Blog
Jason Connor's Upcoming Events
Subscribe to In the Interim...
No items found.
Jason Connor will be participating in several upcoming events, including presentations on Bayesian adaptive trials and findings at various conferences and teaching a class at Johns Hopkins School of Public Health from May 31 to June 28.
I wanted to update everyone on a few upcoming events that I will be participating at:-May 31: Presenting the READAPT teams' Bayesian adaptive redesign of the ALLHAT hypertension trial to the ALLHAT steering committee.-June 11: Presenting "A Bayesian Adaptive Cardiology Device Trail Leading to Approval" at the International Chinese Statistical Association meeting in Bethesda.-June 23: Presenting the READAPT findings at the Academy Health Meeting in Baltimore.-June 24-28: Teaching a week-long class at the Johns Hopkins School of Public Health.-June 25: Presenting a short course on Bayesian adaptive trials at the Academy Health meeting in Baltimore.For more information on any of these, please feel free to contact me (jason@berryconsultants.net).Jason Connor Statistical Scientist Berry Consultants
Download PDF
View
Statistical Plan for the MOST Trial
This is a plan for a Phase Ill trial to explore the efficacy of Argatroban and Eptifibatide in combination with rt-PA in treating stroke patients. Two doses of each combination will be compared to a control (rt-PA only) arm with respect to their ability to improve subjects' 90 day scores on the modified Rankin scale (mRS).
Precision Promise Adaptive Platform Trial update
The Precision Promise platform trial is an adaptive study exploring multiple potential therapies for pancreatic cancer, with pamrevlumab recently advancing to the next stage after demonstrating a predictive probability of at least 35% for improved overall survival, highlighting the trial's innovative approach to efficiently identify effective treatments in a field with limited options.
Comments on the draft FDA master protocol guidance
Kert Viele's blog discusses the FDA's draft guidance on master protocols for drug and biological product development, highlighting key sections on trial design, randomization, control groups, informed consent, and regulatory considerations, while encouraging feedback from experts to enhance the guidance's effectiveness before the comment deadline of February 22.
Is early stopping biased? Maybe, maybe not….
The blog by Kert Viele discusses the potential biases in clinical trial results, particularly focusing on early stopping trials and the implications of only publishing successful outcomes, emphasizing that while biases exist, their significance varies based on the true response rate and the context of the trial.
Time Trends in Clinical Trials (related to 2023 ASA Biopharm panel)
On September 28, 2023, Kert Viele will moderate a panel at the ASA Biopharmaceutical Section Regulatory-Industry Statistics Workshop, discussing time trends in ongoing platform trials using real interim analyses from the PRINCIPLE and REMAP-CAP trials, focusing on their impact on clinical trial analyses, modeling adjustments, and the complexities of additive versus interactive time trends.
Prior Practicum: Interpretable Priors for CRM Designs
Joe Marion's blog discusses the challenges of designing phase I dose-finding studies in oncology using the Continual Reassessment Method (CRM) and Bayesian approaches, emphasizing the importance of selecting appropriate prior distributions to balance patient safety and effective dose escalation, while suggesting that re-parameterizing models can simplify the design process.
If Bayesian inference doesn’t depend on the experimental design, then why does “Bayesian optimal design” exist?
In his blog, Kert Viele discusses the importance of trial design in Bayesian analysis, emphasizing that while conclusions drawn from completed experiments remain consistent regardless of interim analyses, the design of the trial significantly impacts expected utilities, and optimal designs can enhance trial performance.
The use of synthetic or external data in clinical trials
The blog by Kert Viele discusses the tradeoffs of using external or synthetic data in clinical trials, highlighting how aggressive use can save patient resources but risks scientific robustness, and emphasizes the importance of understanding the agreement between synthetic and actual trial data to optimize inferential performance while minimizing patient enrollment.
HOW TO GET CONTROL? CONCURRENT VS CONTEMPORARY VS HISTORICAL VS SYNTHETIC CONTROLS
The discussion highlights the growing role of real-world evidence in clinical trials, particularly as a potential substitute for control arms, while emphasizing the need to address biases associated with various control methods and advocating for a future dominated by platform trials that balance cost savings with reduced bias risks.
WHEN SHOULD YOU BORROW HISTORICAL DATA (OR REAL-WORLD EVIDENCE)?
Kert Viele discusses the concept of historical borrowing in clinical trials, highlighting its potential benefits and risks, particularly in relation to FDA guidance and the importance of assessing "drift" to determine when it is appropriate to utilize historical control data for improving trial efficiency and accuracy.
IMPROVING PROGRAM RESULTS THROUGH BETTER PHASE 1 AND 2 TRIALS
Kert Viele discusses the challenges and probabilities of success in a drug development program, highlighting that a standard approach often leads to a high rate of failure due to poor dose selection in early trials, but suggests that a revised strategy of continuous patient allocation and dose escalation can significantly improve the chances of successfully bringing an effective therapy to market.
HYPOTHESIS TESTING, CLINICALLY IMPORTANT EFFECTS, AND DO WE PAY TOO MUCH FOR CLINICAL TRIAL INSURANCE?
Highly powered clinical trials are costly and often yield statistically significant but clinically meaningless results due to large sample sizes designed to mitigate random errors, suggesting the need for alternative approaches like flexible sample sizes and group sequential designs to optimize resource use and improve trial efficiency.
DESIGNING A COLLECTION OF TRIALS
The article emphasizes the importance of optimizing clinical trial designs by investigating multiple therapies simultaneously and utilizing strategies like Bayesian thinking and platform trials to significantly reduce the time and resources needed to identify effective treatments for difficult medical conditions.
Some Intuition Behind Hierarchical Modeling
Hierarchical modeling is an advanced statistical approach used in clinical trials to make inferences across multiple patient groups, enhancing power and reducing sample sizes while requiring careful implementation to account for variability and potential biases in observed data.
Should I use a Bayesian trial?
This week, we published an article in JAMA titled “Bayesian Analysis: Use of Prior Information in Clinical Trials,” which explores the nuances of Bayesian analysis in clinical trials, emphasizing the importance of transparency and community consensus when using informative priors to avoid bias and enhance trial efficiency.
Todd Graves: let me introduce myself
Todd Graves, who joined Berry Consultants in January 2012, plans to regularly blog about innovative clinical trial designs and his statistical modeling for college football team ratings, sharing insights and updates on both topics.
Jason Connor's Upcoming Events
Jason Connor will be participating in several upcoming events, including presentations on Bayesian adaptive trials and findings at various conferences and teaching a class at Johns Hopkins School of Public Health from May 31 to June 28.
ASA's new section for Medical Devices and Diagnostics
SIGMEDD, a statistics interest group within the American Statistical Association focused on medical devices and diagnostics, is seeking ASA members' support through signatures for its transition to a full section, with the current Chair-Elect encouraging participation via a petition.
SIGMEDD - Statistical Interest Group
As the Chair of the Statistical Interest Group in Medical Devices and Diagnostics (SIGMEDD), I am seeking support from 100 ASA members to transition our group to a Section, and invite interested members to sign our petition via the provided Survey Monkey link.
November Webinars and Conferences
Join us for three upcoming events this month focused on clinical trial simulation and adaptive trial design, including two webinars on November 14 and 15, and a conference on November 29-30, where experts will share insights and benefits of these innovative approaches.