A Conceptual Framework to Reduce Inpatient Preventable Deaths.

Background: Efforts to reduce preventable deaths in the in-hospital setting should target both cardiopulmonary arrest (CPA) prevention and optimal resuscitation. This requires consideration of a broad range of clinical issues and processes. A comprehensive, integrated system of care (SOC) that links data collection with a modular education program to reduce preventable deaths has not been defined. Methods: This study was conducted in two urban university hospitals from 2005 to 2009. The Advanced Resuscitation Training (ART) program was implemented in 2007, incorporating hands-on resuscitative skills and in-hospital-specific training with an institutional resuscitation database. Linkage between the database and training modules occurs via the ART Matrix, which classifies all CPA events into the following etiologies: sepsis, hemorrhage, pulmonary embolus, heart failure, tachyarrhythmias, bradyarrhythmias, acute respiratory distress syndrome, non-intubated pulmonary disease, obstructive apnea, traumatic brain injury, ischemic brain injury, and intracranial mass lesions. This taxonomy was validated using descriptive statistics, before- and-after analysis evaluating CPA incidence, and multivariate logistic regression to predict CPA survival. Results: A total of 336 inpatients suffered a cardiopulmonary arrest during the study period-187 in the pre-ART period and 149 in the post-ART period. The vast majority of CPA events were categorized using the ART Matrix with high inter-observer reliability. As anticipated, changes in CPA incidence and survival were observed for some Matrix categories but not others following ART implementation. In addition, multivariate logistic regression revealed strong independent associations between taxonomy classifications and outcome. Conclusion: A novel SOC using a unique taxonomy for arrest classification appears to be effective at reducing inpatient CPA incidence and outcome.