Rates and Impact of Potentially Preventable Readmissions At Children's Hospitals

Sunday, October 27, 2013
Regency Ballroom S (Hyatt Regency Orlando, formerly the Peabody)
James C. Gay, M.D., Vanderbilt University School of Medicine, Nashvile, TN, Rishi Agrawal, MD, MPH, Lurie Children's Hospital, Chicago, TN, Katherine A. Auger, MD, University of Michigan School of Medicine, Ann Arbor, MI, Mark Del Beccaro, MD, FAAP, Seattle Children's Hospital, Pirooz Eghtesady, MD, PhD, Washington University in St. Louis, Evan S. Fieldston, MD, MBA, MS, Children's Hospital of Philadelphia, Philadelphia, PA, Justin Golias, RHIA, CHDA, Nationwide Children's Hospital, Columbus, OH, Paul Hain, MD, Monroe Carell Jr. Children's Hospital, Children's Medical Center, Dallas, TX, Richard E. McClead Jr., MD, Division of Neonatology, Department of Pediatrics, Nationwide Children's Hospital, Columbus, OH, Rustin B. Morse, M.D., UT Southwestern Medical Center & Children's Medical Center, Dallas, Mark I. Neuman, MD, MPH, Emergency Medicine, Children's Hospital of Boston, Boston, MA, J. Mitchell Harris, PhD, Children's Hospital Association, Alexandria, VA, Hal Simon, MD, MBA, Emory University, Javier Tejedor-Sojo, MD, Children's Healthcare of Atlanta, Atlanta, GA, Ronald Teufel, MD, Pediatrics, MUSC, Charleston, SC and Samir Shah, MD, MSCE, Pediatrics, Cincinnati Children's Hospital Medical Center, Cincinnati, OH

Purpose: State and federal agencies have begun to levy financial penalties on hospitals with high readmission rates. Many readmissions, however, may be planned or unavoidable; therefore distinguishing potentially preventable readmissions (PPRs) from all-cause readmissions is critical. 3M has developed software for this purpose (3M-PPR), currently in use by some state Medicaid programs; however, data, validation, and relationship to all-cause readmission rates are lacking for pediatric populations. We sought to: 1) Compare readmission rates at children's hospitals as determined by 3M-PPR against all-cause rates; 2) Determine which hospitalizations account for the largest volume of readmissions by both methods. 

Methods:  1,749,747 hospitalizations for 1,207,884 unique patients in 58 Children's Hospital Association member hospitals for the calendar years 2009-2011, excluding normal newborns, deaths, and left against medical advice. Readmission rates were calculated for all 314 APR-DRGs using 3M-PPRs and all-cause readmissions.

Results:  At 7-, 15-, and 30-day intervals, overall 3M-PPR rates were 2.5%, 4.1%, and 6.2%, respectively; all-cause rates were 4.9%, 8.5% and 13.0%. At 30 days, 60.2% of all-cause readmissions were excluded using the 3M-PPR algorithm, 36.0% of which were related to malignancies. Sickle Cell Crisis, Seizures, Bronchiolitis, Ventricular Shunt Procedures and Pneumonia topped the 3M-PPR list. Of the top 20 all-cause readmission APR-DRGs, 3 were excluded by the 3M-PPR algorithm (Chemotherapy, Acute Leukemia and Cystic Fibrosis) and in 10 others >40% of 30-day readmissions were not considered potentially preventable by 3M-PPR (including Seizures, Gastroenteritis, Central Line and Urinary Tract Infections and Failure to Thrive). In contrast, 80% or more of all-cause readmissions were considered 3M-PPRs for Sickle Cell Crisis, Bronchiolitis, Ventricular Shunt Procedures, Asthma and Appendectomy.

Conclusion: 3M-PPR rates are lower than all-cause readmission rates, and likely exclude many planned and unavoidable readmissions. However, 3M-PPRs also exclude some diagnoses which may be amenable to quality improvement and include some diagnoses of uncertain preventability. Future research should validate the applicability/appropriateness of PPR inclusions and exclusions and caution is warranted when applying these algorithms for determining penalties in pediatric populations.