HOME
QUICK SEARCH
AUTHORS
KEYWORDS
HELP
BUGSPHUSE EU 2023 Papers - Index
190 papers. Source: PHUSE (Pharmaceutical Users Software Exchange) EU Connect conference proceedings.
| # | Title× | Authors× | Abstract× | Keywords× | Extras | Pages |
|---|---|---|---|---|---|---|
| 1 | "It's ARM, Jim, But Not As We Know It" - Our Journey to Automating Output Delivery Through Metadata | Edwin van Stein, Warwick Benger |
To improve traceability, transparency and consistency GSK is implementing automation of analyses and output creation through standardized metadata. Tool-agnostic Analysis Results Metadata (not to be confused with CDISC ARM!) is used to produce Analysis Results Data. Following the WORM principle (Write Once Read Many) tool-agnostic Analysis Output Metadata then uses those results to render them to different displays (tables, figures and listings, both post-text and in-text), other outputs (e.g., XML files for disclosure) and systems (in whatever format those systems require) without reproducing the numbers (and potentially getting them wrong!). This leads to a process where it's easier to validate and reuse results and to build tools in a multilingual setting. This paper focuses on the metadata, the structure of the Analysis Results Data and how it all fits together on our journey to standardizing our | cdisc | source presentation source |
11 |
| 2 | 'Beep Beep': A Metadata Code Writer to Generate SAS Programs | Giuseppe Di Monaco | Datasets a nd TFL programs are written by many different programmers , each using their own different programming techniques and styles . Often these programs are not built using available metadata and standardized business rules, making it difficult to modify, share , and automate program creation, all at a significant cost in both time and money. To facilitate the sharing of standardized code , we have created the Navigator application. This is a metadata-driven code writer that generates standard, well-structured SAS programs that u se validated SAS macros to create datasets and TFL. Specifically, N avigator allows the user to associate ADaM (or TFL ) metadata specifications with their respective source data and metadata descriptions. These relationship definitions (mappings) provide the information required for Navigator to write SAS programs | macro definition, data validation rules, dataset, response variable, proc sql, data cleaning, raw data, cdisc | source presentation source |
10 |
| 3 | 3P to V - How People, Process and Product Generate Mutual Business Value: Review of an SCE Journey | Javeed Paray, Marc Jantke |
Abbott EPD embarked on a journey to building an integrated, traceable and compliant e2e patient data center (PDC) to manage, analyze, report & gain insights from Clinical (ongoing and historical) & Real World Data. Abbott and Entimo teamed up to implement entimICE FastTrack as the core of the Abbott EPD PDC platform. We would like to share insights and means to resolve roadblocks, which can be valuable and applicable for many other industry projects built around a sponsor/vendor relationship. The project team used its open and problem-solving mindset (People), and found new ways of taking transparency and efficiency in their collaboration to a higher level (Process) to create a platform built around a COTS product optimized for customer demands and ready for future extension (Product). Understanding the collaboration as a partnership brought significant benefits for both sides (Value) | clinical data, data cleaning | source presentation source |
3 |
| 4 | A Beginner's Guide to Overcome the Attrition Bias in Clinical Research: The Method of Inverse Probability of Attrition Weighting (IPAW) | Parichay Srivastava | Longitudinal studies play a crucial role in understanding changes over time, but they are often subject to attrition bias. Attrition bias occurs when participants drop out from a study or are lost to follow -up and the characteristics of those who drop out differ from those who remain, leading to biased estimates. To address this issue, Inverse Probability of Attrition Weighting (IPAW) has emerged as a powerful technique. IPAW aims to adjust for attrition bias by reweighting the observations in the sample. It involves assigning weights to participants based on the probability of remaining in the study at each time point using predictive modelling technique accounting participant characteristics and other relevant factors over time. A competing alternative method, namely propensity score matching involves matching participants who have similar propensity scores | clinical research, propensity score, sample, logistic regression, regression analysis, missing value, baseline characteristics, treatment effect, logistic distribution, clinical trial, propensity score matching, population, treatment arm, adverse events | source presentation source |
9 |
| 5 | A Beginner's Journey with SAS: One Small SAS for Man... | Dale Armstrong | Four years ago, I had never even heard of SAS. After leaving my job in physical sciences and transitioning into a statistical programming role, I would like to share some of my insights and learnings from this process. Why did I choose SAS out of the multitude of other languages and career choices? How did I get started, what was my mind set, what could I bring to the table? How was my first year, my first project, my first large project; what have I learnt over my current career? What went well, what went terribly? What's next? | sas programming, statistical programmer | source presentation source |
8 |
| 6 | A Data Lake is not a CDR and a CDR is not a Data Lake... Understanding the Differences and the Value of Each | Chris Decker | According to Wikipedia a CDR is a database that consolidates data from a variety of clinical sources to present a unified view of a single patient while a 'Data Lake' is a centralized repository that allows organizations to store all their structured and unstructured data at any scale in its native format, without the need for preprocessing or modeling. Within our industry we struggle to align on the definition of these two things and how both can be leveraged together to meet the needs of an organization. This presentation will provide a foundation for understanding each and how they can be used within your organization as well as a look towards the future. | data lake, clinical trial, clinical data | source presentation source |
5 |
| 7 | A Dive Into the World of Biomedical Concepts | Berber Snoeijer | What are biomedical concepts and how can we practically use them? This paper describes the aspects and use of biomedical concepts from various dictionaries. This includes CDISC COSMoS and it's representation in CDISC USDM. One of the difficulties is the linkage (or mapping) of biomedical concepts from one dictionary to another. This can be handled by a combination of automation for exact matches and manual one-off review of non-exact matches using automated documentation and verification in these cases | clinical trial, cdisc controlled terminology, cdisc | source presentation source |
5 |
| 8 | A Green CRO - Why We Should All Be on a Journey of Continuous Improvement | Jessica Hirsch | In 2021, the issue of climate change came to the forefront with the COVID-19 pandemic inadvertently allowing us to reach some climate targets for the first time. We realised that the pandemic would irreversibly change our work lives and despite the doomsday predictions, the world could be saved. We want to be a carbon-neutral company that focuses on sustainability. We have already taken real steps towards this goal, ready to improve further. Our team take sustainable practices into their private lives and have become ambassadors for our mission. By committing to carbon neutrality and sustainability, Contract Research Organisations along with other pharmaceutical companies are joining the green economy. Together we provide a blueprint for green transformation to the pharmaceutical industry. No matter what sector of the pharmaceutical industry (including CROs), we can all become more sustainable. We invite you to join us in the mission to stop climate change. | pharmaceutical industry, service provider | source presentation source |
4 |
| 9 | A new standard of excellence; introducing the annotated Screen Report | Franklin E. Menius | With the adoption of electronic Clinical Outcome Assessments (eCOA) and electronic Patient Reported Outcomes (ePRO) for collection of questionnaires and patient clinical data, there is a need for detailed data lineage from data collection to analysis and submission. This need can easily be met using the growing CDISC QRS supplements library and annotated eCOA/ePRO device screen reports. Like an annotated Case Report Form (aCRF), the annotated Screen Reports (aSR) document the location of collected data on a participant or clinician device to corresponding domains and variables included in the intermediate and the submitted datasets. When utilized, these annotated Screen Reports (aSR) can be a vital tool for anyone accessing these data including programmers, statisticians, and reviewers at regulatory agencies | crf, clinical trial, cdisc | source presentation source |
8 |
| 10 | A Privacy Methodology Toolkit to Improve Cross-Industry Clinical Data Reuse | Boris Grimm, Diana Ye |
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 |
| 11 | A Robust Framework to Provide and Validate Real-Time SDTM Data | Bremer Louw | Clinical trials rely heavily on continuous data monitoring for both patient safety and data quality surveillance. However, using SDTM data for data monitoring has always been a challenge given the obstacles associated with re -running complex SAS/R programs and re-validating the resulting SDTM data on a frequent basis. This white paper presents a framework that explores the potential solutions to make it a reality. We will explore the major components of a framework required to implement a real -time SDTM data solution and present an implementation used on production studies which include the following - Data Acquisition. How to routinely gather the source data needed for SDTM. Data Processing. Building a robust conversion mechanism. Data Publishing. Routinely disseminate the SDTM data to downstream consumers. Data Validation. Automatically running validation checks using both regulatory/CDISC conformance and user- specific rules. * Issue Detection |
data cleaning, raw data, independent qc, data validation rules, completeness, edc, cdisc | source presentation source |
11 |
| 12 | ADTTE: Through a Beginner's Eye ... In Under a Minute! | Daya John | Poster | source | 1 | |
| 13 | Advanced Analytics for Data Quality Assessment in Countries and Regions Affected by Crisis | Archana Nayak, Mohammad Umarfaruque |
Due to unprecedented events in recent years, the conduct of clinical trials has been increasingly affected by crisis at country and regional level. The inability to conduct study visits, difficulties in recording high quality data and delivering effective quality oversight have the potential to significantly impact trial integrity and patient safety. In this paper, we highlight internal methodology implemented as part of Central Statistical Surveillance. Our tool identifies and compares viewpoints among different countries signals and assesses their risk magnitudes to indicate potential areas of high, medium, and low risk at various levels of the study data. These insights provide opportunities to apply appropriate corrective and preventive actions, thereby ensuring higher data quality and dynamic monitoring of ongoing study data | clinical trial, data cleaning, clinical data, subject visits, insurance risk score | source presentation source |
13 |
| 14 | An Agile Approach to Onboarding | Ben White | A discussion on how we can better onboard people new to the life sciences industry whether they be Industrial Placement students, Graduates, or Career Changers by taking an Agile approach. When first entering the pharmaceutical industry, there are many new and complex areas to learn and understand in order to perform in your new role. For example, CDISC standards, ICH guidelines, Numerous SOPs, and Working instructions. It can happen that people new to industry become overwhelmed or de-motivated in their ability to be successful. I will explore how an Agile mindset and adapting tools from the Agile methodology can keep new industry starters more engaged and get them contributing to deliverables with reduced downtime | life sciences, adamig, cdisc | source presentation source |
3 |
| 15 | An Automated Compliance in Project is possible! | Elisa Natale, Thierry Paquet |
In the pharmaceutical industry, quality and compliance are the two pillars of submissions to authorities. Global and local SOPs design our ways of working to ensure rigor, correct ness, and reproducibility. Ensuring compliance (in programming flow, specifications, documentation, testing, etc.) may be challenging for statistical programmers working in big teams and under time pressure. Program0status tool automates the project compliance checks based on SOPs expectations. It scans the development folder's structure to identify potential discrepancies in files locations, timestamps, content, and naming conventions. A report is created to help the statistical programmer to take corrective actions and make the project fully compliant. This standard tool ensures consistent compliance across projects | data cleaning, statistical programmer | source presentation source |
9 |
| 16 | An Experience of Creating Questionnaire ADaMs Using the Admiral ADQS Package | Luke Thompson, David L. Ward |
{admiral} R packages are going to revolutionise the way in which we program ADaMs and provide us the tools required to produce these datasets using simple vignettes. The {admiral} ADQS Vignette is a new vignette that can manipulate questionnaire data into ADaMs. The number of questionnaires used in neuroscience is vast, but typically there are some that are more common. With the development of this new vignette, a change in the way we work and lack of templates available, the ADQS vignette could be used to develop some templates. Determination of which templates could be created were based on most commonly used, difficulty of programming and therapeutic areas of most interest. Once determined, initial templates were created and then SPRINTS were set up to develop more. Some difficulties were encountered, but overall the experience was very positive and we now have a basis to reliably create questionnaire ADaMs using R | clinical trial, proof of concept | source presentation source |
4 |
| 17 | Analysis and Reporting of Immunogenicity Anti-drug Antibody Data | Nicola Perry | Immunogenicity is a measure of the immune response to a therapeutic drug (e.g., a monoclonal antibody) resulting in generation of anti-drug antibodies (ADAs, also known as anti-therapeutic antibodies (ATAs)). It is becoming an increasingly common and important constituent of clinical trial reporting and the FDA is focusing on this topic more and more. This paper aims to provide information on the type of data provided after clinical sample testing through a sequence of assays and a possible approach to converting IS SDTM into CDISC ADaM structure, discussing data handling techniques to support the various analysis and reporting requirements | sample, follow-up period, adverse events, cdisc | source poster source |
6 |
| 18 | Analytical Risk-Based Monitoring (ARBM) - How Central Monitors Detect the Ripple in the Dataflow that Hides Danger On Site | Juergen Zech | Central monitors analyze the technology-driven, real-time picture of data quality provided by modern analytical risk based monitoring (ARBM) tools. By focusing on critical data that represent the greatest risk to the study objectives a central monitor is able to view data from different and more holistic perspectives. On-site monitors (CRAs) in contrast often only see the impact of new data on individual subjects at one timepoint in comparison with subjects at one or several similar sites. However, switching between a study-wide, country-, site- or subject-focused perspective and observing data changes in real time can amplify risk signals. Here we present a case in which central monitoring detected previously overlooked frequent changes of the eligibility status of patients. A further investigation in this signal showed that several sites had problems to confidently calculate a relevant score and therefore continued to change its value. Consequently, armed with this knowledge the study team was able to mitigate the issues and prevent the potential enrolling of not eligible patients and wrong stratification before the end of the recruitment period. | clinical trial, audit trail | source presentation source |
4 |
| 19 | Applying Central Statistical Monitoring to Find Anomalies and Address Quality Issues in Clinical Trial Data | Wiktor Majewski | Finding anomalies and correcting errors in clinical trial data as the trial is ongoing is a way to ensure consistent high- quality data output, that is critical to an accurate safety and efficacy assessment. This paper describes how Central Statistical Monitoring (CSM) is employed in Risk -Based Quality Management (RBQM) operation framework. Our application of CSM aims to identify trial sites and individual subjects with potential data quality issues and takes into account both time series and distribution of clinical parameters and events - on site and subject level. The findings - which include atypical data patterns, outlying values and low variability of variables (e.g. vital signs and laboratory test results) - complement the central monitoring targeted reviews and review of operational metrics like Key Risk Indicators and Quality Tolerance Limits | clinical data, clinical trial, data governance, data cleaning | source presentation source |
5 |
| 20 | Automated Pipelines for Non-CRF Value Level Metadata, A Meta-programming Approach | Selena Baset, Philipp Ernst, Didier Clément |
Like every step in any clinical trial, data collection is a process that lends itself to harmonization. In Roche, collected data should conform to collection standards that are based on the CDISC-controlled terminologies or their Roche sponsor extension. This means that every variable's permissible value is bound to fall within pre-established enumerated value domains. While dictating the permissible values for a given variable in isolation from others is readily achievable using the existing standard models, dictating the permissible combinations of values for variables collected together is more problematic as it requires maintaining Value Level Metadata (VLM) according to schemas that vary dynamically following the different data collection settings and the applicable standard version. We propose an end-to-end solution to the VLM management problem by designing a generic VLM ontology for capturing machine-readable permutations of permissible values combined with a meta-programming approach to automatically create authoring tools and pipelines to propagate changes following updates to the underlying data collection standards. We demonstrate the end-to-end solution applied to VLM schemas of two different non-CRF models: 1) the ophthalmologic grading parameters and 2) digital measures passive monitoring. | clinical trial, controlled terminology, cdisc | source presentation source |
10 |
| 21 | Automating ADaM Dataset Generation via Definitive Specification Writing with A Domain-Specific Language | Kang Zhao | ADaM datasets are standardized assets in supporting clinical trial analysis and is required for regulatory submissions to FDA and other health authorities. To develop ADaM datasets, study programmers first need to specify analytical variables based on a provided Statistical Analysis Plan. Computer programs are then written as implementations. The final step is batch execution with explicit execution orders declared by study programmers. The automation of ADaM development has been challenging. Solutions resort to modularized library development together with metadata repository to achieve some level of automation. Although it enables reusing partial codes from previous studies, it instigates duplicated efforts in specification writing and coding. In this paper, we introduce a novel solution to make the specifications as single source-of-truth while achieving a completely automated deriving process | adam dataset, bds, imputation, response variable, cdisc | source presentation source |
7 |
| 22 | Automation and Orchestration of Data Science Applications Using OpenAPI | Dmitry Zakharov, Alexander Lüders |
The complexity of clinical research processes is rapidly growing. Today's data science requires working with a variety of tools and applications. Interaction with those applications is tedious, error prone and inefficient if done manually. Modern applications often provide built-in means to automate their functions. However, across different applications such means of automation are not homogeneous, and as such are not well suited for overall orchestration of end-to-end processes. OpenAPI specification aims to help solving these challenges. It defines a unified way of specifying how applications can be controlled from the outside. Such Open APIs allow generation of client code in different languages, enabling automation from Python, R, SAS or graphical user interfaces that can visualize the generated code. This approach provides a reliable, secure and convenient way to manage orchestration and automation for seamless cross-application processes, with the flexibility to choose the optimal tool/language for it. | data science, api, clinical research, rest api | source presentation source |
3 |
| 23 | Avoid Major Pitfalls When Using Real-World Evidence (RWE) in Regulatory Submissions | Alex Ide, Mark Lambrecht |
The cost of bringing a medicinal product through all phases of drug development is upwards of USD $1.5 to 2.5 billion, with clinical trials making up a significant proportion of related R&D spend.¹ It's complicated and expensive. Submitting regulatory dossiers to authorities, incl. outcomes of clinical trials, increasingly requires a combination of evidence generated from harmonized data assets (RCTs) along with RWD. Access to global, row level RWD coupled with clinical evidence, usage and benefit/risk ratio from analyzing RWD is fundamental to drug development, programmatic risk reduction, commercial expansion, and patient outcomes. It is of increasing value when making submissions to regulatory authorities. SAS advanced analytics, combined with BC Platforms' data access to diverse global patient cohorts, delivers highly controlled, bespoke and fit-for-purpose data, evidence, and insights. Tapping into the combined power of secure RWD sources and analytics helps to avoid major pitfalls when using RWE in regulatory submissions. | regulatory submission, clinical research, clinical trial, data science | source presentation source |
3 |
| 24 | Be Future-Minded While Designing Your Next Generation SCE | Susanne M. Andersen, Olivier Bouchard |
In current context, life science has more than ever to be more efficient by developing new & innovative drugs. Efficiency is all about Automation, Acceleration, Collaboration, High Value Tasks and Governance. For years, life science companies have invested in customized solutions with high maintenance costs. A SCE is a key solution to look for when optimizing and reducing time to develop, submit to approval for new medicines. A next generation SAS SCE, leveraging SAS Viya is built with that future in mind. One goal: Improve clinical development work in a GxP environment. What is key focus: Maintain data integrity, effectively assess safety and efficacy of compounds, collaborate across trials, and ultimately streamline and accelerate the drug development process. In this paper and presentation, we will present how a modular, multi-language, cloud ready content repository and statistical compute environment is key to achieving that transformation. Page | 2 | life sciences, clinical trial, clinical research, clinical data | source presentation source |
12 |
| 25 | Becoming an Integrated Part of the Open-Source Pharma Community | Philip Bowsher | clinical research, clinical trial | source presentation source |
11 | |
| 26 | Bite-Size Learning - Statistics for Programmers | Alina Striha, Rachel Dunk |
Poster | source | 1 | |
| 27 | Boosting SAS Programming Efficiency with ChatGPT: A Clinical Trials Perspective | Illia Skliar | As the field of clinical trials continues to evolve, there is increasing interest in leveraging artificial intelligence (AI) and natural language processing (NLP) technologies to streamline programming tasks and improve efficiency. One promising technology in this space is ChatGPT, an advanced language model that has shown impressive results in a variety of applications. In this session, we will explore the potential of integrating ChatGPT into SAS/STAT programming workflows for clinical trials. We will begin by providing a simple explanation of what ChatGPT is and how it works. From there, we will investigate how it can be integrated into SAS programming IDEs, including both data-based and non-data-based approaches. Next, we will examine the specific tasks that ChatGPT can help with, including creating ADaM datasets based on CDISC standards, generating figures and tables using SAS 9.4 programming language, and more |
clinical trial, sas programming, sas programmer, clinical data, data science, macro variable, output delivery system, artificial intelligence, box plot, clinical research, data cleaning, cdisc | source presentation source |
11 |
| 28 | Building a Large-Scale Clinical Trial's IT System with In-House Software and OpenClinica | Sonja Grotjahn, Allen Young |
For one of the large scale clinical trials run at Oxford Population Health department of Oxford University, the AscendPlus trial, we combined third-party cloud based platform OpenClinica with our in-house built software. OpenClinica is a clinical data management and electronic data capture system which offers many features to enable easy and fast development of software for a clinical trial. However, we found that due to the nature and the scale of our study, there were several software components that needed to be built in-house, such as treatment supply management, clinics administration and handling postal and email communication with participants and responses. For the smooth running of the trial it was important to provide timely exchange and synchronization of data between OpenClinica and the in-house software. OpenClinica provides several methods for interacting programmatically with its in-the-cloud data | clinical data, adverse events, clinical trial | source presentation source |
7 |
| 29 | Building a Scalable Utility Service to Make Multi lingual Applications Available to the Masses! | Aksel Thomsen | To reduce the workload from traditionally manual tasks in our clinical trials, we have developed a language independent utility service with a webpage, that integrates with our clinical data repository (SAS Life Science Analytics Framework). The service enables the developer to build smaller containerized applications to process these tasks in their favourite language (R, Python, or any other), and make them readily available to their colleagues, without them being required to have any knowledge of the inner workings or language of the application. The service is hosted in MS Azure and consists of several container apps. The primary being a Shiny app that provides a webpage which works as a graphical user interface and distributor. Each of the individual services are hosted as separate orchestrated app containers | clinical data, rest api, life sciences | source presentation source |
3 |
| 30 | Building an integrated pre-clinical and clinical data platform to enable rapid translational data review | Brenda Finney, Ali Dootson Instem |
With the FDA announcement that it is no longer a requirement to test on animals prior to human trials; there is potential for a paradigm shift in the conduct of regulatory investigations. Using integrated databases containing public clinical and legacy preclinical data, and the opportunity to add anonymized clinical trial data we show how visualization and modelling tools could aid in planning testing programs. This presentation will look at examples on how a simple search in the integrated database can pull results from multiple preclinical and clinical studies from several disparate databases. We will show how event counts within the MedDRA system organ class (SOC) terms can overlap with pre-clinical SOC results across the databases and how the nature of these potentially translationally relevant findings was investigated by reviewing the results | system organ class, clinical trial, group by, clinical data | source presentation source |
12 |
| 31 | CDISC Conformance Rules and the CORE Engine: Continuing the Road to Adoption | Nick De Donder | 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 |
6 |
| 32 | CDISC's ADTTE Domain: Time-to-Event Analysis From a Programmer's Point of View | Vitali Gering, Sophia Praeger |
Poster | survival analysis, kaplan-meier estimator | source | 1 |
| 33 | Challenges in Oncology Modular studies: The Fortrea approach | Francesca Dall'Acqua | An Umbrella study is a type of master protocol that is defined as "a unifying study design that includes multiple subgroups and sub-studies, with patients having same or different diseases and that employ one or multiple drugs to treat it."[1]. Such master protocol allows for testing multiple agents simultaneously and may include specified modifications while the trial is ongoing | clinical trial, macro variable, csr, adam dataset, analysis set, sdtm domain, excel, data science, cdisc | source presentation source |
7 |
| 34 | Clinical Graphs Using R | Duong Tran | Poster | source | 1 | |
| 35 | CodeGeneration: Turning Metadata into SAS, R or Python | Stuart Malcolm | clinical trial, clinical data | source presentation source |
6 | |
| 36 | Comparing Analysis Method Implementations in Software (CAMIS): An Open Source Repository to Document Differences in Statistical Methodology Across Software | Christina Fillmore | data science, confidence interval, clinical data, proc phreg, clinical trial, log-rank test, pharmaceutical industry, rank test, likelihood function, missing value | source presentation source |
7 | |
| 37 | Comparison of power between different recurrent events analysis methods using simulation | Soumitra Kar | Recent research in Cardiovascular diseases has resulted in significant advancements in statistical methodologies which have played crucial role to improve the understanding of CVD risk factors and treatment of CVD and to find out effective strategies to reduce burden of CVD. Although Cox proportional hazards regression models have traditionally been used to analyse CVD data, the chronic nature of the disease necessitates addressing complicated scenarios such as recurrent events which provides a more comprehensive picture of disease progression and quality of life. In RCT with recurrent events the power calculation methods are less straightforward and different approaches can be taken | hazard ratio, clinical trial, competing risks, p-value, proc phreg | source presentation source |
7 |
| 38 | Concomitant Medication Duration Calculation Using the SAS DS2 Procedure | Yirong Cao | Concomitant medications are medications other than investigational study drug which are taken at any time on- treatment. In practice, t he calculation of overall duration can cause a lot of challenges due to the combination of complex scenarios including sequential treatment, on and off treatment and overlapping among these treatments. This paper solves these problems by using three different methods from the SAS DS2 procedure. All are available in SAS BASE and demonstrate enhanced programming efficiency: arrays, multi-threading, and packages. The DS2 procedure is the method of choice because of its advantages in re-usability, portability in SAS Viya, multithreading, numerical precision, and viability from multiple databases |
concomitant medications, cpu time, data step, sas base, sas viya, temporary array | source presentation source |
10 |
| 39 | Connecting the Dots to Shape Our Future. The Exciting Journey of an Evolving Statistical Programming Team | Alberto Montironi | Historically, the recruitment of young professionals in Statistical Programming departments of pharma companies has primarily focused on the creation of the next generation programmers or replacement of existing roles with individuals possessing similar skillsets. In todays landscape, as academic institutions increasingly adopt open-source languages and code hosting collaboration platforms (e.g. GitHub), a mismatch emerges between the programming skills offered by early career candidates and those already present within companies. This could potentially become a critical issue. At UCB, we see the evolving landscape as an opportunity to connect the dots, unlocking the full potential from early career candidates joining our organization and embracing a change in how we support our company | clinical programming, csr | source presentation source |
4 |
| 40 | COSA Updates - The Latest From the CDISC Open-Source Alliance | Samuel Hume, Charles Shadle |
Poster | source | 1 | |
| 41 | Creating applications with mash-ups: user written functions, stored sql queries, and interfaces | Stephen Prust | In the pharmaceutical development world the processes of clinical trial reporting, responding to regulator queries and so on it can be difficult to create applications. Combining techniques from the tools and concepts available to us can be a fruitful area of improving both speed of delivery and ease of development. This paper looks at "mashing-up" some of these suitable techniques | proc sql, sample, proc step | source presentation source |
7 |
| 42 | Creating CORE Rules from Biomedical Concepts | Johannes Ulander | The work on improving our clinical data standards to utilize modern web development stack and graph technology is happening. It impacts how we access and manage standards, providing better control and opportunities to change the data flow for collecting and reporting clinical data. Now that the CDISC Open Rules Engine (CORE) and the first Biomedical Concepts are available from CDISC, it is time to start using them to take the next step in improving the quality of the data delivered in clinical studies and easing the burden of QC for programmers. This presentation will go through the steps used to automatically create Core rules matching Biomedical Concepts accessed via the CDISC Library API and show how easy it is to start doing this on your own. It will show how this also impacts how we automate creation of forms and discuss learnings and future improvements | sdtm domain, clinical data, data cleaning, clinical trial, cdisc | source presentation source |
8 |
| 43 | Cross-Team Collaboration: Tech Solutions for Bridging the Gap Between Biostats and Data Management | Alice Perry, Jenn Fukuyama |
Poster | source | 1 | |
| 44 | Customisable Smart Search Tool for Project Leads | Bibas Timilsina | The FILENAME statement has been around a long time. It has been traditionally used to reference an external file on disk. However, there are many "not so well-known" functionalities, for which the FILENAME statement is a boon. This paper discusses, the various tricks to develop the customizable search tool using the power of FILENAME statement together with well-known PIPE option to search the text token in the various file types. This paper is based on real time usage of FILENAME statement with PIPE option to search the data point from Raw, SDTM or ADaM database, identifying the common possible data issues from database, developing customizable personal standard tools such as log checker, QC automation, derivation count etc |
filename statement, macro definition, proc step, cdisc | source presentation source |
6 |
| 45 | Data Managers and R Programming: The Unlikely Pair | Anaelle Perretti (Bonetta) | Let's imagine stone age, Data Manager s did not have any tool available to perform data review on extracted SAS datasets. We were laboriously using EXCEL files and playing with the vlookup function. Data filters were manually created at each review round and precious time was lost. At Idorsia, we have sought help from our Data Science team. They agreed to train us on the R programming language and introduced us to the tidyverse. Under their supervision, we are replacing old-fashioned EXCEL work with R scripts to program monthly metrics, data comparisons, data cleaning cross-checks and to perform our oversight activities. In this paper , we will discuss the experience of learning R for people with little or no programming background and share examples of how we are applying our new skills to transform the way we work with data |
clinical trial, excel, clinical database, dataset, edc, data validation rules, protocol deviations, data science, data review, quality control, r programming | source presentation source |
9 |
| 46 | Data Profiling Made Shiny - Accelerating Clinical Study Delivery and Submission | Marc Hughes, Hao (Harry) Chen |
Poster: The Motivation Before a programmer wishes to work with some data, it's methodical to begin by gaining an understanding about the basics of that data. NULL values, variable names/types and number of observations are key concepts to consider before any quality programming can begin. Some of this information can be viewed by simply opening the dataset and scrolling back and forth, up and down - a tedious process should you forget this information and have to go searching again. SAS macros can be created to try and automate this process but they will likely be created every time a user needs to profile their data. It was deemed appropriate to have a centralised program that can run within a few seconds and produce a report summarising the results that the user could refer to again… The Value This tool provides functionalities that facilitate swift reviews and queries of data profiles, while also enabling documentation for future use | source | 1 | |
| 47 | Dataset-JSON | Samuel Hume, Jesse Anderson |
Presentation | data science, cdisc | source | 26 |
| 48 | Design Considerations for Building a System for Automated Statistical Analysis | Nicolai Skov Johnsen, Christian Olsen |
data science, safety population | source presentation source |
6 | |
| 49 | Detroit, Michigan: A Case Study for DEI Via Decentralised Clinical Trials (DCTs) | Mark Rothe | In April 2022, the FDA issued a draft guidance entitled Diversity Plans to Improve Enrollment of Participants From Underrepresented Racial and Ethnic Populations in Clinical Trials (FDA 13-April-2022)-which recommends Sponsors submit enrollment goals to the FDA as part of Race and Ethnicity Diversity Plans. Furthermore, the FDA issued a draft guidance in May 2023 entitled Decentralized Clinical Trials (DCTs) for Drugs, Biological Products, and Devices: Guidance for Industry, Investigators, and Other Stakeholders (FDA 01-May-2023). This document advocates for the implementation of DCTs, when appropriate, where all or some study activities occur at patient homes, facilities of local healthcare providers (HCPs), etc. Taken together these guidance documents are a catalyst for Sponsors to proactively locate diverse and marginalized patient populations while simultaneously connecting them with improved opportunities for study enrollment | estimate, machine learning, unsupervised learning, missing value, histogram, false discovery rate, clinical trial, dimensionality reduction, imputation | source presentation source |
26 |
| 50 | Developing a Shiny Web Application to Address Suboptimal Treatment of Urology Patients Through Predictive Modelling and Improved Communication of Science | Federico Concas, Juan Manuel Palacios-Moreno |
Around 20% of men over the age of 55 experience urinary symptoms due to an enlargement of the prostate (LUTS/BPE) and are at risk of disease progression. Most of these men are sub -optimally treated, contrary to published guidelines. More effective communication of scientific information is needed to support physicians' data-driven decisions and improve patient management. To contribute to this field, we analysed large GSK trial datasets and developed state -of-the-art statistical models that incorporate patient and treatment characteristics to predict symptoms and outcomes. We then developed an educational web application in Shiny that presents the study information in a more digestible manner. This application uses the models to facilitate visualisation and understanding of predicted outcomes for individual profiles. We have translated it into several languages and made it available to physicians worldwide at www.bphtool.com | data science, proof of concept | source presentation source |
5 |
| 51 | Developing and Implementing CDISC CORE rules: from zero to hero | Marisa Wyckmans | In clinical research, we have one mission: creating high quality data packages for statistical analysis, reporting, and ultimately submission to regulatory agencies. Becoming a CDISC CORE volunteer was an exceptional opportunity to contribute to this. The enormous amount of conformance rules for SDTM, ADaM, and SEND makes this CDISC project a huge challenge. From project kick off with the volunteers in September 2021 until today, the CDISC CORE project evolved from basic concept into established validation system. This paper will share experiences of a volunteer in the CDISC CORE Rules Development team. Examples will demonstrate the rule editor functionality and illustrate how easy developing your own rules can be, even for someone lacking solid programming skills. In addition, time has now come to start implementation. Will we integrate the CDISC CORE engine in current systems or use it as a separate tool? And how about the company's own specific rules? | data cleaning, clinical research, cdisc | source presentation source |
8 |
| 52 | Development Safety Update Report (DSUR)/Periodic Safety Update Report (PSUR) | Phillip Spowart | Development Safety Update Reports (DSUR) are annual safety reports that provide a review and evaluation of the safety data collected during the reporting period a compound is under investigation. Periodic Safety Update Reports (PSUR) are a periodic safety report that provides a comprehensive safety report of a marketed compound. DSURs and PSURs follow a similar process and are often confused as being the same thing. This paper will discuss what DSURs and PSURs are, why DSURs and PSURs required and how DSURs and PSURs are developed. The differences between the two reports should remove the confusion | adverse events, clinical trial, serious adverse event, sas enterprise guide | source presentation source |
8 |
| 53 | Discovering the Power of Graph Data | William McDermott, Jonathan Deacon |
Poster | source | 1 | |
| 54 | Drafting Your Plain Language Summaries: Hurdles to Overcome and How to Overcome Them | Eliza Gray | As the global regulatory landscape changes due to the European Union (EU) Clinical Trial Regulation (CTR) No. 536/2014, sponsors must now submit a plain language (also referred to as lay language) summary (PLS) of results for trials conducted in the EU[1]. All EU trials will now require a PLS to be submitted to the Clinical Trials Information System (CTIS) at the time of results disclosure, i.e. within one year following the study completion for adult trials, and within six months following the study completion for paediatric trials conducted in the EU. Since the requirement for a PLS was detailed under EU CTR, other countries have similarly followed suit. While PLS provision was already required in the Netherlands and Moldova for example, the United Kingdom (UK) have recently detailed a legal requirement for PLS provision for UK-run trials | clinical trial, data science | source presentation source |
8 |
| 55 | Drive Results with Emotional Intelligence | Jess Anderson | Hard skills may get your foot in the door, however soft skills enable you to open the doors to come. Hard skills (usually technical) are measurable and trained or gained through education. In contrast, soft skills (such as emotional intelligence) are inexact, developed, and ironically often harder to attain. As driven individuals we continuously work on building our hard skills so why do we often neglect to similarly enhance our soft skills? Leading a project team forward demands more than giving corporal orders to hit targets. It requires building strong relationships within and across teams to create a positive, collaborative, and trusting environment. People do their best work when they feel supported, and people drive processes. This paper advises how you can elevate your soft skills as a team leader which alongside hard skills will power a project team forward to a successful outcome. Both skills complement one another and are necessary to unleash your full potential | source presentation source |
5 | |
| 56 | Driving Better Health for More People - SAS Life Science Analytics | William Kuan, Olivier Bouchard |
SAS Life Science Analytics Framework (SAS LSAF) is a single, GxP-validated environment that provides reproducibility, traceability, and governance to support the clinical development process which requires both internal and external collaboration. To meet the productivity and confidence demands for delivering regulators quality statistical reports, you need to embrace data from multiple data sources (e.g., EDC, wearables), use new analytical approaches, real-world data as well as artificial intelligence in a controlled, governed and audited way. At the same time, you need to accelerate and automate the end-to-end process via the integration of 3rd party solution without increasing risk. As the process of drug development becomes more complex - SAS LSAF along with direct access to the power of the SAS Viya platform, brings order, connects the components, mitigates the risks, and accelerates the delivery of your evidence to the regulators and onto the market | sas viya, clinical data, clinical trial, life sciences, clinical research, statistical programmer | source presentation source |
8 |
| 57 | Dynamic Testing For Empty Datasets | Timothy Lachlan-cope | Dynamic data requires a dynamic approach. Notes, warnings and errors regarding the existence of data and datasets should not be ignored, although can occur given a changing stream of data. When developing code we should be paying close attention to the log, however, false positives can hide real messages of concern or disrupt automatization for no reason. A developer should have this in mind when creating robust code and consider when it might be appropriate to apply techniques to conditionally select objects. Within this presentation I will detail several techniques for checking datasets including a null DATA STEP and the SASHELP/Dictionary location to check observations, and then detail how this can be combined with SAS macro language to create a truly dynamic program with less pesky notes, warnings and errors |
data step, macro variable, dictionary tables, &variable, proc sql, call execute, proc step, macro definition, %if-%then-%else | source presentation source |
5 |
| 58 | Eat, Sleep, R, Repeat | Miriam Amor | Not only are statistics in clinical research evolving in an exciting way, but also the open software R is gaining popularity for study submissions in clinical trials. This combination opens a new area of professional growth for statisticians and programmers. Over the last decades, there has been a growing interest in analysing recurrent and repeated occurrences of an event. With time-to-event analysis being one of the most common analyses in clinical trials, it should come as no surprise that survival analysis techniques are evolving to provide robust estimators for time -to-event in the presence of recurrent events. This paper aims to show how to use R and SAS to program different models for time-to-event with recurrent events. These models will include the conditional approaches proposed by Anderson -Gill (AG) [1] and Prentice, Williams and Peterson [2], in addition to the marginal model described by Wei, Lin and Weissfeld [3] | proc phreg, regression analysis | source presentation source |
9 |
| 59 | Efficacy Analyses with Imputations: Methods, Limitations and Lessons Learned | Amine Khemiri | imputation, missing value, clinical trial, efficacy summary, multiple imputation, sensitivity analysis, mean imputation, data science, regression analysis, proc princomp | source presentation source |
3 | |
| 60 | Efficacy Endpoints in Adjuvant and Neo-Adjuvant Oncology Trials | Yuliia (Piddubna) Cherednichenko, Anna Amosova |
Depending on the type, location, and grade of cancer, different combinations of disease therapy exist. Each particular case requires a special approach for analysis. For instance, RECIST 1.1 criteria remain the gold -standard outcome measure for solid tumors when the investigator assesses the anatomical tumor burden and its change in the study follow-up. However, for study designs where there are different combinations of Adjuvant and Neoadjuvant treatments coupled with cancer surgery, pure RECIST 1.1 could be not as indicative. For such cases, alternative endpoints might help to predict a clinical benefit to the patient and allow to reduce the trial duration. Pathological complete response (pCR) and time-to-event end-points such as disease-free survival are frequently incorporated into pivotal clinical trials in neoadjuvant and adjuvant settings. The purpose of this paper is to introduce key concepts, study designs, and specific efficacy endpoints in Adjuvant/Neoadjuvant Oncology | clinical trial, surrogate endpoint, clinical endpoint, complete response, overall survival, disease-free survival, correlation, adverse events, waterfall chart, survival analysis, data cleaning, randomization, regression analysis, primary endpoint | source presentation source |
15 |
| 61 | Empowering Paediatric Drug Development: CDISC Standards and ICH Guidelines for Clinical Trials with Little Participants | Ninan Luke | The development of safe and effective medicinal products for pediatric population requires careful attention to ethical, scientific, and practical considerations. The ICH E11 and Clinical Data Interchange Standards Consortium (CDISC) Pediatric User Guide have had a profound impact in the field of pediatric drug development. ICH E11 emphasizes the importance of considering the unique physiological and developmental characteristics of little participants in clinical trial design and provides specific guidance on endpoints, dosing, safety assessment, and ethical considerations. Also, the recently published CDISC Pediatric User Guide provides a standardized approach to data collection, analysis, and submission in the pediatric population, emphasizing the importance of data standardization and quality control. As statisticians and statistical programmers, it is essential that we adhere to these guidelines to optimize the development of pediatric drugs | clinical trial, clinical research, cdisc | source presentation source |
5 |
| 62 | End-to-End Open-source Collaboration Guidance | James Black, Michael Stackhouse |
data cleaning, pharmaceutical industry | source | 23 | |
| 63 | Enhancing the Quality of Data Harmonisation - a Customisable Tool for Validation of Bespoke SDTM Implementation at the Infectious Diseases Data Observatory (IDDO) | Sadie Kelly, Dick Hung |
Poster | sdtm domain, controlled terminology, cdisc | source | 1 |
| 64 | Expanding Visual Clarity: Multi-Page Figure Partitioning and Adaptive Zero-Value Depiction Using PROC SGPLOT | Yeghishe Grigoryan, Arman Melkonyan |
In recent years, the demand for visually appealing and informative graphical representations of data has increased significantly. When dealing with large datasets containing numerous subgroups and visits, it becomes essential to present the information in a clear and accessible manner. This paper introduces a novel approach to enhance visual clarity by partitioning complex figures into multiple pages and adaptively depicting zero-values using PROC SGPLOT in SAS. The proposed method focuses on partitioning figures across multiple pages based on subgroups, such as parameters, and adaptively depicting zero-value on the Y-axis, ensuring a more accurate representation of data. The method dynamically calculates the optimal distribution of visits per page, minimizing the occurrence of underutilized pages that would otherwise display a sparse set of data points |
data visualization, proc sgplot, line chart, macro variable, proc step | source presentation source |
6 |
| 65 | Experiences of Training Teams in R | Alanah Jonas | Several years ago, GSK made the decision, like other pharma companies to begin to transition away from SAS and towards R. In this presentation, I will discuss the learnings from the various ways in which R training has been given at GSK during this transition period. Some of these methods include large in-person and virtual classes, creating documentation and online learning courses, and small tailored study team sessions as teams are about to start programming in R. I will discuss the issues that people ha ve faced from the different methods, the benefits of each of the methods and the ways we have adapted training to help people and teams to thrive | data science, adam dataset | source presentation source |
10 |
| 66 | Experiencing the FDA CBER | Merve Akram, Angelo Tinazzi |
Although both FDA CDER (Center for Drug Evaluation and Research) and FDA CBER (Center for Biologics Evaluation and Research) adhere to the same data standards requirements, there is a unique aspect to submitting data to CBER, particularly for vaccine products. With our upcoming presentation, we will share our recent experiences interacting with and submitting data to the CBER division. Our focus will be on the preparatory work prior to submitting the data packages , specifically in creating the Study Data Standardization Plan (SDSP), and the CBER-specific appendix that details the structure of planned SDTM and ADaM datasets shared with the agency. We will also discuss the agency reviewer feedback and requests following the review of the SDSP and our subsequent interaction with the reviewer, including how we responded to their requests | adverse events, sdtm domain, clinical data, sample, adam dataset, clinical trial, cdisc | source presentation source |
9 |
| 67 | Exploring Antidepressant Side Effects in the UK Through Topic Modelling and Social Media Analysis | Khalid Ismail, Amna Dridi |
Scientists now monitor population health by leveraging social media, where adults discuss adverse drug effects. By analysing social media conversations, scientists gain valuable insights into the public's perceptions with antidepressants, allowing fo r a better understanding of their impact on mental health. Antidepressants' efficacy is controversial, causing diverging opinions among the population, shaping attitudes towards treatment. This study uses Twitter posts in the UK to understand antidepressants side effects through various topic modelling techniques (LDA, BERTopic, Top2Vec). To extract tweets related to antidepressants, we used a list of antidepressants approved by the Food and Drug Administrator (FDA) as keywords. The dataset contains 9,500 English tweets. BERTopic performed the best (coherence score: 0.70\%), evaluated based on human judgment and topic's coherence score | feature extraction, adverse drug reaction, proc freq, machine learning, clinical trial, deep learning | source presentation source |
13 |
| 68 | Eyes On the Prize - The Development of {admiralophtha}, an R Package for Ophthalmology ADaMs | Edoardo Mancini | adam dataset, development age, cdisc | source presentation source |
5 | |
| 69 | Figures are Basic(ally Overlooked) | Kaleb Mead | Many people seem to think that Figures (or Graphs) are more complicated than they actually are, resulting in programmers developing a lack of confidence or never producing any in their career. But why? SAS makes the production of Figures rather straight forward and hugely customizable. Identifying and exploring the common reasons as to why programmers may avoid Figures, will encourage more programmers to seek out Figure development and in turn see the industry adopt the forgotten child i.e., the Figures in "tables, Figures and listings". With Figures being this secondary thought, it's no surprise we see very little thought put towards the colour palettes used. The colours always seem quite arbitrary, when colour should be seen as a key element in a Figure. Colour can denote many things (red=bad, green=good, etc.), but poor colour choice can introduce unconscious bias and or accessibility issues for people with colour-blindness | proc step, series plot, proc template, scatter plot, box plot, clinical trial, clinical data, outlier, sashelp | source poster source |
17 |
| 70 | From Legacy to the Cloud: Novo Nordisk's Journey Towards a Modern Statistical Computing Environment | Marianne Carames | Novo Nordisk A/S has embarked on a journey to replace their legacy UNIX-based SCE with a more modern, multi-language, cloud-based solution. We will present our journey towards a new SCE and the challenges we have encountered along the way. Taking points in our current setup and how our thoughts and ideas changed over time. As a result, our ambition of having an SCE covering our current needs and requirements changed towards getting a modular and modern solution, which could enable fast adoption of new tools, automation, and innovation. To this end, Novo Nordisk scored the 8 different solutions and vendors based on the Request for Quotation material received. We ended up selecting two vendors to participate in a staged proof of concept | audit trail, proof of concept | source presentation source |
3 |
| 71 | From Molecules to Effects: Exploring Pharmacokinetics with a Clinical Pharmacology Data Scientist | Kézia Kobana | Poster | source | 1 | |
| 72 | Further Adventures in SCE Alchemy | Eleanor Sparling | It's sometimes said that a year is a long time in politics, but it's definitely a long time in the development of a new SCE. In 2022 in Belfast, we presented on the development of METEOR (Multilingual Environment to Transform, Enable, and Optimize Reporting); GSKs new language agnostic SCE solution powered by Domino Data Science Platform with GitHub Enterprise integration, leading up to the initial Minimal Viable Product (MVP) release. With an ambitious target of complete adoption by the end of 2023, we will share our continuing story of study team onboarding, training and end-user support and migration of legacy studies whilst continuing to work using Agile methodologies to improve the user experience | adam dataset, regression analysis, cdisc | source presentation source |
6 |
| 73 | Gear Up SDTM with Digital Health Devices | Vicky Poulsen | The advance of digital health technologies, such as continuous glucose monitoring (CGM), wearable actigraphy devices, wearable ECG monitors…to name a few, in recent years has significantly broadened the prospect of real-time, high-quality health data collection. This has been witnessed by the growing number of clinical studies that engage the use of voluminous device and/or omics data. Getting large volume of device data into SDTM, however, is not without its challenges. Worsening performance, validation issues and use of Device Domains, all require extra attentions. In this paper, we will share Novo Nordisk's experience and learnings from preparing submission ready SDTM data that include multiple millions of CGM readings | clinical trial, cpu time, cdisc | source presentation source |
5 |
| 74 | Generalisable Analysis Pipelines for Handling Large HTA Submissions | Nicolai Skov Johnsen, Matthew Phelps |
Poster | analysis set, data standard | source | 1 |
| 75 | Generative AI: A Case Study Highlighting the Potential and the Risks | Nikolay Manchev, Michael Harnish |
In this paper we look at how the use of permissive-licensed languages like Python and R has been steadily increasing in the pharmaceutical industry, often at the expense of proprietary technologies like SAS. We consider the large amount of technical know-how that is currently implemented in SAS and investigate how with the help of Large Language Models (LLMs) such existing code can be migrated to more open and widely used alternatives like R and Python. We cover some key challenges around source to source translation in detail and provide a blueprint for building a modern translation pipeline, which is flexible, cost efficient, compliant, and produces reproducible results. 1 The changing technology landscape When it comes to statistical analysis and data management, the Statistical Analysis System (SAS) software suite has long dominated the pharmaceutical industry | training set, data cleaning, pharmaceutical industry, data science | source presentation source |
23 |
| 76 | Get the Full Picture of Your Data with SAS | Andy Bayliss | How do you turn a billion data points into one actionable insight, how do you improve the visibility and understanding of your data, and how do you provide trustworthy insights to a range of different audiences with data visualization? This paper explore how SAS can be used to deliver the full picture of your data | sas viya, machine learning, correlation, data visualization, proc hpsplit, artificial intelligence, insurance claims | source presentation source |
13 |
| 77 | Getting ready for R | Kieran Martin, Welwyn Garden |
In 2023 Roche PD Data Sciences is moving to a new open source backbone, with packages co-developed with other companies, and a new focus on a continually evolving set of tools that are owned and contributed to by the users. In this paper, I will discuss what that transition looks like, both the infrastructure we have built to support this new approach, and how we ensure that our talent are ready for this transition. We will look at some of the R packages we have developed, and how we are using open source as a way to drive innovation in what we do, and collaborate with our colleagues across the industry | data science, clinical data | source presentation source |
6 |
| 78 | GPT (Generative Pretrained Transformers) and LLM (Large Language Models) have generated a lot of attention since their inception in early 2023 with the advent of OpenAI GPT3.5, then GPT4.0 | Jerome Vetillard, Mark Lambrecht |
Generative artificial intelligence (AI) shows transformative potential in life sciences. Models like generative adversarial networks, variational autoencoders, and transformers synthesize data to expand scientific frontiers. They generate medical images, m olecules, proteins, and text at scale. Key applications include accelerating drug discovery, personalizing medicine, powering clinical trials, and automating data analysis. GANs develop novel molecules efficiently. VAEs augment healthcare datasets. Transformers contextualize genomic and clinical data. Data scarcity, regulations, and ethics present challenges and safeguards. Collaborative solutions address quality and representation issues. Documentation and oversight ensure privacy, transparency, and fairness. Harmonizing global standards balances innovation and safety. Opportunities include pharmacogenomics revolutionizing care through genetic profiles | life sciences, clinical data, clinical trial, machine learning, sas viya, data cleaning, neural network, artificial intelligence, data science, data governance, deep learning, training set, excel, clinical research | source presentation source |
16 |
| 79 | Graph vs Table - A Complex Question | Catharina Dahlbo, Eva Kelty |
Poster | source | 1 | |
| 80 | Guidance on SDTM for Liquid Oncology Endpoints | Swaroop Koduri, Tyagrajan Swaminathan |
Poster | clinical trial, disease response | source | 1 |
| 81 | H2H Portal - A Shiny Automation of Reproducible and Traceable Analysis and Reporting Using a Metadata Framework | Julian Schmocker, Qian Wang |
The process to generate reproducible and traceable statistical outputs is complex. Typically, it requires inputs from cross-functional stakeholders, captured in various documents, and implemented potentially with different programming software, all within a tight timeline. This paper shares the journey and lessons of developing an R Shiny app, H2H Portal, an internal end-to-end automation to provide additional clinical statistics beyond the clinical study reports in support of economic evaluation for Health Technology Assessment dossiers. The app provides a collaborative space for multiple stakeholders over the course of request collection, dataset and analysis specification, output generation (using both SAS and R), | csr, adam dataset, clinical trial, excel | source presentation source |
8 |
| 82 | Handling Missing Data in Exploratory Analyses Using RWD: 'Collect the Data You Love or Love the Data You Get' - A Case Study | Jean-Marc Ferran | Presentation | missing value, insurance risk score, concomitant medications, sensitivity analysis, adverse events, analysis set, cross-sectional data | source | 22 |
| 83 | Handling Multiple Screenings Using the New Domain DC (Demographics for Multiple Participations) at Merck - Implementation of a New Mapping Process | Kristina Zweier | Presentation | data cleaning, sdtm domain, unique key, randomized controlled trial, adam dataset, cdisc | source | 21 |
| 84 | Harmonising Clinical Reporting Through the {falcon} Initiative: A Collaborative Leap Forward | Jürgen Böhl, Vincent Shen |
The {falcon} initiative is an industry collaborative effort under {pharmaverse} that unites Boehringer Ingelheim, Moderna, Roche, and Sanofi with the aspiration of building and open-sourcing a comprehensive catalog of harmonized tables, listings, and graphs (TLGs) for clinical study reporting. By leveraging existing open-source R packages, {falcon} aims to simplify the process of output review, comparison, and meta-analyses, while fostering efficient communication among stakeholders in the pharmaceutical industry. In addition, the initiative aligns with CDISC's ARD/ARM effort to promote a standardized approach to data presentation. Drawing inspiration from the FDA Standard Safety Tables and Figures Integrated Guide, the {falcon} Initiative has successfully developed and open- sourced 18 templates to date. The collaborative effort focuses on improving the clarity, consistency, and accessibility of TLGs by addressing variations and redundancies in their creation and use | clinical trial, pharmaceutical industry, cdisc | source presentation source |
3 |
| 85 | Heading to Base Camp On Our Way Up the Open-Source Mountain... | Andy Nicholls | machine learning, clinical research, data cleaning | source presentation source |
4 | |
| 86 | Heart: The Must-Have for a Good Project | Colin Brown | pharmaceutical industry, where statement | source presentation source |
6 | |
| 87 | Heat Map Visualisation for Cardiovascular Risk Prediction | Kanika Malhan, Usha Krishna Thamattoor |
Poster | source | 1 | |
| 88 | How Can Python Be Used in Clinical Trials? | Andrés Käsa, Matt Davies, Pierre Dostie |
Once upon a time, PythonTM opened the door and independently from the clinical trial field, enthusiastic programmers made several data manipulation and statistical packages in Python for their own purposes. Although Python is greatly utilised in the field of AI and ML, with data evangelists appropriately convinced of its supernature, it is fairly underutilised in the clinical research field. We did some experiments and explored how hard it would be to use these Python packages for data manipul ation, statistics calculation, output creation purposes , or for exploratory purposes such as creating dashboards. Another interesting aspect is that Python is an unvalidated language, meaning its packages are constantly evolving which can be a concern sinc e packages are heavily dependent on each other. We explored this aspect as well and tried to come up with a solution | adverse events, concomitant medications, data science, sas programmer, data visualization, medical history, excel | source presentation source |
21 |
| 89 | How to Get Started in Pharma and How We Can Help Those Starting Out | Saskia Leserer, Lara Suckut |
Poster | data science, pharmaceutical industry | source | 1 |
| 90 | How to Work When You're Not Working: Coaching Effectively to Develop Your Programming Team | Zach Dorman | Presentation | source | 19 | |
| 91 | Human BioGeography: predicting clinical trials outcome using a 3D human body simulation with a GIS framework | Diego Caso, Rocio Matesanz Sanchez, PhD Student |
Due to the limited success rate in clinical trials, with less than 20% of drug candidates progressing through all phases for marketing approval, novel innovative approaches remain essential to improve these statistics. This paper proposes a novel system combining relevant databases from different Open Data sources with a Geographic Information System (GIS) that will contribute to confront this challenge. By integrating information from several biorepositories, including previous reports of drug development processes, into a 3D simulation of the human body our system aims to identify determinants restricting clinical trial success. We call this "Human BioGeography" and is based on a hierarchy of anatomy => disease => therapeutic target => drug => chemical structure. Th is approach provides crucial insights for drug developers, offering data on aspects that influence drug approval rates, such as potential adverse events or drug viability and metabolism | clinical trial, clinical research, adverse events | source presentation source |
6 |
| 92 | ICH M11 Clinical Electronic Structured Harmonized Protocol (CeSHarP) and CDISC: Making the Electronic Protocol a Reality | John Owen, Peter Van Reusel |
Presentation | trial summary, clinical trial, ich, clinical trial protocol, tolerance interval, output delivery system, cdisc | source | 33 |
| 93 | Implementation of Estimands Using Data Standards | Chris Price, Alexandra Cochinaire |
Presentation | intercurrent event, adverse events, adam dataset, concomitant medications, sdtm programming, while-on-treatment, bds, cdisc | source | 26 |
| 94 | Implementation of MedDRA upgrade in SDTM: a SAS-based approach | Mahmoud Abu Azoum | It is often necessary to upgrade the values of coding variables such as AELLT, AEDECOD, etc. from the Medical Dictionary for Regulatory Activities (MedDRA) in SDTM datasets, e.g., for pooling purposes. We will present a SAS - based approach to upgrading the AE, MH, and PR domains by using two integrated programs. First, a program compares the values of the SDTM coding variable (e.g. AELLT, AEDECOD, etc.) with the new dictionary version. It flags the terms that are still current. For the outdated ones, the programs suggest automatic recoding where possible. Then, after a manual review and completion o f the appropriate values for the coding variables in an EXCEL file by a coding specialist, the updated recoded terms are merged into main datasets. A comparison report is generated for user review. We will explain our process with SAS. We will also show ho w we resolved some issues and potential enhancements. Keywords: SAS-Approach, SDTM dataset, MedDRA |
sdtm domain, excel, adverse events, cdisc | source presentation source |
5 |
| 95 | Improving Clinical Operational Insights Delivery to different Levels of Leadership Seniority Using Shiny-Based Automation | Peyman Eshghi, Jose-Felipe Golib-Dzib |
Conducting clinical trials involves complex, costly, and time-consuming processes. Analytical tools can provide insights to identify areas to focus on throughout clinical operations, contributing to effective planning. However, it is difficult to keep emer ging business needs and timely engineering of existing analytical tools synchronized. Moreover, communication at distinct levels of leadership increases administrative burden. Therefore, we develop a Shiny app, coupled with a Continuous Integration / Continuous Deployment system, as a minimal viable product to produce regular executive summary reports while reducing cycle time on reporting, reducing manual burden on staff, and increasing report quality by content refinement and process simplification. This automation effectively eliminates bottlenecks in communication between operation and leadership teams | clinical trial, self-service bi | source presentation source |
7 |
| 96 | Increasing Trend of Biomarker Analysis in Clinical Trials | Anjali Ranga | Clinical trials are often designed to confirm efficacy of the investigational drug or to answer key development questions. In most trials, the data are subjected to exploratory analysis. This exploratory analysis could serve as a basis for explaining or supporting their findings, for improving the design of new trials, and for generating hypotheses for later research. One of the key exploratory analyses that is increasingly used in clinical trials is biomarker. In clinical research, biomarkers are used for multiple purposes like, to explore whether markers can be identified that will be predictive for the effect of a drug treatment, or to help elucidate the mechanism of action of a drug and the biology of the disease. Some clinical biomarker assessments are intended for more immediate impact or decision making. Combinations of clinical and research biomarker are used throughout the conduct of Phase I-III trials | clinical trial, sample, biomarker, change from baseline, clinical research, interval variable, pharmacodynamic biomarker, phase i, raw data, data science | source presentation source |
11 |
| 97 | Inference on Weighted Subjects: Comparing Weighted Individual Patient Data with Results from the Literature | Andrea Parodi, Dimitri Petratchenko |
Poster | type i error, baseline characteristics, sample | source | 1 |
| 98 | InnerSource: A Stepping Stone Towards Open Source in Statistical Programming | Jeppe Juul | clinical programming, clinical trial, pharmaceutical industry | source presentation source |
3 | |
| 99 | Innovating the Leadership Toolkit Through Design Thinking | Marten Krijgsman, James McDermott |
Presentation | phase ii, phase i, phase iii, phase iv, stage 1, stage 2, stage 3 | source | 65 |
| 100 | Integrating Isolated Clinical Trial Systems with RESTful APIs: Enhancing Data Accessibility and Collaboration | Sumesh Kalappurakal, Nicolas Dupuis |
Isolated clinical trial systems are often designed as independent software applications, which make it difficult for the same team using different systems or different team try to access or share data. However, RESTful APIs can be used to facilitate the integration of these systems, allowing for seamless data access between applications. RESTful APIs offer a standardized way to communicate between systems, making it easier for developers to create integrations using the right language, even if the systems were not originally designed for integration. By implementing RESTful APIs, clinical trial systems can become more connected, enabling teams to access data from multiple sources and improve collaboration | rest api, clinical trial, clinical research, proc http, proc step, excel | source presentation source |
7 |
| 101 | Integrating Privacy By Design to Enable Innovative Clinical Trials and Post-Trial Data Sharing | Muhammad Oneeb Rehman Mian | As clinical trial design evolves to incorporate newer technologies and innovative methodologies, identifying appropriate privacy measures for the safe collection and use of sensitive personal information has become a growing challenge. Often, privacy considerations are overlooked or are difficult to predict during the design phase of innovative trials, which necessitates pharmaceutical companies to employ reactive measures to address these concerns during the implementation phase | clinical trial, data protection | source presentation source |
5 |
| 102 | Interaction with the FDA During and After a Type C Meeting, From the Perspective of a Statistical Programmer | Dieter Hartz | The Type C meeting was used to coordinate the electronic submission requirements for a pediatric submission in type 2 diabetes mellitus with the FDA. The planned deliverables were detailed in an esub (electronic submission) plan. FDA's response to that esub plan was quite different to what was proposed concerning the programs for analysis datasets, tables, listings and graphs as well as analysis results metadata. They also detailed specific requirements for the datasets including laboratory units that we did not meet. They asked for a specific listing of the main statistical analyses that we had not foreseen. The details of the BIMO package required extensive discussions with FDA. It will be explained how constructive communication with the FDA after the Type C meeting can lead to a clear understanding of the electronic submission requirements and thereby avoid much unnecessary work and alleviate the risk of rejection | primary endpoint, adam dataset, data science, randomization, cdisc | source presentation source |
5 |
| 103 | Is Every Programmed Report a Part of Your CSR? Stay a Step Ahead with Lean SAP and Digitalisation | Sowmya Singuluri | Poster | source | 1 | |
| 104 | It's Easy Peasy Submitting Data to the China NMPA - Really? | Nigel Montgomery, Chenkai Lv |
Submission of a drug can be challenging at the best of times and submission to the CDE (Center for Drug Evaluation) at the NMPA (National Medical Products Administration) we've found to be no different. 2 Contents 1. INTRODUCTION ................................ ................................ ................................ ................................ ... 3 2. CHINA FILING PREPARATION AND REQUIREMENTS ................................ .......................... 3 3. GLOBAL AND CHINA FILING INTERACTION ................................ ................................ ............ 3 4. GLOBAL AND LOCAL TEAMS SUPPORT FOR CHINA FILINGS ................................ ........... 4 5. ESUBMISSION PACKAGE TRANSLATION ................................ ................................ ................... 5 6. EARLY PLANNING AND SMOOTH DELIVERY ................................ ................................ ........... 5 7. ESUBMISSION PACKAGE ISSUES ARISING FROM TRANSLATION | data science, annotated crf, cdisc | source presentation source |
10 |
| 105 | Janssen DATA4YOU Platform Approach for Clinical Data Review | Erik Lemmens, Paul Fioole |
The paper discusses our initiative to transform the clinical data review environment through the DATA4YOU program. In response to the challenges posed by increasing demand and the complexity of clinical data, Janssen undertook a phased approach to modernize their data review processes and tools. The paper highlights the key objectives of the initiative, which start from the needs of the different main user groups: Clinical Data Managers, Medical Reviewers, Central Monitoring/Risk Management, and Clinical Programmers. It describes the gaps in the existing processes, underscoring the importance of near real-time data accessibility, automation, and improved user experience. Then, it highlights how the collaborative and user-centric approach was adopted in developing the DATA4YOU application. The authors outline the strategic approach taken based on a cloud-based best-of-breed solution | clinical data, data review, data lake, gold layer | source presentation source |
7 |
| 106 | Know Your Environment | Will Greenway | Often when we write SAS programs there is a tendency to think that if we just use basic SAS statements then the program will be easily transportable across programming environments. Rarely do we take a step back and consider what operating system and SAS options we have taken for granted that may be impacting the program in the background, influencing how it was written. For example, OPTIONS VALIDVARNAME is one of my favourite options, often overlooked but its impact can be quite large on how programs are written. Rarely is this option changed, instead programmers write for the environment they are currently in and the behaviour they see. This paper will discuss several environment considerations and SAS options that, when accounted for, can make your SAS macros or solutions more platform independent | proc step, proc power | source presentation source |
6 |
| 107 | Knowledge Graphs for Clinical Trials | Eva Kelty, Anders Högberg |
As clinical trials are conducted vast amounts of data is generated, this includes data describing for example personal attributes, medical conditions, treatments, and procedures conducted together with outcome measurements. Often outcomes/variables differ between different trials but for each specific patient data forms a web of information that is connected and may change over time, a mind-map. Property graph databases are an efficient way of managing this information where measurements and the relations between the measurements are explicit. By appending metadata to this structure, a knowledge graph is created. In the presentation, we will discuss how the knowledge graph is a FAIR way to store harmonized and standardized clinical trial data, but also how it can be used to explore and understand data objectively. We will also show how complex questions can be answered instantly by utilizing indexing structures in a graph database | clinical trial, adverse events | source presentation source |
3 |
| 108 | Let's Compare Them All - TFL Comparisons | Katja Glass | In our daily work, even minor changes can sometimes have unexpected consequences. To ensure nothing unexpected has changed, we often rely on comparison methods. This paper will focus on the comparison of outputs such as tables, listings and figures. We will also explore comparison options and provide helpful hints on how to perform these comparisons effectively | proc step, proc groovy, dataset | source presentation source |
7 |
| 109 | Leveraging Real-World Data to Provide Deeper Insight into Treatment Effects in Clinical Trials | Soundarya Palanisamy, Sherrine Eid |
clinical trial, machine learning | source presentation source |
6 | |
| 110 | Leveraging the Analysis Results Standard (ARS): The Cytel PRISM Experience | Nicolas Rouillé, Sebastià Barceló |
While waiting for the final Analysis Results Standards (ARS) to be finalized and released by CDISC, many companies have created and implemented their own ARS over the years. Following our presentation at PHUSE 2022 titled "From Artifact to Metadata: A Divergent Approach to Automation Through Standardization," we would like to share our experience as a CRO in implementing ARS in our internal solutions. Conceptually, our model is very similar to what CDISC is developing, with the Analysis Results Logical Model built around two components: "metadata" and "results data." From an application standpoint, as we explained in 2022, we developed PRISM, an application that automatically captures and curates "tagged" metadata directly from study table mock shells created by our statistician and generates SAS or R code from that metadata. Currently, by following our internally developed standards, we can cover about 60% of the outputs needed for a study | mock shell, analysis set, cdisc | source presentation source |
8 |
| 111 | Look to the Sky: Introducing COMET - the COncomitant MEdication review Tool | Neville Cope | The review and evaluation of concomitant medication use by medical teams can be a time-consuming manual task. Study teams frequently create their own programs and procedures, often in isolation, leading to wheel- reinvention and inconsistent cross-study approaches. The outcome of these reviews is critical to the identification and reporting of disallowed medication use and potential protocol deviations. COMET is a SAS tool that compares current data to the previously reviewed iteration, and presents the results in a clear, concise, and easy to use spreadsheet. Spreadsheet functionality includes filters, persistent comments, new/amended record identification, highlighting of changed values and required re-review indicators - which together make the review process quick and efficient. This paper will show the life cycle of a COMET report and discuss some of the tricks and hacks utilised in order to fully customise an EXCEL spreadsheet created by SAS |
excel, concomitant medications, protocol deviations, data cleaning, dataset, data integrity, character variable | source presentation source |
7 |
| 112 | Machine-Learning Approaches to Survival Analysis of Clinical Data | Dorothee Childs, Christoph Bergen |
survival analysis, support vector machine | source presentation source |
2 | |
| 113 | Management impact on Employee Engagement - Statistical Programmers in Focus | Wickus Bronkhorst | The term management has a long list of different definitions, components and principles; but at the heart of effective management are employees. For managers to be successful, the needs of employees need to be met, both individually and as part of a collective, in order to achieve high levels of employee engagement and motivation. Unfortunately, the needs of two individuals or two separate teams are often not the same and managers have their work cut out for them to successfully identify and meet the needs of their teams and employees. Similarly, the needs of people from different industries or departments may differ, and self-help books may not always be universally applicable. This paper will show the effects that management has on employee engagement and motivation with reference to performance management, respect and recognition and overall perception of manag ement within a Stat istics and Programming department | clinical programming, data science | source presentation source |
7 |
| 114 | Mastering the Art of Adaptation: Innovative Strategies for Effective Change Management | Mark Bynens | key performance indicator, clinical trial | source presentation source |
13 | |
| 115 | Metadata-Driven ADaM R-Code Generation by Leveraging Automation Potential | Thiago Figueiredo, Waldemar Miller |
data pipeline, clinical data, out=, raw data, sql server, excel, proc sql | source presentation source |
31 | |
| 116 | Metadata-Driven Population of Timepoint Variables | Rasmus Kragh Nielsen, Addie Nina Olsen |
Poster | source | 1 | |
| 117 | Migration to SAS Grid for 300 Users - From Preparation to Production | Cédric Marchand, Steven Wong |
To remain competitive and achieve growth, CROs need to keep up with technological advancements and adopt to new ways of working, which are ever more resource intensive. In line with this, Cytel moved all its SAS computing to an SAS Grid platform with the aim of utilizing computing power and support staff more efficiently. As with previous migrations, we have prepared ourselves thorough investigations and documentation, by ensuring that our standard SAS macro catalogs were working properly, by adopting new work habits and by replacing SAS Display Manager with SAS studio and Enterprise Guide. We also implemented quality control scripts and QA intervention and made difficult decisions such as the operating system of the Grid server. From our experience as SAS users, our recommendation is to review the many preparation steps we took and the unexpected issues we faced when we moved to the production environment | sas studio, sas grid, sas enterprise guide, work library, stress testing, proc step | source poster source |
5 |
| 118 | MINT+ Automation and Digitization of SDTM specification | Magdalena Krochmal | MINT+ is a web application that automates the creation of study SDTM specifications, including SDTM mappings and Controlled Terminology. A set of generic algorithms to map collected data to SDTM lies at the center of the solution. By comparing study metada ta to the metadata in the Roche MDR, MINT+ streamlines the process and reduces the likelihood of errors. Built with React UI, MINT+ enables users to add SDTM mappings for non -standards through an interactive user interface. Study-level mappings are then stored in Saffron Document DB, facilitating metadata sharing and enabling further reuse across other studies using the same data collection elements. This innovative solution significantly reduces the workload and increases the accuracy of SDTM mapping, addressing a significant challenge in the industry | clinical trial, clinical data, sdtm domain, data governance, sdtm programming, rest api, data cleaning, cdisc | source presentation source |
5 |
| 119 | Mr. ight - Why R Should Be Used for Validating TLFs in Clinical Trials? | Monika Ludwinska | SAS is widely used in validating TLFs in clinical trials but over the past few years R has emerged as an alternative software to produce high quality outputs. This open source language created for statistical computing is gaining importance in pharmaceutical industry because of its flexibility and variety. Quality control process is needed to ensure the accuracy of outcomes by comparing results prepared by two independent programmers based on given specification. This paper will look at R and its strengths in the validation process of tables, listings and figures. It will focus on advantages of using another language in the quality control process and will practically show how validation is performed. Moreover, it will present important R functions which every programmer should be familiar with in daily work to efficiently lead the quality process. Furthermore, pitfalls will be exposed while validating SAS outputs with R software on examples | data cleaning, quality control, clinical trial, data science, adam dataset, pharmaceutical industry, missing value, cdisc | source presentation source |
10 |
| 120 | Multiple Myeloma: Challenges in identifying regimen modification for a given line of Chemotherapy | Mrityunjay Kumar, Ephicacy Lifescience, Shailesh Gupte |
The main types of drug therapies used to treat multiple myeloma are proteasome inhibitors, immunomodulatory drugs (IMiDs), steroids, antibodies, B-cell maturation antigen (BCMA)-targeted therapies, and chemotherapies. In the treatment of multiple myeloma, chemotherapy is most often used in preparation for a stem cell transplant. We have seen use of multiple lines of therapy where drugs are given in combination like triplet, quadruplet etc. Occasionally, it becomes challenging to correctly identify part of individual treatment modification (addition, deletion, discontinuation, or change) in given lines of therapy. Disposition form provides patient level treatment discontinuation information and not when a part of given line of regimen was discontinued. Statisticians often request to present regimen modification duration in summary table | proc means, raw data | source presentation source |
7 |
| 121 | Navigating R Package Management in a Validated Environment: Maintaining and Implementing the Latest R Packages with Different Versions of R | Ali Dootson, Rakesh Gotiwale |
Presentation | data cleaning, proc freq | source | 11 |
| 122 | Navigating Unprecedented Challenges: Journey Through a Pandemic and International Conflict | Daniil Shliakhov | After expanding operations into Ukraine and opening an office in the city of Kharkiv, Veramed's head of the Ukrainian division faced events that required swift adaptation to rapidly changing circumstances. This presentation explores the unique challenges faced that started with a global pandemic and overlapped with an international conflict. The presentation not only covers transitioning the team to safe remote working whilst maintaining productivity and client satisfaction, but it also delves into the strategies employed such as conducting daily meetings, addressing employee needs and fostering remote social connections. Furthermore, it discusses the manager's efforts to ensure business continuity during the conflict despite power outages and limited internet connectivity. Through this case study the presentation highlights the importance of adaptability, communication, and resilience in managing teams effectively in unusual and challenging circumstances | offer recommendation, pharmaceutical industry | source presentation source |
4 |
| 123 | No Adobe Pro, No Worries | Catriona O'Hare | Adobe Pro is a very handy tool when working with Tables, Listings, Figures (TFLs), and Patient Profiles using SAS. But it can be a high cost to the company or can't be installed on the server where you use your SAS. I will present how you can combine all your Tables, Listings, and Figures which are in RTF format into one PDF file with a Table of Contents (TOC) and bookmarks using SAS, VBScript, and Ghostscript. And how to create Patient Profiles straight into PDF with bookmarks using ODS Document and ODS PDF in SAS |
proc report, proc step, ods pdf, macro variable | source presentation source |
9 |
| 124 | Oak Garden - Metadata-Driven SDTM Automation in R | Magdalena Krochmal, Amy Franklin |
SDTM mapping is often considered a simple process but remains a challenge across the Pharmaceutical industry. OAK is a metadata-driven SDTM automation solution that uses simple algorithms to transform source data into the target tabulation model. Built on an R-based approach, OAK automates study standards, while the same core algorithms can also be used to prepare non-standard mapping definitions based on study requirements. Its flexible architecture and modular design make it easy to program non-standards, even for complex scenarios. OAK leverages metadata and global data standards to automate approximately 80% of SDTM domains with 22 reusable algorithms. In this presentation, we will showcase Roche's implementation of OAK SDTM automation. We are also working with COSA and industry partners to make this an open-source solution, which will be covered as a separate topic | sdtm programming, clinical trial, sdtm domain, clinical data, data governance, pharmaceutical industry, cdisc | source presentation source |
7 |
| 125 | Office-based or Home-based: Are We Remotely Close to the New Steady State? A Working Session | Michael S Rimler | Working sessions were conducted in both the 2023 PHUSE EU Connect (Birmingham, UK) and the 2024 PHUSE US Connect (Bethesda, MD). At the EU Connect, as part of the People Leadership & Management (PM) stream, we conducted a 60-minute, interactive workshop exploring the recent shift to remote and hybrid working models within Clinical Data Analytics in our industry. At the US Connect, we facilitated a 90-minute Interactive Discussion on the same topic. In both instances, the scope and agenda were identical, with the added element of comparing US findings to EU findings while in Bethesda. Whether one identifies as a programmer, statistician, data manager, data scientist, or other role - they have seen their work impacted by remote working in the past 3-4 years. We want to share previous experience, define what the norm might be, and predict the future of remote working and the impact of it becoming embedded in our daily lives | order by, logistic distribution | source presentation source |
4 |
| 126 | Open-Source MDR and SDR - Managing Your Standards with the OpenStudyBuilder | Nicolas De Saint Jorre, Mikkel Traun |
The OpenStudyBuilder (OSB) is an open-source project for clinical study specifications. This tool is a new approach for working with studies that once fully implemented will drive end-to-end consistency and more efficient processes - all the way from protocol development and CRF design - to creation of datasets, analysis, reporting, submission to health authorities and public disclosure of study information. The OpenStudyBuilder contains: Standards and templates used for study specification Individual versioning of standard elements in the library Real-time team collaboration around study design by study team Central storage of study design data for multiple down-stream usages ('one source of truth') * Versioning and full audit trail of study definitions In the library section you can browse all kinds of standards, like for CDISC the SDTM, ADAM, SEND, CDASH and many more. Also, other dictionaries like SNOMED, LOINC, UCUM are available. MedDRA could be loaded as well of course | crf, cdisc controlled terminology, cdisc | source presentation source |
13 |
| 127 | Open-Source Protocol Automation with the OpenStudyBuilder | Nicolas De Saint Jorre, Mikkel Traun |
controlled terminology, crf, cdisc | source presentation source |
9 | |
| 128 | Optimising a (Supposedly) Reusable Program to Reduce the Need for Manual Changes | Panteleo Nacci | macro variable, sas function, macro definition, dataset, proc cport, reference data, clinical programming | source presentation source |
23 | |
| 129 | Optimising Efficiency and Improving Data Quality Through Meaningful Custom Fix Tips and Explanations | Julie Ann Hood, Jen Manzi |
Poster | source | 1 | |
| 130 | Optimizing Clinical Trial Design with Real-Time Sample Size Calculation and Enrollment Projection: A Shiny Dashboard Solution | Dony Sabu | The clinical trial design is a complex process that requires careful planning and decision to ensure the validity and reliability of the results. Determining the appropriate sample size and enrollment projections is a crucial component of trial design. Recent developments in using real -time data to optimize clinical trial data improve efficiency and reduce costs. This paper explores the potential of utilizing Shiny dashboards to opt imize clinical trial design through real-time sample size calculation and enrollment projection. The dashboard will integrate the clinical trial database and provide users with real -time information on the expected number of participants and the required s ample size. The ability to make more informed decisions using this tool can lead to a shorter overall timeline for completing the clinical trial | sample size, sample, clinical trial, effect size, proc power and sample size, primary endpoint, clinical research, resource allocation | source presentation source |
5 |
| 131 | Oversight and Insight within a Fast-Moving Biotech Company | Stijn Rogiers, Isabelle Vandriessche |
"Build the plane while flying it" is probably a good quote for the journey ahead of us. Argenx is a fast-moving and growing Biotech company. This means we need to rapidly scale for future success while delivering on the ongoing clinical trials and submission milestones. In this paper, we start with an overview of argenx. Then we will look at the journey so far , building a strong foundation to integrate, automate and drive efficiencies. Next, we will focus on two important Data Visualization projects for improved trial oversight and Insights generation. Finally we will look back at our Critical Success Factors and potential Challenges ahead of us. 2 | life sciences, adverse event summary, clinical trial | source presentation source |
15 |
| 132 | Patient Profiles v2.0: Easy and Quick Clinical Trial Insights in Tableau | Roman Radelicki, Annelies Van Zeveren |
The patient profile is an important tool during the conduct of clinical trials for many stakeholders. Monitoring adverse events, medications, study dosings, medical history, vital signs, electrocardiograms and laboratory results can lead to meaningful insights. A picture is worth a thousand words, data visualisations are more powerful than reviewing data in tabular format. Using Tableau software, we created a patient profile dashboard that combines this CDISC data on a single timeline. Possible relationships between adverse events and medication or other domains are made clearly visible. The dashboards are always up-to-date with the latest changes as monitored at the clinical sites because of our existing automated raw to CDISC SDTM conversion pipeline. Us ing filters, users can focus on a specific period of time, or quickly switch to a different subject, all dynamically on the same page | clinical trial, adverse events, data science, data visualization, data cleaning, concomitant medications, vital signs, clinical research, raw data, cdisc | source presentation source |
6 |
| 133 | Practical Operations of Electronic Study Data Submissions | Priyanka Pujari | Regulatory Authorities are continuously updating their electronic study data submission requirements, and this greatly affects timeline and cost for the New Drug Application (NDA). Therefore, in terms of the global drug development, it is more efficient to implement streamlined process for e-study data submission that would meet different requirements from multiple authorities (FDA and PMDA, for example) in order to achieve simultaneous submissions. The purpose of this presentation is to share experience on challenges, poi nts to consider and solutions to implement when submitting e-study data to multiple authorities. In particular, understanding the key requirements, validation rules, documents for consultations and other aspects across Health authorities which should be considered at the development stage of each clinical study | data cleaning, data validation rules, data science, data standard, raw data, adam dataset, cdisc | source presentation source |
9 |
| 134 | Quality Tolerance Limits - More Critical Aspects of Clinical Trials with Much Less Visibility | Saritha Chamarthi, Mahesh Balagonda |
We will be covering about Quality Tolerance Limits (QTLs) in this session to bring in awareness along with implementation of standards with its importance, focus areas, considerations, significance, framework, the primary differences between KRIs and QTLs along with an automation methodology to implement these QTLs in faster turnaround to be compliance with ICH-GCP guidelines | clinical trial, data cleaning | source presentation source |
2 |
| 135 | R Validation: Getting Started | Christian S. Christensen | 'How do we start to use R on our clinical trials?' is a question currently relevant to many companies in the pharma- ceutical industry. Along with this question comes considerations for validation, change management, infrastructure and generally handling of open-source languages within a clinical trial context. The goal of this paper is not to provide an answer. Instead, it is to describe our ongoing journey at Lundbeck, a mid- sized pharma company, in the hope that our experiences and considerations may benefit you should you choose to pursue the answer yourself | data cleaning, clinical trial, clinical data | source presentation source |
7 |
| 136 | R You Ready to Join the RBQM Revolution? Introducing rbqmR | Jeremy Howells, Rohan Palanganda |
adverse events, proc means | source presentation source |
4 | |
| 137 | Real world observational studies and the challenges in data collection | Kavitha Devadiga | Clinical trials are conducted in a very controlled setup. However, to see the treatment effect in real world population Real world observational studies are conducted and the data used will be called as Real-world data (RWD). In Novo Nordisk these studies are called as Non-interventional studies (NIS) and the data will be either from primary or secondary database. Since NIS are conducted in uncontrolled setup, there are many practical challenges seen during data collection ex: data fluctuation or unstructured data. However, by taking certain measures and striking balance between data collection and management, we can utilize the Real-world data to generate further evidence by complementing the results from clinical trials. This paper will describe the challenges seen while working with data from primary database and how addressing these challenges aid the organization to collect, process and use Real-world data in a responsible, efficient, and effective manner | clinical trial, visit | source presentation source |
7 |
| 138 | Real-Time Safety and Medical Review by Integrating Data Visualisations with an SCE | Shivoy Pandita, Santhoshi Madeswaran |
A Statistical Computing Environment is a widely used tool for accessing clinical data from different sources, and the development and management of reporting deliverables. A platform with an integrated visualization tool which is designed specifically for clinical data review, will allow programmers to understand CDISC SDTM data better and make these deliverables readily available to the people who understand best - safety and medical groups. They are the ones who want and need reliable insights from the data sponsors collect so assiduously. A seamless integration of an SCE with such a tool eliminates the need to create and distribute standard and most commonly used TLFs by statistical programming teams, thus simplifying data exchange and ultimately making data review proactive, faster and efficient | adverse events, data review, clinical data, sdtm domain, clinical trial, concomitant medications, clinical programming, sas visual analytics, cdisc | source presentation source |
8 |
| 139 | Real-World Outcomes and Biomarker Testing in Cancer Patients: Exploration of a Novel Genetic Database from Routine Clinical Practice in England | Alexandrina Lambova, Fiona Ingleby |
sample, overall survival, proc means, clinical endpoint | source presentation source |
5 | |
| 140 | Red Pill or Blue Pill? Assessing the Impact of Artificial Intelligence on Pharmaceutical Programming | Paul Shannon | data step, quality assurance, macro definition, artificial intelligence, clinical research | source presentation source |
6 | |
| 141 | Regulating the Future: Navigating the Complexities of 3D Printing in the Pharmaceutical Industry | Burger Lubbe | The use of 3D printing in the pharmaceutical industry has the potential to revolutionize drug formulation and medical device manufacturing, but it is also subject to strict regulatory requirements. This paper will provide an overview of the current regulatory landscape surrounding 3D printing and the practical impacts on the data and the analysis thereof. We will examine the regulatory frameworks and guidelines set forth by regulatory authorities such as the FDA and the EMA, as well as international standards organizations such as the International Organization for Standardization (ISO). Furthermore, we will discuss the challenges associated with implementing 3D printing technology in a regulated environment, including quality control, validation, and documentation requirements. In addition, we will explore the implications of 3D printing on how trial data is structured and analyzed | pharmaceutical industry, clinical trial, quality control | source presentation source |
10 |
| 142 | Revolutionising Data Exploration: An Interactive Journey Through R Shiny | Valentin Legras | Presentation | prediction interval, clinical trial, ecrf, clinical data, data science, pharmaceutical industry, phase iii | source | 30 |
| 143 | Risk-Based Monitoring in Clinical Trials with Machine-Learning-Powered Anomaly Detection | Zaid Al-Jubouri, Ece Kavalci |
Risk-basedmonitoring(RBM)is a criticalcomponentin clinicaltrials,ensuringdataquality, patientsafety, andregulatorycompliance.In thisconference,wepresentaninnovativefeatureaddedtoourclinicaltrialmanagementplatformthatemploysanomalydetectiontechniquesforenhancedRBMinclinicaltrials.Ourapproachistoidentifyanomaliesin trialdatausingmachinelearning,whichmayindicatepotentialrisksthatcannotbe detectedbyhard-codedrules.Thisenablescontinuousmonitoringandearlyriskmitigation,improvingtrialoutcomes.Thecontributionofourresearchisthedevelopmentofarobust,automatedRBMsystemintegratedwithintheclinicaltrialmanagementplatform,whichefficientlyidentifiesandaddressespotentialissuesinclinicaltrials.WewilldiscussthechallengesoftraditionalRBMandthemethodologyofourapproachusingademoofthisdetectionmodelintegratedin our trial platformCitrus.Theintegrationof an automatedanomalydetectionfeatureintothe clinicaltrialmanagementplatformrepresentsa significantadvancementinthefield,withthepotentialtoenhancehowcliniciansmo | outlier, clinical trial, machine learning | source presentation source |
11 |
| 144 | Road to Rome or Roaming - A Reprise | Francis Kendall, Christian Mueller |
Presentation | data science, clinical programming, clinical trial, data analyst, clinical data, type i error, big data | source | 32 |
| 145 | Same Same but Different: Leveraging R for Comparing Clinical Study | TLF Versions | Large amounts of outputs are generated during a clinical trial. Whenever there are changes in the underlying data, expected or unexpected, the outputs need to be re-generated. In this paper we present a set of R functions that offers a powerful and flexible solution for comparing and reviewing clinical study output versions. Using these functions one can speed up the review of the outputs by quickly identifying the outputs that have and haven't changed. Hence, no time is wasted for reviewing the outputs which have not been affected by data changes. The functions can be used via R-shiny App providing a user-friendly interface for running the functions and discussing the findings with different stakeholders from clinical trial and medical writing teams | clinical trial, csr | source presentation source |
3 |
| 146 | SAS Arrays in R Way | Jagadish Katam | This paper will demonstrate a comparative analysis of data from an Array between R and SAS. This is a better way of learning any new programming language. The industry has been using SAS for quite some time and recently there is a new trend of R in clinical trials. Considering the same, we are slowing moving to use R parallelly in generating datasets and outputs, so for data manipulation we need to know how to use R for Array type of programming. A SAS Array is a temporary grouping of SAS variables under a single name. The use of Arrays typically reduces the number of lines of code needed to perform a task, resulting in code that is less error-prone and easy to understand by programmers. In this paper, we will explore the different ways of doing Arrays in R vs SAS with some examples | numeric variable, character variable | source presentation source |
8 |
| 147 | SAS Macro Development Using ChatGPT | Karma Tarap | The emergence of Large Language Models (LLMs) within the Pharmaceutical Industry brings forth a wealth of opportunities alongside notable challenges. This paper sheds light on the potential of LLMs to significantly impact various aspects of the industry, particularly in data analytics and coding endeavors. We present the Analytics Assistant, an innovative web application leveraging the prowess of LLMs, aimed at offering targeted support to clinical programmers, while also discussing potential hurdles like data privacy and computational resources | clinical programming, pharmaceutical industry | source presentation source |
8 |
| 148 | SAS Packages for Clinical Programming | Eldho Alias, Prasanth Prabhakaran, Rahulraj V R |
A collection of SAS scripts can be reused by using a single, automatically generated stand -alone file, and this is what a SAS package is. This package can be used to load the SAS scripts into the environment. Without having to open the program / macro, or its documentation, we can also quickly get the documentation in the log window for any program / macro inside the package. Clinical research can benefit from SAS Packages. Methods for creating SAS packages are the primary topic of this paper. Here, we will go overloading the package, using it with the help file and creating helpful auxiliary programs for it. Additionally, we demonstrate how the programs included in the package can be used by anyone who has the package | sas packages, clinical research | source presentation source |
7 |
| 149 | SAS Programmers Need AI | Craig Parry | When technology advances two things happen, people mention it in every presentation regardless of said technology being relevant and or people worry this is the end of their job. The first point is very true, AI is referenced in almost every presentation I have seen since 2016. But is the second point, redundancy, something to fear? It really depends on how we embrace the advantages and control the disadvantages of AI. I see AI as the super modern version of ask Jeeves, or for people born after the year 2000, Google search with a brain. I have been programming for over a decade, slowly curating a personal portfolio of work, good macros, bad macros, hard to read macros. Yet even after a decade I still find myself with a great idea but limited knowledge of how to implement such an idea, or I am posing a question which sounds correct to me yet it returns something totally irrelevant | proc sql, proc step | source poster source |
10 |
| 150 | SAS Task: Three Ways | Simon Woods | Poster | clinical programming, box plot | source | 1 |
| 151 | SDTM CRT Packages and the Submission Space | Joshua Glaser, Charlotte Prior-Chappell |
The main goal of the Submission Space and generating Case Report Tabulation (CRT) Packages is to ensure "Submission Readiness" of a study by assessing the Datasets, Define and Validation Reports. For all current and ongoing studies, the aim should be for a high-quality package to be produced at a study level, with little to no changes required for the CRT Package. This poster aims to outline what it means for the SDTM side of a CRT package to be "Submission Ready" and how this is achieved, whether we are starting from SDTMs created from the most recent Implementation Guide or from Datasets not in the format of SDTMs yet | crf, data cleaning, cdisc | source poster source |
7 |
| 152 | Size matters - sample size calculation with SAS and R | Fenja Seither | Sample size determination is a very important step in planning a clinical study. This paper intends to provide the reader with details on how to use SAS and R to estimate the number of patients to be enrolled | sample, sample size, significance level, proc power, proc power and sample size, proc step, effect size, standard deviation, clinical trial, dose finding | source presentation source |
4 |
| 153 | Slicing Codes: A Modular approach to ADaM programming | Mahendran Venkatachalam | A significant amount of time and resources are required to complete the submission process for a phase 3 project with multiple trials. A lot of this effort is spent aligning the programming across several trials. In this presentation, we demonstrate a new ADaM programming experiment at Novo Nordisk that helps study teams to efficiently produce trial and submi ssion deliverables, including DSUR, PSUR, ISS, ISE, and country-specific submissions based on a single set of ADaM programs. Our approach to developing a single set of ADaM programs uses stacked SDTM datasets for multiple Phase 3 trials within the same pro ject area. We describe how we took advantage of R programming to slice the ADaM codes for each individual trial based on the program for the pooled ADaM datasets, resulting in reduced code duplication and streamlined programming across all Phase 3 trials | phase iii, adam dataset, sdtm domain, cdisc | source presentation source |
10 |
| 154 | Standardisation in a fast growing environment; MDR, EDC and other abbreviations | Louella Schoemacher, Jasmine Kestemont |
With a fast growing portfolio, the importance and need for standardisation increases. In order to keep up with the growing number of compounds, indications and studies the sponsor Data Standards team implemented a metadata repository (MDR) in 2021. By having a single source of truth for our metadata we aim to spend less time and resources during study set-up phase. Standardisation however is a dynamic process. New vendors, new guidances and new insights are all in place to improve our processes and data quality, but can be a challenge in standardisation. How many versions do we need and how to best govern different versions? When is something study-specific and when do we decide to standardise? Subsequently, how can we use the metadata repo sitory to further expand standardisation across other processes? What is the impact on upstream and downstream processes? With standardising our CRFs and Define-XML files we aim to have more standardised raw data | controlled terminology, sdtm domain, crf, protocol amendment, raw data, cdisc | source presentation source |
6 |
| 155 | Statistical Programming - Hiring Mission Made Possible | Christoph Ziegler, Donald Paterson |
Finding candidates in Clinical and Statistical Programming with the right skills and experience, who are a good fit for your team, is not an easy task. Therefore, when you need to hire >100 permanent programmers, at different pay grades, during a short period of time, the mission becomes almost impossible | clinical programming, calendar year | source presentation source |
4 |
| 156 | Statistical Programming 101 for the Fourth Industrial Revolution | Ritika Aggarwal | The Fourth Industrial Revolution is the advent of "cyber -physical systems" involving new capabilities for people and machines. It builds on the fusion of technologies from physical, digital, and biological spheres. According to a survey conducted by the Economic Intelligence Unit, the healthcare sector would gain the most from th e merging of these systems. Consequently, the method for collecting comprehensive health and fitness data is gradually making headway. For example, Novartis developed a digital inhaler that could enable COPD patients to observe data on their own inhalation use in real time. Inevitably, the constantly shifting complexity and diversity of healthcare data will cause major disruptions and surprises for the statistical programming function of the pharmaceutical "industry 4.0" | statistical programmer, clinical trial, clinical programming, clinical data, pharmaceutical industry, data governance, data integrity, data lineage, data standardization, clinical research, data science, data cleaning | source presentation source |
10 |
| 157 | Stay Ahead of the Curve: Learn About Leapfrogging as a Statistical Programmer | Jitendra Singh, Kavita Dargan |
statistical programmer, clinical data, clinical trial, data science, clinical programming, clinical research, data cleaning | source presentation source |
4 | |
| 158 | Stoic OS: Installing Ancient Wisdom for Holistic Leadership in the Digital World | Sascha Ahrweiler | We embark on a philosophical journey to combine the timeless wisdom of Stoic philosophy with the progressive principles of holistic leadership, all within the ever-evolving digital environment. The VUCA (Volatility, Uncertainty, Complexity, and Ambiguity) environment characterized by the digital landscape today provides a perfect opportunity for us to examine how ancient Stoic tenets can enhance the holistic leader's operating system to enhance self- awareness, rationality, adaptability, and ethical decision-making abilities. With an emphasis on Stoic-based holistic leadership, we provide a user-friendly guide for digital organizations to navigate the complexity of today's challenges, tackle the threat of VUCA head on, and embrace sustainable innovation. In addition, we discuss potential challenges and future directions as we envision a world where leaders can balance the virtues of the past and the advancements of today's digital age, while thriving in the VUCA era | resource allocation, statistical programmer | source presentation source |
9 |
| 159 | Streamlining Clinical Data Analysis and Reporting with R: Best Practices, Tips and Techniques for Efficient Workflow Management and Code Optimisation | Vaibhavi Govind Bhide, Perry Meeks |
Effective data analysis and reporting are critical in clinical research. R is a great tool for these tasks, but poorly optimized code can cause delays and inefficiencies. The paper demonstrates efficient ideas and approaches for speeding clinical data analysis and reporting with R. We discuss about profiling tools to identify and address performance bottlenecks in R code, followed by code optimization to achieve readability, maintainability, and re-usability. We will see how different packages such as but not limited to dplyr, data.table, tidyr, janitor, ggplot2 can be used throughout various stages of workflow management, data cleansing, wrangling, feature engineering, model creation, visualization and for faster data manipulation. We will learn the application of other packages like Rmarkdown, knitr to build reproducible reports and automate report generati on | data science, clinical data, feature engineering, clinical research, clinical trial, parallel processing, life sciences, data cleaning | source presentation source |
12 |
| 160 | Structural Equation Modelling of Quality of Life in Adults with Down's Syndrome in Nigeria | Ogochukwu Ijezie, Philip Davies |
Presentation | confirmatory factor analysis, proc means, factor analysis, service provider, proc freq | source | 15 |
| 161 | Submission Standards for Real-World Data: Gaps, Limitations and Recommendations | Jeffrey M. Abolafia, Sarah Ferko, Ingeborg Holt |
There is a rapid increase in the use of RWD to support marketing applications. In 2020 over 75% of NDAs and BLAs submitted to FDA included a RWD study to support safety and/or efficacy claims. However, submission standards and related documentation required for RWD are still in development and not well understood. This can make both regulatory submission and review more difficult. As a result, sponsors must develop an appropriate strategy for submitting RWD | raw data, clinical trial, controlled terminology, data science, clinical research, cdisc | source presentation source |
8 |
| 162 | Synthesise Your SDTMs! Realistic Simulated SDTMs When You Need Them | Matthew Travell | In this paper, we will demonstrate how a prototype process was developed to simulate SDTMs on an aggressive timeline first-in-human Phase I study to support the dose-escalation review process using the study protocol, EDC specs, readily available R packages, metadata libraries and prior study data as inputs leveraging a combination of open-source technologies of R, Python and Neo4j. We will show how this successfully enabled programming activities on the SAP mock shell, ADaMs, static TLFs and interactive exploratory Shiny application to be completed prior to study enrollment | adverse events, phase i | source presentation source |
7 |
| 163 | Taking down the fence between Biostatistics and Medical writing | Rob Wartenhorst | Even though we distribute files digitally, biostatistics still share the results with medical writing in the same format as we did decades ago. The only change being a transition from paper to PDF. Medical writers then need to interpret and turn this into text for the CSR. How would it look if biostatistics could share the analysis results as data instead? By leveraging a scientific writing copilot, it would eliminate the potential for manual errors in the CSR text and remove repetitive writing tasks | database lock, proc report | source presentation source |
3 |
| 164 | TFL Automation the CDISC 360 Way | Alison Macnab, John McDade |
Traditional programming practices lock vital metadata away within the code. Once this happens, even within macros, it is very difficult to recycle this efficiently and automation is lost. Keeping the metadata distinct from the codebase allows smart metadata management within and across studies. This paper looks at a journey starting in an EXCEL driven process, similar to CDISC 360 TFL proof of concept, and how this evolved into a user interface to provide ultimate flexibility for building TFLs. By combining the modularised CDISC 360 approach with a modern web application, we can overcome some of the limitations with EXCEL storage and manage the required metadata much more efficiently. Using enriched TFL metadata, we can write open code TFL programs capable of generating most study CSR outputs |
proof of concept, metadata management, excel, proc means, clinical research, cdisc | source presentation source |
12 |
| 165 | TFL Designer: Automating Tables, Figures and Listings Design and Generation for Clinical Trial Analysis | Bhavin Busa | Presentation | adam dataset, clinical trial, macro loop, excel, cdisc | source | 10 |
| 166 | The Curious Case of Observational Studies - A Programmer's Dilemma | Vikas Shokeen | There are two major types of studies in clinical research: Observational Studies (OS) and Randomized Clinical Trials (RCT). While the Randomized trials have been titled as the 'Gold Standard' in terms of producing reliable outcomes, Observational studies have also proved beneficial in various cases especially, when there is little to no evidence of the mechanism of a disease. However, these studies have put t he statistical programmers in a state of conundrum due to its various challenges such as 'No pre-defined CDISC Data Standards' etc. Despite these obstacles, programmers navigate through this turmoil with utmost care and following the CDISC guidelines of RCT's. This paper will outline the programming deterrents of observational studies, highlight the key takeaways with the help of an example of a non-experimental cohort study and suggest how methodology studies could be enhanced to pave way for a smoother clinical journey | adam dataset, data science, statistical programmer, trial arms, analysis set, clinical trial, data cleaning, sdtm domain, cdisc | source presentation source |
11 |
| 167 | The Digital Protocol Is Just the Beginning. Or Is It? | Thierry Philippe, Bill Qubeck |
Presentation | clinical trial, data science, customer experience, data cleaning | source | 17 |
| 168 | The Evolving Role of Statistical Programmers in the Pharmaceutical Industry: The Impact of Artificial Intelligence | Cindy Marabotti | In the pharmaceutical industry, the role of the statistical programmer is critical to transforming clinical trial data into reliable insights thus facilitate quality evidence-based decision making. Indeed, statistical programmers are responsible for creating and providing statistical results at each interim decision - making point, as well as for the conclusion of each clinical trial. They are also responsible for electronic submission of data (and statistical results) to regulatory authorities. With rapid changes in technologies (multilingual, visualization) and the regulatory landscape, this role is already undergoing significant changes. The emergence of artificial intelligence tools has the potential to further transform this job by automating some of his tasks. This should result in increased efficiency and speed. However, it i s important to recognize that these tools also present potential limitations and risks | statistical programmer, artificial intelligence, clinical trial, clinical research, machine learning, data science, pharmaceutical industry, data cleaning | source poster source |
8 |
| 169 | The Forgotten Outputs - Data for ClinicalTrials.gov and EudraCT | Erik Brun | As programmers we are generally familiar with study outputs like CDSIC data packages and TLFs. But besides the study report, it is also a requirement for many studies that the results must be published at ClinicalTrials.gov and/or EudraCT (insert link). This is sometimes referred to as the plain language summary. Despite being called plain language summary, a considerable amount of this reporting are Adverse Event tabulations. These results can be uploaded as an xml-file. With a programmer's point of view, this paper will give a short introduction to these result outputs and some of the challenges the author has encountered in creating them. One of the quirks with these outputs that can lead to some frustration is that the results aren't necessarily going to match the study report. I will give examples in the paper to explain why there can be a miss-match | proc means, error message | source presentation source |
5 |
| 170 | The Hassle of Validating Global SAS Programs: Made Easy(ier) | Michael Thomsen | Efficiency through automation is a growing trend in statistical programming, and we in Novo Nordisk have recognized the need to streamline our global validation process for SAS programs. Currently, this process is manual, comprehensive, and time-consuming, leading to a low turnover rate and frustrated end users who resort to making their own adaptations. This undermines the goal of having a global validated program and limits the number of programs that can achieve this status. To address these challenges, we have developed an automated test framework using Azure DevOps CI/CD pipelines, SAS programs, and Git. This approach has enabled us to make rapid updates to the programs, while ensuring their functionality and providing full transparency throughout the workflow. By automating the validation process, we have reduced the workload on maintainers and increased the turnover rate, resulting in a more efficient and effective system | sas log, imputation, in=, macro definition, macro variable, excel, data cleaning | source presentation source |
12 |
| 171 | The Human in the Loop: Accelerating AI/ML Efficiencies for Clinical Data | Achilleas Zaras | Clinical development increasingly recognizes the value of AI/ML. This area within life sciences has shown intent to shift from hype to reality in recent years as companies strive to operationalize their strategies. Still, many are relatively early in the adoption curve. A few years ago, the buzz around AI included many predictions as well as concerns over job security, regulatory guidance and oversight. Now as AI becomes more established, questions remain, but there are also real use cases to pursue, and potential efficiency gains that are much needed. AI adoption means getting more specific about the human's role - which includes providing necessary supervision and reducing risk, but also fine-tuning models, actioning decisions and finding new, efficient ways to work smarter using this intelligence. The Data Manager is a key human in the loop for many of clinical development's practical AI/ML applications | data review, machine learning, data governance, data cleaning, life sciences | source presentation source |
4 |
| 172 | The Journey of Preparing an ISS | Lieke Gijsbers | Presentation | source | 18 | |
| 173 | The Open-Source SCE | Caroline Phares | Withtheprevalenceofcollaborativeeffortsanduseofopensourceontherise,whatnewopportunitiesdoesthisopen?Formanyyears,organizationshavetriedtosolvethesameproblem:implementingtheidealSCE.Afewyearsago,thePHUSESCEwhitepaperstandardizedwhattheidealSCEshouldlooklike.Today, inspiredbythatwhitepaper, andfollowingthetideoftheopensourcemovement,theOpen-SCEcommunityhasbeenlaunched.ThispresentationwilldescribethevisionofOpen-SCE,reviewtheactivitytodate,describethecurrentcontributions,andextendaninvitationtocomealongthejourney | source presentation source |
5 | |
| 174 | The Reality Behind Data Preparation for Statistical Programmers | Megan Carnell, Leon Lombaard |
Poster | data cleaning, data wrangling | source | 1 |
| 175 | The role of FHIR Resources in ensuring Semantic Equivalence in EHR2EDC direct data capture | Patrick Genyn | Collecting study data directly from EHRs is attracting increasing attention as sites offer access to subject records using FHIR methodologies1. EDC practice requires the determination that specific data is accurate to be made by site staff prior to completing the eCRF. With direct data capture this determination is automated and relies on the EHR coding system for this purpose (e.g., LOINC, SNOMED, etc.). Selecting data using these codes is an integral part of querying using FHIR API calls. This may result in the same biomedical concept being coded differently, with the result that the data may not be semantically equivalent (i.e.: the data from the 2 vocabularies does not have similar meaning). The practical implications for studies will be discussed and FHIR Observation Resources used to illustrate acceptable and unacceptable semantic equivalence, how it can be recognised and operationally mitigated | cdisc controlled terminology, clinical trial, cdisc | source presentation source |
4 |
| 176 | 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 |
| 177 | Traceability Confirmation at Your Fingertips | Shafi Chowdhury | Presentation | dataset, data cleaning, raw data | source | 16 |
| 178 | Transparent is the New Black - Improving Data Findability Through Better Clinical Metadata | Lukas Kniola | The secondary use of clinical trial data can significantly increase its value beyond the original analysis and help find new insights, signals, and ultimately, treatments. A major hurdle to data reuse is the lack of transparency and discoverability. SDTM and ADaM datasets have reliable, strict naming conventions and consistent structure within and across studies. These qualities make it possible to create generalized programs which stack and process datasets in an automated manner, regardless of their contents. The output is descriptive metadata that is useful, searchable, and not constrained by privacy concerns. Such metadata can be made open and searchable while access to data remains restricted. It allows for finding tests, events, visits, and characteristics across datasets and studies before requesting access to relevant data assets | adam dataset, clinical data, proc means, proc freq, character variable, numeric variable, cdisc | source presentation source |
16 |
| 179 | Unleashing the Power of Large Language Models for Clinical Data Approaches: Will AI chatbots like ChatGPT/Bard replace statistical programmer and statisticians over time? | Ercan Suekuer | The use of natural language processing (NLP) techniques has been gaining significant momentum in clinical research and healthcare delivery. With the advent of large language models such as ChatGPT, there is now an unprecedented opportunity to leverage these models for analyzing and extracting insights from clinical data.ChatGPT is a state-of- the-art language model that has been trained on a massive corpus of text data and has demonstrated remarkable capabilities in generating human-like responses to natural language queries. In the context of clinical data, ChatGPT can be used to extract valuable insights from electronic health records, patient surveys, clinical notes, and other sources of unstructured data. In this presentation, we will discuss the benefits and opportunities of using ChatGPT for clinical data approaches | clinical data, clinical research, regulatory submission, training set, data science | source presentation source |
9 |
| 180 | Unleashing the Role of a Statistical Programmer in Expediting a Submission, Novartis | Sowmya Singuluri | Being an integral part of pharma world, we know the magnitude of impact created when the drug is made available to patients with critical ailments even by a day ahead. As statistical programmers, we strive hard in all possible ways to plan study activities and deliver the analysis reports well ahead of planned timelines. There are a few key areas to focus on as a statistical programmer in each phase of the study conduct which paves way to deliver the study package to health authority with high quality and double the speed. To mention a few a) Providing correct annotations especially for study specific case report forms ensuring CDISC standards are rightly implemented which will avoid unnecessary post-production changes. b) Staying a step ahead with lean statistical analysis plan to ensure all the reports generated would be a part of your CSR and using R shiny apps for exploratory analysis. c) Drafting a clear and clean define compliant programming data specifications | statistical programmer, adverse events, clinical trial, define-xml, data science, proc sql, cdisc | source presentation source |
7 |
| 181 | Unlock Privacy-Sensitive Data with AI-Generated Synthetic Data | Wim Kees Janssen | artificial intelligence, data cleaning, quality assurance, privacy | source presentation source |
6 | |
| 182 | Using Machine learning for building a demonstrative model to counter the under-diagnosis of dementia | Ashapriya Sebastian | Early and accurate diagnosis of dementia is critical for timely intervention and improved patient outcomes. Neuroimaging techniques like MRI facilitate the identification of structural brain changes associated with dementia pathology. However, manual analysis is limited by inter-rater variability and time intensity. Hence, this work investigates deep learning models for automated MRI segmentation to enable quantitative analysis and aid early dementia detection. Specifically, convolutional neural networks (CNN) like U-Net, SegNet, DeepLab, and Attention U -Net are implemented and evaluated on a brain MRI dataset. T he Attention U-Net emerges as the top performer, achieving a Dice coefficient of 0.9262 for segmentation accuracy. Further experiments with Focal and Focal Tversky losses yield coefficients of 0.9297 and 0.9627 respectively, demonstrating the model's adapt ability | data cleaning, deep learning, neural network, imbalanced data, training set | source presentation source |
12 |
| 183 | Using Python to Automate the Reviewer's Guide | Matt Metherell, Simon Purkess |
Writing Reviewer's Guides can be a very time-consuming part of the submission process, and a lot of what we want to include is already stored in SAS datasets, or other structured documents such as spreadsheets. Python has the ideal package to update Word documents with this data automatically, as it edits Word documents at an XML level. So, taking the CSDRG template as a starting point we can use the python-docx package to easily find a specific paragraph and enter specific text or a table | dataset, excel, data cleaning | source presentation source |
14 |
| 184 | Validating Outsourced Programming Tasks by Comparing SAS Datasets Without PROC COMPARE | Herman Ament | Sponsor companies often outsource programming tasks to a CRO, whereby the SAS programs will be executed in the SAS environment of the company itself. An important requirement is that the delivered programs should give the same results in the SAS environment of the CRO and in the company. Danone Nutricia Research (DNR) has outsourced some SDTM tasks to a CRO. DNR has an LSAF environment, the CRO is using SAS Studio. The programming structure at the CRO mimics that of DNR, including study specific metadata and global SAS macros. Deliverables shared with DNR consist of SDTM programs, study specific macros and SDTM SAS datasets. A special component of the deliverables is a hash dataset. This dataset is used to verify that the SDTM SAS datasets created in both environments by the delivered programs and study specific macros are equal | dataset, proc compare, proc step, sdtm domain, cdisc | source presentation source |
5 |
| 185 | Visualising Your Protocol Objectives | Varun Tripuraneni | With the increasing complexity of study designs and the use of composite endpoints, we often struggle to effectively communicate the objectives of their studies beyond a table of numerical results. To better visualize the most important outcomes of a study and show how they relate to individual endpoints, it is important to develop and utilize effective figur es that can provide a clearer representation of the study's objectives. By doing so, researchers can better showcase the relevance and significance of their trial results to stakeholders. Effective use of visual aids can also help to improve the overall un derstanding of the study outcomes, making it easier for stakeholders to draw meaningful conclusions and make informed decisions based on the results | sample, bar chart, clinical trial, scatter plot, clinical data | source presentation source |
6 |
| 186 | When the Post(er)man Rings Twice - Make Your Science Sexy | Josua Böser | Scientific posters are essential for communicating research findings to a broader audience. Unfortunately, their importance tends to be overlooked, and many researchers struggle to create informative and visually appealing posters. A lack in this skillset often results in hard work going unnoticed, which poses a significant loss to the scientific community. This paper presents a comprehensive guide to creating more captivating scientific posters. It provides practical tools for researchers to improve their overall presentation and design skills, including tips on layout, color schemes, and data visualization. Additionally, we review classic and modern poster designs, offering recommendations on presentation style and facilitation. Using the recommendations outlined in this paper, researchers can acquire tools to morph their posters into a dazzling work of art, capturing their Audience's attention and ensuring their research is understood and appreciated | data visualization, offer recommendation | source presentation source |
11 |
| 187 | When You're Helping Create Life-Saving Treatments, Who's Looking After You? | Jessica Shipley-Burgess | Poster | source | 1 | |
| 188 | Where Are You on the Blob Tree? How to Maintain Team Health and Monitor Project Teams in a Work-From-Home World | Florence Ratcliffe | Poster | source | 1 | |
| 189 | Will AI Tech Steal My Job? | Brittany Long | Poster | source | 1 | |
| 190 | Will R/Shiny Provide a Novel Way for Study Teams to Review Clinical Data? | Gustav Söderström | The open-source programming language R is experiencing increased popularity among statistical programmers in the pharmaceutical industry. Pharmaverse [1] is an excellent example, illustrating the industry's keen interest in the language where multiple companies are joining forces to create and enhance R packages for clinical reporting purposes. The ability to create interactive visualizations makes R/Shiny a popular choice for clinical and statistical programming. R/Shiny is a tool that has the potential to streamline the work of reviewers. This paper aims to provide an overview of the tool and discuss the challenges involved in implementing it for clinical data review | clinical data, data review, statistical programmer, data science | source presentation source |
3 |
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.