Currently Viewing:
The American Journal of Managed Care April 2016
Single- Versus Multiple-Tablet HIV Regimens: Adherence and Hospitalization Risk
S. Scott Sutton, PharmD; James W. Hardin, PhD; Thomas J. Bramley, RPh, PhD; Anna O. D'Souza, BPharm, PhD; and Charles L. Bennett, MD, PhD, MPP
Medicaid Managed Care Reduces Readmissions for Youths With Type 1 Diabetes
Kathleen Healy-Collier, CSSBB, DHA; Walter J. Jones, PhD; James E. Shmerling, DHA, FACHE; Kenneth R. Robertson, MD, MBA; and Robert J. Ferry, Jr, MD, FAAP
Currently Reading
Assessing the Impact of an Integrated Care System on the Healthcare Expenditures of Children With Special Healthcare Needs
Mircea I. Marcu, PhD; Caprice A. Knapp, PhD; David Brown, PhD; Vanessa L. Madden, BSc; and Hua Wang, MS
Lost in Translation: Healthcare Utilization by Low-Income Workers Receiving Employer-Sponsored Health Insurance
Bruce W. Sherman, MD; Wendy D. Lynch, PhD; and Carol Addy, MD, MMSc
Patient Safety Intervention to Reduce Unnecessary Red Blood Cell Utilization
Scott Hasler, MD; Amanda Kleeman MS; Richard Abrams, MD; Jisu Kim, MD; Manya Gupta, MD; Mary Katherine Krause, MS; and Tricia J. Johnson, PhD
Impact of Clinical Pharmacy Services on Outcomes and Costs for Indigent Patients With Diabetes
Marissa Escobar Quinones, PharmD, CDE; Margaret Youngmi Pio, PharmD, BCPS, CDE; Diem Hong Chow, PharmD, CDE; Elizabeth Moss, PharmD, CDE, BCACP; Jeffrey Lynn Hulstein, PharmD, CDE; Steven Micheal Boatright, PharmD, CDE; and Annie Mathew, PharmD, CDE
Adding Glucose-Lowering Agents Delays Insulin Initiation and Prolongs Hyperglycemia
Courtney Hugie, PharmD, BCPS; Nancee V. Waterbury, PharmD, BCACP; Bruce Alexander, PharmD; Robert F. Shaw, PharmD, MPH, BCPS, BCNSP; and Jason A. Egge, PharmD, MS, BCPS
Costs for a Health Coaching Intervention for Chronic Care Management
Todd H. Wagner, PhD; Rachel Willard-Grace, MPH; Ellen Chen, MD; Thomas Bodenheimer, MD, MPH; and David H. Thom, MD, PhD, MPH
Four Steps for Improving the Consumer Healthcare Experience Across the Continuum of Care
Keith Roberts, MBA
Patient Perceptions of Clinician Self-Management Support for Chronic Conditions
Peter Cunningham, PhD

Assessing the Impact of an Integrated Care System on the Healthcare Expenditures of Children With Special Healthcare Needs

Mircea I. Marcu, PhD; Caprice A. Knapp, PhD; David Brown, PhD; Vanessa L. Madden, BSc; and Hua Wang, MS
This study analyzes the effect of a managed care program on Medicaid expenditures for children with special healthcare needs using a quasi-experimental design.

Objectives: The Children’s Medical Services Network, a carved-out fee-for-service healthcare system for Florida’s children with special healthcare needs (CSHCN), chose to develop an integrated care system (ICS) for its enrollees. The goals of this study were to analyze the effects of a managed care program on the Medicaid expenditures of CSHCN and to evaluate the performance of econometric models used to analyze healthcare expenditures.

Study Design: We used administrative data from 3947 CSHCN enrolled in Florida’s Medicaid program between 2006 and 2008 for 2 treatment and 2 control counties. The 2 treatment counties were subject to the new managed care ICS.

Methods: To account for the unique nature of healthcare expenditures data, 5 econometric models were constructed. Using a difference-in-differences approach, these models were used to estimate differences in healthcare expenditures between CSHCN in the reform and control counties.

Results: The ICS program decreased outpatient, inpatient, pharmacy, and total costs. These effects were statistically significant for 1 of the reform counties. Emergency department costs increased slightly, though not significantly. Among the econometric models, the generalized linear models outperformed the ordinary least squares regressions.

Conclusions: This analysis provides evidence that managed care programs such as Florida’s ICS have the potential to reduce healthcare expenditures.

Am J Manag Care. 2016;22(4):272-280
Take-Away Points
Using a quasi-experimental difference-in-differences design, we analyzed the impact that a managed care program implemented in Florida, aimed at children with special healthcare needs, has on healthcare expenditures. Further, we used various econometric models to account for the unique nature of expenditures data. We found that the program decreased outpatient, inpatient, pharmacy, and total costs. These effects are systematically statistically significant for 1 of the 2 reform counties. 
  • The literature on the impact of managed care programs on healthcare expenditures is limited. 
  • We provide evidence that a managed care integrated care system reduced the healthcare expenditures of a high-cost group of children.
Children with special healthcare needs (CSHCN) are defined as those who have, or are at increased risk of having, a variety of conditions that require a higher degree of health services than those generally required by children.1 Approximately 11.2 million (15.1%) children aged between 0 and 17 years in the United States are identified to have special healthcare needs.2 It has been shown that healthcare costs for CSHCN are about 3 times that of children without special needs; further, CSHCN account for more than 40% of total medical care costs for children.3

Considerable pressure has been placed on organizations to reduce costs while providing high-quality care.4 The growing number of CSHCN places additional pressure on these organizations to contain costs of providing care to this high-cost group. In an attempt to contain expenditures and improve outcomes, a number of state Medicaid programs have adopted managed care models for CSHCN.5 Several studies have analyzed the impact of managed care models targeted at CSHCN on the quality of and access to care5,6; however, little work has been dedicated to determining the impact of managed care on healthcare expenditures for CSHCN.

Moving to a managed care program is expected to lead to cost savings through improvements in the coordination of care and incentives for cost reductions within the managed care organizations. A growing literature illustrates such cost savings when considering the entire Medicaid population, and, in particular, these studies found that the cost savings are highest for the aged, blind, and disabled.7 One study, that analyzed Florida’s managed care reform for the entire Medicaid population, attributed the observed cost savings to reductions in the number of nonemergency visits and in average cost per hospital visit.8 Two studies investigated the effect of managed care on healthcare expenditures for CSHCN.9,10 The first took on an observational (non–quasi-experimental) approach to analyze the performance of Ohio’s “Access to Better Care” program.9 Alternatively, the second study used a quasi-experimental approach relying on California’s medical managed care expansion in the 1990s.10 These studies found no statistically significant change in expenditures.

In Florida, CSHCN enrolled in public insurance programs are primarily served by the Children’s Medical Services Network (CMSN), which serves more than 135,000 CSHCN enrolled in Medicaid in the state.11 Florida uses the Maternal Child Health Bureau’s definition of CSHCN, as “those who have or are at increased risk for a chronic physical, developmental, behavioral, or emotional condition and who also require health and related services of a type or amount beyond that required by children generally.”1 Historically, the network’s services were purchased on a fee-for-service basis.

In 2005, CMS approved Florida’s waiver to implement several reforms to its Medicaid program. In 2006, these changes began in 2 pilot counties: Broward and Duval. The CMSN chose to participate in the Medicaid reform and developed an integrated care system (ICS) in Broward and Duval counties for its enrollees. Under the ICS, there were 3 important ways the pilot counties differed from the non-reform counties. First, the ICS closed the network of providers that was available to CMSN enrollees; therefore, enrollees were only able to receive care from providers enrolled in the ICS network. Second, a third-party administrator (TPA) was established to oversee claims prior to transferring them to the state Medicaid agency. Before the existence of the ICS, CMSN providers submitted claims directly to the state Medicaid agency. Although the plan was engaging in utilization review prior to the implementation, the TPA made it easier to get more timely and detailed reports, and it should result in better medical management of its enrollees. Third, the ICS imposed additional prior authorization procedures on the providers.

Research on the impact of Florida’s ICS on CSHCN is limited. Two studies found that the ICS pilot program did not reduce satisfaction and quality of care for CSHCN, as perceived by the children’s parents,6 and the utilization of inpatient and outpatient services decreased for CSHCN.12 These studies suggest that the ICS program may reduce costs without altering the CSHCN patient’s experience.

Our study makes novel contributions to the literature on managed care, modeling healthcare expenditures, and CSHCN. We used a quasi-experimental design to estimate the impact of a managed care program on healthcare expenditures for CSHCN. Further, unlike prior literature, we used an array of statistical models to account for the unique nature of healthcare expenditures. A growing array of literature is providing alternative methods to account for the unique properties of expenditures data.13-16 Our study contributes to the methodological debate over the appropriate model when dealing with skewed expenditures data.

Sample and Design

All children in the study were enrolled in Florida’s Medicaid program and the CMSN during the 2-year study period. All children in the health plan have had a special healthcare need (SHCN) identified by their primary care physician. Individual-level data were extracted from the Medicaid encounter, pharmacy, and enrollment files for the CMSN ICS enrollees. In total, 3947 CSHCN, ranging in age from 1 to 21 years, were included in the analysis.

The ICS pilot program was implemented in 2 counties: Broward and Duval17; for each treatment county, we chose a control county (Palm Beach and Orange, respectively) that closely reflected the reform county in its health and sociodemographic characteristics prior to the treatment. In particular, we chose as a control the closest county on a metric constructed from the 2005 values of all the healthcare and sociodemographic county-level numeric variables in the Area Resource File published by the Health Research and Services Administration.18 The metric was constructed by weighing all the variables with the inverse of their variance, which is akin to standardizing the variables by translating them onto a common scale.

Start dates for Broward and Duval were staggered. The pre-period for Broward and its control (Palm Beach) was January to December 2006, and the post period was January to December 2007. For Duval and its control (Orange), the pre-period was May 2006 to April 2007, and the post period was May 2007 to April 2008. Only children enrolled for at least 6 months in Medicaid and CMSN, both before and after the implementation of the ICS were included. Children may have gaps in enrollment due to a loss of coverage, so these children were not dropped from the analysis.

Statistical Analysis

We used a difference-in-differences (DID) methodology to estimate the impact of the implementation of the ICS on healthcare expenditures for CSHCN. To implement this DID approach, we considered an array of 2-part econometric models. For each of the reform counties, we compared the difference in costs before and after the implementation of the ICS with the difference in costs before and after the start date of the ICS in the control county.

Explanatory Variables

To implement the DID methodology, we included in all our models an indicator variable for the county (1 for the ICS county and 0 for the control county), an indicator of time (1 for post-ICS implementation and 0 for pre-), as well as an ICS indicator variable that equals 1 for CMSN children in reform counties after the implementation of the ICS, and 0 otherwise. The latter captures the impact of the ICS on the dependent variables of interest.19

Several factors were used to control for observable differences of the children, including race/ethnicity (ie, white, black, Hispanic, and other), age, gender, an indicator of Supplemental Security Income disability, and a measure of the child’s health status. We included the number of months of CMSN enrollment pre- and post reform as an exposure variable to control for the fact that children were likely to have more expenditures.

To assess the children’s health status, the Clinical Risk Groups (CRGs) were used.20 The CRGs use over 2000 diagnoses and procedure codes from all healthcare encounters to assign children to 1 of 5 health status categories: a) nonsignificant, nonacute; b) significant acute conditions; c) minor chronic conditions; d) moderate chronic conditions; and e) major chronic conditions.

Healthcare Expenditures

Healthcare expenditures were based on the annual costs of CSHCN in Broward, Duval, and the control counties for 5 categories of health services: inpatient, outpatient, emergency department (ED), pharmacy, and total expenditures. Two characteristics of healthcare cost data make their estimation difficult: a significant fraction of individuals have zero healthcare costs and the health cost data are skewed to the right. Although all children in CMSN have a special healthcare need, there are several reasons for which they might have zero costs: a) perceived adequate receipt of healthcare prior to joining the plan, b) poor case management, c) parental perception of good health, or d) missed appointments.

Economists have dealt with the issue of zero health costs by estimating 2-part models. These models exploit the decomposition of expected health costs into the probability of nonzero costs multiplied by the expected costs conditional on them not being zero:

E(yit|Zit) = Pr(yit>0|ZitE(yit|yit>0,Zit)                                                                 (1)

For the first part, we used logit models to estimate the probability of nonzero healthcare costs (Pr(yit>0|Zit)) as a function of the ICS, county, and time dummy variables while controlling for potential confounders. For the second part, to estimate E(yit|yit>0,Zit), several alternative ways of dealing with the skewedness of nonzero health costs were proposed.15 In particular, there were 5 potential models that could be used: 3 were ordinary least squares (OLS) models of log costs that differ only in the way the retransformation to the original unit of US dollars is made, and the other 2 were generalized linear models (GLMs). For each type of cost category, we used the model with the lowest root mean square error (RMSE).

OLS Models

The method most frequently used to mitigate the impact of observations with very high healthcare cost is OLS regression on the logarithm of nonzero costs:

Copyright AJMC 2006-2020 Clinical Care Targeted Communications Group, LLC. All Rights Reserved.
Welcome the the new and improved, the premier managed market network. Tell us about yourself so that we can serve you better.
Sign Up