Introduction: Science pertains to be self-critical, self-correcting, and ultimately replicable. As 1.7% scientists admitted fabrication or modification of data, and instances of authorship, citation, publication and peer review manipulations have been discovered, we sought to explore the integrity of biomedical publications, especially regarding authorship, nonpublication and means of reducing dissemination bias. Methods: Using three observational studies we analyzed: a) answers to the question: “Why do you think you should be the author on this manuscript?”; b) funding, study type and differences between research presented at peer review conferences and their subsequent publication; c) opinions and experiences of clinical trialists and systematic reviewers regarding dissemination bias. Results: Only 54% of authors in our first study satisfied international authorship criteria. We found differences in number and byline of authors in 44% cases in our second study, as well as 39% of non-publication between research presentation and full publication. In our third study, we confirmed the non-publication rate (36%), and showed that authors regard current publication practices as the main culprits for today’s state of science. Conclusions: Our results indicate that current authorship criteria and publication practices in biomedicine need to be revised and restructured if science is to preserve its integrity.