P-2033. Documentation of Candidemia Risk Factors for T2 Ordering: A Diagnostic Stewardship Intervention

TL;DR

A new system requiring doctors to select risk factors before ordering a Candidemia test reduced test use by 47.5%, saving money and improving test accuracy.

Contribution

A clinical decision support system mandating risk factor selection before test ordering significantly reduced T2 Candida test use and improved diagnostic stewardship.

Findings

The CDSS reduced T2 Candida test orders by 47.5%, saving an estimated $540k annually.

Test positivity rates increased slightly, and testing decreased most in lower-risk floor units.

Overrides were rare, and most positive tests had negative blood cultures, suggesting improved diagnostic accuracy.

Abstract

Candidemia is a highly morbid condition that prompt initiation of antifungal therapy can decrease by as much as 50%; however, there is low incidence in patients without specific underlying risk factors. The T2Candida diagnostic measures Candida-specific DNA in whole blood, providing results in ∼4 hours, but cannot differentiate between viable (living) Candida and circulating DNA, which can persist for 14 days. We assess the utility of a clinical decision support system (CDSS) comprised of a popup requiring selection of risk factors before allowing a test to be ordered.

A prior informational CDSS was modified to require identification of multiple risk factors for Candidemia, otherwise it canceled the order. A deliberate override was included, but made to be challenging enough to dissuade most users. A retrospective analysis of all inpatient T2 tests ±6 months from CDSS implementation was conducted (n=3,842). Blood culture positivity and micafungin usage were analyzed near T2 order. Blood culture positivity was used as a diagnostic gold-standard at ±12h (false negative) and ±7d (false positive).Univariate and multivariate analyses were performed.

The new CDSS launched in August 2024, resulting in a rapid 47.5% reduction in T2 orders, equivalent to $540k annual savings (Fig. 1). The overall test positivity rate was increased, but not significantly (5.6% vs. 4.5%). The distribution of tests ordered among inpatient unit types (ICU vs. Floor vs. Post-Acute) changed significantly with the highest reduction in testing from Floor units (Fig. 2, p< .0001).103 popups were overridden, yielding 88 completed tests. 5 of these were positive; 100% of which had ≥ 2 sets of negative blood cultures drawn ±12h. 15-20% of patients received empiric micafungin at time of T2 with a higher proportion of appropriately discontinued therapy in the post-CDSS cohort (Fig. 3, p =.018). Differences in identified species between positive T2 and blood cultures are shown in Fig. 4.

A CDSS that requires correct identification of underlying risk factors decreased the number of ordered tests by almost 50%. A significant decrease in tests ordered for patients on the floor compared to those in an ICU (a risk factor addressed in the CDSS) was observed, indicating test reduction in a lower risk population.

Todd P. McCarty, MD, Basilea: Grant/Research Support|Cidara: Grant/Research Support|F2G: Grant/Research Support|Mundipharma: Grant/Research Support|Pfizer: Advisor/Consultant|Scynexis: Grant/Research Support Peter G. Pappas, MD, Astellas: Grant/Research Support|Basilea: Advisor/Consultant|Basilea: Grant/Research Support|F2G: Advisor/Consultant|Gilead: Grant/Research Support|Melinta: Advisor/Consultant

Figures