SCIP to the Loo?

Many general and specific tactics for prevention of SSI have been brought together in a “bundle“ known as the Surgical Care Improvement Project (SCIP). A predecessor program, the National Surgical Infection Prevention Project (SIP), focused primarily on the process of administering antibiotic prophylaxis, including choice of agent, timing of administration, and duration of prophylaxis [1,2]. A national audit found that the agents being prescribed for prophylaxis were often inappropriate, that the effectiveness of prophylaxis was decreased because the timing of administration was suboptimal, and that only 40% of patients administered surgical antibiotic prophylaxis had the antibiotic discontinued within 24 h, risking adverse events such as superinfection or the development of bacterial resistance [1]. It was advised that antibiotic administration should occur within 60 min before incision, and that prophylaxis should continue for no more than 24 h [2]. Implementation demonstrated improved adherence to process measures [3], but whether reduced numbers of SSIs were the result remained obscure.

The SIP was incorporated into SCIP as reviewed in this issue of Surgical Infections by Rosenberger et al [4]. with additional process measures added, including recommendations for agents to be used for prophylaxis in specific circumstances. As a U.S. federal program, SCIP includes reporting mandates, with financial incentives for compliance that will eventually become penalties for non-compliance [5].

In a 2003 editorial, Lee [6] anticipated with some irony that in focusing on processes moreso than outcomes, the quality improvement bureaucracy might not have it right, or at least, might be taking too simplistic an approach, it being “… easy to overlook a bedrock principle in quality improvement. That principle says ‘no work performed by humans can be always perfect.’ Humans perform the work of quality improvement. Logically, healthcare improvement cannot be always perfect, and its believers should be willing to submit thinking and plans to scrutiny.” That scrutiny has come.

Perhaps not unexpectedly then, five studies have reported that the incidence of SSI has not decreased as a result of SCIP or SCIP-like interventions [7–11], despite improved compliance with execution of the bundle components. Edmiston et al., also in this issue of Surgical Infections, outline the reasons why SCIP may not have been to be “believed” [12]. The reasons may be several [13,14]. Baseline infection rates may have increased as a result of improved reporting, masking any decrease from process improvement. The assumption inherent, that a focus on process improvement will result in an improved outcome, may itself be defective. The bundle may be imperfect, either for what it includes, or for something (known or unknown) it does not. The bundle may need to be executed flawlessly each time, which has not been measured; only compliance with components has. Perhaps the incremental benefit is so subtle that we have yet to power a study sufficiently to observe it.

The causes and prevention of SSI are complex and multi-factorial, and compartmentalization by SCIP may be an over-simplification. Moreover, SCIP is not a smorgasbord of tactics from which the clinician may pick and choose; prevention of SSI requires the flawless execution of an ensemble of prevention tactics [6], not all of which are included in SCIP. For example, correction of patient-related risk factors for SSI is notable among the missing.

Nonetheless, all of the SCIP measures are supported by ample, good-quality evidence. So, what exactly is (are) the problem(s)? Questions abound. Does SCIP deserve to be flushed down the proverbial toilet? Should the search continue for processes that lead to meaningful improvements in outcome, or are these efforts misdirected and therefore doomed? Is the substantial expense of further quasi-experimental trials justifiable? Is more rigorous methodology (e.g., withdrawing items sequentially from the ensemble to see at what point the SSI rate goes up) even ethical such that it would pass muster with institutional review boards? Should there be proof of concept prior to implementation, especially when sanctions may be in the offing? Finally, and especially, are penalties appropriate if compliant behavior so far makes no difference?