• J T Street, N P Thorogood, A Cheung, V K Noonan, J Chen, C G Fisher, and M F Dvorak.
• Blusson Spinal Cord Centre, Division of Spine, Department of Orthopedics, University of British Columbia, Vancouver, British Columbia, Canada. John.Street@vch.ca
• Spinal Cord. 2013 Jun 1;51(6):472-6.

Study DesignObservational cohort comparison.ObjectivesTo compare the previously validated Spine Adverse Events Severity system (SAVES) with International Classification of Diseases, Tenth Revision codes (ICD-10) codes for identifying adverse events (AEs) in patients with traumatic spinal cord injury (TSCI).SettingQuaternary Care Spine Program.MethodsPatients discharged between 2006 and 2010 were identified from our prospective registry. Two consecutive cohorts were created based on the system used to record acute care AEs; one used ICD-10 coding by hospital coders and the other used SAVES data prospectively collected by a multidisciplinary clinical team. The ICD-10 codes were appropriately mapped to the SAVES. There were 212 patients in the ICD-10 cohort and 173 patients in the SAVES cohort. Analyses were adjusted to account for the different sample sizes, and the two cohorts were comparable based on age, gender and motor score.ResultsThe SAVES system identified twice as many AEs per person as ICD-10 coding. Fifteen unique AEs were more reliably identified using SAVES, including neuropathic pain (32 × more; P<0.001), urinary tract infections (1.4 × ; P<0.05), pressure sores (2.9 × ; P<0.001) and intra-operative AEs (2.3 × ; P<0.05). Eight of these 15 AEs more frequently identified by SAVES significantly impacted length of stay (P<0.05). Risk factors such as patient age and severity of paralysis were more reliably correlated to AEs collected through SAVES than ICD-10.ConclusionImplementation of the SAVES system for patients with TSCI captured more individuals experiencing AEs and more AEs per person compared with ICD-10 codes. This study demonstrates the utility of prospectively collecting AE data using validated tools.

13 keywords...

hide…