Background: Asymptomatic bacteriuria is often treated inappropriately in hospitalized patients, leading to unnecessary antibiotic exposure. Reflexive urine cultures (Ucx), sent only in response to a positive urinalysis (UA) (Figure 1), were instituted at the University of Chicago Medicine in August 2013 to attempt to decrease the rates of contaminated and unnecessary Ucx. We evaluated the effects of implementation of a reflexive Ucx on antibiotic usage and Ucx attainment.
Methods: We evaluated encounters with a Ucx sent (February 2013-July 2013), prior to the implementation of the reflexive Ucx policy. Post-intervention, we evaluated encounters with a UA with reflex Ucx, or Ucx sent for a condition allowed as an exception to the policy (September 2013-February 2014). Antibiotic utilization was evaluated within 7 days of Ucx collection, calculated as days of antibiotic therapy (DOT)/1000 patient-days. Statistical analysis of Ucx result and DOT included 95% Confidence Intervals and Incidence Rate Ratio (IRR) where appropriate.
Results: Pre-intervention, 6448 Ucx were performed, with 729 positive Ucx [11.3%, 95%CI +/-0.77(0.53-12.07)]. Post-intervention, 6010 reflexive UA (+/- a subsequent Ucx), or Ucx exception were performed, with 945 positive Ucx [15.72%, 95%CI +/-0.92(14.8-16.64)]. Of reflexive UA/Ucx sent, 1208 [33%, 95%CI +/-1.19(31.81%-34.19%)] did not have an accompanying Ucx performed based on the policy. Midstream-collected specimens represented 4264 (66.13%) and 4084 (67.95%) of the total pre- and post-intervention Ucx respectively, representing a similar proportion of samples in both time periods (p=0.51). Subgroup analysis for specimens collected midstream demonstrated a decrease in antibiotic therapy between the pre- and post-intervention cohorts (107.1 DOT/1000 patient-days vs 89.4 DOT/1000 patient-days, IRR 0.834, [95%CI (0.804-0.865), p<0.0005]).
Conclusion: The implementation of reflexive Ucx based on presence of pyuria +/- leukocyte esterase on UA was associated with decreased antibiotic DOT in the midstream collection group. These results do not take into account patient symptoms at time of collection, so we cannot comment on appropriateness of therapy or concurrent infections.
N. Pettit, None
A. Charnot-Katsikas, None
V. Tesic, None
K. G. Beavis, None
J. M. Pisano, None