Aims and objectives. To examine whether the relationship between specialty and patient safety is mediated by safety culture.
Background. Research has shown that patient safety in hospitals varies by specialty. The safety culture among healthcare professionals in hospital units is believed to influence safety performance. If there is a mediation effect of safety culture in the relationship between specialty and safety, then safety culture could explain why units vary in performance.
Design. Cross-sectional observational study in 28 units of 20 hospitals in the Netherlands. Units were of three specialties: emergency medicine, surgery and internal medicine.
Methods. Safety culture was measured with the Dutch version of the Hospital Survey on Patient Safety Culture with 11 culture dimensions (n = 542; response 56%). Safety outcomes were types of unintended events (six types). Unintended events were collected through staff reporting (n = 1885 events). Data were examined using multilevel regression analysis.
Results. The overall safety culture in the units did not mediate the relationship between specialty and the safety outcomes (event types), but three of the 11 dimensions showed significant mediation on one or more event types: non-punitive response to error, hospital management support and willingness to report.
Conclusions. Only a few safety culture dimensions mediated the relationship between specialty and some of the outcomes, with ‘willingness to report’ as the most important mediating factor. Our study did not give strong evidence that specialties differ in performance because of their safety cultures. More research into the causes of variation in patient safety between hospital units is needed.
Relevance to clinical practice. Our study could not give support for the claim that safety culture is a key factor affecting patient safety.