Data Needs for the State Children's Health Insurance Program (2002)

Many states are experiencing high rates of disenrollment from their SCHIP programs, yet it is not clear if these rates reflect dissatisfaction with the program by participants, changes in income limits that cause participants to lose SCHIP eligibility, participants moving to private insurance, or deficiencies in the administration of the program. Improved data on the insurance status of disenrollees and reasons for disenrollment would allow policy makers to better understand which of these factors results in disenrollment.

A number of states reported low levels of retention in the SCHIP program. Data for these studies are primarily administrative. Ian Hill, in his Urban Institute survey of five states, found the approval rate at redetermination to range from a low of 26 percent in Michigan to a high of 65 percent in New York. He found that the SCHIP retention rate was less than 50 percent in four of the five states. However, these data provide little information about whether or how a child was insured after leaving SCHIP or the reasons for disenrollment. Hill’s data do show that referral rates to Medicaid ranged from 9 percent in Colorado to 32 percent in Michigan. (The length of time from initial coverage by SCHIP until renewal—or reapplication for continued coverage—varies among states from a few months to one year.)

Enrollment, retention, and disenrollment from the SCHIP program represent a dynamic process, since the target population can be classified as being in one of four statuses at any given time: uninsured, Medicaid en

rolled, enrolled in a separate state SCHIP program, or privately insured. Currently, little information is available about the insurance status of children leaving the SCHIP program, who may leave for many reasons, including that they are no longer eligible because of income or age. They may be enrolling in Medicaid because their family income has fallen, or they may be covered by private insurance because their family income has risen. Alternatively, they may become uninsured either because they are no longer eligible for SCHIP or Medicaid and have no private insurance option, or because they were unable to complete the paperwork needed for renewal.

It is important to distinguish among these reasons for disenrollment, because policy makers would interpret coverage under private insurance as a success, but failure to renew due to paperwork burden as a failure. Hill listed some of the factors that hinder the redetermination process. Among them are the reliance on a mail-based system, rather than on personal contact, that many times left the recipient confused; the requirement to resubmit documentation that had already been submitted on initial enrollment; lack of coordination with Medicaid when income dropped, leaving the family eligible for Medicaid rather than for SCHIP; and automatic disenrollment in some state programs when there was no response to the renewal notice.

In Hill’s five-state survey, the major reason for discontinuation at the time of redetermination was failure of the family to respond to renewal notices. This reason accounted for as much as 41 percent of all discontinuances in North Carolina. In Michigan, in contrast, renewal forms are sent out 50 days before they are due. Denise Holmes found that only 6.3 percent (77 families out of 1,219) failed to return their forms and most of these families had obtained private insurance. A major reason for the high rate of return was that candidates for renewal received application forms that were already filled out from previous information so that the applicant merely had to indicate any changes that had occurred and sign the form.

Several strategies were mentioned by participants in the workshop for increasing retention in SCHIP. They included:

• Sending out notices well in advance (Hill mentioned 60-90 days).

• Simplifying the language on the form—for example, using the word “renewal” rather than “redetermination” or “reenrollment.”

• Sending forms that are already completed and requesting only changes and a signature.

• Using “passive reenrollment”—that is, if no response is received, the child is automatically reenrolled.

The opportunity to transfer learning about SCHIP retention from one state to another is lost through the lack of consistent definitions of insurance categories and reasons for lack of reenrollment across states. Lack of consistent definitions impedes understanding of how differences in these rates relate to state policies, such as passive reenrollment. We have indicated above how the simplified reenrollment process has played a role in increasing reenrollment in Michigan. Another factor that Michigan found to be instrumental in increasing the rate of reenrollment was allowing self-declaration of income rather than requiring the submission of pay stubs or copies of income tax statements.

The SCHIP population is continually changing, both in terms of residence and income, and because eligibility for Medicaid coverage depends on the age of the child. Thus, there is much movement among the insurance statuses. Marilyn Ellwood cautioned that, while assessing retention in SCHIP is important, the problems of retention in Medicaid should not be ignored. She points out that there are seven times as many children on Medicaid than are covered by SCHIP. The churning of applicants among insurance statuses is not tracked by most states, and longitudinal record systems do not appear to be prevalent. Holmes reported that Michigan has begun selecting monthly samples of new enrollees and then following them over time to be able to track changes in their insurance status and changes in factors that affect insurance eligibility, such as increases or decreases in income.

Administrative data often provide little information about the reasons for disenrollment from the program. Some states have merged their Medicaid and SCHIP files to gain more information about the movement between the two programs. Ellwood reported that, beginning in 1999, all states were required to submit to the Centers for Medicare and Medicaid Services monthly information on all of their Medicaid-eligible children, including those on the traditional Medicaid program as well as the children eligible for the Medicaid expansion through the Children’s Health Insurance Program. This information is included in CMS’s Medical Statistical Information System (MSIS). CMS also offered the states that had separate SCHIP programs the option of including their SCHIP data in the MSIS. As of 2001, only nine of the seventeen states with separate SCHIP programs had submitted their data. For the states whose data are included in

the MSIS, analysis of the data in the system can provide a picture of movements in and out of SCHIP and transfers between SCHIP and Medicaid. Several of the speakers at the workshop commented that, while some data indicate a net flow out of SCHIP, no consistent data exist that indicate reasons for such movement.

The need for qualitative data (also discussed in Chapter 5) was mentioned by several speakers as a means for states to learn why children do not reenroll, so that they can alter their procedures to improve reenrollment. Hilary Bellamy described plans for a series of 52 focus groups in 9 states to obtain information that will be helpful at both the state and national levels in improving enrollment and retention in both SCHIP and Medicaid. Of these, 15 of the groups will involve families whose children had recently disenrolled from SCHIP and were not enrolled in any health insurance program at the time of the focus group meeting. Discussions will center on reasons why SCHIP participants have not reenrolled in the program.

Although Medicaid and SCHIP participation are often seen as distinct in state data systems, they are often conflated in household responses to surveys. Population-based surveys, such as the National Survey of America’s Families and the State and Local Area Integrated Telephone Survey, rely principally on household reports. Because household respondents are often poor reporters of type of coverage, other sources of data are needed. Ideally, a combination of data from surveys and administrative records would be useful as a way of obtaining more accurate information.

Several states are trying to unravel the reasons for SCHIP disenrollment by conducting sample surveys of disenrollees. However, it is difficult to draw conclusions about how program characteristics, such as whether it is integrated with Medicaid or a stand-alone program or whether the state has passive reenrollment, relate to retention in SCHIP because there is little uniformity in the categories that states use to classify reasons for disenrollment.

On issues related to disenrollment raised by workshop participants, the panel concluded:

• Longitudinal studies measuring gross flows across the four insurance categories—Medicaid, SCHIP, private coverage, and no coverage—would be helpful to state and national planners. Cross-sectional surveys can provide, at best, limited information on transitions among insurance categories. Short-term retrospective studies may be yet another approach to measuring these transitions.

• Data from multiple sources, including qualitative data, are needed to fully understand the dynamics of the enrollment/ disenrollment process.

• Separating reasons for disenrollment from SCHIP—ineligibility, application burden, disinterest—is important. More uniform categories would facilitate comparisons across the states.

• Knowing the subsequent insurance status of children who leave SCHIP and the reasons for their disenrollment would allow policy makers to better understand how state policies, such as integrated Medicaid/SCHIP program and passive reenrollment, affect retention rates in the program.