<img alt="" src="https://secure.lote1otto.com/219869.png" style="display:none;">
Skip to content

How to Determine if Your Clinical Trial Has Sufficient Data?


It can be difficult to estimate just how much time and data you need to address the multitude of considerations that underpin the success of your clinical programs. However, beyond a certain point there are scientific, rational and ethical reasons not to continue enrolling patients onto a clinical trial. Defining this conceptual point, Cytel experts have developed the Sufficient Information Threshold to aid decision-makers in determining whether their trial has accumulated sufficient data for their various goals.

Cytel’s new position paper explains the concept of Sufficient Information Threshold, why it is important and how it can be facilitated by using methods (e.g., Bayesian Predictive Probability, Conditional Power) that enable unplanned early stopping.

Factors defining the Sufficient Information Threshold

Sufficient Information Threshold is the theoretical point at which all the information possibly relevant to a study has been collected and there are reasons to stop a clinical trial. This includes information about the primary endpoint, any secondary endpoint, and safety data to construct a risk-benefit profile. It could also include early phase information relevant to designing optimal late-phase studies. Therefore, it is a conceptual point that can enable us to begin asking questions about the efficiency, ethical quality, and effectiveness of a clinical study.

According to the authors of the position paper, we can define the Sufficient Information Threshold using these three factors:

  1. Any additional scientific insights collected after this point will not justify the allocation of possibly scarce resources towards the trial
  2. Continuing the study after this point is unlikely to alter a business or regulatory decision about a new therapy
  3. The data collected is enough to answer every question required by a clinical study and it can be adequately determined if the new intervention is superior to standard of care

Use of Bayesian methods for stopping at Sufficient Information Threshold

The Sufficient Information Threshold is typically unknown during trial planning as it is difficult to tell ahead of time exactly when there will be sufficient information to justify stopping a clinical trial. However, we can use methods that enable unplanned early stopping and thereby facilitate stopping at said threshold.

This was a topic that Cytel co-founder Nitin Patel and Chief Scientific Officer Yannis Jemiai contributed to an award-winning book Bayesian Methods in Pharmaceutical Research. The authors note that at each point on the clinical development journey, when a Go/No-Go decision must be made, a trial sponsor is essentially engaging in a Bayesian update about beliefs regarding probability of success. [1]

Bayesian methods are popularly used for interim analyses, where Bayesian predictive power calculates the probability of reaching a specific trial outcome based on the data at hand. This means that the trial can be stopped the moment the Sufficient Sample arrives. Bayesian techniques also allow you to make a direct comparison against all available treatments, across clinical trials. A trial can be stopped early if the new therapy is shown to be comparatively not so effective thereby providing ethical benefits to the patients who can now enroll in other trials that might have more effective therapies.

The position paper explores the significance of this for both early and late phase trials.

Download Position Paper


[1] https://bookauthority.org/book/Bayesian-Methods-in-Pharmaceutical-Research/1032241527

About Pantelis Vlachos

Pantelis Vlachos photo on black 2018Pantelis is Vice President, Customer Success for Cytel, Inc. based in Geneva. He joined the company in January 2013. Before that, he was a Principal Biostatistician at Merck Serono as well as a Professor of Statistics at Carnegie Mellon University for 12 years. His research interests lie in the area of adaptive designs, mainly from a Bayesian perspective, as well as hierarchical model testing and checking although his secret passion is Text Mining. He has served as Managing Editor of the journal “Bayesian Analysis” as well as editorial boards of several other journals and online statistical data and software archives.