a) Pharma identities the type of patients they need (e.g., 25-50 female, not pregnant, with specific ailment if Phase III), specific tests, and measurements required throughout the study.
b) pharma contacts third party (3P) to manage study patients.
c) 3P has relationship with dozens or even hundreds of doctors' offices, knowing what office can fulfill the test & measure requirement, and has the potential trial patient pool.
d) 3P has existing contract with these doctors' & hospitals. They get patients onto the study. ( <--- #1 reason this is farmed out in my opinion)
e) Doctors & hospitals perform the study and collect the data.
f) doctors & hospitals pass the data to the 3P
g) 3P passes it to the pharma
h) repeat e) through g) as many times as the study requires it. This can be once, or many times over years.
i) pharma pays 3P, 3P pays doctors & hospitals, and they pay the trial patients - each taking their cut along the way.
There are variations on how this is done, sometimes no 3P, sometimes pharma will have their own pool of patients and 3P. Also this is a very rough flow as there are often checks, audits, and validations (should be) done during the study.
Each study is different and therefore run differently. Many, many, factors determine how a study is run, analyzed, and published.
Most studies are very small, using only one site and a few volunteers. Most of these never see the light of day, as the results aren't publishable or are uninteresting. Think power law distributions with studies, not normal distributions. These studies are often so small that the collector and interpreter are the same person, typically they are also the grant writer and admin. If lucky, they may get some nurses or undergrads to help out. Again, I'd say this is ~80% of studies.
The really large studies that places like Pfizer run will separate out nearly all parts of a study. So consenters, nurses, intake, data admin, funding admins, stats guys, etc are all different people. These are very expensive studies to run so it's really only for FDA approval, not scientific inquiry and case studies.
Generally, most studies are very small and not publishable. They don't need to separate out everyone. Everyone kinda trusts that everyone else is doing their best. If something snazzy is found, then follow up studies will build on it's findings. Most of the time though, nothing is really found.
Here's one quick way to rig any clinical trial: Anyone and everyone that has any kind of negative reaction or health condition gets disenrolled. Since it's 'double blind' it appears on the surface level that there's no way to know who's in what group. Naturally, the end result is always the same: The test group had the same number of reactions as the control by the end of the study.
Do you have any sources for this? I'm rather disbelieving of it, but would love to be proven wrong. I can't imagine that _some_ major news outlet wouldn't love to stick it to the status quo with a whistleblower, unless the "whisteblower" made false claims about their proximity in the company to the dangerous/illegal actions they are trying to bring attention to.
We can also see that the FDA does nothing to investigate the integrity of the trials. They just accept whatever the manufacturers tell them.