An analysis of root cause identification and continuous quality improvement in public health H1N1 after-action reports

摘要

Objective: To identify the extent to which the Homeland Security Exercise and Evaluation Program's (HSEEP) After Action Report/Improvement Plan (AAR/IP) template was followed by public health entities and facilitated the identification of detailed corrective actions and continuous improvement. Design: Data were drawn from the US H1N1 Public Health Emergency Response (PHER) federal grant awardees (n = 62). After action report/improvement plan text was examined to identify the presence of AAR/IP HSEEP elements and characterized as "minimally complete," "partially complete," or "complete." Corrective actions (CA) and recommendations within the IP focusing on performance deficits were coded as specific, measurable, and time-bound, and whether they were associated with a problem that met root cause criteria and whether the CA/recommendation was intended to address or fix the root cause. Main outcome measures: A total of 2619 CA/recommendations were identified. More than half (n = 1480, 57%) addressed root causes. Corrective actions/recommendations associated with complete AARs more frequently addressed root cause (58% vs 51%, χ = 9.1, P < 0.003) and were more specific (34% vs 23%, χ = 32.3, P < 0.0001), measurable (30% vs 18%, χ = 37.9, P < 0.0001), and time-bound (38% vs 15%, χ = 115.5, P < 0.0001) than partially complete AARs. The same pattern was not observed with completeness of IPs. Corrective actions and recommendations were similarly specific and measurable. Recommendations significantly addressed root cause more than CAs. Conclusions: Our analysis indicates a possible lack of awardee distinction between CA and recommendations in AARs. As HSEEP adapts to align with the 2011 National Preparedness Goal and National Preparedness System, future HSEEP documents should emphasize the importance of root cause analysis as a required element within AAR documents and templates in the exercise and real incident environment, as well as the need for specific and measurable CAs.