Medical Support Orders: Potential Fiscal Effects of Matching Wisconsin Insurance and Child Support Data
Thomas Kaplan and Ingrid
Rothe
Institute for Research on Poverty
University of Wisconsin-Madison
January 2003
Introduction
This report, written in response to a contractual agreement between the Wisconsin Bureau of Child Support (BCS) and the Institute for Research on Poverty (IRP), is part of a project designed to promote data sharing between BCS and the Bureau of Health Care Systems and Operations (BCSO) in the Wisconsin Department of Health and Family Services. The purpose of the data sharing is to identify Wisconsin children who have a noncustodial parent with access to affordable health insurance that provides coverage for dependents, and to assure that such children are covered under that plan.
A premise of the data sharing effort is that some children covered by Medicaid have a noncustodial parent with health insurance, but that the family connection between the parent and child is unknown to the Medicaid program. In these situations, Medicaid savings could be realized if BCSO data on who is covered by insurance were regularly matched against BCS data identifying the noncustodial parents of children with a child support order. The same data matches could also help children not participating in Medicaid and with no other health insurance, if it were discovered that the noncustodial parent has access to health insurance. This project was intended to test the feasibility of such data matches and to estimate some of their fiscal effects.
Methodology
To test some of the effects of a routine data match, the Department of Workforce Development created, with advice from IRP, a data file containing most open (as of March 2002) IV-D cases with an identifiable minor child and noncustodial parent.[1] The file contained enough data for each case to permit matching, including the Social Security numbers and the names, birth dates, and addresses of the noncustodial parents and the associated children.[2] The file then went to EDS, which holds the contract with the state to manage the Wisconsin Medicaid Management Information System. EDS matched the children in the file with various Medicaid files to determine
- if the children were ever on the Medicaid rolls;
- if the children were on the Medicaid rolls in a sample month;
- whether the Medicaid program spent any money on the children during sample months;
- and whether the Medicaid program was pursuing third-party liability "cost avoidance" from an insurance carrier for the children in the sample month.[3]
EDS also matched the adults in the files to tapes received from two insurance companies--Blue Cross/Blue Shield and WPS--showing policyholders covered by the companies in September 2002.[4] The two carriers were selected for this test match because they are large Wisconsin carriers and write many fee-for-service policies.
IRP then reassembled the two files received from EDS (one matching children to Medicaid files and one matching adults to insurance company files), reconnecting the noncustodial parents to the children for whom the parents were obligated to provide child support. IRP also computed how many children (a) received Medicaid services or were on the Medicaid rolls, and (b) were not part of a third-party liability "cost avoidance" case being pursued by the Medicaid program, and (c) had a noncustodial parent who maintained health insurance with one of the two companies. This calculation allowed IRP to estimate broadly the maximum potential Medicaid savings if the files were routinely matched.
Findings
In the first nine months of 2002, Medicaid payments (either for a fee-for-service billing or for a monthly capitation payment) were made for 173,651 children covered by child support orders. The noncustodial parent for 2,527 of these children had private health insurance with WPS or Blue Cross/Blue Shield. No coordination of benefits code had been activated--indicating that a third party liability "cost avoidance" case was not being pursued by the Medicaid program--for 1,472 of these children.
A total of 162,852 of the children covered by child support orders were "Medicaid-eligible" in March 2002. The phrase "Medicaid-eligible" includes children for whom monthly capitation or fee-for service bills were paid in that month plus children on fee-for-service Medicaid who were eligible but did not receive a paid service in the month.[5] The noncustodial parent for 2,279 of these children had private health insurance with WPS or Blue Cross/Blue Shield, and no coordination of benefits code had been activated for 1,168 of these children.
It is not entirely clear whether the best estimate of potential savings would come from the children for whom a Medicaid payment was made in the first nine months of 2002 or children who were Medicaid-eligible in March 2002 (the sample month used to determine children with a child support order). Because we wanted to be conservative in our rough estimate of potential savings, we used as the basis for our estimate the 1,168 children eligible in March 2002, rather than the 1,472 children for whom a Medicaid payment was made over the nine-month period.
The composite capitation rate in 2002 for Medicaid was $128.68 per person per month. If all the savings for the 1,168 children had been captured by the state and federal governments, then the savings generated by matching child support files against the files from Blue Cross/Blue Shield and WPS would have been 1,168 X 128.68 X 12 = $1,803,579 all funds ($721,400 GPR annually, if captured by the state).[6]
In 2001, the latest year for which the Office of the Wisconsin Insurance Commissioner reports data, a total of $7 billion of premiums were written for group and individual health insurance in the state. Blue Cross/Blue Shield wrote $564.1 million of these premiums, and WPS wrote $261.1 million of the premiums. The two companies together wrote 11.7 percent of all health insurance premiums.[7]
If the above findings for Blue Cross/Blue Shield and WPS are representative of what would be found for the other carriers, and if all potential savings were realized for every case in which a child on Medicaid could be claimed under a noncustodial parent's policy, the savings would be $1,803,579 / .117 = $15,415,200 all funds ($6,166,100 GPR annually, if captured by the state).
It would likely be many years, however, before the effort to match data would encompass every health insurance company doing business in Wisconsin. In practice, in the first several years, the data matching effort would probably focus on six of the largest carriers--Blue Cross/Blue Shield, WPS, United Health Care, Compcare, WEA Insurance, and Dean Health Plan--which together wrote $3 billion worth of the Wisconsin premiums (42 percent of the statewide total).[8] A somewhat more realistic estimate of maximum potential annual Medicaid savings in the reasonably near future would thus be $15,415,200 X .42 = $6,474,400 all funds ($2,589,800 GPR annually, if captured by the state).
Even this, however, is probably an overestimate, for at least five reasons.
1. Because Medicaid benefits are more comprehensive than are the benefits of many commercial health insurance plans, Wisconsin law requires the Medicaid program to cover eligible people who have private insurance and to bill the private carrier for services covered by the carrier. In these situations, the state's savings are not the full cost to the state (represented by the capitation rate) but instead the share of Medicaid expenses that the insurance carrier will pay. This provision applies only to Medicaid and not to the BadgerCare program. Children who can be covered by a noncustodial parent's health insurance are ineligible for BadgerCare.
2. Some health insurance plans of noncustodial parents may cover only the individual and not dependents. In these situations, county child support agencies would have to investigate to determine if the noncustodial parent could purchase a family plan for a cost deemed affordable under standard child support formulas.
3. Approximately 80 percent of children covered by Medicaid in Wisconsin are served by health maintenance organizations. Under the state's contracts with those organizations, the HMOs rather than the state capture all collections from private insurance carriers for the people they insure. This contract provision exists because, until a few years ago, the state calculated capitation payments based on costs for equivalent fee-for-service clients, and these costs already netted out collections from private insurance carriers for the fee-for-service clients. The state no longer calculates capitation rates that way, however, and it may be appropriate to reconsider the contract provision with HMOs if routine data matches were to be launched. The state could capture the savings from the data matches either by reducing the capitation rate to reflect the new savings anticipated from them or by requiring HMOs to reimburse the state when they receive collections from insurance carriers.
4. Many Wisconsin residents covered by commercial insurance carriers receive their coverage through an HMO. Because HMOs typically develop their provider panels in particular geographic areas, it could be the case that a child who lives in one part of the state would not be covered under the commercial policy of a noncustodial parent living in another part of the state. For the majority of children with recent child support orders, however, the custodial and noncustodial parents reside in the same county, as Table 1 shows. As a result, the geographic character of HMO coverage would not present a problem in many cases.
Table 1 Residence of Noncustodial Parent When Custodial Parent Resides in a Known Wisconsin County, November 2002 N = 39,346 |
|
Same County as Custodial Parent |
64.2% |
Different Wisconsin County from Custodial
Parent |
24.7% |
Different State |
10.7% |
Address Unknown |
0.3% |
Source: IRP analysis of data
from the Wisconsin KIDS data system. Note: Data are for children with a first order in 2000. Because the residence information was collected in November 2002, the orders were, on average, approximately 17 months old at the time of the investigation. Noncustodial parents with older orders may tend to live farther from their children. |
|
5. The projected savings do not take account of administrative and programming costs that could be incurred by the state child support program to develop the matching process. The savings estimates also do not consider local staff costs that would be incurred to accomplish these savings. Although the process could be automated in the KIDS system, enhanced enforcement of medical support orders would require local child support agency staff to have additional interactions with employers, parents, and the courts.
Although these five factors suggest that our estimate of potential Medicaid savings may be generous, it is also important to note that one of the insurance companies used for this sample match (Blue Cross/Blue Shield) reports only policyholders with insurance in a particular month, rather than all those who had held policies in the preceding year. If the carrier had reported in the format used by other carriers (or if we had selected a different carrier for the sample match), our estimates would probably have been larger.
Conclusion
Although our investigation was unable to determine a precise fiscal estimate for Medicaid savings that would flow from routine data matches of the type described here, it is clear that some children now covered by the Wisconsin Medicaid program have noncustodial parents not known to that program who have health insurance, and that savings would be possible if those carriers could be identified and billed. In addition, it is likely that some children not on Medicaid and with no insurance coverage could receive coverage as a result of this kind of data match.
[1] The file excluded the following case types: Kinship Care Arrears Only (ARRK), Arrears Only (ARRN, ARRP), Kidnapping (CNAP), Child Custody (CUSY), FFP Substitute Cares Only (FPAR), Locate Only (LOCO), Non-FFP Substitute Care Arrears Only (NFAR), and Quick Locate (QLOC).
[2] Approximately 250 individuals who appeared to be children in the original files were older than age 19 and were dropped from the analysis.
[3] The phrases "third party liability," "cost avoidance," and "coordination of benefits" all refer to efforts by the Medicaid program to maximize the contributions of private insurance to the health care costs of Medicaid-eligible people.
[4] The monthly file that Blue Cross sends to EDS contains only individuals with coverage in the current month. In contrast, the WPS monthly report contains individuals with coverage in the current month or any time in the preceding year. If Blue Cross reported in the same way that WPS (and most other insurance carriers) report, it is likely that the Blue Cross tapes would have included additional adults matched to children with child support orders in March 2002.
[5] The general budget presumption is that, in any given month, about 20 percent of children covered by Medicaid are paid on a fee-for-service basis and 80 percent are covered by a health maintenance organization.
[6] The federal government pays approximately 60 percent of Medicaid costs; the state, or General Purpose Revenue (GPR), share is approximately 40 percent.
[7] These data are from Wisconsin Insurance Report: Business of 2001, Table E, Financial and Statistical Data, Madison, WI: Office of the Commissioner of Insurance, 2002.
[8] Some of these carriers are health maintenance organizations, which present additional complications for this estimate, as discussed under point 4, below.