I'm a bit sceptical here. As it happens, my partner was taking Effexor (a branded formulation of venlafaxine) and was switched to the generic Enlafax. She complained to me that she was feeling a little bit worse starting around this point, and when patient groups started sharing stories of many people who began feeling worse following the switch (which is what led, eventually, to the media coverage), she requested switching back to Effexor (which required paying out of pocket, as it's no longer covered), resumed taking it, and started feeling better.

She will no doubt be part of the statistics that paper is using to suggest a nocebo effect, as she only complained publicly after the media coverage started, but she has notes showing she had noticed the effect prior to seeing anyone else complain about the generic formulation.

To be clear: her experience is absolutely not proof that Enlafax is in any way inferior or different than Effexor (for my partner specifically, or for people in general)! It's very possible her entire experience was a nocebo effect (triggered, perhaps, by the Enlafax packaging or something) followed by a placebo effect when she switched back to the branded drug. However, it could equally have been caused by Enlafax being actively inferior for her.

What I think is fairly clear from my partners experience, however, is that the study is not correctly designed to reject the null hypothesis here. It's assuming that the people who publicly complained about Effexor after the media coverage were fine with it until the media coverage, whereas it's just as plausible that they simply didn't publicly complain until they realised other people were also having an issue.

Ultimately, any study here that's not a proper double blind test is going to struggle pretty hard here, and I think this is no exception.