Development of Hospitalization Resource Intensity Scores for Kids (H‐RISK) and Comparison across Pediatric Populations. Issue 9 (25th April 2018)

• Record Type: Journal Article
• Authors: Richardson, Troy; Rodean, Jonathan; Harris, Mitch; Berry, Jay; Gay, James C.; Hall, Matt
• Is Part Of: Journal of hospital medicine. Volume 13:Issue 9(2018)
• Page Start: 602
• Page End: 608
• Publication Date: 2018-04-25
• Subjects: Hospital care -- Periodicals; Clinical medicine -- Periodicals; 610
• DOI: 10.12788/jhm.2948
• Languages: English
• ISSNs: 1553-5592

Abstract

Abstract : BACKGROUND: In the Medicare population, measures of relative severity of illness (SOI) for hospitalized patents have been used in prospective payment models. Similar measures for pediatric populations have not been fully developed. OBJECTIVE: To develop hospitalization resource intensity scores for kids (H‐RISK) using pediatric relative weights (RWs) for SOI and to compare hospital types on case‐mix index (CMI). DESIGN/METHODS: Using the 2012 Kids' Inpatient Database (KID), we developed RWs for each All Patient Refined Diagnosis Related Group (APR‐DRG) and SOI level. RW corresponded to the ratio of the adjusted mean cost for discharges in an APR‐DRG SOI combination over adjusted mean cost of all discharges in the dataset. RWs were applied to every discharge from 3, 117 hospitals in the database with at least 20 discharges. RWs were then averaged at the hospital level to provide each hospital's CMI. CMIs were compared by hospital type using Kruskal–Wallis tests. RESULTS: The overall adjusted mean cost of weighted discharges in Healthcare Cost and Utilization Project KID 2012 was $6, 135 per discharge. Solid organ and bone marrow transplantations represented 4 of the 10 highest procedural RWs (range: 35.5 to 91.7). Neonatal APR‐DRG SOIs accounted for 8 of the 10 highest medical RWs (range: 19.0 to 32.5). Free‐standing children's hospitals yielded the highest median (interquartile range [IQR]) CMI (2.7 [2.2–3.1]), followed by urban teaching hospitals (1.8 [1.3–2.6]), urban nonteaching hospitals (1.1 [0.9–1.5]), and rural hospitals (0.8 [0.7–0.9]; P < .001). CONCLUSION: H‐RISK for populations of pediatric admissions are sensitive to detection of substantial differences in SOI by hospital type.