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

Vicki Grant, Southern Institute on Children and Families

Title: Managing by Eligibility Outcomes Data

Funded by the Robert Wood Johnson Foundation, the Covering Kids National Program Office provides direction and technical assistance to 51 statewide lead organizations and over 170 local pilot programs. These programs received grants to conduct outreach to find and enroll low-income, uninsured children, to simplify the eligibility process, and to coordinate coverage programs. Based on lessons learned during the first year, the foundation created the Supporting Families after Welfare Reform program to provide technical assistance and funding to states experiencing declines or stagnation in Medicaid and SCHIP caseloads for children and adults. A primary focus is to assist states in understanding the causes for stagnation or decline and to provide resources to address the causes.

The presentation focused on reasons why applicants are denied Medicaid and SCHIP coverage and why some participants leave the program. Children should be denied Medicaid and SCHIP or leave the program primarily for economic reasons or because the child is too old for the program. However, there are many other procedural reasons why eligible children are not receiving Medicaid or SCHIP benefits. Some typical procedural reasons applicants are denied Medicaid and SCHIP coverage are

failure to provide verification, failure to show for an interview appointment, and being uncooperative. A couple of reasons that participants leave the program are that they withdraw or they fail to comply with procedures. Examples of the latter are failure to provide verification, failure to show for interview appointment, failure to return a report, being uncooperative, and failure to apply for other benefits.

Linda Bilheimer, Robert Wood Johnson Foundation

Title: Data Needs for Tracking Children’s Health Insurance Coverage

Low-income families, especially those with income between 100 and 200 percent of the poverty level, have volatile health insurance coverage. Despite expansions of public coverage for children through Medicaid and the State Children’s Health Insurance Program, evidence from the states suggests that turnover and churning persist among children who enroll in public programs. In this presentation, Linda Bilheimer discussed the types of data needed to understand these phenomena better and the data that are currently available from national surveys and state administrative data systems. Longitudinal surveys, such as the Survey of Income and Program Participation (SIPP), are the best tools for tracking coverage changes. The SIPP data are not sufficiently timely to guide current policy decisions, however, nor can they produce state-specific analyses. Nonetheless, they provide important insights into the volatility of insurance status. For example, an analysis by Mathematica Policy Research of the 1992 panel shows that if all children who were uninsured at a point in time became insured, within one year half of that number of children would be uninsured (Czajka and Olsen, 2000). Other national surveys provide snapshots of who participates, who does not, and the reasons why. The presentation suggested a comparison of point-in-time data on enrollees to ever-enrolled data to indicate the degree of stability in SCHIP. Administrative data from the states can throw light on the outcomes of enrollment and eligibility redetermination processes. But many states have difficulty producing data on enrollment outcomes, definitions vary widely among the states, and linking procedural policies to outcomes is difficult. Medicaid eligibility systems were not designed to be management tools. Also, major investments in eligibility data systems are unlikely to be priorities.

A few reasons why children leave SCHIP were also discussed. Some children rotate between Medicaid and SCHIP. Some states can track this behavior, others cannot. It is very difficult for states to track people who

drop out of public coverage entirely. The National Survey of America’s Families suggests a significant percentage of people who drop out of public coverage become uninsured. Surveys of disenrollees would be helpful with this problem.

Pamela Paul-Shaheen, Center for Advancing Community Health

Title: Covering Michigan’s Kids: Using Information to Inform Policy and Practice

As the lead agency for the Robert Wood Johnson Foundation-funded Covering Michigan’s Kids initiative, the Center for Advancing Community Health works with its state and community partners to optimize the implementation of the state’s SCHIP Program: MIChild. The program is a health coverage program that covers children up to age 19 in families with incomes between 150 and 200 percent of the federal poverty level. The program provides subsidized, low-cost medical and dental coverage for uninsured children across the state. Parents pay a monthly premium of $5.00 per family. Coverage is for a 12-month period. Since the initiation of the Covering Kids effort, those engaged in covering Michigan kids have worked diligently to enroll children and provide them access to needed health care services. The Covering Michigan’s Kids initiative has developed a comprehensive evaluation strategy that has used qualitative and quantitative data to monitor enrollment and retention trends, identify barriers, and recommend policy changes to promote the goal of covering eligible children. The presentation described the structure of the effort, discussed the areas in which qualitative and quantitative information has been utilized, and discussed the challenges and lessons learned.

The presentation noted some barriers to SCHIP assessment. Some Maximus data, which evaluates MIChild, is not broken down by county. Acquiring Healthy Kids data, which evaluate Medicaid, is cumbersome and there is no electronic version of data available in Michigan. Also, the Healthy Kids data are for enrollment only; there is no attempt to track the transfer of enrollees to other programs. Another problem is that choosing comparison counties or sites requires analysis of many factors and will serve only as an estimate of comparison, never an exact match. One barrier to qualitative SCHIP data is that it is expensive and time-consuming to collect.

SCHIP data analysis influences SCHIP policy. Data analysis has verified that a simplified mail-in application is successful in enrollment efforts.

Co-pays, such as the MIChild $5 per month per family charge, increase the likelihood of utilization and the reduction of stigma associated with public programs. Analysis of enrollment data has shown a dramatic increase in application processing and enrollment since the advent of the self-declaration of income on the application. Analysis of application processing and call volume data has shown increases in application submissions and call volumes associated with statewide campaigns of MIChild and Healthy Kids. As a result, Michigan has refined its efforts toward media blitzes and coordinates efforts among different departments.

Mary Alice Lee, Children’s Health Council

Title: Connecticut’s HUSKY Program: Using Data to Improve Enrollment and Retention

As lead agency for the Robert Wood Johnson Foundation-funded Covering Kids initiative, the Children’s Health Council tracks enrollment in Connecticut’s children’s health insurance program, Healthcare for UninSured Kids and Youth (HUSKY). HUSKY Part A is a Medicaid managed care program that covers children under 19 in families with income under 185 percent of the federal poverty level. HUSKY Part B is a separate SCHIP plan that provides subsidized, low-cost coverage for uninsured children in families with income between 185 percent and 300 percent of the federal poverty level; higher-income families with uninsured children can buy in at state-negotiated group rates. Key features of HUSKY include a single point of entry; a single simplified application and renewal form; application or renewal by mail; 12 months of continuous eligibility, regardless of changes in family income; and presumptive eligibility (for HUSKY Part A). Despite intensive outreach and simplification of enrollment, net enrollment increases since July 1998 have been lower than expected.

The Children’s Health Council developed a systematic, ongoing approach to evaluation of enrollment trends and identification of enrollment barriers. Using enrollment data and reports from families and outreach partners, the Children’s Health Council showed that retention is a major problem in the HUSKY program. In the first two years of the program, 78,000 children were newly enrolled in HUSKY Part A, but the net enrollment increase was just 14,500. In one city, net enrollment actually decreased. Many children whose families reported increased income lost coverage at the end of one-year eligibility periods. Renewal forms for HUSKY

Part A were not always forwarded for HUSKY Part B eligibility determination, so some families had to complete the application again. A survey of Hartford families, conducted by the Children’s Health Council, revealed that the main reason children were not reenrolled in HUSKY was that their families received employer-sponsored insurance; in fact, 67 percent of formerly enrolled children (n = 225) were insured at the time of the survey. However, the survey also showed that many parents of enrolled children (n = 478) did not know what determines eligibility for HUSKY, did not know how long coverage would last, and did not know that children must be reenrolled every year.

The Children’s Health Council and the Connecticut Department of Social Services have used information about enrollment and retention to improve outreach and to design, implement, and evaluate interventions aimed at increasing retention. For example, a special mailing was sent out to 34,000 families whose children lost coverage but might still be eligible; few responded. Monthly mailings on family-friendly HUSKY stationery have been more effective in informing families about “renewal” when their children near the end of continuous eligibility periods. Parents and care-taker relatives in families with income less than 150 percent of the federal poverty level are now eligible for HUSKY coverage; thousands have enrolled since January 2001. Self-declaration of income replaced the need for submitting proof of income, and the percentage of incomplete applications dropped.

The Children’s Health Council also recommends that family coverage be expanded to 185 percent of the federal poverty level, that eligibility determinations for Parts A and B be coordinated, and that HUSKY applications be coordinated with applications for other income-based programs, like the subsidized school lunch program. With enrollment data from HUSKY Part B, the Children’s Health Council will be able to track enrollment and retention as children move between HUSKY Parts A and B.

Lisa Dubay, Health Policy Center, The Urban Institute

Title: Assessing SCHIP Impacts Using Household Survey Data: Promises and Pitfalls

The objective of this presentation was to describe how household surveys can be used to assess the impacts of the new State Children’s Health

Insurance Program, review methodological issues associated with household survey data, and propose solutions for dealing with these issues. While evaluating SCHIP using household surveys has some challenges, if conducted carefully such analyses will provide important information on the impact of the program that cannot be obtained elsewhere.

In assessing SCHIP’s impact, eligible children must be identified using a detailed simulation model. Analyses, like the Current Population Survey, that use either a simple eligibility model or examine only children with incomes targeted by the SCHIP program will not accurately identify SCHIP eligible children. Under these circumstances, estimates of the impact of SCHIP will be biased downward.

Extensive individual and family information is necessary to accurately simulate eligibility. Some examples of necessary information are earned income, Supplemental Security Income, Temporary Assistance for Needy Families income, general assistance income, pension income, Social Security income, other income, child support, assets, family structure, family size, welfare history, child care expenses, and employment status and history. Eligibility rules specific to each program also need to be understood in order to estimate eligibility. Examples of these rules are eligibility thresholds, asset limits and disregards, categorically eligible groups, unemployed parent rules, work disregards, earned income disregards, child care disregards, child support disregards, waiting periods, deeming of stepparent and grandparent income, and definition of family unit. The presentation went on to explain that more detailed simulation models are able to find more and more eligible children.

In order to assess the impacts of expansion in coverage via SCHIP, it is necessary to examine both the change over time for the eligible population (treatment group) and the change over time for comparison populations. The experience of the comparison population should be used as a counterfactual for what would have happened in the absence of the policy change. Multivariate methods control for differences in the composition of the treatment and comparison groups and changes in the composition of the treatment and/or comparison groups between the pre- and post-policy change period. There is no perfect comparison group. The comparison group should be similar to the treatment group, but it should not be affected by the policy change. Alternative comparison groups should also be used so that the researcher can test the sensitivity of the results to the use of alternative groups.

In addition, analyses must rely on the same survey in the pre- and post-

SCHIP period in order to obtain reliable estimates. Moreover, the survey must attempt to obtain data on separate SCHIP programs and analysts must consider the implications of the likely increasing underreporting of public health insurance coverage. Finally, analysts should be cautious about evaluating SCHIP’s success before the program is mature. Only small impacts are expected in 1999. Real assessment will not be possible until 2002 or 2003. Moreover, it is likely that Medicaid drives the uninsurance rate for the low-income population.

Thomas M. Selden, Jessica S. Banthin, and Julie L. Hudson, Agency for Healthcare Research and Quality, U.S. Department of Health and Human Services

Title: New and Improved Eligibility Simulation Methodology using MEPS National Survey Data

The Medical Expenditure Panel Survey (MEPS) offers a valuable resource for tracking national trends in insurance coverage and program eligibility among children. The presentation provided a general introduction to the MEPS data and a description of the eligibility and enrollment research under way at the Agency for Healthcare Research and Quality. The authors are currently refining the eligibility simulation model and applying that model to successive waves of the MEPS from 1996-2000. This research should offer insights into the impacts on children’s health insurance coverage of Medicaid, SCHIP, welfare reform, the (now-fading) economic boom, and more.

The presentation began by noting successes and challenges with SCHIP and Medicaid. First, Medicaid expansions have helped more children recently. Under 1987 rules, 15.9 percent of children were eligible for Medicaid, whereas under 1996 rules, 29.5 percent of children were eligible. Expansions are still continuing. By 1996, uninsurance began to decline among children in families under 200 percent of the federal poverty level. There were 4.7 million eligible uninsured children, but 60 percent of uninsured children still remained ineligible. Since 1996, SCHIP has been able to provide coverage for children from families up to and, in some cases, more than 200 percent of the federal poverty level. Outreach has also improved.

Some research goals are to obtain national estimates of uninsured children, to track at-risk children through welfare reform and program expansions, and to obtain measures of access to care and burden of expenditures.

MEPS includes detailed data on coverage, expenditures, and sociodemographics. MEPS is nationally representative. It oversamples high expenditure and low-income children, and it has representation for a mix of states for eligibility and enrollment analyses. Unfortunately, it is currently impossible to estimate eligibility and enrollment by state. MEPS does have longitudinal potential, however. MEPS has problems studying SCHIP for a few reasons: the program’s family definitions are broad in many states, waiting periods also cause problems, and not all uninsured children are eligible for SCHIP.

MEPS has found that progress on uninsurance continues. The trend is minus 1.5 percent per year among children from families less than 200 percent of the federal poverty level. The biggest drop occurred in 1999. There has been no discernible trend above 200 percent of the federal poverty level. Targeted outreach programs are having an effect on enrollment. As of 1999, 20 percent of uninsured children from families under 200 percent of the federal poverty level were still uninsured.

Stephen Norton, New Hampshire Department of Health and Human Services

Title: State-Specific SCHIP Estimates

Since SCHIP is a state-based program, it is critical to move away from a national model for evaluating it to a more state-based approach. SCHIP, at the state level, has an evaluation component that provides evaluation much sooner than a national-level evaluation could. The state legislators want to know not only that SCHIP has expanded coverage to a certain number of children; they also want to know if SCHIP has actually improved children’s access to care or their health status, which is very difficult to demonstrate and tends to be done anecdotally. The simulation models presented by Dubay and Kenney provided incredibly precise estimates for a small state. The models also provided geographic specificity across the state. New Hampshire has developed its own survey and has worked on comparing it with Dubay’s and Kenney’s simulation models. Norton spoke specifically about estimating whether a person has been uninsured for the previous 6 months. The New Hampshire survey did not ask that question specifically, so Norton’s agency is now developing methods of estimating the number of uninsured children. The survey did ask “How many months were you uninsured for the past 12 months?” The use of the 6-month criterion has made a fairly large impact on eligibility and potential eligibility. New Hampshire’s estimation model is the best that could be developed

with the resources available. New Hampshire is beginning to use the survey as an outreach tool. The state has been able to identify variation in insurance rates around it. Significantly higher uninsurance rates occur in rural areas, so the state is working to develop more effective outreach for those areas. State-specific estimates will have a huge impact on the ability to evaluate what is happening in each state.

Gestur Davidson, Minnesota Department of Health and Human Services

Title: Finding the True SCHIP Enrollment Rate

This presentation commented on the paper by Lisa Dubay, Assessing CHIP Impacts Using Household Survey Data: Promises and Pitfalls, and reviewed a Minnesota health insurance survey. The study discussed in Dubay’s paper does not have a true control group. The researchers must compare changes in insurance rates before and after implementation of SCHIP. The comparison group used consists of children just above the eligibility level in the SCHIP states. The uninsurance rate drops after implementation of SCHIP. How do we know that the uninsurance rate did not drop for some other reason? Did SCHIP really have an effect? Perhaps a study of the experience of higher-income groups during the same time period could shed light on what was happening in groups that are not eligible for SCHIP.

Since 1990, Minnesota has conducted a large statewide health insurance survey every four to five years. In 1999, the survey showed that a large number of those known to be on Medicaid responded that they were on SCHIP, whereas most of those known to be on SCHIP replied that they were covered by SCHIP. This is a problem that must be accounted for when analyzing the survey.

Ian Hill and Amy Westpfahl Lutzky, Health Policy Center, The Urban Institute

Title: There’s a Hole in the Bucket. Understanding SCHIP Retention

Research Objective: During early SCHIP implementation, considerable policy attention was directed at state efforts to enroll eligible children; comparatively little attention was focused on how states conducted eligibility redetermination and whether strategies have been implemented to maxi

mize retention of SCHIP enrollees. This study attempts to build the knowledge base in this area by examining state eligibility redetermination processes under SCHIP and administrative data on retention rates and reasons for denial at redetermination. For comparison purposes, Medicaid redetermination processes and outcomes were also explored.¹

Study Design: Information and data were collected from a sample of nine states. Information on state redetermination processes were collected via telephone interviews with program officials during spring/summer 2000; standard protocols were used to ensure consistency. Administrative data regarding redetermination approvals, denials, and reasons for denial were collected during summer/fall 2000. Data were requested for two points in time—May 1999 and May 2000—to permit longitudinal comparisons.

Population Studied: Children enrolled in SCHIP/Medicaid.

Principal Findings: Findings suggest that while states have primarily focused their energies on maximizing enrollment under SCHIP, some effort has also been made to streamline redetermination processes and maximize retention—most of the study states have 12-month continuous eligibility for SCHIP, do not require face-to-face interviews at redetermination, only verify income at determination, and use forms that are somewhat simpler than initial program applications. However, several additional strategies that would further simplify the process were adopted by a smaller number of states, including use of joint SCHIP/Medicaid redetermination forms, preprinting redetermination forms with information already on hand, and passively approving continued eligibility when families miss deadlines for submitting redetermination information.

Difficulties obtaining data on redetermination outcomes, especially in Medicaid programs, suggest that state data systems are limited in their ability to report on retention indicators. When available, data were reported inconsistently, making cross-state comparisons difficult. However, data provided by four states showed rates of retention ranging from 35 to 50 percent. Leading reasons for denial at redetermination included failure to

meet eligibility criteria and failure to comply with procedures. Importantly, the number of children “lost” at redetermination (i.e., families never successfully contacted) appears to be high, ranging from 22 to 40 percent.

Conclusions: States have implemented various strategies to maximize retention under SCHIP by simplifying the redetermination process. However data, when available, suggest high rates of turnover among SCHIP enrollees and large proportions of children being denied ongoing eligibility either for procedural reasons or because they were lost to the system.

Implications for Policy, Delivery, or Practice: While states have taken some initial steps to simplify eligibility redetermination under SCHIP, more effort may be needed to enhance rates of retention. Difficulties obtaining administrative data suggest that state data systems lack capacity to provide needed indicators, resulting in large gaps in what is currently known about the outcomes of the redetermination process.

Hilary Bellamy, Health Systems Research, Inc.

Title: Exploring Disenrollment from Medicaid and SCHIP Through Focus Group Research

The presentation discussed the congressionally mandated Medicaid and SCHIP Focus Group Study sponsored by the Office of the Assistant Secretary for Planning and Evaluation, U.S. Department of Health and Human Services. A total of 51 focus groups were to be conducted in 9 states throughout summer 2001. Focus groups were to be conducted with parents of children who were eligible for Medicaid or SCHIP but not enrolled, those who were enrolled, and those who disenrolled from these programs. The study also included focus groups with low-income privately insured families. The complex study design was presented, emphasizing the focus groups convened with parents of children who have disenrolled from Medicaid and SCHIP programs. The number and type of focus groups were related to each state’s SCHIP program design. There was a mix of geographic regions and populations. Additional populations of interest, such as adolescents, welfare leavers, recent immigrants, and higher income families, were included in the focus group design. The project’s approach to identifying and recruiting disenrolled families for the focus groups was also discussed, as well as the issues of enrollment, retention,

access, and quality of care. The focus groups had not yet been conducted as of the time of the workshop.

Denise Holmes, Michigan Department of Community Health, Medical Services Administration

Title: Using Data to Focus Outreach and Improve Enrollment and Retention in Michigan’s SCHIP Program

The Michigan Department of Community Health began its SCHIP program, which is called MIChild, in May 1998. Since the program’s inception, the department has made extensive use of focus groups, beneficiary surveys, and administrative reports to make improvements in the program. These tools have allowed the department to target limited outreach resources into those methods that most effectively reach families. Survey feedback from beneficiaries who successfully complete the application process as well as from people who request an application but never complete it has resulted in several application and enrollment policy changes. Survey data from beneficiaries who fail to renew coverage when the eligibility period expires have been used to make policy changes that improve retention.

The various tools, the results of analysis of data from these tools, the subsequent policy changes, and the effect of these changes on program enrollment and retention were reviewed in this presentation. Some lessons have been learned about outreach. Media are substantially more effective than other outreach tools. Outreach is a continuous process given the turnover in the target population. MIChild outreach has also been an effective tool to increase Medicaid enrollment. After studying the eligibility process, MIChild made some changes to its application. One of the main reasons that people did not apply was documentation problems. MIChild used to require an applicant to provide a copy of the child’s Social Security card and pay stubs from family members. Since MIChild dropped the requirement for a copy of the Social Security card and began to allow self-declaration of income, incomplete applications have fallen dramatically. MIChild has changed aspects of the renewal process as well. The program has achieved higher retention rates by simplifying the renewal process and following up on any failure to return a renewal form. MIChild has also studied reasons for denial. The main reason for denial is that the family has too much income or they have become eligible for Medicaid instead of SCHIP. There has been a steady increase in MIChild enroll

ment. As the program learns more about enrollees, more people are enrolled and retained in it.

Marilyn Ellwood, Mathematica Policy Research, Inc.

Title: SCHIP Retention and Data Issues

This presentation focused on retention issues. Retention is not a new problem; it has plagued the Medicaid program for years. It is important not to focus entirely on retention in SCHIP. Greater numbers of lowincome children are affected by Medicaid retention problems and, since Medicaid-eligible children are even poorer than SCHIP-eligible children, they deserve at least as much attention. Some states claim that they are only enrolling 50 percent of children at point of redetermination. This may be a reasonable outcome and expectation for SCHIP. Some of these children may be enrolling in the Medicaid program or in private insurance; there is no way to determine this using the current data. Standardization of reasons for disenrollment should be developed for the states.

The Medicaid Statistical Information System (MSIS) data system of the U.S. Centers for Medicare and Medicaid Services now offers some potential for analyzing disenrollment patterns across states. Beginning in FY 1999 all states were required to submit to the Centers for Medicare and Medicaid Services complete month-by-month eligibility information for all Medicaid-eligible people, including children enrolled in Medicaid expansion SCHIP programs. In addition, some states submit MSIS information on children in separate SCHIP programs. One of the advantages of MSIS is that children are given unique permanent identifiers, which allow researchers to track children’s public insurance status over time. MSIS also provides researchers with a way to compare enrollment and retention across states. State SCHIP officials would like MSIS to gather information about the reasons for disenrollment.

Genevieve Kenney, The Urban Institute

Title: Using Other Government Programs to Reach Uninsured Children

This presentation examined the potential of certain federal programs (including the National School Lunch Program, the Special Supplemental

Nutritional Program for Women, Infants, and Children (WIC), and the Food Stamp Program) for reaching the families of uninsured children. It used the 1999 National Survey of America’s Families to update the information provided in “Most Uninsured Children Are in Families Served by Government Programs” (Kenney et al., 1999). Findings suggest that about 70 percent of all low-income uninsured children live in families that participate in one of these programs. The National School Lunch Program— serving families with almost 60 percent of all low-income uninsured children—appears to be a particularity promising vehicle for identifying uninsured children who are eligible for Medicaid or SCHIP coverage. The presentation considered the potential barriers and gains associated with targeting uninsured children through these programs.

David Hanig, Washington Department of Social and Health Services

Title: Nutrition and Health: Matching Data from Two Systems

In this presentation, David Hanig described Washington State’s efforts to coordinate Medicaid enrollment with the school Free and Reduced Price Meals program. He began by reviewing the structure of the Medicaid program in Washington state:

• Medicaid covers children in families between 0 and 200 percent of the federal poverty level since 1994;

• SCHIP covers children from families between 200 and 250 percent of the federal poverty level since 2000;

• The state covers noncitizen children up to 100 percent of the federal poverty level;

• The state sponsors a community outreach and media campaign;

• The state has simplified its SCHIP application down to one page. Simplification includes:

• 12 months of continuous eligibility;

• Self-declaration of income;

• Elimination of the asset test;

• No face-to-face interview; and

• Applicants can apply by phone or on paper.

The state has been using post-TANF funds, which are almost exhausted, to fund its outreach campaigns. In order to avoid caseload decline, the state has developed an online application and is coordinating

enrollment with schools and WIC programs. The state began its coordinated enrollment efforts by including a box on the school lunch application for the parents to check if they wanted an application for medical assistance. The school then forwarded the requests to the appropriate state office. This caused a few problems:

• Schools had to go through all of the applications to find ones requesting medical assistance;

• The Medicaid agency had to manually look up thousands of forms to remove those already on Medicaid;

• Ultimately, the state found that the actual yield from this burdensome process was 3 percent new cases, so the program ended.

Washington is now attempting to electronically match data to identify students on the school lunch program who may also be eligible for medical coverage. Washington’s next step will be automation. The state is going to try to develop a scannable school meal application and obtain software to scan hard copy applications and create databases and generate reports.

Robert Gellman

Title: Will Computer Matching Law Affect SCHIP?

The first computer match ever done matched District of Columbia welfare rolls to the federal payroll in order to catch people who might be welfare cheaters. The results were unfortunate. The sponsors did not consider some important factors. For example, some people previously on welfare were working at the time of the match, and the difference in timing produced incorrect results. Some working people were still poor enough to receive welfare legitimately. Because of situations like this, the Computer Matching and Privacy Protection Act of 1988 amended the Privacy Act of 1974 by adding a series of procedural requirements for computer matching. The law regulates computerized comparisons of records for specified purposes as long as any of the records used in the match are subject to the privacy act. The law may apply when federal records are used to identify children eligible for enrollment in the State Children’s Health Insurance Program. Beyond the specific requirements of the law, general privacy concerns are relevant whenever personal records are used in ways not anticipated or disclosed when the records were originally collected. There are currently no laws to govern matching in private industry; the law governs

only federal records. State records are also generally not covered, but the same types of privacy concerns arise nevertheless.

Heidi Smith, Office of New Jersey Family Care, Division of Medical Assistance and Health Services

Title: Data Matches in the New Jersey FamilyCare Program

The New Jersey FamilyCare (NJFC) program began without a media campaign, with limited resources, and with limited staff with competing priorities. It has been using existing databases to find families eligible for Medicaid and SCHIP. When NJFC attempted to use food stamp datasets to identify eligible families, it found that the data were too raw to be very useful. It is now working to improve its data systems so that they can use this type of data. The WIC program is another potential source of useful data, but it cannot share its data directly with NJFC. NJFC was able to mail letters out through the WIC office to each WIC family offering them assistance. The response was poor. NJFC has now hired outreach workers to work in each WIC office in the state. Those workers talk to people applying for WIC and provide them with information and applications for Medicaid and SCHIP. NJFC has also begun working with the Free and Reduced Price Lunch Program. The application includes a box for parents to check if they would like an application for public health insurance. This has caused some administrative problems for the schools, but the schools are very interested in helping their students receive health insurance, so New Jersey will continue this outreach.

Lynn Blewett, School of Public Health, University of Minnesota

Title: Final Comments

There are currently several types of data on the uninsured. Administrative data include information on state enrollment, eligibility, and claims files. Monthly enrollment files of the Health Care Financing Administration are also useful administrative data. Federal survey data are available to the states. There are a few foundation-supported household surveys. The National Survey of America’s Families covers 13 states, and community

tracking covers 60 communities. State household surveys and qualitative research are also available. Among national surveys of the uninsured, some include state estimates and others do not. The Current Population Survey, insurance component (Employer Survey Data) of the Medical Expenditure Panel Survey (MEPS) and the Behavioral Risk Factor Surveillance System include state estimates. The Survey of Income and Program Participation and MEPS’s household component do not include state estimates. The

National Health Interview Survey allows states to do their own estimates. National surveys should strive to help state programs in several ways. They should offer state benchmarks, trends over time, cross-state comparisons, macro-level analyses, and guide national policy initiatives. States would like national surveys to collect state representative data, take a large enough sample size to allow for valid and reliable state estimates, develop a good survey design to produce policy-relevant information, provide timely and routine release of data, and provide access to microdata or public use tapes for additional state-specific analysis.

In response to a lack of information from national surveys, states have begun to develop their own surveys. In all, 27 states are doing household surveys, and at least 10 states are conducting employer surveys. The Health Resources and Services Administration, U.S. Department of Health and Human Services, has a state planning grant program, which is stimulating even more state survey and data collection activity. The Robert Wood Johnson Foundation is funding a State Health Access Data Assistance Center to support states in their data and survey activities. State surveys will aid policy development by simulating policy options. Program design and development, such as marketing and outreach, premium levels, and willingness to pay, will be studied. State surveys can also provide details on subpopulations of interest, such as specific geographic areas, race and ethnicity, and county or region.

Kristen Testa, Children’s Partnership, Sacramento, CA

Title: Challenges of Data Collection for States

Current surveys, such as the Current Population Survey, offer opportunities for meaningful analysis, but states must not rely solely on these national-level datasets. State reporting data in these surveys are not sufficient to base policy decisions upon. State quantitative and qualitative data and monitoring are necessary to identify what is affecting enrollment and disenrollment.

There is a marked lack of data on retention, particularly in Medicaid. In the absence of a tracking system, California is intending to conduct sample case reviews each year parent coverage waiver requests. These types of data should be made public to provide another tool in the absence of data from existing data sources at the state level.

One of the major problems with data collection is that there are two separate programs, SCHIP and Medicaid, with two different systems of reporting. Since SCHIP was an entirely new program, it allowed state workers to build an entirely new system using the knowledge gained from mistakes made in the Medicaid program. Another challenge in California is how to measure the effects of recently enacted policies to see in fact if they have had the intended effect of increasing enrollment. It is also difficult to discern what is affecting what when there are so many policies interacting at the same time. That information is helpful not only for a state to evaluate its own programs but also to identify best practices and effective policies for other states to adopt.

On a positive note, the U.S. Centers for Medicare and Medicaid Services, as part of a statutorily mandated evaluation, has been compiling information available from the states in a manner that allows researchers to make comparisons. Through the federal evaluation, it has also commissioned individual state surveys and focus groups, which can provide some information at the state level as to the effectiveness of their programs.

Cynthia Shirk, Division of State Children’s Health Insurance Program, U.S. Department of Health and Human Services

Title: State Data Needs

Remarkable progress has been made in the SCHIP program over the past four years. At the end of fiscal year 2000, 3.3 million children were enrolled in SCHIP; 4.5 million children may be enrolled in the program at the end of fiscal year 2001. The scope of coverage provided is increasing. Currently, 36 states provide coverage to at least 200 percent of the federal poverty level or more. Four years ago, only four states provided coverage that high.

Retention is an important issue for SCHIP. It is important that families not lose coverage unnecessarily. Families are more likely to get comprehensive, coordinated, and preventative services if they remain in the program for the entire time they are eligible. Retention efforts can also help

ensure that states are reaching enrollment goals, obtain more value for managed care dollars, and avoid administrative costs that are associated with families enrolling and disenrolling repeatedly.

States need comparability of data, and systems are needed to support that data. Common definitions are imperative to quality research. The states would welcome some program standardization, but the federal government wants to be sure that the states maintain a certain amount of flexibility in their programs. Both the states and the federal government need technical expertise and funding for adequate data systems to analyze SCHIP data.

Quality of and access to care are two areas of SCHIP that need to be studied. Now that the program is well established, this is even more important.

The federal government’s focus has been primarily on enrollment data. It has developed the Statistical Enrollment and Data System (SEDS) specifically for SCHIP. The system provides data on the number of ever-enrolled children on a quarterly and yearly basis. Recently, it has been used to obtain point-in-time data.

SCHIP is currently at a critical point. The economic situation at both the state and federal level is not as good as it was a few years ago. It is more important than ever to show that this program is a good one and helps many people so that progress can continue.