108
Abstracts
NOSOCOMIAL BACIEREMIA IN PATIENTS IN A PEDIATRIC INTENSIVE CARE UNIT IS RELATED TO THE DURATION OF CENTRAL VENOUS CATHETER PLACEMENT. D. Jacobsen.* MPH, D. Uden, Phatm D. C. Leach, MLT. N. Payne, MD. Minneapolis Children’s Medical Center. Minneapolis. MN. The purpose of this study was to identify predisposing factors for and rates of bactcremia associated with central venous catheter (CVC) placement in patients receiving care in a pediatric intensive care unit (PICIJ). This study pmspectively monitored 448 central venous catheters placed in 436 PICU patients over a 12month period. During the study period, 1070 patients were admitted to the PICU which serves as both a surgical and a medical intensive care unit for children. There were 15 nosocomiai bacteremias (1.4% of admissions) and 13 (87%) were associated with CVC. The incidence of CVC-related bacteremias was 13/448 catheters (2.9%). or 5.4 infections per loo0 catheter days. The only factor that was significantly associated with CVC-related bactenemia was duration of CVC placement. Catheters that became infected had been in place longer (mean?SD=9.31-3.8 days) than those that did not become infected (5.3t3.8 days, P=O.o004, t-test). The relative risk for nosocomial bacteremia in patients with a CVC left in place for seven or more days was 7.5 times that for lines left in place for less than seven days. Infected patients also tended to be younger than uninfected patients (age= 1.1 + 1.0 vs. 3.3 +4.2 years) but this difference was not signiiicant (P=O.lO, MannWhitney). Other factors investigated and found not to be related to bactetemia in this study were placement of the CVC in the operating room vs. PICU. type of line. type of dressing applied to CVC site, frcquency of dressing changes, and she of line placement. We conclude that the duration of CVC placement contributes significantly to the risk of nosocomial bactercmia in PICU patients.
PERIPHERALLY INSERTED CENTRAL CATHBTERS (PICC): EXPERIENCE IN A COMMUNITY HOSPITAL (CH). T. Barre”. Alta Bates Herrick Hospital. Berkeley, CA. PURPOSE:
To describe the complication 575.bed CH
rate for PICC lines used for adult> in a
METHODS:
In the Spring of 1989, a small group of IV RNs were trained to insesen and maintain a through-the-needle percutaneously inserted silicone 20 ga catheter (Ven-A-Cath). Any patient needing vascular access for >lO days was eligible for PICC line placement Fhtids including hypcral and medications were give” through the lines but the lines were not used for blood drawing. All lines were placed by RNs who bad cot”#ted a qualification and competency pmgram. The Infection Control (IC) Department was given a copy of the insertion sheet, and basic demographic data were collected on each patient. PICC line compiicadons were rcportcd to the IC RN by the IV RN, MD, or nursing unit. IV site infections were detected by existing IC surveillance. phlebitis was a clinical diagnosis made by the attending MD. infection was established by written standardized criteria. RESULTS:
Complication rates for the first 100 lines placed arc depicted below.
CONCLUSIONS:
AN INFECITON CONTROL-QUALITY IMPROVEMENT (Ql) APPROACH TO NOSOCOMIAL BACTEREMIA IN NEONATES. L. Karanfil.* RN, MA, CIC. A. Iosephso”. RN, MPH, CIC. PhD. H. Alonzo, RN, MA. SUNY-Health Science Center at Brooklyn. Brooklyn. NY. University Hospital of Brooklyn is a 376&d teniary care hospital with a high risk nutscry (HRN). As a participant in NNIS, our Infection Control Program has been utilizing a separate surveillance component for the past two years to examine HRN nosocomial infection data. In addition to documenting all nosocomial infections in neonates this process involved the daily collection of specific denominator data on catheter and ventilator use by birthweight category. This has allowed us to develop infection rates by catheter days and to stratify these rates by birthweight category. The birthweights were broken down into three groups: (A) 51500 gnr. (B) lSOI-25CiO gtn. (C) 12500 gm. The catheters involved in the assessment were central and umbilical catheters. We used comparative data fmm the NNIS program to evaluate the appropriateness of our rates. After one year of patticipation, it became apparent that our birthweight specific rates of IVassociated bactcremia were higher than those of other NNIS hospitals. This information was communicated to the HRN nursing and medical leadership. In 1990 the data were Fe-examined and while there was a” overall 25% rate decrease, we were still a” outher according to NNIS. In September of 1990 a QI project was begun u) achieve a” additional rate reduction. A task force was fotmed. IV care practices were changed and goals were set. A chart indicating the cttrrent birthweight specific rates and our goal rates was placed on the unit and periodically updated to follow our progress. On close examination of the data the task force found that 71% of ail infected nwnates fell into the <1500 gm category. Sixty-nine percent of those in the s15tM gm group wctc below ICOO gm with 15% below 750 gm. The task force decided that we should further stratify the rates below 15CB gms in,order to clarify the extent to which infection variation in the 51500 grit group might be due to birthweight differences within that group. Preliminary data for the latler pan of 1990 showed all bactetemias in the s1500 gm gmup to be in those 91Mx) grants. The
I) The complication rate for in-patients with PICC lines was 17%. 2.) Complications rates were higher for HIV-infected patients than “on HIV infected patients. 3) These rates should be compared wifh device-specific complication rates for other vascular access devices when selecting the device of choice. 4) Considering the relativeIy high rate of complication rates with PICC lines, 0nIy a few well qualified RNs should insert the lines so careful monitoring can be awtred.