Variation in Prevalence and Patterns of Peripherally Inserted Central Catheter Use in Pneumonia

Background: Successful treatment of bacterial pneumonia often requires intravenous antibiotic therapy. Although peripherally inserted central catheters (PICCs) are often used to deliver such treatment, hospital practice patterns related to PICC use have previously not been reported.

Methods: Premier’s Inpatient database was used to identify patients hospitalized with pneumonia between 1 July 2007 and 30 November 2011. We compared the characteristics of patients that received a PICC during hospitalization to those that did not, and developed generalized estimating equations (GEE) to identify the patient-, provider- and hospital-factors independently associated with PICC placement. To assess practice variation, we computed hospital risk-standardized rates of PICC insertion using hierarchical generalized linear models (HGLM).

Results: 545,250 patients with a median age of 71 years (range 57-82) were included. Most patients were Caucasian (69.2%), female (53.0%), insured by Medicare (67.9%), and had a median length-of-stay of 5 days. A total of 41,849 (7.7%) patients received a PICC during hospitalization (median day of receipt: hospital day 4). Compared to those that did not receive PICCs, recipients were younger (median age 69 vs. 71 years), had higher levels of comorbidity (Gagne score median: 4 vs. 2) and were more often diagnosed with healthcare-associated pneumonia (43.1% vs. 29.9%).

Weight loss (OR 2.03, 95% confidence interval [CI]: 1.97-2.10), sepsis on admission (OR 1.80, 95%CI: 1.75-1.85), and ICU-status on hospital day one or two (OR 1.70, 95%CI: 1.64-1.75) were independently associated with receipt of PICC. Admission by geriatricians and critical care physicians was also associated with increased likelihood of receiving a PICC (OR 1.81, 95%CI: 1.62-2.03 and OR 1.14, 95%CI: 1.05-1.24, respectively). Risk-standardized rates of PICC utilization varied from 0.3% to 41.7% with a median (25^th-75^thpercentile) rate of 11.4%, (5.3% – 17.1%). Nearly 70% of the variability in PICC use remained unexplained by patient-, provider- or hospital-factors.

Conclusions: In adults hospitalized with pneumonia, PICC use is common and directed towards those with severe illness. Substantial variation in PICC use between hospitals was noted and could not be explained by differences in clinical characteristics. Future research examining hospital-level determinants of PICC use and impact on patient outcomes is warranted.