Introduction: The sharing of biomedical research data is important for reasons of transparency and potential re-use in further research. Funding agencies and scientific publishers play a key role in mandating and enabling sharing and are also committed to increasing transparency and promoting reproducibility of data published in research literature. Despite the fact that all signs point to tightening requirements for sharing at the time of publication, research data management planning in Swiss academic institutions is slow in coming. This poster presents an overview of the current data sharing practices at the University hospital Lausanne (Switzerland). Methods: We extracted from the institutional repository of the Faculty of Biology and Medicine the list of original articles published in 2014 (around 2,000 publications). We then focused on 300 peer-reviewed publications which explicitly referred to data analysis (clinical trials, cohort study and experimental research). These research articles were published in around 200 journals. We investigated the current data policies of these journals. We also investigated if any access to the data could be found in the online versions of these articles. For additional input, we organised interviews with researchers about their commitment to making the datasets publicly available. Results: About 10% of the articles offered access to primary raw data online, essentially in journals which had specific guidelines concerning research data. Conclusion: Both publication practices and interviews prove that researchers are not strongly active in giving access to primary data. Libraries cannot promote new scientific practices alone but they can assist in raising awareness and in organising training.