Achievable Benchmarks of Care for Pediatric Readmissions
BACKGROUND: Most inpatient care for children occurs outside tertiary children’s hospitals, yet these facilities often dictate quality metrics. Our objective was to calculate the mean readmission rates and the Achievable Benchmarks of Care (ABCs) for pediatric diagnoses by different hospital types: metropolitan teaching, metropolitan nonteaching, and nonmetropolitan hospitals.
METHODS: We used a cross-sectional retrospective study of 30-day, all-cause, same-hospital readmission of patients less than 18 years old using the 2014 Healthcare Utilization Project National Readmission Database. For each hospital type, we calculated the mean readmission rates and corresponding ABCs for the 17 most common readmission diagnoses. We define outlier as any hospital whose readmission rate fell outside the 95% CI for an ABC within their hospital type.
RESULTS: We analyzed 690,949 discharges at 525 metropolitan teaching hospitals (550,039 discharges), 552 metropolitan nonteaching hospitals (97,207 discharges), and 587 nonmetropolitan hospitals (43,703 discharges). Variation in readmission rates existed among hospital types; however, sickle cell disease (SCD) had the highest readmission rate and ABC across all hospital types: metropolitan teaching hospitals 15.7% (ABC 7.0%), metropolitan nonteaching 14.7% (ABC 2.6%), and nonmetropolitan 12.8% (ABC not calculated). For diagnoses in which ABCs were available, outliers were prominent in bipolar disorders, major depressive disorders, and SCD.
CONCLUSIONS: ABCs based on hospital type may serve as a better metric to explain case-mix variation among different hospital types in pediatric inpatient care. The mean rates and ABCs for SCD and mental health disorders were much higher and with more outlier hospitals, which indicate high-value targets for quality improvement.
© 2019 Society of Hospital Medicine
METHODS
Data Source
We conducted a retrospective analysis of patients less than 18 years of age in the Healthcare Utilization Project 2014 Nationwide Readmissions Database (NRD). The NRD includes public hospitals; academic medical centers; and specialty hospitals in obstetrics and gynecology, otolaryngology, orthopedics, and cancer; and pediatric, public, and academic medical hospitals. Excluded are long-term care facilities such as rehabilitation, long-term acute care, psychiatric, alcoholism, and chemical dependency hospitals. The readmissions data contains information from hospitals grouped by region, population census, and teaching status.14 Three hospital type classifications used in this study were metropolitan teaching hospitals, metropolitan nonteaching hospitals, and nonmetropolitan hospitals. These three hospital type classifications follow the reporting format in the NRD.
Study Population
Patients less than 18 years old were included if they were discharged from January 1, 2014 through November 30, 2014 and had a readmission to the index hospital within 30 days. We limited inclusion to discharges through November 30 so we could identify patients with a 30-day readmission as patient identifiers do not link across years in the NRD.
Exposure
We included 30-day, all-cause, same-hospital readmissions to the index acute care hospital, excluding labor and delivery, normal newborn care, chemotherapy, transfers, and mortalities. Intrahospital discharge and admissions within the same hospital system were not defined as a readmission, but rather as a “same-day event.”15 For example, institutions with inpatient mental health facilities, medical unit discharges and admission to the mental health unit were not identified as a readmission in this dataset.
Outcome
For each hospital type, we measured same-hospital, all-cause, 30-day readmission rates and achievable benchmark of care for the 17 most commonly readmitted pediatric discharge diagnoses. To identify the target readmission diagnoses and all-cause, 30-day readmissions based on their index hospitalizations, All-Patient Refined Diagnosis-Related Groups (APR-DRG), version 25 (3M Health Information Systems, Salt Lake City, Utah) were ordered by frequency for each hospital type. The 20 most common APR-DRGs were the same across all hospital types. The authors then evaluated these 20 APR-DRGs for clinical consistency of included diagnoses identified by the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes within each APR-DRG. Three diagnosis-related groups were excluded from the analysis (major hematologic/immunologic disease except for sickle cell, other anemia and disorders of blood and blood forming organs, and other digestive system diagnoses) due to the heterogeneity of the diagnoses identified by the ICD-9-CM codes within each APR-DRG. We refer to each APR-DRG as a “diagnosis” throughout the article.
Analysis
The demographic characteristics of the patients seen at the three hospital types were summarized using frequencies and percentages. Reports were generated for patient age, gender, payer source, patient residence, median household income, patient complexity, and discharge disposition. Patient complexity was defined using complex chronic condition (CCC) and the number of chronic conditions (CCI).16,17 As previously defined in the literature, a complex chronic condition is “any medical condition that can be reasonably expected to last at least 12 months (unless death intervenes) and to involve either several different organ systems or one organ system severely enough to require specialty pediatric care and probably some period of hospitalization in a tertiary care center.”16 Whereas, the Agency for Healthcare Research and Quality’s Chronic Condition Indicator (CCI) defines single, non-CCCs (eg, allergic rhinitis).17
