Jay G Berry1,2, Matt Hall3,4, Eyal Cohen5,6, Chris Feudtner7,8, Vincent W Chiang2,9, Paul J Chung10, James C Gay11, Samir S Shah12, Elizabeth Casto13, Troy Richardson3. 1. Division of General Pediatrics, Boston Children's Hospital, Boston, Massachusetts; jay.berry@childrens.harvard.edu. 2. Harvard Medical School, Harvard University, Boston, Massachusetts. 3. Children's Hospital Association, Lenexa, Kansas. 4. Department of Pediatrics, Children's Mercy Kansas City and School of Medicine, University of Missouri-Kansas City, Kansas City, Missouri. 5. Paediatric Medicine, The Hospital for Sick Children, Toronto, Ontario. 6. Department of Paediatrics, University of Toronto, Toronto, Ontario. 7. Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania. 8. Children's Hospital of Philadelphia, Philadelphia, Pennsylvania. 9. Division of Emergency Medicine, Boston Children's Hospital, Boston, Massachusetts. 10. Department of Pediatrics and Health Policy and Management, Health Systems Science, Kaiser Permanente School of Medicine and UCLA RAND Health, RAND Corporation, Los Angeles, California. 11. Monroe Carell, Jr. Children's Hospital at Vanderbilt, Nashville, Tennessee; and. 12. Division of Hospital Medicine, Cincinnati Children's Hospital Medical Center, Cincinnati, Ohio. 13. Division of General Pediatrics, Boston Children's Hospital, Boston, Massachusetts.
Abstract
BACKGROUND AND OBJECTIVES: Hospitals treating patients with greater diagnosis diversity may have higher fixed and overhead costs. We assessed the relationship between hospitals' diagnosis diversity and cost per hospitalization for children. METHODS: Retrospective analysis of 1 654 869 all-condition hospitalizations for children ages 0 to 21 years from 2816 hospitals in the Kids' Inpatient Database 2016. Mean hospital cost per hospitalization, Winsorized and log-transformed, was assessed for freestanding children's hospitals (FCHs), nonfreestanding children's hospitals (NFCHs), and nonchildren's hospitals (NCHs). Hospital diagnosis diversity index (HDDI) was calculated by using the D-measure of diversity in Shannon-Wiener entropy index from 1254 diagnosis and severity-of-illness groups distinguished with 3M Health's All Patient Refined Diagnosis Related Groups. Log-normal multivariable models were derived to regress hospital type on cost per hospitalization, adjusting for hospital-level HDDI in addition to patient-level demographic (eg, age, race and ethnicity) and clinical (eg, chronic conditions) characteristics and hospital teaching status. RESULTS: Admission counts were 383 789 (23.2%) in FCHs, 588 463 (35.6%) in NFCHs, and 682 617 (41.2%) in NCHs. Unadjusted mean cost per hospitalization was $10 757 (95% confidence interval [CI]: $9451 to $12 243) in FCHs, $6264 (95% CI: $5830 to $6729) in NFCHs, and $4192 (95% CI: $4121 to $4265) in NCHs. HDDI was significantly (P < .001) higher in FCHs and NFCHs (median 9.2 and 6.4 times higher, respectively) than NCHs. Across all hospitals, greater HDDI was associated (P = .002) with increased cost. Adjusting for HDDI resulted in a nonsignificant (P = .1) difference in cost across hospital types. CONCLUSIONS: Greater diagnosis diversity was associated with increased cost per hospitalization and should be considered when assessing associated costs of inpatient care for pediatric patients.
BACKGROUND AND OBJECTIVES: Hospitals treating patients with greater diagnosis diversity may have higher fixed and overhead costs. We assessed the relationship between hospitals' diagnosis diversity and cost per hospitalization for children. METHODS: Retrospective analysis of 1 654 869 all-condition hospitalizations for children ages 0 to 21 years from 2816 hospitals in the Kids' Inpatient Database 2016. Mean hospital cost per hospitalization, Winsorized and log-transformed, was assessed for freestanding children's hospitals (FCHs), nonfreestanding children's hospitals (NFCHs), and nonchildren's hospitals (NCHs). Hospital diagnosis diversity index (HDDI) was calculated by using the D-measure of diversity in Shannon-Wiener entropy index from 1254 diagnosis and severity-of-illness groups distinguished with 3M Health's All Patient Refined Diagnosis Related Groups. Log-normal multivariable models were derived to regress hospital type on cost per hospitalization, adjusting for hospital-level HDDI in addition to patient-level demographic (eg, age, race and ethnicity) and clinical (eg, chronic conditions) characteristics and hospital teaching status. RESULTS: Admission counts were 383 789 (23.2%) in FCHs, 588 463 (35.6%) in NFCHs, and 682 617 (41.2%) in NCHs. Unadjusted mean cost per hospitalization was $10 757 (95% confidence interval [CI]: $9451 to $12 243) in FCHs, $6264 (95% CI: $5830 to $6729) in NFCHs, and $4192 (95% CI: $4121 to $4265) in NCHs. HDDI was significantly (P < .001) higher in FCHs and NFCHs (median 9.2 and 6.4 times higher, respectively) than NCHs. Across all hospitals, greater HDDI was associated (P = .002) with increased cost. Adjusting for HDDI resulted in a nonsignificant (P = .1) difference in cost across hospital types. CONCLUSIONS: Greater diagnosis diversity was associated with increased cost per hospitalization and should be considered when assessing associated costs of inpatient care for pediatric patients.