Objectives Preparation of injectable anticancer drugs in hospital pharmacies, in particular of cytotoxics, is a high-risk activity. We used Preliminary Risk Analysis (PRA) to analyse the risks in the different steps of our anticancer drug circuit, including the preparation step (PRA1). Then, to prepare an important change in management of the circuit with the software Chimio® (pooling of three databases for subcontracting), we repeated the analysis of preparation step (PRA2). PRA is known to be time and resource consuming. To overcome this, we developed a strict organisational framework to perform the analysis within a reasonable amount of time. We present the PRA method including its practical implementation, and its application to the anticancer drug preparation process, before and after pooling of Chimio® databases. Methods PRA has two main stages, PRA “system” and PRA “scenario”. A multidisciplinary working group is created for the entire PRA process. PRA “system” is an exploratory and qualitative stage. PRA “scenario” requires the creation of risk assessment tools and decision tools before actually developing, analysing and treating scenarios, with risk reduction actions structured in an action plan. For PRA2 we used the same working group, assessment and decision tools as for PRA1 and we only analysed dangerous situations (DS) that appeared or changed towards more risk, requiring a new action plan. The different PRA only required four 2 h meetings thanks to the investment of a coordinator who is expert in the method. Results In PRA1, the riskiest phase was production while it was the verification and delivery of the finished product in PRA2. The risks were mainly related to management, human and technical dangers in PRA1. Human danger was found to be the main danger in PRA2, followed by organisational danger. Among the 264 scenarios described in PRA1, six of criticality 3 and 69 of criticality 2 have been associated with risk reduction actions. These actions mainly involved managing the risk of human error, with the control system Drugcam® and the standardisation of the pharmaceutical assistants’ training program. In PRA2, 11 scenarios were analysed, including three of criticality 3 and 4 of criticality 2 for which risk reduction measures were taken. Conclusions PRA allowed us to perform an in depth analysis of the highly specific and technical process of anticancer drug preparation. Human danger was one of the most important dangers identified, and it should always be taken into consideration, whatever the measures taken to prevent it. PRA2 was extremely useful to plan the organisation that would result from the new Chimio® database, while involving the team and winning its commitment. It allowed an exhaustive and structured anticipation of this major change. Practical aspects of PRA method implementation we have adopted facilitate its application and can help to deploy it on many areas in our hospitals. Indeed, besides an exhaustive analysis of the risks, this approach promotes collaboration, develops a quality culture and is an excellent tool for team and project management, as well as communication.