Evaluation of an interdisciplinary re-isolation policy for patients with previous Clostridium difficile diarrhea

Diarrhea caused by Clostridium difficile is increasingly recognized as a nosocomial problem. The effectiveness and cost of a new program to decrease nosocomial spread by identifying patients scheduled for readmission who were previously positive for toxin was evaluated.

The Memorial Sloan-Kettering Cancer Center is a 410-bed comprehensive cancer center in New York City. Many patients are readmitted during their course of cancer therapy. In 1995 as a result of concern about the nosocomial spread of C difficile, we implemented a policy that all patients who were positive for C difficile toxin in the previous 6 months with no subsequent toxin-negative stool as an outpatient would be placed into contact isolation on readmission pending evaluation of stool specimens. Patients who were previously positive for C difficile toxin were identified to infection control and admitting office databases via computer. Admitting personnel contacted infection control with all readmissions to determine whether a private room was required.

Between July 1, 1995, and June 30, 1996, 47 patients who were previously positive for C difficile toxin were readmitted. Before their first scheduled readmission, the specimens for 15 (32%) of these patients were negative for C difficile toxin. They were subsequently cleared as outpatients and were readmitted without isolation. Workup of the remaining 32 patients revealed that the specimens for 7 patients were positive for C difficile toxin and 86 isolation days were used. An additional 25 patients used 107 isolation days and were either cleared after a negative specimen was obtained in-house or discharged without having an appropriate specimen sent. Four patients (9%) had reoccurring C difficile after having toxin-negative stools. We estimate (because outpatient specimens were not collected) the cost incurred at $48,500 annually, including the incremental cost of hospital isolation and equipment.

Our policy to control the spread of nosocomial C difficile required interdisciplinary cooperation between infection control and the admitting department. By identifying patients who were positive for toxin through admitting, we were able to place all potentially infected patients into isolation. Our positivity rate of 15% on readmission demonstrates the importance of this policy. The cost of controlling C difficile can be significantly lowered by clearing patients who were previously positive for toxin before hospital readmission.