Our previous multicenter study suggested that a bundled intervention was associated with lower rates of complex S. aureus surgical site infections (SA SSIs) among patients undergoing cardiac or orthopedic operations in community hospitals. We aimed to evaluate the effect of this bundle in patients undergoing neurosurgical (NSG) operation, cardiac operation, or hip/knee arthroplasty at an academic health center.
This pragmatic quasi-experimental study included adult patients who underwent one of the procedures between 7/1/2012 and 9/30/2015 except those whose operations were done to treat infection. The bundle involved screening patients for SA nasal carriage, decolonizing carriers with intranasal mupirocin and chlorhexidine-gluconate bathing, and perioperative prophylaxis with vancomycin and cefazolin for patients who carried MRSA. The primary outcome was complex SA SSIs. To analyze changes in SSI rates, we used Poisson regression in time-series analysis. We used breast operations as a non-equivalent control group.
141 complex SA SSIs occurred after 23,920 operations during the pre-intervention period (7/2004 - 6/2012) and 28 occurred after 11,588 operations during intervention period (7/2012 - 9/2015) (rate ratio [RR] 0.41; 95% confidence interval [CI] 0.27 - 0.61; Fig 1). During the same period, the complex SA SSI rate after breast operations did not decrease (RR 1.96; 0.82 - 4.65). Neurosurgeons implemented other interventions before implementing the full bundle in period 4 (Fig 2). The rate of complex SA SSIs after NSG operations decreased significantly only after the bundle was implemented (period 1 vs 4, RR 0.22; 0.11 - 0.46).
During the intervention period, 53% of patients received all bundle elements appropriate for their carriage status and 39% received some bundle elements. The complex SA SSI rate decreased significantly among patients who fully adhered (RR 0.23; 0.09 - 0.57) and among patients who partially adhered or not adhered (RR 0.56; 0.39 - 0.81).
Despite suboptimal adherence, the complex SA SSI rate decreased after implementing the evidence-based bundle but did not decrease in the non-equivalent control. Implementation science could help improve bundle adherence.
H. Y. Chiang,
M. Ward, None
N. Noiseux, None
J. Greenlee, None
M. Bashir, None
D. Diekema, None
A. Haleem, None
R. Nair, None
L. Herwaldt, None