Should we just skip SCIP?

There is a new paper and editorial on SCIP (Surgical Care Improvement Project) in this week's JAMA. SCIP is a Medicare project to improve compliance with practices that have been shown to reduce surgical site infections. These metrics include correct antimicrobial selection and timing, and appropriate hair removal. In this study the authors correlate compliance with SCIP metrics to surgical site infection rates, and the conclusion is that SCIP compliance doesn't correlate with reduction in SSIs. However, in my opinion, there's a huge problem with this study. The outcome data on SSIs were obtained using administrative data (ICD-9 codes), which are notoriously inaccurate. Kurt Stevenson showed that the positive predictive value of ICD-9 codes when compared to NHSN surveillance methods for 8 different categories of surgical site infections was as low as 14% and only as high as 51%. Thus, I don't think this study should sway our opinion about SCIP one way or the other. Since nearly all hospitals collect SCIP data and some states mandate NHSN SSI surveillance, the data should be available to replicate the study with better outcomes data. And if that shows similar findings, SCIP should be put to pasture.