The cultural divide between scientists and clinicians has been described as undermining the advance of medical science, by hindering the production of practice-relevant research and of research-informed clinical decisions. Here, I consider the field of post-marketing risk assessment of drugs as an example of strict interdependence between basic biomedical research, clinical research, and clinical evaluation and show how it would benefit from a closer collaboration between scientists and clinicians. The risk assessment of drugs after their marketing relies on spontaneous adverse effect reports to drug agencies and on peer-reviewed case reports. I emphasize the importance of qualitative analysis of such reports for the improvement of mechanistic understanding of harmful effects of drugs. I argue that mechanistic explanations of drug effects are at least as important as determination of their frequency, in order to establish causation. An ideal risk assessment, then, verifies not only the frequency of undesired effects but also why and how the harm happens. For this purpose, the frequency or novelty of the unintended outcome, although contextually indicative, should not determine the epistemic value of a report.Details about the context that generated an unexpected outcome, instead, can offer the chance of improving causal understanding about how the intervention works. This is illustrated through examples from medical research. Mechanistic understanding is a domain of joint collaboration among (1) clinicians, in charge of detailed, qualitative reporting of patient stories about side effects, (2) qualitative clinical researchers, in charge of analyzing clinical contexts or harmful effects and formulating explanatory hypotheses, and (3) basic biomedical researchers, in charge of verifying such hypotheses. In addition, direct information flow can on one side focus clinicians' attention on knowledge gaps about drugs/effects where more research is needed, while on the other side create a more contextualized concept of mechanism among scientists.