191. Long-term Utility of Clinical Nudges within Clostridioides difficile Diagnostic Stewardship

TL;DR

This study shows that using computer alerts to remind doctors about proper testing for a specific infection significantly reduced unnecessary tests and improved diagnostic accuracy over time.

Contribution

The study demonstrates the long-term effectiveness of clinical decision support nudges in reducing inappropriate Clostridioides difficile testing and improving diagnostic stewardship.

Findings

An informational alert decreased CDI testing and lowered test positivity rates in patients at low risk for infection.

Overrides of alerts were most common in ICU settings and among certain provider types.

A small group of providers was responsible for a large proportion of alert triggers and overrides.

Abstract

IDSA recommends that, prior to testing for Clostridiodes difficile in patients with new-onset diarrhea, clinicians should consider other potential causes, such as recent laxative use. Clinical decision support systems (CDSS) can nudge providers to reduce inappropriate CDI testing and 2-step testing algorithms can optimize diagnostic accuracy. We assessed the long-term clinical utility of clinical nudges for CDI diagnostic stewardship.

A retrospective analysis of inpatient CDI tests (n=44,677) ordered for 10 years after the introduction of 2-step diagnostic testing (toxin→DNA-NAAT) in the 2nd quarter of 2015 (Q2 2015) to evaluate the long-term impact of a CDSS nudge for providers to reconsider testing in patients receiving laxatives in the 48 hours prior. Univariate and multivariate analyses were performed via Pearson’s chi-squared test.

An informational popup (Q3 2016) produced a marked decrease in CDI testing (Fig. 1). It was transitioned to an interactive popup prompting either cancelation or acknowledgment that a “positive result will be of no value” (Q2 2017). The test positivity rate was significantly lower (7.0% vs. 8.5%, P < .001) for inpatients who triggered the popup (n=9231) versus those who did not during the same period.

No significant change in overall propensity to override the CDSS popup was observed over the study period (µ=57%). Popups were most frequently overridden in ICUs (62.0%), compared with Floor and Post-Acute Units, at 56.1 and 58.0%, respectively (p< .0001). There was a significant difference in rate of override between ordering provider types (APP vs. Attending vs. Resident) with rates of 59.4%, 57.2% and 55.5%, respectively (p< .01). In looking at individual providers, 2.5% of providers (n=50) were responsible for 20% of popup triggers, with a negative correlation between frequency of override and ordered test positivity (Fig. 2).

While it has been shown that CDSS nudges can decrease diagnostic ordering, here we show that implementation of a CDSS yielded a statistically significant decrease in test positivity for a patient population at low-risk for CDI. Subgroup analyses looking at provider, unit, and level of care provide opportunities for targeted educational intervention to further improve diagnostic stewardship.

All Authors: No reported disclosures

Figures