Implementing a Clostridium difficile testing algorithm and its effect on isolation duration and treatment initiation: a pre- and post-implementation study

Abstract

A proportion of patients suspected of Clostridium difficile infection are unnecessarily placed in contact isolation. By introducing a random-access glutamate dehydrogenase (GDH) test for C. difficile, we aimed to reduce isolation time. In addition, we investigated whether the result of the toxin A&B enzyme immunoassay (EIA) was associated with the decision to initiate antibiotic treatment against C. difficile. This retrospective pre- and post-implementation study was from June 3, 2016, to June 4, 2018. Pre-implementation, only a NAAT was performed. In the post-implementation period, a GDH test was performed; if positive, a toxin A&B EIA followed the same day and subsequently a NAAT. Contact isolation for CDI was discontinued when the GDH test was negative. Median time in isolation was 50.8 h pre-implementation (n = 189) versus 28.0 h post-implementation (n = 119), p < 0.001. The GDH test had a negative predictive value of 98.8% (95% CI 97.9–99.4). In 7/31 (22.6%) patients with a positive NAAT and GDH test and a negative toxin A&B EIA, no antibiotics against C. difficile were initiated versus 4/28 (14.3%) patients who were NAAT, GDH and toxin A&B EIA positive. Introducing a random-access screening test resulted in a significant decrease in patient isolation time. The GDH test had a high negative predictive value making it suitable to determine whether contact isolation can be discontinued. Furthermore, the result of a toxin A&B EIA had limited added value on the percentage of patients in whom antibiotic treatment against C. difficile was initiated.