Accurate Urinary Catheter Surveillance Requires Individual Chart Review

Background: Non-payment from Medicare for catheter-associated urinary tract infections (CAUTI) has created interest in lowering rates of CAUTI.  One of the most effective ways to reduce CAUTI is to reduce urinary catheter use, but few hospitals are doing comprehensive surveillance for urinary catheter use.

Objective: We compared 3 methods of surveillance for urinary catheters: bedside visits, chart review of nursing notes in an electronic medical record, and reports from infection control.  We also determined the rates of use of various types of urinary catheters and the incidence of bacteriuria (> 10³ cfu/ml) associated with each.

Methods: We compared urinary catheter data obtained from chart review to catheter data obtained by bedside visits over a 5 month period on 10 inpatient wards at a tertiary care VA medical center.  We studied indwelling (transurethral) catheters, condom catheters, intermittent catheterization, and supra-pubic catheters.   Infection control records of indwelling catheters were obtained for 1 month on 4 of these wards for comparison.

Results: For 575 cases we obtained urinary catheter data by both chart review and bedside visit.  Of the 184 catheter-days detected by chart review, 9 were not present on bedside visit (4.9 %).  Chart review had a sensitivity of 100%, a specificity of 98% and an agreement of 96% (kappa coefficient) with bedside assessment of urinary catheters.  In the month of July 2010, 7742 patient bed-days were assessed by chart review.  Catheters were used on 3264 (42%) of these bed-days. Of the catheter-days, 44% were indwelling catheters and 54% were external catheters. The rate of positive urine cultures was 19/1000 catheter-days.  Of the positive urine cultures from catheterized patients, 57% were from indwelling and 39% were from condom catheters. When the indwelling catheter data by chart review was compared to the data reported by infection control, 117 of 1365 indwelling catheter days were missed by infection control (8.6%), while 55 of 2265 days without catheters were erroneously reported as with catheters (2.4%).  Infection control reports had a 91% sensitivity, a 98% specificity, and 89.8% agreement (kappa) with the indwelling catheter days detected by chart review.

Conclusions: Patient-level review  of nursing notes in the electronic medical record provided very accurate urinary catheter data at our institution.  Urinary catheter use was high and was associated with a high incidence of positive urine cultures.  Condom catheters contributed significantly to catheter-associated bacteriuria and should not be ignored in surveillance for CAUTI.   Infection control data had significant inaccuracies, probably because infection control personnel must rely on catheter reports from wards rather than on individualized review of patients or of charts.