OBJECTIVE: Determine differences in patterns of percutaneous injuries (Pis)(Pls) in different types of hospitals. DESIGN: Case series of injuries occurring from 1997 to 2001. SETTING: Large midwestern healthcare system with a consolidated occupational health database from 9 hospitals, including rural and urban, community and teaching (1 pediatric, 1 adult) facilities, ranging from 113 to 1,400 beds. PARTICIPANTS: Healthcare workers injured between 1997 and 2001. RESULTS: Annual injury rates for all hospitals decreased during the study period from 21 to 16.5/100 beds (chi-square for trend = 22.7; P = .0001). Average annual injury rates were higher at larger hospitals (22.5 vs 9.5 PIs/ 100 beds; P = .0001). Among small hospitals, rural hospitals had higher rates than did urban hospitals (14.87 vs 8.02 PIs/100 beds; P = .0143). At small hospitals, an increased proportion of injuries occurred in the emergency department (13.7% vs 8.6%; P = .0004), operating room (32.3% vs 25.4%; P = .0002), and ICU (12.3% vs 9.4%; P = .0225), compared with large hospitals. Rural hospitals had higher injury rates in the radiology department (7.7% vs 2%; P = .0015) versus urban hospitals. Injuries at the teaching hospitals occurred more commonly on the wards (28.8% vs 24%; P = .0021) and in ICUs (11. 4% vs 7.8%; P = .0006) than at community hospitals. Injuries involving butterfly needles were more common at pediatric versus adult hospitals (15.8% vs 6.5%; P = .0001). The prevalence of source patients infected with HIV and hepatitis C was higher at large hospitals. CONCLUSIONS: Significant differences exist in injury rates and patterns among different types of hospitals. These data can be used to target intervention strategies.

Original languageEnglish
Pages (from-to)731-736
Number of pages6
JournalInfection Control and Hospital Epidemiology
Issue number10
StatePublished - Oct 2003


Dive into the research topics of 'Differences in percutaneous injury patterns in a multi-hospital system'. Together they form a unique fingerprint.

Cite this