What's new in critical illness and injury science? Preventing surgical infections requires the right antibiotic for the right duration.

The bounded rationality model of cognition describes how thinking processes can become distorted during clinical decision-making.[1] When time is short and available information limited, conclusions are often made instinctively — based on what feels right, from experience. These conclusions are not necessarily incorrect, but unconscious, systematic biases in thinking place limits on the human capacity for objective reasoning. Clinical practice guidelines provide a rational foundation for rapid decision-making. However, despite their rigorous evidence-based formulation, adherence to recommendations is often low and this can have adverse consequences for patients.

For example, surgical site infection (SSI) prevention recommendations for elective surgery have been standardized by the US Surgical Infection Prevention Project (SIPP). In a baseline study, recommendations for antibiotic administration and duration were followed in only 48 and 41% of elective surgical procedures, respectively.[2] Following introduction of the SIPP, adherence improved and SSI rates fell by 27% in the first year.[3] A large follow-up study of 400,000 surgeries, observed overall adherence improving to more than 80%. Low adherence remained a significant predictor for subsequent SSI.[4]

Guidelines for antibiotic prophylaxis in traumatic open extremity fracture surgery have been available from the Eastern Association for the Surgery of Trauma (EAST) since 1998. In this issue, Barton et al. report their findings from a retrospective study of adherence to the EAST recommendations. They found that adherence was generally low — fully compliant in only 28.5% of surgeries, and this non-adherence correlated significantly with a need for longer hospital stays, and higher morbidity. They also identified a non-significant trend toward more surgical site (4.9 vs. 9.8%) and hospital-acquired (11.5 vs. 17%) infections when antibiotic prophylaxis recommendations were not followed. Given the relatively small number of events, it would be interesting if these findings could be replicated in a larger study.

The most common reason (71%) for non-adherence to the EAST guidelines was antibiotic prophylaxis continued for longer than recommended. Interestingly, Barton et al. observed that adherence was lowest when the risk of SSI was greatest. That is for higher Gustilo grade fractures, and following trauma from road traffic collisions. A cognitive bias might help explain this behavior. The omission bias describes the greater discomfort felt from doing harm through action rather than allowing harm through inaction, even when consequences are the same. As all open-extremity fractures receive antibiotic prophylaxis, the decision regarding when to stop antibiotics postoperatively is actively made by the managing team. In high risk patients where a SSI is expected in at least 10% of patients, it may be that this SSI, which develops despite continuing prophylactic antibiotics (‘allowing harm’) is less discomforting than the SSI that develops after prophylactic antibiotics have been stopped (‘doing harm’). In other words, it feels better to ‘err on the safe side’ having ‘done what was possible’ despite the lack of recommendations — and evidence — that this offers any benefit.

Instead, evidence is accumulating that prolonged use of antibiotics postoperatively causes harm. This includes not only predictable adverse effects from drug toxicities and Clostridium difficile–associated diarrhea, but also an increased rate of SSIs.[5] The ecological niche formed by continued antibiotic selection pressure prevents a reconstitution of host flora, and allows hospital-acquired pathogens with multiple drug resistance mechanisms to invade, and clinical infection results.[6] The line between preventing and causing an infection with prophylactic antibiotics is closer than has been appreciated.

Of course there is no simple explanation for why behavior deviates from guidelines. Clinicians may disagree with recommendations, or simply lack awareness of what is advised.[7] It is hoped that neither is a significant caveat to the findings of this study. Despite Barton et al. using the 1998 EAST publication for comparison, an updated review in 2011, did not find contradictory evidence to change the recommendations although new high quality studies are lacking.

Guideline recommendations can also be difficult to incorporate into clinical practice when standardized protocols do not appear to correspond with individual patient scenarios. This can result in clinician distrust, and the perception of cookbook medicine - a conflict between the science and the art of clinical practice. For prophylactic antibiotics, this problem should be less important, as the strategy for preventing surgical site infections follows a simple, standardized rule: obtain tissue concentrations of the antibiotic effective against organisms expected to contaminate the surgical site and maintain these levels for the period of risk, which typically ends no more than 24 hours after wound closure. Deviating from this either reduces protection or over-compensates, resulting in an increased risk of SSI. The evidence from Barton et al. is that adherence to EAST recommendations improves patient outcomes, and rationally the next step is for guideline writers and clinical decision-makers to remove the barriers to their acceptance into practice.