BACKGROUND: Open and direct communication (\"speaking-up\") about errors, bypassed safety rules and risky behaviours among hospital staff is required to avoid patient harm, and it is an essential characteristic of an established safety culture. In German-speaking countries, little is known about speaking-up behaviour among health care professionals (HCPs) in hospitals.
METHODS: Safety concerns and speaking-up behaviours among HCPs of nine oncological units of eight hospitals were assessed using a self-administered survey. A vignette was embedded to assess hypothetical speaking-up and its predictors. The association of hierarchical position and speaking-up was investigated. 1,013 physicians and nurses completed the survey (65 % response rate).
RESULTS: 53 % of the HCPs reported having concerns about patient safety at their unit, \"sometimes\", \"frequently\", or \"very frequently\". Colleagues bypassing important safety rules at least \"sometimes\" were reported by 30 %. A considerable fraction of responders reported episodes of withholding of voice. Nearly 20 % said they did not communicate safety problems at their unit at least sometimes. 73 % of higher-ranking staff and 60 % among those at lower ranks said they had never withheld information which could have reduced threats to patients (OR=1.8, p≤0.001). Many responders felt that speaking-up is often difficult and challenging. 32 % responded that they would not speak-up about a missed hand disinfection towards a colleague assessing the wound of a recently operated oncological patient.
CONCLUSIONS: HCPs in hospital frequently experience safety concerns and often withhold them. An important resource for better patient safety is lost. The development of interventions to improve speaking-up culture is warranted.

Critical incident reporting and learning systems (CIRS) have been recommended as an instrument to promote patient safety for a long time. However, both their scientific value and their actual impact have been disputed. The nationwide German CIRS for primary care has been in operation since September 2004. Incident reports are available online, and the question is how to make use of this large database to promote patient safety. A descriptive analysis of the content was performed, classifying, in particular, types of error and contributing factors. Its usage is presented for the period from 2004 to 2013 where a total of 483 complete reports have been recorded. Their severity ranges from 35.6 % with no tangible harm to patients to 14.6 % with important harm (or errors contributing to mortality). The majority of them (74.2 %) were process errors, compared to 25.8 % knowledge/skills errors. The main areas involved were treatment/medication (54.2 %) and diagnosis/tests (16.4 %). The results of the analysis of the CIRS cannot be used as an epidemiological data source. And yet they will generate hypotheses for further research in the field of patient safety. Moreover, they will enable practice teams to make themselves familiar with and learn from critical incident analysis. In spite of the specific difficulties in ambulatory care, CIRS should be promoted in this sector to enable learning. Participation in CIRS can be increased by enhanced feedback.