I read with interest the Point-Counterpoint column of Dr. Arjun Srinivasan and Dr. Michael Edmond ("Hospital-acquired infections: Is getting to zero the right medicine?" May 2013, p. 4).
I have to agree with Dr. Srinivasan that, on one level, it is hard to identify any target lower than zero for which we should strive. However, I must also agree with Dr. Edmond that try as we may, we will never actually get to zero. Calling a "median" rate of zero "zero" is playing with words and acknowledges that up to 50% of facilities reporting have rates higher than zero.
I would like to propose a compromise that I believe is compatible with the points being made by both of these experts. We can define the steps that we all agree should be taken to make the risk of HAI as low as possible. One of the areas where this has been studied and written about extensively is surgical site infections (SSIs).
The American College of Surgeons’ Surgical Care Improvement Project (SCIP) recommends giving appropriate antibiotics within a specific time interval before the incision. SCIP recommendations are good but not sufficient. Multiple studies have indicated that following SCIP recommendations alone does not result in a significant reduction in SSI rates. In addition to SCIP, we need to make sure that our doses are adequate for our generally overweight patients and that prophylactic antibiotics are being redosed during long operations.
Above and beyond SCIP, we should ensure intraoperative normothermia for all patients (not just colectomy patients) and intraoperative and postoperative normoglycemia (no glucose levels above 180 mg/dL) for all patients (not just cardiac surgical patients). We should ensure optimal teamwork and communication in the OR among all members of the team, aided by the use of checklists. We should ensure understanding and optimization of surgical technique, aseptic technique, good postoperative incision care, preoperative optimization of nutritional status, and other elements that readers of this publication would recommend. SCIP is not sufficient, even though it is very important in addition to the things I have listed above and undoubtedly others I have not listed. Similar lists could be constructed for targeting all HAIs. A list for eliminating central line–associated bloodstream infections (CLABSI) would include use of wide area drapes; full mask, gown, and glove attire for the inserting provider; chlorhexidine skin preparation; use of a checklist; proper dressing of inserted lines; and daily queries regarding whether the line can be removed.
I would like to suggest a goal different from zero HAIs. I propose that we strive to achieve zero "potentially preventable HAIs." A pioneer in surveillance of SSI, Dr. James T. Lee of the University of Minnesota proposed this concept (Infect. Dis. Clin. North Am. 1992;6:643-56). Dr. Lee, an ACS Fellow, defined a "potentially preventable SSI" as one in which investigation of the circumstances demonstrated that the medical/surgical team had not done all that they could and intended to do that would reduce SSI risk. On the other hand, an SSI that occurred in the setting in which every known preventive measure had been taken would be termed an "apparently unpreventable SSI." This conceptualization keeps us on track to follow infections and to investigate each one as a potential failure. It also allows us to focus on where we have done what we intended and where we have failed in our intentions. This approach also gives us the opportunity to improve.
In the Surgical Infection Prevention Committee at the University of Washington Medical Center (UWMC) we started over 10 years ago to follow this practice in our surveillance and focus on SSIs. We started with a focus on the proper delivery of prophylactic antibiotics. We utilized a simple definition of potentially preventable SSI.
An SSI in a case that did not get an appropriate antibiotic within the 60-minute window before incision and that was shaved before the operation was termed potentially preventable. Later we added to the definition those cases that did not get a repeat dose for cases that extended more than two half-lives beyond the time of administering the preoperative dose.
We strived to have no potentially preventable SSIs. When our system achieved the goals of giving preoperative doses and intraoperative redoses such that we had essentially no potentially preventable SSIs (although we continued to have SSIs, of course), we raised the criteria for our definition of preventable SSI.
Next, the definition of potentially preventable SSIs included those diagnosed in any patient who did not get the correct antibiotic at the correct time, was not redosed when necessary, or who arrived in the recovery room with a temperature below 36o C. When we perfected our system to the point where we no longer saw patients arriving in the recovery room with low temperatures, we then focused on glucose.