Overall Similar Infection Rates Reported in the Physician-reported Scoliosis Research Society Database and the Chart-abstracted American College of Surgeons National Surgical Quality Improvement Program Database.

STUDY DESIGN: This is a retrospective cohort study.

OBJECTIVE: To compare the infection rates in the Scoliosis Research Society (SRS) morbidity and mortality database to those in the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database.

SUMMARY OF BACKGROUND DATA: Surgical databases have demonstrated value in orthopaedic surgery. However, few studies in the literature have used the SRS database relative to other available databases. To validate the surgeon-reported data of the SRS database, we compared the rate of postoperative infection in this database to that in the chart-abstracted ACS-NSQIP database.

METHODS: International Classification of Disease, ninth revision (ICD-9) codes were used to group patients from the 2012 and 2013 ACS-NSQIP database into groups corresponding to spinal deformity diagnoses delineated by SRS. Postoperative infection rates after surgical correction of these deformities compared with those reported in the 2012 and 2013 SRS database using χ tests.

RESULTS: The overall rate of acute postoperative infection was slightly lower in the SRS database than in the ACS-NSQIP database (1.21% compared to 2.05% in ACS-NSQIP, a difference of only 0.84%, statistically significant, P < 0.001).The infection rates of most diagnoses did not differ; these included scoliosis (idiopathic <10 yr of age/congenital/neuromuscular/other scoliosis), spondylolisthesis (isthmic/dysplastic), and kyphosis (congenital and other). The only differences noted were in Scheuermann kyphosis (difference of 5.49%, P = 0.003), degenerative spondylolithesis (difference of 0.69%, P = 0.003), and idiopathic scoliosis in the age 10 to18 yr and adult age ranges (differences of 1.10% and 1.28%, P < 0.001 for both).

CONCLUSION: Although some statistical differences were found in infection rates between the surgeon-reported SRS database and the chart-abstracted ACS-NSQIP database, these small differences likely reflect differences in surgical cases and data collection methods. This finding supports the validity of the surgeon-reported SRS morbidity and morality database.

LEVEL OF EVIDENCE: 3.