[throwaway98031]

> pharmas are able to start clinical trials and then shut them down without reporting result is a problem.

As you've worked in the industry, I'm curious: What do you think could remedy this situation?

Forcing all registered clinical trials to release all of their data seems like the obvious solution. We still need the ability to shut down clinical trials if unexpected side effects such as increased suicide rates appear, but allowing companies to keep the results private is obviously problematic.

[acveilleux]

Mandatory release is useful. At least Phase 3 and 4 trials conclusions should probably require publication. Phase 2 is less essential, drug will never go for sale without a Phase 3. Just saying "nope, won't continue with drug" in a press release doesn't really explain why.

However, I think Phase 2 trials for new use of already approved compound should have a minimum publication requirement if they don't follow on to a phase 3. The safety (side effects and adverse events) should at least get aired.

[interesting_att]

More than mandatory release, we need to:

1) Have full transparency of the placebos being used. Many studies that are published don't fully disclose the placebo being used. Full release should really mean full release.

2) Movement way from the for-profit journal system to a not-for-profit system, to prevent obvious conflicts of interest. Full release means nothing if the means in which information travels is biased.

3) Reining in the benefits pharmaceutical companies can give researchers and doctors. A number of studies have shown even the smallest of gifts can change a doctor's propensity to prescribe a specific drug. Surprisingly this can be as powerful as a journal article.

4) Passing real legislation to reduce the revolving door between the FDA and the rest of the medical industry to prevent a number of other problems.

[andreyf]

One option: https://en.wikipedia.org/wiki/Contract_research_organization

[acveilleux]

How is that a solution? I worked for a CRO and contractually, all we did belonged to the client, including all publishing rights on all results.

[andreyf]

Sorry for not expounding, that was indeed a bit opaque. I don't have first hand experience with the industry and so would appreciate any corrections to my understanding of it.

When research is done via contract between two companies, it can be more effectively regulated to prevent hiding or discarding of results. I imagine there are already laws in place saying a company can't hide research results like this and CROs make compliance more likely because they do not benefit from hiding bad results as pharmaceutical manufacturers do. I don't think publication rights are the important bit here, but compliance with rules about disclosure to the FDA.

[acveilleux]

That's not how CROs work usually work. Effectively, as far as the FDA is concerned, the CRO does not exist as a distinct entity, it's an organic part of the sponsor (i.e.: the reporting pharma.)

The sole organization reporting to the FDA is the trial sponsor.

A CRO that felt like it could just tip off the FDA could soon find itself without further contracts.

In addition, in a properly blinded trial, most of the players (whether CRO, sponsor or site/treating physician) have no clue who's being treated or not. Nor do they know much more than they need to conduct their business.

The safety of a clinical trial is normally the job of a data monitoring committee or something similar. They are normally independent and empowered to monitor the result and, most importantly, unblinded so they can actually use stats taking treatment into account to spot actual anomalies. They report to and are paid for by the sponsor. The sponsor is legally required to act on the findings and adverse events are required to be reported to the FDA. At least that's the theory.

Gory details here: http://www.fda.gov/downloads/RegulatoryInformation/Guidances...

[andreyf]

Got it. Thank you for the thorough explanation!

[tsotha]

The obvious response for companies if they're forced to release all data is to stop doing any trials on approved drugs.