Date of Award


Document Type

Open Access Dissertation


College of Nursing

First Advisor

Rita Snyder


The ability to successfully rescue patients from complications has been shown to be a good measure of quality care processes in hospitals. Failure-to-rescue (FTR) has been defined using secondary International Classification of Disease (ICD-9 CM) codes. Studies of FTR using these codes have demonstrated satisfactory accuracy when compared to clinical events documented in the medical record. However, a subset of the original codes for FTR, thought to be sensitive to nursing care, have failed to show the same level of accuracy. This study examined the possibility of using clinica lpredictors to identify failure-to-rescue.

Secondary analysis of a previous dataset was used to establish and improve diagnostic accuracy of FTR using ICD-9 CM codes as compared to the gold standard of record review. These ICD-9 CM codes performed poorly in terms of diagnostic accuracy (sensitivity and specificity) when compared to record review. A variety of clinical predictors were then tested for accuracy in the measurement of FTR compared to record review. Transfer to a higher level of care in combination with a variety of clinical predictors as well as complications following a procedure demonstrated strong sensitivity and fair specificity. Combining these clinical predictors with secondary ICD-9 CM codes did not enhance diagnostic accuracy. While specificity for clinical predictors was not robust, high levels of sensitivity for certain predictors warrants an increased level of surveillance for patients who exhibit these signs and symptoms.


© 2013, Amelia Marie Joseph

Included in

Nursing Commons