Objectives: Billing and claims data can potentially be used to identify surgical site infections (SSIs) for quality improvement initiatives. We investigated the effect that variation in procedure coding by providers and facilities has on the calculated SSI incidence.
Methods: We established a retrospective cohort study of individuals aged 6 months – 64 years with ICD-9-CM procedure or CPT-4® codes from facility and/or provider claims for umbilical, femoral/inguinal, or incisional/ventral herniorrhaphy from 1/1/2004-12/31/2010 using private insurer claims data. SSIs within 90 days were identified by ICD-9-CM diagnosis codes, with censoring for other surgeries within 90 days. Complex surgeries with additional procedures performed on or before the herniorrhaphy date during a hospitalization were excluded.
Results: 155,748 non-complex herniorrhaphy procedures were initially identified based on distinct procedure dates > 7 days apart. The number of distinct procedures was reduced to 144,220 after removing procedures with no supportive evidence for operation (e.g., anesthesia, operating room revenue codes, pathology; n=4,609) and surgeries coded for > 1 hernia site or unclassified (n=6,919). The percentage of procedures complicated by SSI was compared according to the stringency of identification of hernia site. Using all claims coded for herniorrhaphy (facility and/or provider), 1.23% (363/29,582) of umbilical, 0.48% (433/90,231) of femoral/inguinal, and 4.00% (976/24,407) of incisional/ventral procedures were complicated by SSI within 90 days. In contrast, the percentages of procedures complicated by SSI were 1.17% (297/25,323) for umbilical, 0.46% (367/79,063) for femoral/inguinal, and 4.18% (752/17,981) for incisional/ventral herniorrhaphy when agreement between the provider and facility procedure coding was required to define the hernia site.
Conclusions: Use of claims data to determine SSI rates requires careful classification of procedures, particularly when characteristics of the surgical procedure are important risk factors for infection.
