U.S. Department of Health and Human Services

Children with Severe Chronic Conditions on Medicaid

Executive Summary

Brian Burwell, William Crown, Ph.D., and John Drabek, Ph.D.

The MEDSTAT Group

November 1997

This report was prepared under contract #HHS-100-92-0013 between HHS's ASPE/DALTCP and the MEDSTAT Group. For additional information about this subject, you can visit the DALTCP home page at http://aspe.hhs.gov/_/office_specific/daltcp.cfm or contact the ASPE Project Officer, John Drabek, at HHS/ASPE/DALTCP, Room 424E, H.H. Humphrey Building, 200 Independence Avenue, S.W., Washington, D.C. 20201. His e-mail address is: John.Drabek@hhs.gov.

The opinions and views expressed in this report are those of the authors. They do not necessarily reflect the views of the Department of Health and Human Services, the contractor or any other funding organization.

EXECUTIVE SUMMARY

Many children with special health care needs depend on the health and supportive services provided by Medicaid. It is important that policymakers have a sound understanding of these children's patterns of health care use, especially given the shift to managed care. Previous studies of children on Medicaid with severe disability and/or chronic illness have focused on "SSI-related" children, or children with extremely high health care costs. This study provides further analysis of Medicaid expenditures for SSI children, and adds to the literature by applying diagnostic and utilization-based criteria to claims data to identify children with severe chronic illness. This methodology enables analysis of the service use patterns and expenditures for children with severe chronic conditions who are receiving Medicaid, but are not enrolled in SSI.

Three data sources were used in the analysis:

• Medicaid administrative records extracted from HCFA's "Tape-to-Tape" database provided data on Medicaid service utilization and expenditures. Tape-to-Tape files for calendar year 1992 from the states of California, Michigan and Georgia were used as the source of data, along with Tape-to-Tape files from the state of New York for calendar year 1991.

• The 1992 MarketScan data base, a proprietary database of integrated claims and population data of individuals enrolled in private insurance plans, was used to compare the prevalence of children with chronic conditions with private health insurance coverage to the prevalence of children with chronic conditions enrolled in Medicaid. The MarketScan database is constructed from the aggregate claims experience of employees and dependents of approximately 80 large employers representing the health insurance experience of approximately 7 million employees and dependents.

• For SSI children, data on the principal disabling condition (as identified by the disability determination process) was obtained from Social Security Administration records.

The principal findings of the study are:

• There are many more children with severe chronic conditions receiving Medicaid coverage than simply those who become eligible for Medicaid through the SSI program.

  SSI children accounted for 1.7% to 4% of children on Medicaid in the four states. Diagnostic criteria identified an additional 1.4% to 2% of children on Medicaid with severe chronic conditions.

• The prevalence of severe chronic conditions among children receiving Medicaid is much higher than the prevalence among privately insured children.

  Using the study criteria, the prevalence of severe chronic conditions among Medicaid-covered children was three to five times higher among Medicaid children across the four states than among privately insured children. Thus, managed care plans which enroll Medicaid children will encounter a much higher prevalence of chronic conditions than they have encountered in their commercial populations, even if SSI-related children are excluded.

• Most SSI children did not meet the study criteria and most children meeting study criteria were not receiving SSI.

  SSI children tended to be older than the non-SSI children identified by the study criteria. In addition, while diagnoses on medical claims do well in identifying children with medical needs, they are much less successful in identifying children with functional disabilities. For example, the diagnosis of mental retardation may be significantly under-reported on medical claims, except for health care services which are specifically related to that condition. Consequently, in many cases it is difficult to identify the condition causing disability in SSI children solely by using medical claims data.

• Medicaid expenditures for SSI children vary significantly by type of disability.

  Although SSI children have expenditures which are higher on average than those of AFDC children, the overall distributions have a similar shape. Thus, the health insurance risk profile of the SSI child population is markedly similar to that of the general population--a relatively small percentage of the population accounts for the majority of total health care costs. In addition, the relative Medicaid costs of SSI children with varying types of disabilities was remarkably consistent across the study states. The implication is that a single capitation payment for all SSI children will significantly overpay plans for SSI children with certain conditions and underpay plans for other conditions.

• The average Medicaid expenditure for non-SSI children meeting the study diagnostic criteria was higher than the average for SSI children.

  This is not surprising since the study criteria tended to identify children with significant medical needs. However, the average expenditure for SSI children who also met the study diagnostic criteria was 2-3 times the average expenditure for children who met the study's diagnostic criteria but were not receiving SSI. Thus, within the group of children who met the same criteria, SSI children had the highest medical needs.

• Inpatient hospital services account for most Medicaid expenditures for children with severe chronic conditions.

  Inpatient hospital expenditures accounted for about two-thirds to three-quarters of total expenditures for non-SSI children who met the diagnostic criteria. Such expenditures accounted for a somewhat lower proportion of total expenditures for SSI children. Data are also presented for the following services: ICF-MR and nursing facility, physician, occupational, physical, and speech therapy, skilled nursing and home health, prescribed drugs, other.

This study has implications for state Medicaid programs as they expand managed care initiatives to enroll more Medicaid-covered children, including children receiving SSI. First, Medicaid expenditures for children with severe chronic illness are predictably greater for children with specific types of conditions than for others. This suggests that states should use risk adjustment methods to account for predictable variation in Medicaid expenditures across diagnoses and conditions. Otherwise, to the extent that Medicaid managed care plans can influence enrollment of children with severe chronic illness, plans will have an incentive to selectively enroll children with predictably lower health care costs.

Further, while SSI children, on average, have higher medical and supportive service needs than non-SSI children, this study also demonstrates the relatively high incidence of severe chronic conditions among Medicaid children not receiving SSI. Managed care plans with experience in primarily serving children with private insurance will encounter a more challenging caseload of children with complex medical needs in their Medicaid contracts, even if SSI children are excluded from enrollment. States and plans need to work together to ensure that provider networks and care management practices in Medicaid plans are adequate to serve this caseload, and that payments to plans fairly reflect the risk profiles of these populations.

Claims data provide a useful data base for the development of risk adjustment models. However, for certain disabling conditions the diagnosis coding on claims is far from complete. This is to be expected since providers are reporting the medical conditions they treat, which do not necessarily represent the underlying condition causing disability. This distinction shows up quite clearly when one compares SSA and Medicaid coding for children with mental retardation. Diagnosis coding may become much more complete if risk adjustment provides an incentive to report additional diagnoses. The implication is that condition-specific capitation rates derived from existing Medicaid claims databases represent a starting point in implementing a risk adjustment system. However, once a risk adjustment system is implemented, it is also likely that the reporting of conditions, and therefore reported expenditures for various conditions, will change in turn, as plans and providers adapt to the financial incentives of risk adjustment.

The Medicaid program plays a critical role in national health care policy in providing medical assistance and supportive services to children with severe disabilities and chronic illness.^{1, 2} First, through its links with the Supplemental Security Income (SSI) program, Medicaid provides health insurance coverage to most children who receive disability benefits under SSI. In 1996, there were approximately 1 million children with disabilities receiving cash assistance under the SSI program.

Second, Medicaid is the primary source of health insurance coverage for children living below the poverty threshold, and previous studies have shown that children living in poverty have significantly higher rates of disability and chronic illness than children living in non-poor households. ^{3, 4}

Third, through a number of special eligibility provisions, Medicaid also provides health insurance benefits to many children with significant health care needs who otherwise would not qualify for Medicaid. This includes children who qualify for Medicaid under its "medically needy" provisions, children who receive home and community-based services under Section 1915(c) waiver programs, and children who are placed in out-of-home residential settings, such as nursing homes, foster care homes, and institutions for persons with developmental disabilities.

It is also important to note that the Medicaid benefit package provides access to a much broader array of covered services than either Medicare or commercial health insurance plans. Not only does Medicaid provide coverage for basic medical care such as inpatient hospital, physician, and prescription drugs (generally without copayment requirements), in most states Medicaid also provides access to an array of specialized therapeutic services for children with disabilities such as physical, speech and occupational therapies, as well as fairly deep coverage of skilled home care visits, home health aide services, personal care services, case management services, and other socially oriented supportive services under Medicaid home and community-based waiver programs. Thus, even for many children in lower-income households which have private insurance coverage for basic health care, Medicaid provides supplemental coverage for specialized therapies and community-based services that are not covered under private insurance plans.

NOTES

1. Adler, Gerald and Rymer, Marilyn. "Children and Medicaid: The experience in four states." Health Care Financing Review. 9 (1): 1-20. 1987

2. Newacheck, Paul, Dana Hughes and Miriam Cisternas, "Children and Health Insurance: An Overview of Recent Trends." Health Affairs. (Spring): 245-254: 1995

3. 13.4% of AFDC children have a functional disability compared to 8.2% in the general population. Source: Disability Among the AFDC Population: Findings from the 1994 Disability Survey ASPE Research Note (forthcoming).

4. 8.5% of children with mothers on AFDC had a disability compared to 6.3% whose mothers were never on AFDC and 9.9% whose mothers were (but are not now) on AFDC. Source: "DisabilityAmong Women on AFDC: An Issue Revisited." Proceedings of the American Statistical Association: Government Statistics Section. 1993.