Objectives: To evaluate the performance characteristics of urinalysis and urine microscopy parameters for predicting urine culture results and to implement a reflex urine culture program.
Methods: We reviewed the charts of all patients presenting to our clinic January-March 2013 and June-August 2014, excluding those who were catheter-dependent or with urinary diversions. We assessed the association of urinalysis and urine microscopy parameters on urine culture outcomes defining a positive urinalysis as nitrite-positive and/or the presence of ≥5 white blood cells per high-powered field with bacteria and a positive urine culture as ≥10 000 colony-forming units/mL excluding diphtheroids. We carried out logistic regression to assess for predictors of positive urine culture to inform implementation of a reflex urine culture program.
Results: A total of 2764 patients were evaluated. Logistic regression using urinalysis variables identified positive nitrites (odds ratio 18.6, P < 0.001) and large leukocyte esterase (odds ratio 41.8, P < 0.001) as the strongest predictors of positive urine culture. Logistic regression using urine microscopy variables identified >50 white blood cells per high-powered field (odds ratio 13.6, P < 0.001) and moderate/many bacteria (odds ratio 16.8, P < 0.001) as the strongest predictors of positive urine culture. We used our positive urinalysis definition to implement the reflex urine culture program and noted a 60% reduction in urine culture rates over the first 3 months of implementation.
Conclusions: A urine positive for nitrites and/or ≥50 white blood cells per high powered field with bacteria seems to have a strong association with a positive urine culture and the best negative predictive value. A reflex urine culture program is an effective strategy to decrease the rates of unnecessary urine culture and their associated costs.
Keywords: quality control; quality improvement; urinary tract infection; urine; urine assay.
© 2018 The Japanese Urological Association.