Association of neurosurgical work relative value units (RVUs) with objective markers of operative complexity.

OBJECT Relative value units (RVUs) form the backbone of healthcare service reimbursement calculation in the United States. However, it remains unclear how well RVUs align with objective measures of procedural complexity within neurosurgery. METHODS The 2018 American College of Surgeons' National Surgical Quality Improvement Program (ACS-NSQIP) database was queried for neurosurgical procedures with >50 patients, using Current Procedural Terminology (CPT) codes. Lengths of stay (LOS), operative time, mortality, readmission and reoperation rates were collected for each code and a univariate correlation analysis was performed, with significant predictors entered into a multivariate logistic regression model, which generated predicted work RVUs, which were compared to actual RVUs to identify under- and over-valued procedures. RESULTS Among 64 CPT codes, LOS, operative time, mortality, readmission and reoperation were significant independent predictors of work RVUs, and together explained 76% of RVU variance in a multivariate model (R2=.76). Using a difference of >1.5 SD from the mean, procedures associated with greater than predicted RVU included: surgery for intracranial carotid circulation aneurysms (CPTs 61697 and 61700, residual RVU=12.94 and 15.07 respectively), and infratemporal preauricular approaches to middle cranial fossa (CPT: 61590, residual RVU=15.39). Conversely, laminectomy/foraminotomy for decompression of additional spinal cord, cauda equina, and/or nerve root segments (CPT: 63048, residual RVU=-21.30), transtemporal craniotomy for cerebellopontine angle tumor resection (CPT: 61526, residual RVU=-9.95), and brachial plexus neuroplasty (CPT: 64713, residual RVU=-11.29) were associated with lower than predicted RVU. CONCLUSION Work RVUs for neurosurgical procedures are largely predictive of objective measures of surgical complexity, with few notable exceptions.