types of analysis are ever expanded, a substantial increase in the number of packages is certain to follow. The use of R packages for dependency management, such as packrat and renv, is very applicable for personal R environments or where users are permitted to fully customize their collection of packages, but not necessarily validated GCP environments. The R bundle approach discussed is designed to provide a similar high degree of flexibility for dependency management while staying within the constraints imposed by GCP compliance. Additional benefits, such as significant decrease in complexity and improved overall life cycle of the R environment, are further explored through examples | data cleaning, gcp, r package | source presentation source |
8 |
| 181 | Using R for QC of Tables | Marcus Trincia | The R programming language is no longer a tool for the future generation of clinical programmers, rather it has become a staple for this one. Due to the magnitude of open-source libraries available for R, it is difficult to find an area it cannot enrich. In this paper, we combine the familiarity of SAS within the pharmaceutical industry with the versatility of R to provide a method for performing quality control under a double programming model. We use a combination of easily accessible SAS macros, R custom functions and libraries, and simple command-line interface to implement standard processes including dataset comparison between R and SAS files, log checks, and batch runs. We address the benefits and detriments of using both R and SAS as opposed to two SAS programs, the challenges that come with implementation, and the work yet to come regarding this project within our organization | quality control, adam dataset, in=, cdisc | source poster source |
3 |
| 182 | USUBJID Challenges: Representing Multiple Participations for Analysis and Reporting | Emily Bernstein-Hanlon, Sirisha Davuluri |
Multiple enrollments in a study can be challenging to address, particularly with increasing complexity of trial designs, evolving regulatory requirements and without clear data standards. Scenarios range from study designs where re-enrollment by the same participant is allowed to participants enrolling multiple times when a protocol does not allow it. Per the FDA Study Data Technical Conformance Guide, SUBJID should be unique for each screening or enrollment in a study, but USUBJID should remain the same across all studies involving the product. Appropriate and timely data collection is necessary to be able to link all the enrollments together. Challenges can arise when initial data entry is incomplete, and updates are made between multiple submissions. Several approaches can be used to handle re-enrollment, each with its own advantages and disadvantages. This paper will discuss these scenarios and possible approaches, with a focus on ADaM and the corresponding analysis and reporting | adam dataset, sdtm domain, treatment arm, adverse events, study day, unique key, system organ class, treatment period, data cleaning, cdisc | source presentation source |
9 |
| 183 | Utilising Machine Learning Algorithms to Detect the Outlier or Inconsistencies in the Analysis Results in Datasets and Summary Reports | Weiwei Xu, Linshan Yuan |
Safety related summary tables (e.g., AE, LB, VS, EG) are a core part of the Clinical Study Report (CSR) and are usually lengthy reports when we display the statistical results by visits/time points. In this poster we will explain how we applied multiple ML methods to 1) detect the outliers (e.g., extreme values in small Ns) and 2) obtain data pattern related to the derivation of Treatment Emergent Adverse Event (TEAE) flag | proc hpsplit, outlier, machine learning | source poster source |
5 |
| 184 | Validation Unveiled: Striking the Balance for Data Integrity and Compliance in R, Python, and SAS | Sy J. Truong | The Statistical Computing Environment (SCE) has expanded beyond SAS, encompassing languages such as Python and R. While the validation of SAS is widely recognized in Pharma, validating these additional languages is less familiar. This paper outlines new approaches for testing and validating Python and R, drawing upon concepts from established SAS validation best practices. Risk-Based: Python and R exploratory analysis differs from SAS, leading to less formal validation. User Acceptance: Similar to SAS OQ testing, executing user-representative Python and R scripts on user data is effective. Performance Qualification: Unique performance considerations are tested when multiple users run Python, R, and SAS code on large datasets in parallel. AI-Generated Test Scripts: Tools like ChatGPT aid in generating useful Python and R tests while safeguarding proprietary data. Python and R, being open-source tools, have distinct use cases compared to SAS | data cleaning, clinical data, clinical trial, survival analysis, elapsed time, data integrity, excel, sample, data science, regression analysis | source presentation source |
10 |
| 185 | What Role Does ChatGPT Have in your Clinical Programming? | Andy Ratcliffe | This paper explores the potential role of AI bots and chat bots in the creation of statistical analyses for regulatory review. With recent advancements in AI technology, bots are capable of generating essays, blog posts, and promotional copy, raising the question of whether they can also contribute to the creation of statistical analyses. The paper considers the advantages and limitations of using AI bots in this context and examines the feasibility of their application in regulatory review. While bots have the ability to automate tasks and reduce human error, they may not possess the necessary contextual understanding and expertise required for creating accurate statistical analyses. This paper is relevant for anyone interested in the potential applications of AI bots in regulatory review processes | data science, artificial intelligence, data cleaning | source presentation source |
7 |
| 186 | Whodunnit? Automation Tools Accelerate Statistical Analysis in an SCE with Full Compliance | Shivoy Pandita, Jörn Bilow |
Presentation | data science, clinical programming, 21 cfr part 11, clinical data, api, data lineage, audit trail | source | 13 |
| 187 | {admiral} - From Open Source Through to Company Implementation | Ross Farrugia | At Roche we have been on a multi-year journey to introduce an open source primary backbone to our clinical reporting workflows, based on the R programming language and industry collaborations around pharmaverse. In this paper, we discuss our experiences specifically around the ADaM (Analysis Data Model) and building and implementing a solution in this area that traditionally has equated to ~50% of the time spent by our analytical data scientists working on clinical reporting | clinical trial, bds, cdisc | 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.