Accuracy of administrative code data for the surveillance of healthcare-associated infections: a systematic review and meta-analysis.

Administrative code data (ACD), such as International Classifications of Diseases, Ninth Revision, Clinical Modification codes, are widely used in surveillance and public reporting programs that seek to identify healthcare-associated infections (HAIs); however, little is known about their accuracy. This systematic review summarizes evidence for the accuracy of ACD for the detection of selected HAIs, including catheter-associated urinary tract infection, Clostridium difficile infection (CDI), central line-associated bloodstream infection, ventilator-associated pneumonia/events, postprocedure pneumonia, methicillin-resistant Staphylococcus aureus, and surgical site infections (SSIs). We conducted meta-analysis for SSIs and CDIs, where acceptable numbers of primary studies were available. For these 2 conditions, ACD have moderate sensitivity and high specificity, but evidence for detection of other HAIs is limited. With current low prevalence of HAIs, the positive predictive value of ACD algorithms would be low. ACD may be inaccurate for detection of many HAIs and should be used cautiously for surveillance and reporting purposes.