These requirements are most often ignored. The law requires that trials be registered and their results published within a year of completion, but most register

These requirements are most often ignored. The law requires that trials be registered and their results published within a year of completion, but most registered trials never have published results: http://www.bmj.com/content/344/bmj.d7373 http://www.bmj.com/content/344/bmj.d7373 The results are simply buried if they're not sufficiently favorable. There have been a number of meta-analyses which dig up unpublished results (which is very difficult) and demonstrate that medications no longer have statistically significant effects when you include the unpublished data. For example: http://www.bmj.com/content/341/bmj.c4737 http://www.bmj.com/content/341/bmj.c4737 (Both these papers are open access, so you can read them if you'd like.)