In-hospital cardiac arrest (IHCA) is a significant public health concern, afflicting an estimated 370,000- 750,000 patients annually, with survival rates generally below 20%. Over half of these patients are known to display signs of clinical deterioration in the hours leading up to the arrest. Rapid Response Systems (RRSs), designed to respond to patients in the early stages of clinical deterioration, have been surprisingly underwhelming with regards to preventing IHCA and death, leading some policy makers and researchers to suggest failures to identify the signs of early clinical deterioration or to call for help as possible etiologies. One possible solution to this problem is the development of a risk prediction tool that could be used to accurately stratify patients based on their likelihood of impending IHCA or ICU transfer, allowing interventions to be targeted at high risk patients. Several physiology-based scoring systems, which assign point values to abnormal vital signs, have been proposed but their mediocre predictive ability and cumbersome nature have limited their adoption. We have developed a simple, single question, quantitative scale of clinical judgment regarding patient stability that predicts IHCA or ICU transfer within the next 24 hours. We propose to validate that tool in a larger sample of patients and compare it to two physiology-based prediction algorithms, in an attempt to find the most sensitive and specific predictor of impending clinical deterioration. We will then use the best of the three, or a combined measure if better, in order to identify high-risk non-ICU inpatients and target them for a RRS intervention that bypasses the need to identify deteriorating patients and call for help, thereby allowing a targeted assessment of the RRS in high risk patients. RELEVANCE (See instructions): Some cardiac arrests in the hospital may be preventable if the clinical warning signs can be identified and acted upon quickly. Since it is not practical to monitor every hospitalized patient at all times, strategies to determine which patients are at high risk would allow additional resources to be targeted specifically at those patients. (End of Abstract)

K23HL097157

2009-09-01

2014-06-30