Using Tracking Data to Find Complications That Physicians Miss: The Case of Renal Failure in Cardiac Surgery

Health Data Research, Inc (HDR) develops, manages, and maintains clinical registries from physicians and hospitals, including the Merged Cardiac Registry. Quarterly reports indicate data that are inconsistent, out of range, or outside the norms found in other medical centers. In reports on cardiac surgery patients, HDR noted that for the 1992–1996 period, 3 of the 30 contributing centers experienced a significant increase in the incidence of moderate and severe renal failure. One of these three contributors gave HDR access to its detailed clinical database, and HDR ruled out most of the suspected causes for this increase in renal failure. A risk model for renal failure identified 20% of the patients to be at high risk. HDR then isolated a fast-track protocol as the culprit. One of the 30 contributing centers found that the protocol was associated with significant decreases in the intensive care unit (ICU) and hospital lengths of stay. However, as the severity of renal failure increased, charges, average length of stay, transfusions, and ICU times increased. At one of the three sites, after protocol changes were instituted in mid-1995 for the high-risk patients, the rate of renal failure reverted to below baseline levels. Analysis of run charts led to protocol changes for patients at high risk of renal failure, while retaining the positive outcomes associated with rapid extubation and shorter ICU stays for the remaining 80% of the patients. In the case study described here, the increase in renal failure was not obvious to the clinicians. Only the ability to chart the incidence of complications and look at trends allowed them to become aware of the problem and adjust the protocols.