Purpose - Hospitals within the United States consistently have injury rates that are over twice the national employee injury rate. Hospital safety studies typically investigate care providers rather than support service employees. Compounding the lack of evidence for this understudied population is the scant evidence that is available to examine the relationship of support service employees' perceptions of safety and workrelated injuries. To examine this phenomenon, the purpose of this study was to investigate support service employees' perceptions of safety leadership and social support as well as the relationship of safety perception to levels of reported injuries. Design/methodology/approach - A nonexperimental survey was conducted with the data collected from hospital support service employees (n=1,272) and examined: (1) relationships between safety leadership (supervisor and organization) and individual and unit safety perceptions; (2) the moderating effect of social support (supervisor and coworker) on individual and unit safety perceptions; and (3) the relationship of safety perception to reported injury rates. The survey items in this study were based on the items from the AHRQ Patient Safety Culture Survey and the U.S. National Health Care Surveys. Findings - Safety leadership (supervisor and organization) was found to be positively related to individual safety perceptions and unit safety grade as was supervisor and coworker support. Coworker support was found to positively moderate the following relationships: supervisor safety leadership and safety perceptions, supervisor safety leadership and unit safety grade, and senior management safety leadership and safety perceptions. Positive employee safety perceptions were found to have a significant relationship with lower reported injury rates. Value/originality - These findings suggest that safety leadership from supervisors and senior management as well as coworker support has positive implications for support service employees' perceptions of safety, which, in turn, are negatively related to lower odds of reporting injuries. Copyright © 2013 by Emerald Group Publishing Limited All rights of reproduction in any form reserved.