I just wanted to make a comment and start a discussion about a concept that is pervasive on podcasts, here on Meddit, and generally everywhere non-Pharma, which is that industry/Pharma sponsored trials meet their primary efficacy endpoints far more often than investigator-initiated or other randomized, controlled trials, and this makes them seemingly "less valid or lower-quality evidence".
As someone very experienced with translational medicine from early Phase 2 to late Phase 3, there is a likely reason for this which is never discussed and therefore leads to a systemic and extreme bias against industry studies.
The regulatory framework in the US under the FDA is extremely thorough, detailed, rigorous, and governed by Federal statutory laws as well as FDA rules and regulations. In order to even be allowed to run a Phase 3 study for approval (what we call a "pivotal trial"), the FDA has to agree to:
- The overall design
- The inclusion/exclusion criteria
- The study conduct
- The draft protocol
- The power assumptions and calculations
- The primary efficacy endpoint as well as the draft statistical analysis plan
- The choice of dose, including maximal effective dose, minimal effective dose, and maximal INeffective dose
There are many cases where Pharma ignored this, and the FDA issued a "refuse to file" letter and the company could not even submit a New Drug Application. Thus, the studies were not considered "adequate and well-controlled".
The FDA is concerned with exposing patients to a drug with unknown benefit, probable harm, and the risks/burden of participating in a clinical trial. Therefore, the FDA requires that the study design and conduct be rigorously compiled to answer the primary efficacy question. I have seen cases where it has taken years to arrive at a mutually-agreeable study design with the FDA.
So you not only have Pharma scientists, clinicians, and statisticians working on the study (with their enormous resources), you have FDA scientists, clinicians, and statisticians (with their enormous resources and access to thousands of study results) also providing their input and guidance for the study. Hence, because the way the regulatory framework is set up in the US, Pharma studies have far, far, more resources and people to be tapped into.
Lastly, once studies are submitted to the FDA, the FDA writes their own statistical analysis plan and runs their own analyses. They can make different choices about missing data, and adjusting for confounding variables. All in all, if a trial is considered "positive" for labeling, two separate entities have analyzed the data in different ways and come to the same conclusion.
If you want an example of where industry trials have almost never been positive, look at Alzheimer's. Since 2010, there have been over 20 Phase 2 and 3 studies with anti-amyloid agents that ALL have been overwhelmingly negative, despite all of these resources and attempts.
Source: Original link