Ben Vaughn is a statistical scientist who has participated in nearly 20 regulatory submissions and is an expert on CDISC implementation and standards. His work has included serving as lead statistician to complete displays and datasets for several ISS/ISEs (as well as the associated pivotal studies); coproducing the ISE for two opioid products; and statistical consultation, display generation and submission work for four separate products for OA knee pain. He has authored responses to various FDA queries regarding NDAs, PMAs, IDEs, and SPAs. Additionally, he has represented sponsors in FDA teleconferences and face-to-face meetings for both OA knee pain products and opioid products.
A new drug application (NDA) covers information about a product from inception through clinical trials. The integrated summary of efficacy (ISE) is a section of the NDA that provides a focused summary of efficacy data primarily from the pivotal phase 3 trials. This article discusses the primary goals and considerations for developing an ISE.
The primary objective of the ISE is to orient the reviewer to the key measures of efficacy presented in the pivotal studies and the associated results.
Beyond the information available in the individual study results, the ISE analyses should also:
- Obtain a precise estimate of the primary endpoint.
- Obtain precise estimates of key secondary endpoints which provide supporting evidence for the primary endpoint.
- Conduct key a priori subgroup investigations to examine consistency of effect.
Here are a several considerations for developing the ISE:
The ISE should rely on results from the pivotal phase 3 studies. It is possible to include phase 2 results if the primary endpoint is the same as for the pivotal study, but this can be risky. Because phase 3 studies are typically designed by using the phase 2 study with the best outcome (or the best endpoint from a given phase 2), this results in an inflated alpha (alpha, also known as type 1 error, is the likelihood that the study shows a difference between active and control when in reality there is no difference). Therefore, it may not be legitimate to include the phase 2 results with the phase 3 results.
Focus only on primary endpoints for which there was tight alpha control in the pivotal trials.
In the best case scenario, all phase 3 trials had the same primary endpoint and collected and analyzed it in the same way, however, this isn’t always the case. Often, lessons learned in earlier studies are applied to later studies, so there are minor differences. It is acceptable to look back at the data collected in the earlier studies and analyze it in the same way as was done in later studies- as long as you can still show alpha control was maintained and understand that the reviewer’s assessment of success will largely be based on the pre-specified analysis.
It is not helpful to perform many additional analyses that are different than those planned for the pivotal trials in hopes of creating supporting evidence. FDA reviewers are keenly aware that the more analyses are run, the more likely it is to find some significance and will not find that strategy acceptable. They are likely to accept a change in model or analysis when the move is from a significant result with a flawed analysis to another significant result with a better analysis. Anytime there is move from an original analysis with an insignificant result to a new analysis that shows a significant result, reviewers will object.
Include only one or two secondary endpoints. A laundry list of secondary endpoints will not persuade FDA reviewers. Secondary endpoints should be clinically relevant and should support the story told by the primary endpoint.
Conduct analyses of therapeutic response in subgroups of interest. Heterogeneity of response will be of interest to reviewers and it is better to identify any instances of this in ISE rather than leave it to a reviewer to discover. However, subgroup investigations and the interpretations of these results should be done with some caution. Sample sizes are often insufficient to get stable results; therefore, focus on subgroups with sufficient sample size and a scientifically plausible rationale that they would have a meaningful difference in response.
Including everything is not OK: While there is a tendency to want to include any possible tidbit that may look positive for the product, be aware that it will not improve chances of approval. It is far better to have an efficacy summary that tells a clear and concise tale.
FDA reviewers are less likely to spend as much time looking at the ISE as they do the integrated summary of safety (ISS) and the results from the individual phase 3 trials. Use the ISE to make a clear, concise, and focused case that the product is efficacious.