Background: The All Patient Refined-DRG (APR-DRG) system is commonly used for benchmarking and reimbursement. Little is known about the adequacy when applied to pediatric service lines. Cardiac neonates not on ECMO are billed under one of three APR-DRGs, undifferentiated by case type/complexity. Two are not cardiac specific. We hypothesized that differences in pediatric case mix not captured under the DRG/severity system may have large impacts on pediatric cardiac benchmarking and reimbursement.
Methods: We utilized national administrative data from 46 pediatric tertiary hospitals from the Pediatric Health Information System Database, 2014. We included all neonates with APR-DRGs 588, 609, and 630 (Newborn <1500gm w major procedure, Newborn 1500-2499gm w major procedure, and Newborn ≥2500gm w major cardiovascular procedure). Log linear regression was used to compare adjusted cost-to-charge ratio (CCR) costs between cardiac and non-cardiac discharges and across clinical case complexity categories (Risk Adjustment for Congenital Heart Surgery, RACHS-1), controlling for DRG/severity category and clustering standard errors by center. Estimated reimbursements were calculated, multiplying New York State 2014 APR-DRG weights by a range of hospital base rates.
Results: In total, 4,631 neonates met inclusion. Neonates <2500gm undergoing cardiac surgery had 32% higher costs than those undergoing non-cardiac surgeries under the same DRG/severity (CI 20-46%, p<0.001; median $283,000 vs $200,000). Neonates ≥2500gm undergoing high complexity operations (RACHS-1 class 5 or 6) had 44% higher costs than children undergoing lower complexity under the same DRG/severity (CI 26-65%, p<0.001; median $198,000 vs $120,000). Payer mix was similar for cardiac/non-cardiac patients. Assuming 2014 base rates of $6-8,000, average expenses for cardiac neonates <2500gm undergoing major procedures needed to be <45-68% of CCR costs to generate profit (vs <54-80% for non-cardiac); expenses for neonates ≥2500gm undergoing high complexity cases needed to be <42-60% of costs (vs <67-83% for lower complexity).
Conclusions: The APR-DRG system is inadequate for neonatal cardiac benchmarking, and its role in reimbursement has significant potential ramifications for the revenue of pediatric cardiac service lines paid on DRG.