Do the Revised NHSN Surveillance Criteria Really Identify Catheter Associated Urinary Tract Infections?

Background:

Heightened attention has been placed on the detection and prevention of catheter associated urinary tract infections (CAUTI). NHSN surveillance criteria were recently refined and place emphasis on symptoms. Manual chart review to perform CAUTI surveillance is labor intensive for infection preventionists.

Objective:

To assess the performance of the revised NHSN surveillance criteria in detecting CAUTI.

Methods:

Patients with positive urine cultures obtained at least 48 hrs after admission were identified from the institutional data repository of the University of Maryland Medical Center. An automated computer based algorithm was developed to capture cases meeting indwelling urinary catheter and lab criteria based on the revised 2009 NHSN CAUTI surveillance definition. Manual chart review was used to assess UTI symptoms/signs required by NHSN. In the absence of a gold standard for CAUTI diagnosis, we also obtained data on clinical UTI diagnosis (defined as diagnosis and treatment for UTI by a patient’s clinician within 72 h of index urine culture), and presence of a non-UTI infection (defined as infection at a site other than the urinary tract within 72 h of the index urine culture). The agreement of surveillance cases of CAUTI (NHSN-CAUTI) with clinical UTI diagnosis, and with the presence of a non-UTI infection was assessed using Cohen’s Kappa statistic. Multivariable logistic regression was used to evaluate predictors of NHSN-CAUTI and clinical UTI diagnoses.

Results:

Of 642 patients with positive urine cultures between August 1 and September 30, 2010, 130 cases (20%) meeting NHSN urinary catheter and lab criteria were reviewed for CAUTI. Of these, 85 (65%) had fever and met surveillance definition for NHSN. Other symptoms/signs of UTI were noted in only 2 cases. There was only 46% agreement between NHSN-CAUTI and clinical UTI diagnoses [kappa: -0.10, 95% confidence interval (CI): -0.26-0.07]. In contrast, the agreement between NHSN-CAUTI and non-UTI infection diagnoses was higher (61%), [kappa: 0.23, 95% CI: 0.08-0.38]. In multivariable analysis, presence of a non-UTI infection [odds ratio (OR): 3.8, 95% CI 1.5-9.2] and organism other than yeast [OR (95% CI): 2.8 (1.1-7.1)] were independent predictors of NHSN-CAUTI. Pyuria was not significantly associated with NHSN-CAUTI. Statistically significant predictors of clinical UTI diagnoses were absence of a non-UTI infection, pyuria, organism other than yeast, and colony count > 10⁵ cfu/ml.

Conclusions:

Despite recent refinement of the NHSN surveillance definitions, we observed poor agreement between NHSN-CAUTI and clinical UTI diagnoses. In the absence of a gold standard, capture of non-UTI infections and lack of association with pyuria indicate non-specificity of NHSN criteria for CAUTI diagnoses. Finally, in our study, chart review for symptoms other than fever was resource-intensive but did not add any value to CAUTI surveillance.