What is the real rate of surgical-site infection?

171 Background: Surgical site infections (SSI) are associated with patient morbidity and increased healthcare costs. Although several national organizations monitor SSI including the University Health System Consortium (UHC), National Surgical Quality Improvement Program (NSQIP), National Healthcare Safety Network (NHSN) and Agency for Healthcare Research and Quality (AHRQ), there is no standard reporting methodology. We compared SSI rates from these databases to our own chart review.

METHODS: We queried the UHC, NSQIP, NHSN and AHRQ databases from 7/2012-6/2014 for SSI following gynecologic surgery at our institution. UHC and AHRQ rely on ICD-9 coding while NSQIP and NHSN employ trained reviewers. Each organization uses different definitions, inclusion and exclusion criteria for SSI. NSQIP reviews 13-17% of cases at our institution while the other agencies include all cases. The rate of SSI was also obtained from chart review for 5/1/2014-6/30/2014 with SSI defined as an infection of the surgical incision or organ space requiring antibiotics. SSI was classified as superficial, deep or organ space (OS). The rates reported by the agencies were compared to the rate obtained by chart review using Fisher's exact test.

RESULTS: The combined UHC/NSQIP/NHSN/AHRQ SSI rate was 5.1% (78/1,540) while the rate found by chart review was 12% (20/166) (p = 0.001). Overall SSI rates for the databases were: UHC 1.6%, NSQIP 8.8%, NHSN 2.9% and AHRQ < 1%. The combined database reported fewer superficial SSI compared to investigator chart review but did not differ significantly when reporting deep and OS: superficial 1.6% v 7.8% (p < 0.001), deep 2.1% v 1.2% (p = 0.57) and OS 2.9% v 3.6% (p = 0.26). The individual databases had wide variation in rate of superficial (UHC 0.7% NSQIP 1.2% NHSN 0.5% AHRQ 0%), deep (UHC 0.7% NSQIP 4.7% NHSN 1.3% AHRQ 0%) and OSI (UHC 0.03% NSQIP 4.4% NHSN 1.4% AHRQ < 1%). Only 19 cases (24.4%) were included in > 1 database. Only one case was included in three databases (1.3%) and no cases were included in all four.

CONCLUSIONS: There is discordance among national reporting agencies tracking SSI and all agencies reported a different rate of SSI compared to chart review. Adopting standardized metrics across agencies could improve consistency and accuracy in assessing SSI rates.