Eliminating CLABSI, A National Patient Safety Imperative: Neonatal CLABSI Prevention

Participation

A total of nine States participated in the quality improvement initiative (CO, FL, HI, MA, MI, NC, NJ, SC, WI) with a total of 100 units (Figure 1), 98 percent of which were matched to their 2010 AHA annual survey.  The majority of participating units were located in general medical/surgical hospitals (96 percent) although some were located in children’s general medical/surgical facilities (4 percent).  All hospitals had 100 or more beds with the majority of participating facilities having 500 or more beds (47 percent).  Additionally, the majority of participating hospitals were teaching facilities (76 percent), had Level III obstetric units (78 percent), and were in metropolitan settings (70 percent).  On average, hospitals had 26 neonatal ICU beds (minimum, 0; maximum, 105) and 13 neonatal intermediate care beds (minimum, 0; maximum 68).

Figure 1. States and number of units participating in initiative*

* One unit is located in Georgia, a State which did not formally participate in the initiative.

Central Lines

Central lines were classified as: peripherally inserted central catheter (PICC), umbilical artery catheter (UAC), umbilical venous catheter (UVC), Broviac (a surgically placed, tunneled central venous catheter), or "other" (all other percutaneously placed central lines). To date, 17,212 central lines have been placed (Table 2) with 16,067 of these lines having both insertion and removal dates. Lines are more frequently missing removal dates later in the project than earlier in the project (of missing line removal dates, 29.4 percent occur in August compared to 3.6 percent in January), illustrating the "real time" nature of the database system.

Table 2. Central lines placed by type

Approximately 90 percent of central lines were new (8 percent placed outside the unit; 2 percent were repairs or replacements of an existing line).  Of the new lines placed, the procedures followed during placement can be found in Table 3.  Most frequently, skin was prepped using povidone iodine (41.9 percent) followed by chlorhexidine (39.9 percent).  Other skin preps such as alcohol (2.7 percent), sterile water (0.34 percent) or combinations of techniques (e.g. chlorhexidine and alcohol) were infrequently used accounting for approximately 15 percent of skin preparations.  No skin prep was noted for approximately 1 percent of patients.

Table 3. Procedures followed for 15,404 new lines placed

Infections

To date, 166 infections have been reported. Most patients with a positive infection also had a secondary culture noted as being obtained at the time of the initial culture (59.6 percent).  The types of catheters in use among patients with infections can be found in Table 4.  Patients with PICC lines had the greatest number of infections overall (50.6 percent); however, as a proportion of lines placed to date, infections were most frequently found among patients with Broviac (18/577 or 3.1 percent) followed by PICC lines (84/6,783 or 1.2 percent).

Table 4. Types of catheters in use among patients with infections*

*Classification of "Other" includes all other percutaneously placed central lines. Eight infections had no catheter type noted.

Maintenance

As of August 31, 2012, 127,578 line maintenance notes have been entered representing 8,427 unique patients. The majority of patients were noted as having feedings <120 cc/k/d (87.6 percent). During multidisciplinary rounding, lines are typically assessed as necessary (90.3 percent) with 2.8 percent considered unnecessary (for 6.7 percent of lines it was unknown if the line was necessary). When lines were accessed, procedures followed can be found in Table 5.  On shifts when tubing was changed (n=52,077), clean gloves were used without prompting 99.7 percent of the time.

Table 5. Procedures followed when accessing lines (n=102,029)

*When lines were accessed, cleaning was not noted in 0.8 percent of instances.

Team Check-Up Tool

The number of unit teams submitting Team Check-Up Tool data per month has varied from 75 to 38 units over 8 months of data currently available (January—August 2012). Within the Knowledge/Skills section all questions have shown improvement over time. Five out of the six questions in the Attitudes/Beliefs section have shown improvement since baseline. Results for "A junior staff member feels comfortable to question a more senior staff member who is not following the maintenance bundle," showed initial improvement but in the last two reporting months, results dropped below baseline levels. Finally, three out of the five Resources section questions showed improvement with one question remaining fairly stable over time ("Data collection has not been a burden") and one question showing a drop below baseline levels in the last data collection period ("Unit leadership is stable, i.e. low turnover"). The highest scoring questions in each of the categories can be found in Table 6.

Table 6. Highest scoring questions in each of the three NCLABSI team check-up tool domains