Which aspects of safety culture predict incident reporting behavior in neonatal intensive care units? A multilevel analysis

OBJECTIVES: Safety culture assessments are increasingly used to evaluate patient-safety programs. However, it is not clear which aspects of safety culture are most relevant in understanding incident reporting behavior, and ultimately improving patient safety. The objective of this study was to examine which aspects of safety culture predict incident reporting behavior in the neonatal intensive care unit (NICU), before and after implementation of a voluntary, nonpunitive incident reporting system. DESIGN: Survey study based on a translated, validated version of the Agency for Healthcare Research and Quality Hospital Survey on Patient Safety Culture. This survey incorporates two outcome measures, 11 dimensions of patient-safety culture as well as demographic data. SETTING: Eight tertiary care NICUs and one surgical pediatric ICU. SUBJECTS: All unit personnel. INTERVENTION: Implementation of a specialty-based, voluntary, nonpunitive incident reporting system. MEASUREMENTS AND MAIN RESULTS: The survey was conducted before (t = 0) and after (t = 1 yr) the intervention. Primary outcome: number of self-reported incidents in the past 12 months. Overall response rate was 80% (n = 700) at t = 0 and 76% (n = 670) at t = 1 yr. Based on a multivariate multilevel regression prediction model, the number of self-reported incidents increased after the intervention and was positively associated with a nonpunitive response to error and negatively associated with overall perceptions of safety and hospital management support for patient safety. CONCLUSIONS: A nonpunitive approach to error, hospital management support for patient safety, and overall perceptions of safety predict incident reporting behavior in the NICU. The relation between these aspects of safety culture and patient outcome requires further scrutiny and therefore remains an important issue to address in future research