Syndromic Antibiograms and Nursing Home Clinicians' Antibiotic Choices for Urinary Tract Infections

Abstract

Importance: Empirical antibiotic prescribing in nursing homes (NHs) is often suboptimal. The potential for antibiograms to improve empirical antibiotic decision-making in NHs remains poorly understood.

Objective: To determine whether providing NH clinicians with a urinary antibiogram improves empirical antibiotic treatment of urinary tract infections (UTIs).

Design, setting, and participants: This was a survey study using clinical vignettes. Participants were recruited via convenience sampling of professional organization listservs of NH clinicians practicing in the US from December 2021 through April 2022. Data were analyzed from July 2022 to June 2023.

Interventions: Respondents were randomized to complete vignettes using a traditional antibiogram (TA), a weighted-incidence syndromic combination antibiogram (WISCA), or no tool. Participants randomized to antibiogram groups were asked to use the antibiogram to empirically prescribe an antibiotic. Participants randomized to the no tool group functioned as controls.

Main outcomes and measures: Empirical antibiotic selections were characterized as microbiologically (1) active and (2) optimal according to route of administration and spectrum of activity.

Results: Of 317 responses, 298 (95%) were included in the analysis. Duplicate responses (15 participants), location outside the US (2 participants), and uninterpretable responses (2 participants) were excluded. Most respondents were physicians (217 respondents [73%]) and had over 10 years of NH practice experience (155 respondents [52%]). A mixed-effects logistic model found that use of the TA (odds ratio [OR], 1.41; 95% CI, 1.19-1.68; P < .001) and WISCA (OR, 1.54; 95% CI, 1.30-1.84; P < .001) were statistically superior to no tool when choosing an active empirical antibiotic. A similarly constructed model found that use of the TA (OR, 1.94; 95% CI, 1.42-2.66; P < .001) and WISCA (OR, 1.7; 95% CI, 1.24-2.33; P = .003) were statistically superior to no tool when selecting an optimal empirical antibiotic. Although there were differences between tools within specific vignettes, when compared across all vignettes, the TA and WISCA performed similarly for active (OR, 1.09; 95% CI, 0.92-1.30; P = .59) and optimal (OR, 0.87; 95% CI, 0.64-1.20; P = .69) antibiotics.

Conclusions and relevance: Providing NH clinicians with a urinary antibiogram was associated with selection of active and optimal antibiotics when empirically treating UTIs under simulated conditions. Although the antibiogram format was not associated with decision-making in aggregate, context-specific effects may have been present, supporting further study of syndromic antibiograms in clinical practice.

Figure 1.. Traditional Antibiogram and Weighted-Incidence Syndromic Antibiogram Used in the Survey

The same population of 45 g-negative urinary isolates were used to create both the traditional antibiogram and the weighted-incidence syndromic antibiogram. Light blue shading indicates results based on fewer than 30 isolates, which should be interpreted with caution as they are less reliable. R indicates that the organism is intrinsically resistant. Error bars indicate 95% CI. TMP/SMX indicates trimethoprim/sulfamethoxazole. ^aNitrofurantoin had 100% sensitivity for E coli but is not active against other listed species.

Figure 2.. Process Mapping of Methodology Used to Create Bug-Drug Matrices for Active and Optimal Empirical Antibiotic Choice

Figure 3.. Estimated Mean Probabilities of Active and Optimal Therapy for Each Case by Group

Pairwise differences among groups were tested within each outcome and case with Tukey adjustment for the multiple comparisons in each. Error bars indicate 95% CI. NT indicates no tool; TA, traditional antibiogram; WISCA, weighted-incidence syndromic combination antibiogram.