Infection control II: A practical guide to getting to zero

Abstract

Although I did not have this in mind while writing “Infection control: Public reporting, disincentives, and bad behavior,”1 shortly before its publication I attended a meeting at which several hospital attendees presented strategies being employed to reduce hospital-associated infections (HAIs). I was impressed by howmany interesting and creative ideas were discussed, several of which may be valuable to implement at my own institution. However, I was also impressed by how many things I had to learn to game the system, if not overtly then at least to reconsider how to interpret the rules regarding the intended surveillance versus clinical definitions of HAI. One policy that seems to be gaining momentum, and which I subsequently found is sanctioned by the Centers for Disease Control and Prevention as good clinical care (ignoring the surveillance definition), is to screen urine samples submitted for culture using a urine analysis (U/A) and proceed not to culture urine samples when < 10white blood cells/mm3 are detected. The fact that wewere not doing this at my institution and that other hospitals also were not aware of this led me to conclude that various practices are being employed that make for a distinctly uneven playing field. After polling several colleagues, I put together a list, including pros and cons, of many things that can be suggested to hospital administrators to help a facility reach the goal of zero infections (Table 1). Indeed, the use of bundles, more awareness, and better oversight practices have made a significant dent in the rate of HAIs.3 However, as each medical center/system attempts to move to the magic zero or the 100th percentile, some are simply more creative than others. If National Health Safety Network-reported data are going to be used to rate hospitals and determine dollar distribution, then we should all take advantage of the best techniques (unfortunately not always the best patient care) available to lower our HAI rates by sharing information. For instance, when I reviewed my hospital’s catheter-associated urinary tract infection rates for 2015, excluding those with U/A with < 10 white blood cells/mm3, our reported rate would have been 28% lower. I appreciate that presenting the information in a commentary is not conventional, but it is perhaps the simplest way to get the information across. Furthermore, perhaps others may be inspired to step forward and report other ways to improve data. Unfortunately, although some of these practices are reasonable, others clearly do not always represent best clinical practice. That said, some—if not many—hospitals have institutionalized these practices. There are other administrative approaches toward lowering HAI rates that I have not included in Table 1. One is to keep infection control departments woefully understaffed. As a hospital epidemiologist working closely with the infection preventionists (IPs), I have a very good idea of what it takes to review the numerous blood cultures and urine cultures, and follow postoperative procedures on a daily basis. I have visited several hospitals approximately the size of my own that employ 1 or perhaps 2 IPs. In my assessment, it was virtually impossible for these practitioners to perform the required reporting work completely, or accurately, despite the IP’s personal commitment, particularly given the IP’s other teaching and observation responsibilities. Furthermore, if IPs can only review some of the data, it may work in the institution’s favor when reporting to the Centers for Medicare and Medicaid Services (CMS). Playing the odds may be a fairly safe bet. When 10 random charts are pulled for review, one can hope to reach the 80% threshold. Recently, I was somewhat validated when Centers for Disease Control and Prevention and CMS published Adherence to the Centers for Disease Control and Prevention’s Infection Definitions and Criteria Is Needed to Ensure Accuracy, Completeness, and Comparability of Infection Information,4 which implicitly, if not explicitly, acknowledges attempts by some to game the surveillance definitions. I suggest here a few ideas that may help to make the data that are generated by different hospitals more comparable so that institutions that do not or cannotmake some of the reasonable practice changes suggested in Table 1 are not penalized. Furthermore, bad * Division of Infectious Diseases and Immunology, Department of Medicine, NYU School of Medicine, 550 First Ave, NBV 16S5, New York, NY 10016. E-mail address: Harold.horowitz@nyumc.org. Conflicts of Interest: None to report.