Medicaid Managed Care Impact on Pediatric Diabetic Readmissions

According to the Juvenile Diabetes Research Foundation (JDRF), as many as three million
individuals had type I diabetes in the United States in 2011. In addition, each year more than 15,000
children and 15,000 adults - approximately 80 people per day - are diagnosed with type I diabetes. The
rate of type I diabetes incidence among children under the age of 14 is estimated to increase by 3%
annually worldwide (JDRF, 2011). A significant proportion of children who suffer from diabetes are on
a State Medicaid program, approximately 34% of the total in our dataset. Because of the lack of
standardization between the Medicaid programs across states, potentially large variations in access and
quality of care may exist for children with diabetes. These programs have inconsistent levels of patient
access to services, providers and supplies needed to treat this at risk population. Specifically, the
insurance plan structures of the Medicaid programs for each state vary a great deal in terms of their
levels of managed care penetration and use of utilization management tools. This effort will attempt to
determine the extent to which the juvenile onset diabetic patients on managed care plans are more or less
likely to be readmitted to children's hospitals with uncontrolled diabetes if they are on a Medicaid
managed care plan compared to traditional non-managed care Medicaid.


It is clear that more studies are needed to support or refute the use of managed care tools in
health care, especially for the Medicaid population. This research effort will focus on evaluating the relative contributions of the factors involved with readmission rates for diabetic children with a focus on managed care v/s non managed care Medicaid plans using the Pediatric Health Information Systems database from the Children's Hospital Association.
OBJECTIVE: To determine whether the likeliness to be readmitted within 90 days (adjusted for severity on the first admission) for pediatric type 1 diabetes (T1D) differs between Medicaid managed care and non-managed care.
STUDY DESIGN: Retrospective, cross-sectional design over four years using the Children's Hospital Associations (CHA) Pediatric Health Information System (PHIS) data.
METHODS: Definitions of Medicaid managed care and Medicaid non-managed care were determined using the PHIS data dictionary. De-identified patients were selected from the entire PHIS population across 43 hospitals and 26 states for those discharged between 2008 and 2012, receiving Medicaid at the time of service, and had juvenile diabetes as their primary ICD-9 diagnosis. These results were then flagged for readmission within 90 days. Our analysis included any readmission, readmission for

diabetes, and readmission for diabetic ketoacidosis (DKA) for this subset of patients. Multiple factors
and co-variants for readmission were then analyzed by logistic regression, including age, race, gender,
severity of illness (APR-DRG), patient type on discharge, and state of admission.
RESULTS: Our dataset included 14,544 diabetic patients. Of these, 9,633 were not readmitted and 4,985
diabetics were readmitted. Of these 4,985, one cohort of 1,807 was admitted for DKA. Simple
readmission rates by state for diabetes ranged from 4.08-24.82% (mean 13.57%), with readmission rates
for diabetes with DKA ranging between 3.69-25.33% (12.10%). The state where admission occurred
significantly correlated with the number of days between readmissions (P < 0.0001). Readmission per se
was 1.12 times more likely for Medicaid patients on non-managed care plans than those on managed
care (Chi-square 4.3, P = 0.037). Readmission rates for DKA were similar between the managed care
and non-managed care cohorts (P = 0.9997).
CONCLUSION: Pediatric patients with T1D on Medicaid managed care plans appear less likely to be
readmitted within 90 days of discharge. The state in which admission occurred exerted the greatest
impact on variation among readmission rates.

This publication has been peer reviewed.
Publication Type: 
Kathleen Healy-Collier
Year of Publication: 
Journal, Book, Magazine or Other Publication Title: 
ProQuest Dissertations
Date Published: 
Thursday, December 13, 2012
Place Published: 
Publication Language: 
ISBN Number: 
Ann Arbor

Additional content will be provided upon request.