Global Pharma Tek’s Post

For those interested in clinical trial systems. This article about the design of Randomized Clinical Trials (RCT)s by Stolberg, Norman and Trop is particularly interesting for the ways it describes how they may be classified. In my view, capturing the data for a clinical trial is to capture the digital shadow of its procedures in an information system. It is compelling to consider how those procedures may be influenced by higher level aspects of trial designs, and in particular the potential difference in data needs that may exist between any two classifications. Years ago the need to identify and classify use cases was important when I was designing a model-driven software system for supporting data capture use cases in multiple areas of molecular biology. The idea was to generate generic features and support configuration for things that would vary. Looking at how data capture needs of proteomics, genomics and metabolomics (which often reflected that digital shadow idea) were similar and different was useful in assessing what was most likely feasible to generate and what would be best left to aspects of configuration such as the support for different plugins or substitution of field-focused ontologies. In that case, system design was often informed well by knowing the delta needs between use cases. But back to trials! The article is well written and would be of interest to data architects and other system designers. #clinicaltrials #design #classifications #data https://lnkd.in/egu7ApcU.

Simpler but methodological rational platforms for mapping the patient population effects in clinical trials and individualization of dosing in precision medicine In one basic development area we are focusing on mapping population PK variability through simpler rational platforms. We have developed a methodological approach to simplify the covariate mapping of PK-PD profiles. Variations of PK-PD in diverse patient populations is usually observed during the clinical trials and market launch. This probably means dose adjustment decisions or off-label usage. We have developed and calibrated a simpler, efficient, and robust algorithm and computer codes for accurately linking and estimating the model parameters and patient population characteristics. Our aim in this line of research is to enhance the knowledge and understanding of methods used in nonlinear mixed effects and answer following questions: § What is the reliability of statistical check points and finding conditions where the reliability is somewhat sacrificed? § How different algorithms perform? § How selection bias kicks-in and its role on covariate effects? § How can therapeutic decision making be improved by incorporating time-varying effects Integrated population mechanistic approach is based on following steps: 1.    Data sorting & modeling tasks identification 2.    PBPK model structure 3.    Introducing System Parameters 4.    Introducing drug related parameters 5.    Uncertainty & variability of parameters   6.    Study design & sensitivity 7.    Integrating Clinical Data 8.    Population PBPK 9.    Validation 10. Concentration Linking to Effect 11. Clinical predictions 12. Refinements of model structure, parameters and selection bias  

Discover how intuitive decision rules in clinical trials can support regulatory requirements in our latest PharmaLex article by Marco Munda, Associate Director Statistics. Key Takeaways: 🔍 Bayesian framework enhances adaptive trial designs 🔍 Predictive modeling offers clearer decision-making 🔍 Improved flexibility and regulatory confidence Curious to learn more? Read the full article here: https://lnkd.in/dGrVzab5 #ClinicalTrials #RegulatoryAffairs #BayesianMethodology #PharmaLex

This article explores the use of #longreadsequencing in #clinicalgenomics and medical research, highlighting its potential applications, but also highlighting its limitations. #genetictesting #genomictesting #genomicmedicine https://lnkd.in/dqT9vRSE

We are thrilled to announce the launch of Loon, Inc.! Loon is a biotechnology company dedicated to creating the realm of Augmented Intelligence Clinical Discovery and Evidence Synthesis. Visit our new website at loonbio.com to explore a world of reimagined patient-driven clinical research to accelerate innovations in healthcare.   At the core of Loon is our distinguished team of experts: — Dr. Ghayath Janoudi, our Founding CEO and Chief Science Officer, an MD / Ph.D. (Epidemiology) with over a decade of experience in clinical research with over 15 publications, drives our scientific endeavours. He is an expert in the intersection of AI and clinical research, health technology assessment, market access, regulatory affairs, and quality assurance. — mara rada, MA, our Founding Chief Strategy Officer and Chief Operating Officer, is a marketing scientist and peer-reviewed author with a rich background in pharmaceutical marketing and advertising strategy, alongside many entrepreneurial ventures. She orchestrates our commercial, operational, and strategic initiatives. — Josip Ivkovic, MSc, our Chief Technology Officer, a computer technology and AI scientist with widespread experience in hardware and software development, Web Apps, AI, and ML, spearheads our technological advancements.   Historically, many medical advancements were often found by chance. At Loon, we turn these happy accidents into systematic, deliberate, patient-driven quests and methodological approaches that fast-track the path to groundbreaking medical innovations.   Our validated and patent-pending method of Augmented Intelligence Clinical Discovery and Extreme Misclassification Outlier Analysis ensures that no potential therapy, indication, or clinical pathway remains concealed in the vast waters of clinical data. Augmented Intelligence not only elucidates hidden correlations but accelerates the elucidation of mechanisms of action, propelling medical research into a new era of patient-driven innovation.   Our method is substantiated by two peer-reviewed papers, with more underway. And we're just getting started! The forthcoming AI Systematic Reviews platform will further expedite Clinical Discovery, enabling regulator-grade, human-verified systematic reviews to be conducted in days, not years. At Loon, every data point is a steppingstone to potentially life-changing medical innovations. Our venture into the unchartered territories of Augmented Intelligence isn’t merely a quest for knowledge, but a mission for a healthier, patient-driven future.   Join us in this groundbreaking venture at loonbio.com as we redraw the frontiers of Clinical Research with Augmented Intelligence Clinical Discovery, drastically reducing systematic research timelines and unveiling a new horizon of patient-centric medical innovation. #AIClinicalDiscovery #AIEvidenceSynthesis #HeathcareInnovation

We're pleased to introduce our novel Causal Responders Detection (CARD) method, detailed in our latest publication. CARD applies machine learning and causal tree techniques to identify individuals who benefit significantly from treatments, ensuring false discovery control in both RCTs and observational studies. This method refines our understanding of individual treatment responses, providing crucial insights for precision medicine development. Our simulations confirm CARD's effectiveness across various settings, enabling to apply this method across therapeutic areas and clinical phases. #PersonalizedMedicine #MachineLearning #ClinicalTrials #HealthcareInnovation Tzviel Frostig Oshri Machluf Elad Berkman Raviv Pryluk Amitay Kamber

Interesting but flawed article... the author claims that the vast majority of published clinical prediction models are useless. But we don't know that. Could be true, or not (pick your favorite threshold for the "vast majority" - say 95% or 99% - the core argument does not change). The reason is that it takes an extraordinarily long time and a lot of money to establish the usefulness of clinical models. For example, it took ~15 years and many millions of dollars to conclusively prove the utility of one of the successful tests mentioned in the paper, Oncotype Dx. There simply aren't sufficient resources - time and money - to subject the vast numbers of published models to such evaluations. As a result, their usefulness will never be known. The claim we can confidently make is: "We will never find out the usefulness of the vast majority of published clinical prediction models". https://lnkd.in/gf8BVssW

Hierarchical Interaction Network (HINT) model for general clinical trial approval prediction tasks. Research paper which uses HINT to quantifies uncertainty and improves interpretability in clinical trial approval predictions. https://lnkd.in/en78SQsi #clinicaltrial #artificialintelligence

Great review article that provides an overview of WGS in clinical practice - describing the technology and current applications as well as challenges connected with data processing, interpretation and clinical reporting. "We are confident that WGS has the potential to make a difference for patients and we foresee that the clinical use will increase in the coming years." https://lnkd.in/dkHbEKJc

Introducing our new "Dynamic Baseline" feature, available for spectral flow and mass cytometry. Check it out today on app.teiko.bio! This feature allows users to set any time point as the new baseline, instead of just the first timepoint in a series. This provides more flexibility when analyzing longitudinal data across key clinical milestones. Key Features: * Visualization: Select any time point as the baseline to better reflect biological changes during treatment, allowing for clearer insights into immune responses over time. * Statistics: Statistics automatically update when the selected baseline changes, ensuring accurate comparisons. * Guidance: Time points are colored in the dropdown to differentiate those with sufficient samples for statistical analysis. How to Use: * Set your baseline using the “Normalization Baseline” dropdown—choose a specific time point like C1D15. * Data is normalized by subtracting each subject’s baseline value from their subsequent values, aligning population or marker values for better tracking of immune responses. This feature improves the ability to track immune responses at critical treatment points, enabling more precise comparisons between key dates, especially during pivotal events like lymphodepletion, or the start of a second round of treatment. Contact us today if you have a multiple dosing clinical trial, or if you want statistics run on your prior datasets collected by us.