PROGRAM EVALUATION DIVISION Centennial Building -Suite 140 658 Cedar Street -St. Paul, MN 55155 Telephone: 651-296-4708 Fax: 651-296-4712 E-mail:firstname.lastname@example.org Web Site: http://www.auditor.leg.state.mn.us
Photo Credits: The photographs on the MinnesotaCare report cover and those in the body of the report were taken by Legislative Auditor staff.
Program Evaluation Division The Minnesota Office of the Legislative Auditor was established in 1973, replacing the century-old Public Examiner's Office. Its role is to audit and evaluate public programs and ensure accountability for the expenditure of public funds. In 1975, the Legislature created the Program Evaluation Division within the auditor's office. The division's mission, as set forth in law, is to determine the degree to which activities and programs entered into or funded by the state are accomplishing their goals and objectives and utilizing resources efficiently.
The division conducts a variety of studies for the Legislature, including program reviews, which describe state programs; program evaluations, which examine whether programs are implemented efficiently and achieving results; and best practices reviews, which identify the most efficient and effective methods of service delivery. The division also conducts follow-up studies and updates previous research findings.
Topics for study are approved by the Legislative Audit Commission (LAC), a 16-member joint, bipartisan commission. The division's reports, however, are solely the responsibility of the Office of the Legislative Auditor. Findings, conclusions, and recommendations do not necessarily reflect the views of the LAC or any of its members.
The Office of the Legislative Auditor also includes a Financial Audit Division, which annually conducts a statewide audit of the 25 largest agencies, an audit of federal funds, and approximately 40 financial and compliance audits of individual state agencies.
Professional Staff James Nobles, Legislative Auditor Roger Brooks, Deputy Legislative Auditor Joel Alter Valerie Bombach
David Chein Jody Hauer Adrienne Howard Daniel Jacobson Deborah Junod Carrie Meyerhoff John Patterson Judith Randall Jan Sandberg Jo Vos John Yunker
Support Staff Geri Britton Theresa Wagner Barbara Wing
This document can be made available in alternative formats, such as large print, Braille, or audio tape, by calling 651/296-8976 Voice, or the Minnesota Relay Service at 651/297-5353 or 1-800-627-3529.
e-mail:email@example.com Reports of the Office of the Legislative Auditor are available at our Web Site: http://www.auditor.leg.state.mn.us
OFFICE OF THE LEGISLATIVE AUDITOR State of Minnesota James Nobles, Legislative Auditor January 21, 2003 Members Legislative Audit Commission MinnesotaCare is a subsidized health insurance program for lower-income Minnesotans. In April 2002, the Legislative Audit Commission directed the Office of the Legislative Auditor to evaluate certain aspects of the MinnesotaCare program. Legislators' questions centered on how well the Department of Human Services (DHS) and counties are determining enrollment eligibility and how efficiently they are processing applications. Overall, DHS does not adequately ensure that MinnesotaCare eligibility criteria are being accurately applied. Two key criteriahousehold income and access to other insuranceare prone to error and inconsistency because of processing errors and problems with the accuracy of information reported by applicants. Over the past four years, DHS has had trouble keeping pace with rapidly growing workloads. DHS has recently improved the way it manages new applications, but the factors making the process susceptible to large backlogs are still present. DHS is investing in several computer modernization projects to help manage cases and determine eligibility. We recommend that DHS expedite these projects, but also take interim steps to tighten eligibility determination policies and procedures. This report was researched and written by Deborah Parker Junod (project manager), Adrienne Howard, and Dan Jacobson. The Department of Human Services cooperated fully with our review.
Sincerely, /s/ James Nobles /s/ Roger Brooks James Nobles Roger Brooks Legislative Auditor Deputy Legislative Auditor
Room 140, 658 Cedar Street, St. Paul, Minnesota 55155-1603 Tel: 651/296-4708 Fax: 651/296-4712 TDD Relay: 651/297-5353
Table of Contents Page
Health Care Program Administration
Enrollment and Funding
How the Program Has Changed
2. ACCURACY OF ELIGIBILITY DECISIONS
Access to Other Health Insurance
3. CASE PROCESSING
Workload and Productivity
Factors Affecting Performance
SUMMARY OF RECOMMENDATIONS
APPENDIX: Methodology and Sample Design
RECENT PROGRAM EVALUATIONS
List of Tables and Figures Tables Page 1.1 MinnesotaCare Definitions 6 1.2 MinnesotaCare Income Limits, FY 2003 7 1.3 MinnesotaCare Benefits 11 1.4 Examples of MinnesotaCare Monthly Premium Amounts for Families at Various Income Levels, FY 2003 15 1.5 Actual and Projected Health Care Access Fund Balances, FY 1998-2005 17 1.6 Major MinnesotaCare Program Changes, FY 1992-2003 18
2.1 Framework for Evaluating MinnesotaCare's Eligibility Process 20 2.2 DHS Policy for Proof of Earned Income 22 2.3 Errors Determining Earned Income 24 2.4 Errors Determining Self-Employment Income 24 2.5 Errors Determining Wage Income 25 2.6 Impact of Different Wage Calculation Methods on MinnesotaCare Eligibility and Premiums 26 2.7 Income Discrepancies Between MinnesotaCare's Income Determination and Data From Tax Returns and the Unemployment System 29 2.8 Insurance Questions on the Application and Renewal Forms 32 3.1 MinnesotaCare Operations Staffing, FY 1999-2002 55 A. 1 File Review Sample 63 A. 2 Enrollee and Employer Survey Response Rates 67
Figures 1.1 MinnesotaCare Average Monthly Enrollment, FY 1994-2002 12 1.2 MinnesotaCare Spending, FY 1994-2002 13 1.3 Growth in Administrative and Health Care Costs Per Enrollee Since FY 1996 14 1.4 Percentage of MinnesotaCare Funding by Source, FY 1994-2002 15 3.1 Processing a New Application 51 3.2 Application Backlog Time, MinnesotaCare Operations, July 1998-October 2002
Summary Major Findings: In about one-third of cases, state and county staff made errors when determining MinnesotaCare applicants' income. These errors resulted in many enrollees paying the wrong premium and, in a small proportion of cases, incorrect eligibility decisions (pp. 23-27).
MinnesotaCare estimates of annual income frequently did not match income reported to other sources, often because individuals' income changed after the initial estimate. As a result, we estimate that participants may have underpaid premiums by $5-22 million (pp. 28-31).
We also found that many applicants misreport information on insurance available from their employers, a key factor in determining whether a person is eligible for MinnesotaCare (pp. 32-36).
Weaknesses in Department of Human Services (DHS) computer systems, compliance activities, and other means of overseeing MinnesotaCare underlie eligibility errors (pp. 38-43).
Over the past four years, large application backlogs at DHS often delayed health care coverage for eligible applicants. DHS recently hired more staff and improved productivity to process applications more promptly, but remains vulnerable to large backlogs because of staffing issues and heavy reliance on manual operations (pp. 50-58).
Key Recommendations: DHS should tighten its income and insurance eligibility policies and do more frequent compliance reviews to check the accuracy of information reported by applicants (pp. 45, 48).
DHS should expedite the development of new computer systems to help manage cases and determine eligibility, and the department should use the new systems to verify income with tax return and unemployment wage data (pp. 45,47).
The Legislature should change the law to allow mid-year premium adjustments when income increases, as it now allows for changes when income decreases (p. 47).
DHS should consider alternatives to self-reporting of insurance eligibility (p. 46).
The Department of Human Services (DHS) needs to improve the way it determines eligibility for MinnesotaCare.
Report Summary MinnesotaCare is a subsidized health insurance program created in 1992 to help Minnesotans caught in the gap between state-provided health care and affordable private health insurance. To be eligible for MinnesotaCare, applicants must meet a complex set of criteria on income, assets, access to other health insurance, residency, and citizenship. Coverage is available to families and adults without children. Eligibility is determined at the time of initial enrollment and, once enrolled, is redetermined annually. Enrollees pay a monthly premium for coverage.
MinnesotaCare has grown steadily over the past ten years. In fiscal year 2002, average monthly enrollment was 144,000, and total spending was about $390 million. The program is funded from three sources: a state tax on health care providers, federal matching funds, and enrollee premiums. The state share of MinnesotaCare funding is provided through a separate budget account. In recent years, revenues for the account have not kept pace with expenditures, but fund reserves have been sufficient to absorb annual losses. If projections hold true, however, fund reserves will be nearly depleted in fiscal year 2005, and continued funding imbalances may trigger mandatory actions to reduce MinnesotaCare spending.
Our evaluation assesses how accurately agency enrollment staff determine eligibility and how efficiently they process cases. We reviewed a random sample of MinnesotaCare case files processed from January through March 2002, matched income reported by applicants to income data reported to other sources, and surveyed enrollees and their employers about access to employer-based health insurance. We also analyzed data on staffing and the speed of case processing.
Frequent Errors Determining Income Resulted in Many Enrollees Paying the Wrong Premium
For the purposes of MinnesotaCare, state law establishes what income should be counted, and DHS policy sets specific procedures for calculating houshold income. Based on our file review, state and county workers erred in applying DHS procedures in an estimated 32 percent of cases. Errors included using the wrong calculation method and relying on incomplete or unclear documentation. Because of these errors, many enrollees paid the wrong premium. In 63 percent of cases in which we found an income error and the file contained enough information for us to make an independent income determination, the error resulted in a premium difference. The premium differences went in both directions and averaged $295 per year.
Workers' errors in determining income result from a failure to apply correct eligibility policy, lack of clarity in that policy, and reliance on manual eligibility determinations. DHS is counting on computer modernization projectsan online application and an automated eligibility systemto help the agency resolve many of these issues. In DHS' view, automating eligibility decisions is an important part of a long-term solution to problems with consistency and accuracy. We agree and recommend that DHS expedite these projects, but implementation is at least 18 months away. In the meantime, DHS can take interim steps to address problems raised in our report, and we recommend that the department clarify its policies and require workers to take refresher training.
Enrollment staff did not follow income determination procedures in nearly one-third of cases.
Income Estimates for MinnesotaCare Frequently Did Not Match Actual Income, Often Because Income Changed After Eligibility Was Determined
To ensure that MinnesotaCare is targeted to the intended recipients, the information used to assess eligibility should accurately reflect applicants' circumstances during the time they are enrolled in the program. Our analysis showed, however, that annual income estimates used to decide MinnesotaCare eligibility for 2001 often did not match income that was reported on tax returns and to the unemployment system. In about 27 percent of matched cases, actual income exceeded income used for MinnesotaCare by $5,000 or more. In another 10 percent of cases, actual income was $5,000 or more lower than that reported to MinnesotaCare. Many of these discrepancies occurred because MinnesotaCare generally projects wage income based on a four-week snapshot, but individuals' income often changed later in the year. Had premiums been based on actual income, enrollees would have paid an estimated $5-22 million more in annual premiums. The magnitude of this estimate is uncertain for several reasons, including changes in the economy and the limited sample size.
By law, DHS is allowed to make mid-year corrections when enrollees' income declines, but is prohibited from doing so when income increases. As a result, discrepancies between MinnesotaCare income and actual income are less of a problem when MinnesotaCare income is too high. But, MinnesotaCare income was understated three times as often as it was overstated. To better ensure that MinnesotaCare income reflects actual income, the Legislature should amend the law to allow mid-year premium adjustments when income increases. Although DHS computer systems currently cannot match income reported to MinnesotaCare with income reported on tax returns and to the unemployment system, the department should ensure that this capacity is built into its new automated eligibility system. Data from these sources would provide workers with additional information, such as indicators of unreported income, to guide eligibility decisions. DHS is required to use electronic data as the primary means of verifying income, but how DHS will obtain data sufficient to do so is an open question.
Many Applicants Reported Incorrect Information About Availability of Insurance From Their Employers
MinnesotaCare relies largely on applicants' self-reporting their compliance with insurance-related eligibility criteria, but these reports are often unreliable. Based on a survey of enrollees and their employers in September 2002, 22 percent of the time, employers reported offering health insurance benefits to some or all of their employees, but enrollees reported that no benefits were offered. There was a greater degree of mismatch when we compared employer survey responses to what enrollees reported on their applications early in 2002. In this comparison, 52 percent of the time, enrollees did not flag possible employer insurance on the application or renewal form when the employer reported offering health insurance benefits.
Since it is important to restrict MinnesotaCare eligibility to those who do not have access to affordable private health insurance, DHS should reconsider the method by which insurance status is verified. For example, it could require employer verification of insurance status for all employed applicants. This would improve the accuracy of insurance information received by DHS, but would impose additional burden on staff, enrollees, and employers. This burden could be reduced by targeting mandatory verification to cases in which applicants are more likely to have access to insurance, such as those with relatively high wage income. In the future, computerized verification could increase compliance with insurance requirements with less administrative burden.
SUMMARY xi Mid-year income reviews may allow premiums to better reflect actual income.
Inaccurate information from applicants makes it harder for enrollment staff to assess insurance-related eligibility.
Oversight Was Not Sufficient to Ensure Accurate Eligibility Decisions
DHS uses a variety of means to control the accuracy of MinnesotaCare eligibility decisions, but these mechanisms were not sufficient to prevent the level of inaccuracy found in our review. Weaknesses included lack of refresher training for experienced workers, unclear policies, limited supervisory review of eligibility determinations, and use of application and renewal forms that leave gaps in the information enrollment staff need. By investing in a new automated eligibility system, DHS is taking an important step to provide better controls over eligibility determination. But, the new system alone may not be sufficient to ensure program integrity.
DHS also does not adequately test whether households enrolled in the program are eligible. DHS is required by law to use random audits to verify reported income and eligibility, but it did its last compliance audit specific to MinnesotaCare in 1995. Also, while county agencies assign staff to investigate fraud, DHS does not have a formal mechanism in place to detect or investigate allegations about MinnesotaCare applicants or enrollees whose cases are processed by DHS. As DHS managers acknowledge, the department should do more frequent compliance audits and should have procedures for identifying and dealing with applicant fraud and abuse.
DHS Recently Reduced Large Application Backlogs, but the Process Is Susceptible to Delays
Over the past four years, Minnesotans sending a MinnesotaCare application to DHS often had to wait more than 20 days for DHS to begin processing the application and 60 to 90 days in total for coverage to begin. MinnesotaCare workloads increased rapidly in recent years, and DHS responded by using additional funding to hire more staff, reassigning staff to handle peaks in new applications, and improving productivity. Recently, DHS reduced the time to begin processing applica-tions to less than one week. Yet, the underlying factors that have made the process susceptible to large backlogs are still present. These factors include problems attracting and retaining enrollment staff and heavy reliance on manual operations.
DHS is investing in a new electronic case management system that it hopes will provide quicker access to case files, more flexibility in assigning work, and more detailed performance data. We recommend several ways that DHS can capitalize on this project by improving the way it collects and uses performance data to manage the program.
DHS should tighten its eligibility procedures and more closely monitor the accuracy of information reported by applicants.
Rapidly growing workloads and paper-driven processes hinder timely processing of new applications.
Introduction According to recent census data, Minnesota has one of the lowest proportions of uninsured residents in the nation7.8 percent compared to a national average of 14.5 percent. 1 Public insurance plays an important role. According to the Minnesota Department of Health, about 9 percent of Minnesotans are insured through the state's three insurance programs for low-income residentsMedical Assistance, General Assistance Medical Care, and MinnesotaCare. 2 The 1992 Legislature created MinnesotaCare to ensure that health insurance was available to Minnesotans caught in the gap between other public health care programs and affordable private health insurance. MinnesotaCare enrollees must meet eligibility criteria related to income, access to other sources of health insurance, and other factors. They are also required to pay a monthly premium. Since 1992, MinnesotaCare enrollment has grown steadily, and the program has undergone significant changes in eligibility criteria and funding. Originally administered only by Minnesota's Department of Human Services (DHS), some counties started processing MinnesotaCare cases in 2000.
In April 2002, the Legislative Audit Commission directed us to evaluate certain aspects of MinnesotaCare. Legislators' questions centered on how well DHS and counties are determining enrollment eligibility and how efficiently they are processing applications. In addition to providing background information on MinnesotaCare eligibility criteria, program administration, and funding, our evaluation addressed two primary research questions:
To what extent do state and county agencies accurately determine MinnesotaCare eligibility and set premium levels?
How efficiently does the Department of Human Services process MinnesotaCare cases? To describe MinnesotaCare eligibility requirements, program administration, and funding, along with how these program elements have changed over time, we reviewed state laws, legislative research reports, information for applicants and enrollees, and other program documents.
To determine how accurately DHS and counties make eligibility decisions and set premiums, we reviewed a random sample of about 600 new and renewal applications submitted from January through March 2002. To assess how accurately income figures used by MinnesotaCare reflected actual income, we matched income figures used by DHS and counties with third-party sources of income information for a subgroup of the case file sample. For another subgroup of the case file sample, we examined how accurately enrollees self-report insurance information by surveying enrollees and their employers about access to employer health insurance benefits. Finally, we interviewed DHS and county officials and reviewed reports and other documents regarding oversight of the eligibility determination process.
MinnesotaCare is one of three health insurance programs for low-income Minnesotans.
1 This is a three year average for 1999 through 2001. U. S. Census Bureau, Health Insurance Coverage: 2001 (Washington, DC: U. S. Department of Commerce, 2002), 10.
2 Minnesota Department of Health, 2000 Minnesota Distribution of Insurance Coverage (St. Paul, 2002), 1.
To assess the efficiency of MinnesotaCare case processing, we collected and analyzed DHS data on processing times over the past four years. We also obtained available data regarding factors that influence processing times, such as workload and staffing trends, and interviewed DHS officials regarding their efforts to improve case processing.
We focused our study on the application and eligibility determination processes. Thus, several important issues were outside the scope of our review. For example, the program's insurance-related eligibility criteria were specifically designed to discourage individuals and employers from substituting MinnesotaCare for employer-based insurance. We did not assess the extent to which this substitution may be occurring. Also, because the Minnesota Department of Health has a number of research projects underway regarding Minnesota's uninsured population, we did not address the extent to which eligible Minnesotans are participating in MinnesotaCare. 3
The report is divided into three chapters. In Chapter 1, we describe eligibility rules, benefits, and application procedures and present data on enrollment and funding. Chapter 2 addresses the accuracy of MinnesotaCare eligibility decisions. Chapter 3 discusses how efficiently DHS processes MinnesotaCare cases. The Appendix describes in more detail our methodology for assessing the accuracy of eligibility decisions.
3 For example, see Minnesota Department of Health, Minnesota's Uninsured: Findings from the 2001 Health Access Survey (St. Paul, 2002).
Background SUMMARY MinnesotaCare is a subsidized health insurance program for lower-income Minnesotans. To be eligible for MinnesotaCare, applicants must meet a complex set of requirements regarding income, assets, access to other health insurance, residency, and citizenship. MinnesotaCare provides comprehensive medical, vision, and dental benefits through a managed care system. Most people apply for MinnesotaCare through a central office at the Department of Human Services, although residents of some counties may apply through their county social services office. In fiscal year 2002, an average of about 144,000 individuals were enrolled in MinnesotaCare each month, and total spending was about $390 million. MinnesotaCare is funded from three sources: a state tax on health care providers, federal matching funds, and enrollee premiums. Funding has not kept pace with expenditures, and if left unresolved, this imbalance may trigger mandatory actions to reduce MinnesotaCare spending.
With the creation of the Children's Health Plan in 1988, Minnesota became the first state to have a state-subsidized health insurance program to cover uninsured children. Soon after this program was enacted, the Legislature considered how to address the larger population of uninsured Minnesotans. MinnesotaCare, created in 1992 as part of a larger health insurance reform package, was designed to make health insurance available to Minnesotans caught in the gap between state health care programs for those with very low income and affordable private health insurance. 1 The target population was the working poor, including farmers, other self-employed individuals, and small business employees.
As background for our evaluation of MinnesotaCare, this chapter addresses the following questions:
What are MinnesotaCare eligibility criteria and benefits? How is the program administered and funded? To answer these questions, we reviewed state laws, legislative research reports, information for applicants and enrollees, and other program documents. We also interviewed officials from the departments of Human Services and Finance.
1 This bill also initiated health care cost containment strategies, medical malpractice reform, pools for private employers to purchase health insurance, and a specific funding source for the MinnesotaCare program. Laws of Minnesota (1992), ch. 549.
HEALTH CARE PROGRAM ADMINISTRATION MinnesotaCare is one of three comprehensive health care programs in Minnesota, each targeted at different subgroups of Minnesota's lower-income population and each with different eligibility criteria. The other two programs are Medical Assistance (MA) and General Assistance Medical Care (GAMC). MA is a joint federal-state health care program for low-income parents, children, elderly, or disabled people. GAMC is a fully state-funded program providing health care to Minnesotansprimarily childless adultswho do not qualify for MA or other state or federal health care programs. MA and GAMC cases are managed by county social service agencies.
Minnesota's Department of Human Services (DHS) administers MinnesotaCare. DHS is responsible for educating the public about MinnesotaCare, processing applications and determining eligibility, contracting with health care plans, managing premium payments and spending, and adopting rules to administer the program. County social service agencies also play a role by informing the public about the program and, in some counties, by processing MinnesotaCare cases. Eligibility is determined at the time of initial application and, once a participant is enrolled, is redetermined annually.
Historically, day-to-day administration of the application and renewal processes was handled through a centralized unit within DHS (currently called MinnesotaCare Operations). Beginning in January 2000, the Legislature allowed county social service agencies to process MinnesotaCare cases if they chose to do so. 2 This provision was added as an option to increase access to health care coverage through local enrollment and to help ensure continuous coverage for Minnesotans shifting from MA or GAMC (programs administered only by counties) to MinnesotaCare. An applicant or enrollee can, at any time, request that his or her file be transferred from DHS to an enrollment county or vice versa. As of October 2002, 34 counties were MinnesotaCare enrollment sites, but many applicants living in these counties chose to send their applications to DHS. DHS maintained 93 percent of cases active in October of 2002 compared with 7 percent maintained by counties.
Most MinnesotaCare cases are processed centrally at the Department of Human Services, but some Minnesota counties also serve as enrollment sites. Unlike other health insurance programs, which are administered by counties, DHS processes most MinnesotaCare cases centrally.
Whether an enrollee sends an application or renewal to a county or to DHS, the application procedure is basically the same. 3 Minnesota has one Health Care Programs Application that can be used to apply for MA, GAMC, or MinnesotaCare. MinnesotaCare applications may also be processed using forms for other social service programs, such as food stamps. Applicants are required to provide "proof" of income and immigration status, and DHS policy outlines the types of documentation that are acceptable. 4 By law, once an application is submitted, DHS and counties are supposed to determine an applicant's eligibility within 30 days of receipt. Once a decision is made, DHS sends a letter informing applicants of their eligibility. If eligible, the applicant selects a health plan, sends in the first premium payment, and coverage begins on the first day of the month after the premium is received.
Enrollees' eligibility needs to be renewed every 12 months. Several months before the renewal date, enrollees receive a Health Care Programs Renewal form that they must complete and submit along with required documentation of current income. If an enrollee has a lapse in coverage of one month or more, the enrollee must reapply and must meet all eligibility criteria. 5
ELIGIBILITY To be eligible for MinnesotaCare, applicants must meet a relatively complex set of criteria on income, assets, access to other health insurance, residency, and citizenship and follow other program rules. MinnesotaCare does not restrict eligibility on the basis of existing or previous health conditions. Coverage is available to families and to single adults and married couples who do not have children in their households, though eligibility rules vary by type of household. MinnesotaCare enrollees must pay a monthly premium for coverage and can be dropped from the program if they fail to pay it.
Understanding applicants' living situations is an important element of MinnesotaCare eligibility. Household composition and size affect whose income and assets count, the income standards that apply, and the premium amount to be paid. A MinnesotaCare household can be a single adult or a group of individuals who live together and who have a parental or marital relationship. This includes parents, spouses, stepparents, children, and dependent siblings. Table 1.1 shows how certain relationships are defined for the purposes of MinnesotaCare.
BACKGROUND 5 MinnesotaCare eligibility rules are complex.
2 Minn. Stat. (2002), §256L. 05, subd. 1. 3 Minn. Stat. (2002), §256L. 05. 4 Documents that can be used to prove income are discussed in more detail in Chapter 2. 5 Minn. Stat. (2002), §256L. 05, subd. 3.
Household composition can be difficult to discern given the complicated living arrangements in many households, and DHS policy outlines how many of these situations are to be handled. 6 Some individuals are to be excluded from the household. An emancipated minor 7 and his or her spouse and children must be a separate household even if living with the minor's parents. Similarly, a child over age 25 living with his or her parents, even if financially dependent, is considered a separate household. Adults or juveniles residing in a correctional or detention facility are not eligible for MinnesotaCare and are not counted in MinnesotaCare households. 8
Table 1.1: MinnesotaCare Definitions Term Definition Family (1) parents, their children, and dependent siblings residing in the same household; or
(2) grandparents, foster parents, relative caretakers, or legal guardians; their wards who are children; and dependent siblings residing in the same household. Child An individual under 21 years of age, including the unborn child of a pregnant woman, an emancipated minor, and an emancipated minor's spouse. a Dependent Sibling An unmarried child who is a full-time student under the age of 25 years who is financially dependent upon a parent, grandparent, foster parent, relative caretaker, or legal guardian and who has a sibling under age 21 living at home at the time of initial enrollment. Proof of school enrollment is required.
Household Individuals who live together and who have a parental or marital relationship. This includes parents, relative caretakers, spouses, stepparents, and children. A MinnesotaCare household can be a family, childless individual, or a married, childless couple.
a An emancipated minor is an individual under age 18 who is married or has been married, is serving in the military, or has been emancipated by court order.
SOURCES: Minn. Stat.( 2002), §256L. 01; and Minnesota Department of Human Services, Health Care Programs Manual( St. Paul, 2002), §§ 0902.11, 0908.
Eligibility decisions are made on an individual basis, but household composition affects income limits and premiums.
6 For example, except for temporary absences (such as time at camp or visits with noncustodial parents), a child must live in the household at least 50 percent of the time to be counted as a member of the household. An unborn child is counted in a pregnant woman's household size. Two unmarried adults living with a common child will be in one household, as would a married couple each with one child from a previous marriage living with them. Minnesota Department of Human Services, Health Care Programs Manual (St. Paul, 2002), §0908.03; http://www.dhs.state.mn.us/hlthcare/reportsmanuals/default.htm; accessed November 4, 2002.
7 An emancipated minor is an individual under age 18 who is married or has been married, is serving in the military, or has been emancipated by court order. DHS, Health Care Programs Manual, §0902.11.
Income and Assets For MinnesotaCare, income eligibility is based on the household's gross income. Income is defined to include earned income from wages or self-employment and unearned income such as social security benefits, child support, and unemployment benefits. Income limits are set in statute and are linked to federal poverty guidelines (FPG) published annually. 9 For families, gross household income cannot exceed 275 percent of FPG; for single adults and married couples without children, income cannot exceed 175 percent of FPG. 10 Children in families with incomes below 150 percent of FPG 11 and some children enrolled in Minnesota's original Children's Health Plan (a precursor to MinnesotaCare) are granted special eligibility status with more lenient eligibility rules. Table 1.2 shows income limits effective during fiscal year 2003 for families of various sizes and for childless adults.
If a household's income rises above income limits after initial enrollment, the household may be dropped from the program after an 18-month notice period. Before being dropped, the household's income is reassessed several times. Households with incomes that remain above the limit may stay in MinnesotaCare if their income is deemed to be too low to afford individually-purchased health care coverage. 12
Table 1.2: MinnesotaCare Income Limits, FY 2003 Maximum Annual Income Families Adults Without Children Household Size (275 Percent of FPG) (175 Percent of FPG)
1 $24,372 $15,516 2 32,844 20,904 3 41,316 Not Eligible 4 49,776 Not Eligible 5 58,248 Not Eligible 6 66,720 Not Eligible
NOTES: FPG is the federal poverty guideline. For illustrative purposes, this table shows income limits up to a household of six. The poverty guidelines may be applied to a household of any size. A family can have a household size of one if, for example, the application is for a foster child or an emancipated minor.
SOURCE: Department of Human Services MinnesotaCare income guidelines.
Income limits are linked to federal poverty guidelines.
8 DHS, Health Care Programs Manual, §0908.09; and Minn. Stat. (2002), §256L. 04, subd. 12. 9 Annual updates to the federal poverty guidelines go into effect on the date they are published in the Federal Register by the U. S. Department of Health and Human Services. For the 2002 update, see Annual Update of the HHS Poverty Guidelines, 67 Fed. Reg. 6931 (2002).
10 Children under age 21 are considered a family household regardless of whether they live with parents or a legal guardian, and they are subject to the income limit for families.
11 The threshold defining this group of lower-income children will increase to 175 percent of FPG on July 1, 2003. Laws of Minnesota (2002), ch. 220, art. 15, sec. 21-22.
As of July 1, 2002, MinnesotaCare eligibility includes an asset test, but the asset limits do not apply to pregnant women or children. In order to be eligible, a household of one person must not own more than $15,000 in total net assets, and a household of two or more persons must not own more than $30,000 in total net assets. Many assets can be excluded, including: a homestead, household goods and personal effects, assets owned by children, vehicles used for employment, court-ordered settlements up to $10,000, individual retirement accounts, and capital and operating assets of a trade or business up to $200,000. The law specifies that MinnesotaCare applicants self-report on applications and renewals whether they meet the asset requirement, although DHS may require applicants or enrollees to provide information verifying compliance if it has reason to believe that an applicant has assets over the limit. 13
Access to Other Insurance To prevent individuals and employers from dropping private health coverage in favor of MinnesotaCare, the program has several insurance-related eligibility requirements. With some exceptions, Minnesotans who have access to other health insurance coverage cannot enroll in MinnesotaCare. 14 More specifically, MinnesotaCare is not available to those who:
are currently enrolled in another health insurance plan, either through an employer or from another source (including Medicare);
had health insurance coverage in the four months immediately preceding enrollment; 15 have access to subsidized health insurance from a current employer; 16 or had access to subsidized insurance from a current employer in the last 18 months.
MinnesotaCare is intended for people without access to affordable private health insurance.
12 Households with incomes above the FPG limit may stay in MinnesotaCare if an amount equal to 10 percent of the household's annual gross income is less than the annual premium for the $500 deductible insurance policy offered by the Minnesota Comprehensive Health Association (MCHA)a state program offering insurance policies to Minnesotans who have been denied coverage by private carriers. For example, the annual MCHA premium as of July 2002 for two parents age 35 and two children is $6,416. If the family's income at renewal was above the $49,776 limit but below $64,160, the family would be able to retain MinnesotaCare coverage. Eligibility for this exception is reassessed annually. Minn. Stat. (2002), §256L. 07, subd. 1; and Department of Commerce, MCHA Rates, http://www.commerce.state.mn.us/pages/Insurance/InsMCHAQual.htm, accessed November 20, 2002.
13 Minn. Stat. (2002), §256L. 17. 14 Minn. Stat. (2002), §256L. 07. 15 While Medicare is considered "other insurance" under this and the current coverage provision, other government-sponsored health care programs, such as MA and GAMC, are not. But,
individuals cannot be simultaneously enrolled on MinnesotaCare and either MA or GAMC. Minn. Stat. (2002), §256L. 07, subd. 3.
16 Insurance coverage is considered subsidized if the employer pays 50 percent or more of the premium.
The law further clarifies the meaning of the 18-month provision by specifying that if a family or individual lost employer-subsidized coverage during that period because the employer stopped offering the insurance as a benefit, the individual or family is not eligible for MinnesotaCare. 17
Some children enrolled in Minnesota's original Children's Health Plan and children in families with income less than or equal to 150 percent of FPG have less stringent insurance eligibility rules. 18 Children in this group are not subject to the rules regarding access to employer-subsidized insurance. They may also be eligible for MinnesotaCare even if currently enrolled in another health insurance plan or if they had other insurance in the four months preceding application. This exemption applies if:
1. the other insurance lacks two or more of the following types of coverage: basic hospital, medical-surgical, prescription drug, dental, or vision;
2. coverage requires a deductible of $100 or more per person per year; or 3. coverage is limited because the child has exceeded the maximum coverage for a particular diagnosis or the policy excludes a particular diagnosis.
Other Eligibility Rules MinnesotaCare enrollees also must meet residency and citizenship requirements. 19 To be eligible, adults without children must be permanent residents of Minnesota. This means that they must demonstrate that they have a verified address (other than a place of public accommodation) and that they have lived in the state for the past 180 days. Pregnant women, children, and parents do not need to meet this durational residency requirement; rather, they must only demonstrate intent to reside in Minnesota permanently or for an indefinite period. MinnesotaCare is available to all citizens and many immigrants. Undocumented non-citizens and non-citizens who entered the country legally but for temporary purposes are ineligible for MinnesotaCare.
BACKGROUND 9 MinnesotaCare is available to families and individuals who meet eligibility criteria for income, assets, access to other insurance, residency, and citizenship.
17 The 18-month rule does not apply to individuals who were enrolled in MinnesotaCare, dropped coverage, then reapplied within six months and who no longer have employer-subsidized coverage due to the employer terminating health care coverage as an employee benefit. Minn. Stat. (2002), §256L. 07, subd. 2( b).
18 Minn. Stat. (2002), §256L. 07, subd. 3. This threshold is to increase from 150 to 175 percent of FPG on July 1, 2003. Laws of Minnesota (2002), ch. 220, art. 15, sec. 21-22.
19 Minn. Stat. (2002), §256L. 09; and DHS, Health Care Programs Manual, §§ 0906.03, 0906.05.
Applicants also must comply with a variety of other statutory provisions. For example, under the "all or nothing rule," families cannot choose to enroll only certain members of the family. 20 Under this rule, unless other insurance is available: parents who enroll must also enroll their children and dependent siblings (though children and their dependent siblings may enroll without their parents); if one parent in the household enrolls, both parents must enroll; if one child from a family is enrolled, all children must be enrolled; and if one spouse in a household enrolls, the other spouse must also enroll. Adults also must cooperate with the state in establishing third-party liability, paternity, and other medical support. 21 Applicants and enrollees are expected to identify other parties that may be liable for services being provided by MinnesotaCare and to help the state recover payment from them. Parents, guardians, and caretakers also must help DHS and county agencies establish the paternity of enrolled children and to obtain medical care support for them. Children will not be denied coverage if their parents or guardians fail to cooperate; only adults may be denied.
BENEFITS As shown in Table 1.3, MinnesotaCare pays for a broad range of basic health care services, including doctor visits, hospitalization, dental care, and vision benefits. MinnesotaCare enrollees do not pay deductibles, but some adults are subject to benefit limits and cost-sharing for certain services. Children and pregnant women have available a broader set of benefits that matches the benefit set available through MA. Added benefits include the following: access services, such as transportation to appointments and lodging; special education services; services provided at long-term care facilities; orthodontics; personal care attendants; case management services; and private duty nursing.
MinnesotaCare provides dental and vision benefits. Health care services under MinnesotaCare are provided through managed care plans.
20 Minn. Stat. (2002), §256L. 04, subd. 1( b). 21 Minn. Stat. (2002), §256L. 04, subd. 2.
MinnesotaCare enrollees receive health care services through managed care plans. The Department of Human Services contracts with health plan carriers who receive a fixed monthly payment per enrollee for providing a specified set of services. (This monthly payment, which does not vary according to the amount of services provided, is often referred to as a capitated payment.) The same set of managed care plans is used for MinnesotaCare, MA, and GAMC. As of January 1, 2003, DHS had contracts with seven health care plans, though all plans are not available in every county. Most MinnesotaCare enrollees can choose between at least two health care plans. In four counties, enrollees have only one plan available; in 30 counties, enrollees may choose among three, four, or five plans. Although all plans provide the same set of MinnesotaCare benefits, each plan has its own network of health care providers to deliver health services to plan members. These providers may include doctors, hospitals, outpatient centers, mental health clinics, and other specialized services. For those enrollees who
BACKGROUND 11 Table 1.3: MinnesotaCare Benefits Basic Services Ambulance for emergencies Chemical dependency treatment
Chiropractic care Doctor and health clinic visits Dental care preventive services (teeth cleaning, X-rays, oral exams) Emergency room Eye checkups and prescription eyeglasses (adults who are not pregnant pay $25 for glasses) Home care, such as a nurse visit or home health aide Hospice care
Hospitalization (see limits, below) Immunizations Laboratory and X-ray services Medical equipment and supplies Mental health services Prescriptions (adults who are not pregnant pay $3 per prescription) Rehabilitative therapy Additional Services for Children and Pregnant Women Case management Non-emergency medical transportation
Non-preventive dental care Nursing home or immediate care facilities Orthodontics Personal care attendants Private duty nursing Special education
Hospitalization Limits and Cost Sharing No dollar limit for children and pregnant women No dollar limit for adults who have a child under 21 in their home and whose income is equal to or less than 175 percent of the federal poverty guideline All other adults have a $10,000 limit per year per health plan Childless adults pay 10 percent of costs up to $1,000 per health plan per adult have more than one plan available, they may change plans once within the first year of enrollment and once a year thereafter during a 30-day open enrollment period.
SOURCE: Department of Human Services summary of MinnesotaCare benefits.
ENROLLMENT AND FUNDING MinnesotaCare enrollment and related health care expenditures have grown steadily since the program began. As shown in Figure 1.1, average monthly enrollment in MinnesotaCare grew from about 62,000 individuals in fiscal year 1994 to about 144,000 in fiscal year 2002an increase of about 130 percent.
While growth in enrollment among adult-only households (more than a 10-fold increase since this group was added to the program in 1995) has outpaced growth in enrollment for families (an 84-percent increase since 1994), most individuals enrolled in MinnesotaCare are children or parents. According to Minnesota Department of Health data, MinnesotaCare insures nearly 3 percent of all Minnesotans. 22
Spending for MinnesotaCare enrollees has grown with enrollment. As shown in Figure 1.2, total spending (adjusted for inflation) grew from about $60 million in fiscal year 1994 to about $390 million in fiscal year 2002. 23 Program administration costs accounted for a greater share of total spending in the earlier years, primarily because DHS needed to invest in changes to its health care payment system (the Medicaid Management Information System, or MMIS) to accommodate MinnesotaCare.
0 25,000 50,000 75,000 100,000 125,000 150,000
1994 1995 1996 1997 1998 1999 2000 2001 2002 Families Adult-only Households
Figure 1. 1: MinnesotaCare Average Monthly Enrollment, FY 1994-2002
Fiscal Year Number of Individuals
SOURCE: Department of Human Services data.
MinnesotaCare enrollment has grown steadily.
22 Minnesota Department of Health, MinnesotaCare Disenrollee Survey Report (St. Paul, 2002), 1. 23 Dollar values were indexed for inflation using the Bureau of Labor Statistics' consumer price index for all urban consumers (CPI-U). http://data.bls.gov/labjava/outside.jsp?survey=cu;accessed
August 27, 2002.
Most of the spending increase over time is explained by the growing cost of delivering health care benefits. As shown in Figure 1.3, health care costs per enrollee have increased by about 140 percent since fiscal year 1996. 24 Over that same time period, DHS administrative costs per enrollee grew by 25 percent. By comparison, average monthly enrollment grew by 59 percent.
MinnesotaCare is funded from three sources: state appropriations, federal matching funds, and enrollee premiums. The state share of MinnesotaCare funding comes from a tax on gross revenues of health care providers, hospitals, surgical centers, and wholesale drug distributors. The current tax rate is 1.5 percent of gross revenues, but, under current law, the rate is scheduled to increase to 2 percent on January 1, 2004. Also on that date, the state plans to reinstate a second health care taxequal to 1 percent of premiumson health maintenance organizations and other health service businesses. 25
BACKGROUND 13 0 50 100 150 200 250 300 350 400
1994 1995 1996 1997 1998 1999 2000 2001 2002 Health Care Payments MinnesotaCare Administration
Figure 1. 2: MinnesotaCare Spending, FY 1994-2002 Fiscal Year 2002 Dollars in Millions
NOTE: Dollar values were indexed using the Bureau of Labor Statistics' consumer price index for all urban consumers (CPI-U); http://data.bls.gov/labjava/outside.jsp?survey= cu; accessed August 27, 2002.
SOURCE: Office of the Legislative Auditor analysis of data from the departments of Human Services and Finance.
Growing health care costs per enrollee have driven growth in MinnesotaCare spending.
The state funds MinnesotaCare through a tax on health care providers, money from the federal government, and enrollee premiums. 24 We chose fiscal year 1996 as the base year for this analysis because MinnesotaCare administrative expenses were, relatively speaking, unusually high in fiscal years 1994 and 1995 as the program was initiated. In addition, adult-only households were not added to MinnesotaCare until 1995. 25 Prior to 1998, the provider tax rate was 2 percent, and the state had also implemented the 1 percent premium tax. Beginning in 1998, the provider tax rate was reduced to 1.5 percent, and the premium tax was suspended.
Minnesota receives federal funding for MinnesotaCare through two federal programsMedicaid (MA in Minnesota) and the State Children's Health Insurance Program (SCHIP). In 1995, the federal government approved a change to the rules governing Minnesota's MA program that allowed the state to receive federal funding for MinnesotaCare family coverage. Federal funding began in July 1995 for children and pregnant women and in March 1999 for parents, guardians, and relative caretakers. In 2001, the federal government approved federal SCHIP funding for MinnesotaCare parents, guardians, and relative caretakers in families with household income between 100 and 200 percent of FPG. In fiscal year 2002, Minnesota received federal funding for about 75 percent of enrollees.
MinnesotaCare enrollees pay premiums on a sliding scale, with the monthly premium amount determined by household size, the number of people enrolled, and income. 26 Premium amounts range from 1.5 percent to 8.8 percent of gross income, though premiums for children in families with income less than 150 percent of FPG 27 are set at a flat rate of $4 per month. Table 1.4 shows the fiscal year 2003 monthly premiums for two types of families at various income levels. For fiscal year 2003, the highest possible premium for families under the maximum FPG limit (for a family of five or more with at least three people enrolled) is $427 per month. 28
0 20 40 60 80 100 120 140 160
1997 1998 1999 2000 2001 2002 Health Care Costs Administrative Costs Enrollment
Figure 1. 3: Growth in Administrative and Health Care Costs Per Enrollee Since FY 1996
NOTES: Percentage change for costs were calculated using figures adjusted to 2002 dollars. Enrollment figures are based on the average number of individuals enrolled monthly.
Percentage change from 1996
SOURCE: Office of the Legislative Auditor analysis of data from the departments of Human Services and Finance.
Enrollees pay monthly premiums based on their income.
26 Minn. Stat. (2002), §256L. 15. Pregnant women and children under age 2 cannot lose MinnesotaCare coverage for failure to pay premiums.
27 The threshold defining this group of lower-income children will increase to 175 percent of FPG on July 1, 2003. Laws of Minnesota (2002), ch. 220, art. 15, sec. 21-22.
28 If a household has income above the maximum but meets requirements to remain on MinnesotaCare, it pays a higher, unsubsidized premium.
As shown in Figure 1.4, the state has always provided the largest share of funding for MinnesotaCare health care costs, but federal matching funds have become increasingly important. In fiscal year 2002, about one-third of MinnesotaCare health care funding$ 128 millioncame from the federal government. The state provided 55 percent and enrollees the remaining 10 percent, or $205 million and $39 million, respectively.
BACKGROUND 15 Table 1.4: Examples of MinnesotaCare Monthly Premium Amounts for Families at Various Income Levels, FY 2003 150 Percent FPG 200 Percent FPG 275 Percent FPG Monthly Monthly Monthly Monthly Monthly Monthly Household Income Premium Income Premium Income Premium
Parent and Child $1,493 $32 $1,990 $118 $2,736 $240 2 Adults, 2 Children 2,263 65 3,018 177 4,148 363
NOTE: Premium amounts assume that all household members are enrolled. SOURCE: Department of Human Services MinnesotaCare premium tables.
0% 20% 40% 60% 80% 100%
1994 1995 1996 1997 1998 1999 2000 2001 2002 State Appropriations Federal Funds Enrollee Premiums SOURCE: Office of the Legislative Auditor analysis of Department of Human Services data.
Figure 1. 4: Percentage of MinnesotaCare Funding by Source, FY 1994-2002
Fiscal Year Percentage The federal government has provided an increasing share of MinnesotaCare's funding.
29 Beginning in fiscal year 2000, state funding for pregnant women and children under age two was to be paid from the general fund as part of MA. This provision was repealed, effective January 1, 2003. Laws of Minnesota (2001), ch. 9, art. 2, sec. 76.
The state share of MinnesotaCare funding is provided through a separate budget account called the Health Care Access Fund (HCAF). By law, state expenditures for MinnesotaCare are limited to the funds available in HCAF. 29 In addition to appropriations for MinnesotaCare health care payments and program administration at DHS, the HCAF is also used to fund the Department of Revenue's administration of the provider tax and some Department of Health activities. If projected fund balances for the current and following biennium are not expected to cover projected expenditures, the Commissioner of Human Services must make program adjustments to limit expenditures after consulting with the chairs of the House Ways and Means Committee and Senate Finance Committee and the Legislative Commission on Health Care Access. 30 The law spells out the Commissioner's options along with the order in which they must be used. They are:
1. stop new enrollment of single adults and households without children; 2. upon 45 days' notice, stop coverage of single adults and households without children already enrolled in the program;
3. upon 90 days' notice, decrease the premium subsidy amounts by 10 percent for families with gross annual income above 200 percent of FPG;
4. upon 90 days' notice, decrease the premium subsidy amounts by 10 percent for families with gross annual income at or below 200 percent of FPG; and
5. require applicants to be uninsured for at least six months prior to eligibility in the program.
Also, the Commissioner may not hire new staff using HCAF appropriations until the Commissioner of Finance has determined that spending controls are sufficient to bring MinnesotaCare expenditures and revenues back into balance. If these measures are insufficient to limit expenditures to the estimated amount of revenue, the Commissioner must further limit enrollment or decrease premium subsidies. 31
When MinnesotaCare was a relatively new program, fund revenues exceeded expenditures, and the fund built a growing balance from 1993 to 1998. As shown in Table 1.5, though, annual expenditures exceeded revenues in fiscal years 1999, 2001, and 2002 and, based on a November 2002 forecast, are projected to do so through fiscal year 2005. To date, these deficits have not triggered the cost-saving measures described above because fund reserves have been sufficient to absorb annual losses. If projections hold true, however, fund reserves will be nearly depleted in fiscal year 2005, and the fund will run deficits of $81 million and $190 million in fiscal years 2006 and 2007, respectively. These funding imbalances may trigger mandatory actions to reduce spending.
According to the Department of Finance, growth in the provider tax base has not kept pace with growing MinnesotaCare enrollment and health care costs. In fiscal year 2002, MinnesotaCare spending accounted for about 95 percent of spending from the fund, so this program most influences the fund's solvency. From fiscal year 1998 (the year current tax rates went into effect) through fiscal year 2002, HCAF tax revenues increased by 23 percent while MinnesotaCare spending (including DHS program administration and direct health care spending) increased by 142 percent. The provider tax rate increase and addition of the premium tax in 2004 will not prevent the projected decline in the fund reserve. At the time we prepared this report, DHS said that the fund was balanced within the requirements of the law, but that future deficits could be addressed in a variety of ways, including restricting eligibility, increasing enrollee cost sharing, reducing benefits, or changing the program's revenue source.
MinnesotaCare funds are held in a separate budget account called the Health Care Access Fund.
The fund's reserves, built in earlier years, have been sufficient to absorb recent deficits.
30 Minn. Stat. (2002), §256L. 02. 31 Ibid.
HOW THE PROGRAM HAS CHANGED MinnesotaCare began as a relatively small, fully state-funded program to insure low-income children and their parents. In the years since, the Legislature has extended coverage to adults without children, expanded benefits, and modified eligibility requirements. In order to receive federal matching funds, the Legislature also enacted eligibility and benefit changes to meet federal Medicaid requirements. These changes, many of which were discussed in more detail above, are summarized in Table 1.6.
BACKGROUND 17 Table 1.5: Actual and Projected Health Care Access Fund Balances, Fiscal Years 1998-2005 Fiscal Year Totals (Dollars in Millions) Actual Projected 1998 1999 2000 2001 2002 2003 2004 2005 Balance Forward 261 303 274 298 287 247 151 60 Provider Tax 154 138 158 176 190 208 253 331 Gross Premium Tax 0 18 0 0 0 0 17 44
MinnesotaCare Enrollee Premiums a 15 21 29 22 28 24 29 33 Other Receipts b 5 1 16 14 (2) (6) (5) (9) Total Revenues 174 177 203 212 216 225 294 399 MinnesotaCare c 100 145 156 182 242 305 373 442 Other Uses d 31 62 16 40 15 16 12 12
Total Uses 131 207 172 222 256 321 385 454 Annual Surplus/( Deficit) 43 (29) 31 (10) (40) (96) (91) (55) Ending Balance 303 274 305 288 247 151 60 5
a Enrollee premium payments are allocated to state and federal shares. The Health Care Access Fund reflects only the state portion. b Includes federal matching funds for administrative costs and investment income offset by revenue refunds. c Includes payments to health care providers and DHS program administration.
d Uses vary by year, but include appropriations to the Department of Health, Department of Revenue, University of Minnesota, and the Legislature.
SOURCE: Department of Finance data.
Continuing deficits may trigger mandatory actions to reduce MinnesotaCare spending.
MinnesotaCare is undoubtedly more complex now than it was in 1992. As we discuss in more detail in chapters 2 and 3, this complexity has created administrative challenges for DHS and counties. Over time, MinnesotaCare has also been more closely integrated with MA and GAMC, though the transition from one program to another is not transparent or seamless for program participants.
Table 1.6: Major MinnesotaCare Program Changes, FY 1992-2003 Fiscal Year Program Changes 1992 MinnesotaCare Act passed in April 1992 1993 Coverage phased-in for parents and children with family incomes at or below 275 percent of FPG
Provider tax implemented at 2 percent rate 1994 No major changes 1995 Federal government approved federal Medicaid funding for families enrolled in MinnesotaCare
Added childless adults with household incomes at or below 125 percent of FPG 1996 Eligibility changes implemented to comply with federal rules Child age limit changed from 18 to 21 Expanded benefits for pregnant women and children Eliminated 180-day residency requirement for families Started receiving federal funding for pregnant women and children Verification of income and pregnancy required One percent gross premium tax implemented 1997 Childless adult income limit changed to 135 percent of FPG 1998 Childless adult income limit changed to 175 percent of FPG Provider tax rate reduced to 1.5 percent and gross premium tax suspended 1999 Started receiving federal funding for parents Expanded family definition to grandparents, relative caretakers, legal guardians, and foster parents
2000 Implemented policies to end coverage for those whose income rises above FPG limits Eliminated hospitalization cost sharing for parents with incomes at or below 175 percent of FPG and eliminated cost sharing for pregnant women and children 2001 Federal government approved asset test and additional federal funding for some parents
DHS and counties were directed to use electronic income verification 2002 Started receiving enhanced federal funding for parents with income between 100 and 200 percent of FPG
2003 Asset test implemented (applies only to adults who are not pregnant)
NOTE: FPG is the federal poverty guideline. SOURCE: Compiled by the Office of the Legislative Auditor.
To obtain federal funds for MinnesotaCare, the state had to change some eligibility criteria.
Accuracy of Eligibility Decisions SUMMARY In administering MinnesotaCare, the Department of Human Services (DHS) does not adequately ensure that state and county enrollment staff are making correct eligibility decisions. Two key eligibility criteriahousehold income and access to other insuranceare prone to error and inconsistency. Also, MinnesotaCare estimates of annual income frequently did not match income as reported on tax and unemployment records, often because individuals' income changed after MinnesotaCare staff determined their eligibility. As a result, Minnesota may have lost $5-22 million in premium revenue. DHS uses supervisory review, training, and other means to ensure the accuracy of eligibility decisions made by staff, but these efforts were not sufficient to prevent the level of inaccuracy we found in household income determinations. Further, DHS does not regularly monitor the accuracy of information provided directly by applicants. While no process for determining eligibility will eliminate all errors, effective targeting of MinnesotaCare toward the intended recipients requires that eligibility criteria be applied more consistently and correctly.
As Chapter 1 discussed, the Legislature established eligibility criteria and income-based premiums to target MinnesotaCare to lower-income individuals who do not have access to affordable health insurance and to control program cost. Accurately determining individuals' eligibility status and premium levels are important parts of achieving these objectives. This chapter examines MinnesotaCare's eligibility process, addressing the following question: To what extent do state and county agencies accurately determine MinnesotaCare eligibility and set premium levels?
Our study focused on income and insurance, two key aspects of determining eligibility. To evaluate how accurately DHS and counties determine eligibility, we reviewed a random sample of 594 new and renewal applications submitted from January through March 2002. The sample included cases processed centrally by the MinnesotaCare Operations unit at the Department of Human Services and cases processed by county social service offices. Through the file review, we assessed the extent to which MinnesotaCare enrollment representatives followed eligibility determination procedures as set forth in statute and the DHS Health Care Programs Manual. We also interviewed DHS and county officials regarding oversight of the eligibility determination process. 1 We did not evaluate compliance with asset rules because this eligibility test was implemented in July 2002, well after our review was underway.
We used two other methods to assess the accuracy of MinnesotaCare eligibility determinations. For a subgroup of the case file sample, we matched income figures used by DHS and counties with third-party sources of income informationemployer wage reports from the Department of Economic Security's unemployment system and state and federal tax return information from the Department of Revenue. This allowed us to make an overall assessment of how accurately income figures used by MinnesotaCare reflected actual income. For another subgroup of the case file sample, we examined how accurately enrollees self-report insurance information by surveying enrollees and their employers about access to employer health insurance benefits.
Most of the data reported in this chapter are estimates to the population of new and renewal applications based on the file review sample. To reflect the precision of each estimate, we also present a confidence intervala range of values within which we expect the actual value to usually fall. By convention, we used 95 percent confidence intervals. Over time, we would expect 95 percent of such confidence intervals to contain the actual value. Our methodology is described in more detail in the Appendix.
As a framework for interpreting our evaluation results, we considered the factors listed in Table 2.1. These factors reflect objectives set forth in statute as well as generally accepted criteria for assessing public subsidy programs. In some cases, there are tradeoffs among them. For example, how well the program restricts its enrollees to the targeted population must be balanced against the objective of having a streamlined application process. Throughout the chapter, we apply these concepts to help show the impact of problems with the eligibility determination process.
Table 2.1: Framework for Evaluating MinnesotaCare's Eligibility Process 1. Compliance: Eligibility decisions should comply with Minnesota statutes and DHS policy guidelines. We assessed the extent to which enrollment staff followed eligibility policy and the extent to which errors affected equity, targeting, and cost.
2. Equity: Equity should be considered in two ways. First, eligibility and premiums should be based on ability to pay. We assessed whether eligibility and premium decisions accurately distinguished between households with different incomes. A second aspect of equity is consistency. We assessed the extent to which staff made consistent eligibility decisions for people in similar circumstances. 3. Targeting: MinnesotaCare coverage should go to the people whom the program was designed to help. We evaluated the extent to which the eligibility determination process may result in coverage being denied to eligible individuals or granted to those not eligible.
4. Cost: The eligibility process should meet program goals while minimizing public cost. For example, we measured how discrepancies between income reported to MinnesotaCare and income reported to other sources affected premium revenue.
5. Streamlined Application Process: The application process should provide enrollment staff with the information they need to make eligibility decisions while minimizing delays and burden for applicants. In this chapter, we discuss the accuracy of information provided by applicants. We address delays in the application process in Chapter 3.
1 We interviewed MinnesotaCare directors from the following counties: Anoka, Carlton, Dakota, Itasca, Kandiyohi, and Morrison. These counties processed about half of the new and renewal applications received by counties between January and March 2002.
INCOME Household income is used to determine both eligibility for MinnesotaCare and the premium amount an eligible household is required to pay. DHS and county workers determine household income at the time a household initially applies for the program and again at the annual renewal. (Hereafter, we use the term "applicants" to refer collectively to new applicants and enrollees seeking renewal.) For MinnesotaCare, household income includes (1) earned income from wages or self-employment and (2) unearned income such as child support and unemployment benefits. Applicants are required to provide proof to verify their earned income but not their unearned income.
Law and DHS policy set forth the types of documentation that should be used to verify earned income and how workers should use it to calculate annual household income. As shown in Table 2.2, DHS has a standard form of proof for each type of earned income, but accepts some alternatives. According to DHS policy, when applicants submit pay stubs to verify wage income, enrollment workers are supposed to average gross pay per pay period and annualize by multiplying the average by the number of pay periods per year. 2 For example, a household member who is paid weekly should submit four recent pay stubs. The worker should average the gross pay amounts from these four pay stubs and multiply the average by 52 to determine the household member's annual wage amount. When a federal tax return is used to verify income, workers must start with the adjusted gross income amount, then add back certain deductions. 3 Workers are also supposed to adjust tax return information if it is not currentfor example, if it includes wages for a job no longer held or a business that has closed. If self-employed applicants either did not file a tax return or if the business is new or has undergone a substantial change, workers need to obtain other forms of documentation showing gross income and expenses to determine net self-employment income.
ACCURACY OF ELIGIBILITY DECISIONS Applicants must document how much income they earn.
2 The law requires wage income to be verified, but the specific procedures for doing so are set forth in DHS policy. 3 For self-employed individuals in a business other than farming, income should be calculated using adjusted gross income from the previous year's federal tax return and adding back in reported depreciation, carryover loss, and net operating loss amounts. For self-employment income from farming, income is also calculated using federal adjusted gross income as the base, but only depreciation is added back. Minn. Stat. (2002), §256L. 01, subd. 4.
In our review of MinnesotaCare case files, we evaluated the extent to which workers followed these procedures for determining household income. We focused primarily on how well workers determined earned income because this is the only type of income that must be verified. We grouped errors into three categories:
1. Use of insufficient documentation. The file did not contain sufficient proof of income with which to make an accurate income determination. By insufficient, we mean that the income verifications did not meet DHS policy. For example, we noted this error when pay stubs did not cover four weeks of pay, were not current, or did not clearly indicate the pay frequency. For self-employment income, we noted this error when, for example, the worker made an income determination without having necessary tax return schedules. In these cases, because file documentation was insufficient, we could not independently determine household income. When an applicant does not send in sufficient documentation, workers are supposed to request more information. 4
Table 2.2: DHS Policy for Proof of Earned Income Standard Proof of Income Type of Income and Method of Calculation Alternative Documentation Wage or salary income Gross pay from the past four weeks of pay stubs, averaged and annualized Tax forms and W-2s from most recent year, if reflective of current income Employer statement of current earnings Seasonal income Gross pay from the past four weeks of pay stubs, averaged and multiplied by the number of months the person expects to work
Tax form from most recent year if it reflects current seasonal earnings and unemployment insurance, if applicable Employer statement regarding past or anticipated earnings If income is new or changes each year, client's estimate of expected income Self-employment income Adjusted gross income plus certain deductions from most recent tax form and all related schedules If the business is new, has changed substantially, or if tax forms do not reflect current income, the following are acceptable: Financial statement or detailed records of gross receipts and expenses Quarterly reports Computer print-out of gross receipts and expenses Signed statement from an accountant verifying projected business income and expenses
SOURCE: Minnesota Department of Human Services, Health Care Programs Manual, (St. Paul, 2002), §§ 0904.13, 0911.09.03, 0911.09.09, 0911.11.
We assessed the extent to which staff followed procedures for determining earned income.
4 If the applicant does not send required proof of income but the application form contains enough information for the worker to reasonably estimate annual income, the worker can grant eligibility to the household for one month. Eligibility will continue after that month if the household provides the required proof of income. We found that workers rarely used this option.
2. Misapplication of policy. The worker failed to apply DHS policy in calculating annual income. For example, we noted this error when a worker used net pay rather than gross pay for wage income, used the wrong pay period frequency when annualizing income, used one pay stub rather than the average of those submitted, or failed to add back depreciation when calculating self-employment income. In these cases, we recalculated income according to DHS policy. 5
3. Simple Math or Transcription Error. These errors included cases in which a worker erred in summing the components of total household income or erred in copying a number.
Staff errors in following income determination policy may or may not affect an applicant's eligibility or premium. In those cases in which we were able to make an independent income determination (e. g., misapplication of policy or math errors), we estimated the impact on eligibility and premiums. We were not able to make independent income determinations for those cases in which workers relied on insufficient documentation, so the impact of these errors on eligibility and premiums is unknown.
Accuracy of Income Determinations Of the 30,606 cases represented by our sample of new and renewal applications submitted from January to March 2002, about 21,300 had at least one household member reporting earned income. Based on our file review:
Both DHS and counties frequently made errors when determining MinnesotaCare applicants' household income.
As shown in Table 2.3, we estimate that state and county workers erred in determining income in 32 percent of cases with earned income. 6 Incorrectly applying DHS policy for calculating income was the most common type of error, made in an estimated 18 percent of cases. Basing an income determination on incomplete or unclear documentation was the second largest source of error. Workers rarely made simple arithmetic or transcription errors.
Because the income calculation methods differ for self-employment and wage income, we discuss the specifics of these errors separately. Overall, the error rate among cases with self-employment income was higher than the error rate for cases with wage income. Workers made errors in an estimated 40 percent of 5,809 cases involving self-employment income, as shown in Table 2.4. Documentation and policy errors were almost equally prevalent. Most errors in documenting self-employment income resulted from calculating income without the appropriate federal tax schedules. Errors in determining self-employment income included the use of total income rather than adjusted gross income and the failure to add back depreciation.
ACCURACY OF ELIGIBILITY DECISIONS Enrollment staff did not follow income determination standards in nearly one-third of cases.
Error rates were higher for self-employment income than for wage income.
5 We redetermined income using documentation in the case file. We did not contact enrollees directly to reverify their income. 6 We calculated error rates separately for cases processed by counties and DHS. State workers erred in determining earned income in an estimated 32 percent of state cases while county workers made errors in an estimated 34 percent of county cases. The 95 percent confidence intervals are 27 to 38 percent for errors by state workers and 26 to 43 percent for errors by county workers. The difference in error rates between state and county workers was not statistically significant.
Workers made errors in an estimated 28 percent of 18,387 cases that required a wage income determination, as shown in Table 2.5. We estimate that workers misapplied policy in calculating income in 15 percent of these cases and used insufficient documentation in 10 percent of cases. For example, workers misapplied policy by using the wrong pay period frequency when projecting annual income, by using a single pay stub amount rather than the average of 4-weeks worth submitted, or by using net pay rather than gross pay. Errors in documenting wage income included annualizing income from pay stubs covering fewer than four weeks or using pay stubs that were not current. 7
Table 2.3: Errors Determining Earned Income Precision of Estimates (Confidence Interval a ) Estimated Estimated Number Percentage Number Percentage Type of Error of Cases of Cases of Cases of Cases All Errors Determining Earned Income 6,903 32% 5,834-7,971 28-38% Misapplication of Policy 3,749 18 2,888-4,609 14-22 Insufficient Proof of Income 2,821 13 2,050-3,591 10-17
Simple Math or Transcription Error 759 4 327-1,192 2-6
NOTE: Percentages are based on an estimated 21,300 cases received between January and March of 2002 in which at least one person reported earned income. Some cases had more than one error, so the number of cases and percentages do not add to the totals. a Calculated at the 95 percent confidence level (see the Appendix for further explanation).
SOURCE: Office of the Legislative Auditor review of MinnesotaCare case files.
Table 2.4: Errors Determining Self-Employment Income Precision of Estimates (Confidence Interval a ) Estimated Estimated Number Percentage Number Percentage Type of Error of Cases of Cases of Cases of Cases All Errors Determining Self-Employment Income 2,311 40% 1,717-2,904 30-50%
Misapplication of Policy 1,049 18 582-1,516 11-27 Insufficient Proof of Income 1,081 19 607-1,556 12-28 Simple Math or Transcription Error 315 5 45-585 2-12
NOTE: Percentages are based on an estimated 5,809 cases received between January and March of 2002 in which at least one person reported self-employment income. Some cases had more than one error, so the number of cases and percentages do not add to the totals. a Calculated at the 95 percent confidence level (see the Appendix for further explanation).
SOURCE: Office of the Legislative Auditor review of MinnesotaCare case files.
Problems obtaining and using federal tax returns led to many errors determining self-employment income.
While the majority of applicants with wage income submit pay stubsthe form preferred by DHS, some applicants documented wage income with tax returns. In the past, DHS requested tax returns as the preferred form of documentation, regardless of whether individuals earned wage or self-employment income. DHS switched to requiring pay stubs because the agency thought pay stubs would better reflect current earnings. To ease the burden on applicants, DHS still accepts tax returns if they reflect current employment. By accepting both forms of documentation, however, the period over which DHS measures income varies widely (the last 30 days or the preceding year).
The level of inaccuracy in determining income has important implications for program equity. In general, inconsistent and inaccurate application of DHS policy means that applicants in similar income situations are not treated equally, and eligibility determinations and premium amounts are not based on an applicant's ability to pay. In reviewing case files, we observed workers using different wage calculation methods on similar sets of wage documentation. Table 2.6 illustrates the impact of applying these different income determination methods to the same set of information. If a worker followed DHS policy, this individual's annualized income would be $19,474. If a worker used the year-to-date amount on the last paycheck of 2001, annual income would be $14,736. Using the average of the three pay stubs yields a higher annual income. 8 Because premiums are determined on a sliding scale, the premium owed for each of these income levels can be quite different. For example, if the pay stubs were for a parent and child household (both enrolled), the difference in premium based on the high and low income estimates is $780 per year. If the pay stubs were for a single adult household or a childless couple, the differences in the method used to calculate income would actually affect eligibility for the program. DHS policy sets clear income calculation standards, but other sources, such as training materials and case-by-base policy interpretations, discuss circumstances in which workers can deviate from the standard method. 9 Such ad hoc policy-setting has created more disparity in the methods used by workers. While we recognize that workers need some flexibility to react to applicants' unique circumstances, from a program equity and targeting viewpoint, these deviations from policy should be the exception rather than the rule.
ACCURACY OF ELIGIBILITY DECISIONS Table 2.5: Errors Determining Wage Income Precision of Estimates (Confidence Interval a ) Estimated Estimated Number Percentage Number Percentage Type of Error of Cases of Cases of Cases of Cases All Errors Determining Wage Income 5,122 28% 4,169-6,075 23-33%
Misapplication of Policy 2,775 15 2,021-3,530 11-20 Insufficient Proof of Income 1,815 10 1,187-2,443 7-14 Simple Math or Transcription Error 633 3 237-1,030 2-6
NOTE: Percentages are based on an estimated 18,387 cases received between January and March of 2002 in which at least one person reported wage income. Some cases had more than one error, so the number of cases and percentages do not add to the totals. a Calculated at the 95 percent confidence level (see the Appendix for further explanation).
SOURCE: Office of the Legislative Auditor review of MinnesotaCare case files.
Staff often did not follow established procedures in projecting applicants' annual wage income.
7 We actually found more inconsistency in workers' use of pay stubs than is reflected in Table 2.5. In many cases, applicants provided more or less than the four consecutive weeks of current pay stubs required. While DHS policy establishes a standard method for calculating wage income from pay stubs, training materials and other informal communications instruct workers to use their judgment and deviate from the standard if the situation warrants it. We estimated that in 34 percent of cases with wage income, workers deviated from the standard method of calculating income from pay stubs. The 95 percent confidence interval for this estimate is 29 to 40 percent. If the worker documented why the deviation was justified and made a reasonable income determination, we did not treat the deviation from the standard as an error.
Table 2.6: Impact of Different Wage Calculation Methods on MinnesotaCare Eligibility and Premiums To understand how different methods of using pay stub information affect eligibility and premiums, assume that a MinnesotaCare applicant submits three bi-weekly pay stubs with the following information: Pay Stub 1 Pay Date: 12/7/01 Gross Pay: $1,028 Year-to-Date Gross: $13,708 Pay Stub 2 Pay Date: 12/21/01 Gross Pay: $708 Year-to-Date Gross: $14,736 Pay Stub 3 Pay Date: 1/4/02 Gross Pay: $790 Year-to-Date Gross: $790
State and county workers used pay stub information in different ways to estimate annual income, as illustrated in the table below. DHS Policy: Average Two Average Three Most Recent Pay Stubs Use 2001 Pay Stubs Submitted Year-To-Date and Annualize and Annualize Amount Average Bi-weekly Gross Pay $ 749 $ 842 N/A Annual Income 19,474 21,892 $14,736
The table below shows how these different income calculations would affect eligibility and premiums for different types of households. Annual Household Premium if: Income is Income is Income is $19,474 $21,892 $14,736 Parent and Child, Both Enrolled $ 744 $ 1,056 $ 276 a Two Adults and One Child, All Enrolled 456 b 600 b 276 b
Single Adult Not Eligible Not Eligible 564 Childless Couple, Both Enrolled 744 Not Eligible 456
a For a family of two, this income amount is below 150 percent of the federal poverty guidelines. As a result, the child's premium is fixed at $48 per year while the parent's premium is based on the sliding scale.
b For a family of three, this income amount is below 150 percent of the federal poverty guidelines. As a result, the child's premium is fixed at $48 per year while the parents' premiums are based on the sliding scale.
SOURCE: Office of the Legislative Auditor file review data and application of DHS' premium calculation algorithm.
Errors and inconsistency in determining income raise concerns about equal treatment of applicants.
8 By DHS convention and according to MinnesotaCare training materials, if an applicant provides pay stubs covering more than four weeks, workers are to use the extra stubs only if it is to the applicant's advantage. In our example, using all three pay stubs would not be to the applicant's advantage.
For cases in our file review, we evaluated the extent to which errors determining income affected eligibility status and premium amounts. We limited this analysis to those cases in which the file contained income documentation sufficient for us to make an independent income determination. Our analysis showed that:
Among our file review cases, income determination errors resulted in many enrollees paying the wrong premium and, in a small proportion of cases, incorrect eligibility decisions. Income errors often resulted in enrollees paying the wrong premium because premiums are calculated on a sliding income scale in which a small change in income can affect the premium amount. Among cases in which we found errors and were able to independently determine income, the income error resulted in a premium difference in 63 percent of cases. 10 The differences went in both directions, and the average difference between the premium determined by DHS and our premium calculation was $295 per year. 11
Income determination errors affect eligibility for MinnesotaCare only when the household's income is near the income limit and the error is large enough to move the household above or below that limit. We estimate that worker errors in determining income affected eligibility in 4 percent of cases where sufficient proof of income was provided and an error was found. 12 This figure may underestimate the impact of income errors on program eligibility decisions because we excluded from our analysis errors relating to insufficient income documentation because we were not able to independently determine income.
Workers' errors in calculating income result from the combination of failure to apply policy, lack of clarity in that policy, and inadequate program oversight. We discuss these factors later in the chapter as part of a broader discussion of program oversight.
ACCURACY OF ELIGIBILITY DECISIONS Errors calculating income led to incorrect premiums, with some premiums set too high and some too low.
Income errors affected program eligibility in only a few cases.
9 In addition to training materials, DHS also has an e-mail system in which county workers can ask DHS for guidance on the application of policy in specific cases. The Director of MinnesotaCare Operations feels that, without a clear policy on when deviations from the standard method are appropriate, case-by-case decisions can lead to greater inconsistency in the methods used.
10 The 95 percent confidence interval for this estimate is 50 to 74 percent. For this analysis, we included cases in which the household (1) submitted a new or renewal application between January and March 2002, (2) received earned or unearned income, (3) was approved for the program, and (4) had an error in income determination. The difference in premium must have equaled $1 or more to be counted as a change. 11 The 95 percent confidence interval for this estimate is $152 to $438. The net effect of the difference in premiums is unclear because the estimates for premium increases and decreases have large margins of error. 12 The 95 percent confidence interval for this estimate is 1 to 12 percent. For this analysis, we included all cases with earned or unearned income. In our sample, we found income errors that altered eligibility status in five cases. Of these cases, we found that three were approved for coverage in error, and two were denied in error.
Comparison With Actual Income In the previous section, we focused on how well workers followed DHS eligibility policies when measuring income. In this section, we take a broader look at the accuracy of income determination by comparing it with actual income from tax return and unemployment system data. How well MinnesotaCare targets its resources and how equitably it treats its applicants and enrollees depend on how accurately it measures household income for the full year of enrollment. Under DHS policy, wage income is typically based on earnings during a four-week period, but DHS uses this income to set premiums for the next 12 months. For cases that have self-employment income, MinnesotaCare typically bases income on tax returns from the previous year and, in some cases, from two years back. To find out how well MinnesotaCare's method reflected actual income during the time people were enrolled in the program, we compared MinnesotaCare's income determination in early 2001 with actual income for 2001. By actual income, we mean income reported on 2001 state and federal tax returns from the Department of Revenue and 2001 wage data reported by employers to the Department of Economic Security's unemployment system.
Our income comparisons are based on a subset of our file review sample. We restricted our analysis to cases for which we were able to obtain third-party income data and that were on MinnesotaCare from early 2001 through the end of the year. While we believe that our comparisons are generally reliable, our results may understate the extent to which MinnesotaCare income is less than actual income because neither the wage data from the unemployment system nor the tax data are complete. 13
Income Comparisons Based on our analysis of how earned income determined by MinnesotaCare compared with actual income during 2001:
MinnesotaCare estimates of annual income frequently did not match income reported on tax returns or to the unemployment system, often because individuals' income changed after MinnesotaCare staff determined their eligibility.
As shown in Table 2.7, for the eligibility year beginning in early 2001, we estimate that actual income exceeded MinnesotaCare's income determination by more than $5,000 in about 27 percent of matched cases. At the same time, MinnesotaCare's income determination exceeded actual income by over $5,000 for nearly 10 percent of matched cases. Thus, while there are differences in both directions, MinnesotaCare's income determination understated income by at least $5,000 about three times as often as it overstated income by this amount.
Most of the income discrepancies we found involved wage income. Wages and unemployment insurance benefits explained the entire discrepancy in an estimated 74 percent of cases with a discrepancy exceeding $5,000. Self-employment income explained the entire discrepancy in about 9 percent of these cases. A combination of wages, unemployment benefits, and self-employment explained the discrepancy in the other 17 percent of these cases. 14
DHS typically uses a four-week snapshot to project annual income.
MinnesotaCare's income estimate often differed from actual income by $5,000 or more.
13 Limitations associated with these data sources are discussed in more detail in the Appendix.
One reason for these income discrepancies is that actual income frequently fluctuates during the year, but DHS typically projects income for MinnesotaCare households based on a short time interval. In fact:
Seasonal income variation explained about half of the large discrepancies that we found in wage income.
In about half of the cases with large wage discrepancies, households that applied for (or renewed) MinnesotaCare in the first quarter of 2001 had a substantial change in income later in the year. 15 For many of these cases with a substantial change, MinnesotaCare's income determination reflected the actual income during the quarter the application was made but not for the entire year. For example, one MinnesotaCare household earned $2,212 during the first quarter of 2001, but averaged $8,119 per quarter in the last three quarters of 2001. As a result, when the household applied in early 2001, DHS underestimated their annual income for 2001 by more than $10,000 because it based their income on four weeks of income during the first quarter of 2001.
ACCURACY OF ELIGIBILITY DECISIONS Table 2.7: Income Discrepancies Between MinnesotaCare's Income Determination and Data From Tax Returns and the Unemployment System Estimated Precision of Estimates Percentage (Confidence Interval a ) of Cases Percentage MinnesotaCare's Income Determination ( N= 215) of Cases
Exceeded actual income by: More than $5,000 10% 6-15% $1,000 -$4,999 16 11-22 Was within $1,000 of actual income 24 18-32 Was less than actual income by: $1,000 -$4,999 23 18-30
More than $5,000 27 20-34 Total 100%
NOTE: Actual income refers to earned income reported to the Minnesota Department of Revenue on 2001 state or federal tax returns or wage income during 2001 reported by employers to the Department of Economic Security's unemployment system. a Calculated at the 95 percent confidence level (see the Appendix for further explanation).
SOURCE: Office of the Legislative Auditor analysis of MMIS data from the Department of Human Services, MinnesotaCare case files, income tax data from the Department of Revenue, and wage detail data from the Department of Economic Security's unemployment reporting system.
Income discrepancies often occurred because wage income substantially changed later in the year.
14 The 95 percent confidence intervals are 61 to 84 percent for the 74 percent estimate, 4 to 19 percent for the 9 percent estimate, and 10 to 30 percent for the 17 percent estimate. These estimates are based on income discrepancies exceeding $5,000 in either direction. 15 The estimate is 48 percent with a 95 percent confidence interval of 34 to 61 percent.
In the other half of cases with large wage discrepancies, MinnesotaCare's income determination neither reflected actual income in the first quarter nor the rest of the year. In these cases, the discrepancy may be due to a variety of factors, including errors by MinnesotaCare workers or underreporting of income by applicants. Or, it may be that the time period used to determine income (typically four weeks) neither reflected quarterly nor annual income. We could not determine the relative size of these factors because we could not measure actual income on a monthly basis nor could we measure underreporting of income.
While income fluctuations in either direction can result in enrollees paying the wrong premium, DHS and counties by law are allowed to change MinnesotaCare premiums between annual renewals only if income declines. 16 Enrollees may call their MinnesotaCare worker at any time after the annual income determination to report a drop in income and request a premium change. In contrast, if household income increases, DHS waits until the next annual renewal before adjusting the premium regardless of when it discovers the increase. To determine how often premiums are adjusted before the regularly scheduled renewal, we identified cases from our sample in which actual income was lower than MinnesotaCare's determination and the lower income would have resulted in a monthly premium difference of at least ten dollars. We found that in about one-fourth of these cases, the enrollees reported that their income declined and DHS reduced the premium prior to the next renewal.
Premium Impact To assess how income discrepancies affected aggregate collection of MinnesotaCare premiums, we compared the premium set by DHS in early 2001 with premiums based on actual earned income for 2001. Overall, 23 percent of cases had a premium impact of over $50 per month, including 20 percent with premiums that were set too low and 3 percent with premiums that were too high. 17
To estimate the aggregate annual impact on MinnesotaCare premiums for the entire caseload, we assumed that the average impact for cases in our sample was the same as it was for other MinnesotaCare cases. We found that:
MinnesotaCare may have lost about $14 million in premium revenue because of income discrepancies.
We estimate that the program's net loss in premiums due to differences in MinnesotaCare determinations and actual income was about $14 million, or about 5 percent of fiscal year 2001 program spending. For cases in which MinnesotaCare underestimated income, we estimate that MinnesotaCare lost about $20 million in premiums during the year. For cases in which MinnesotaCare overestimated income, we estimate that MinnesotaCare took in roughly $6 million in excess premiums. These estimates may understate the net premium loss to the program because they do not reflect the fact that enrollees may adjust their premium when their income declines. 18
By law, premiums can be lowered, but not increased between annual renewals.
16 Minn. Stat (2002), §256L. 15, subd. 2( a). DHS managers said that the prohibition against mid-year premium increases was included in the law as a way to provide enrollees with a set premium for a year and to encourage stable enrollment in the program.
17 The 95 percent confidence intervals are 17 to 29 percent for the 23 percent estimate, 14 to 26 percent for the 20 percent estimate, and 1 to 7 percent for the 3 percent estimate. The accuracy of MinnesotaCare's premiums varied widely in our sample, ranging from being too low by $563 per month to being too high by $599 per month.
These estimated premium impacts have large margins of error due to our limited sample size. For example, the margin of error for our $14 million estimate was about $8 million. 19 Another source of uncertainty is that we assumed that households that applied for MinnesotaCare in the first quarter had the same income discrepancies as households that applied during the rest of the year. While it is reasonable to assume that the problems we found for cases that applied for renewal during the first quarter exist in other cases, it is not clear whether they occur to the same extent. Also, the accuracy of income estimates may vary from year to year because of changes in the economy. MinnesotaCare income estimates are more likely to understate future income during periods of economic expansion than during periods of economic decline.
Finally, our estimates may be affected by how enrollees would respond if premiums were increased based on higher income. It is not clear how many enrollees would pay the higher premium, transfer to Medical Assistance (MA), switch to private insurance, or become uninsured. Our estimates assume that enrollees would remain in the program at the higher premium level, even if the actual income places the enrollee over the income limit. 20 To the extent that enrollees switch to MA, the state would lose premium revenue. DHS contends that because many MinnesotaCare enrollees are eligible for MA, the apparent gains from raising premiums may often not be realized. However, the enrollees included in our analysis who would be charged higher premiums also had incomes substantially higher than the average MinnesotaCare household. As a result, they would be less likely to meet MA income limits. Also, if enrollees reacted to higher premiums by obtaining private insurance or becoming uninsured, the program savings may be higher than estimated. 21
ACCESS TO OTHER HEALTH INSURANCE With some exceptions for very low-income children, MinnesotaCare is not available to residents who have access to other health insurance coverage. As discussed in more detail in Chapter 1, MinnesotaCare is not available to those who, at the time of application, have health insurance, have access to subsidized insurance from an employer, had access in the past 18 months to insurance subsidized by an employer, or had insurance coverage in the past four months. 22
ACCURACY OF ELIGIBILITY DECISIONS Mid-year income adjustments may allow premiums to better reflect actual income, but the full impact on program costs is unclear.
18 As discussed in the appendix, we do not have a reliable estimate of how much this affects premiums because of our small sample size. 19 The 95 percent confidence intervals are $13 million to $27 million for the $20 million loss in premiums; $5 million to $22 million for the net impact on premiums, and $1 million to $11 million for the $6 million estimate of excess premiums. 20 As discussed in Chapter 1, some enrollees may remain on the program paying an unsubsidized premium for 18 months after DHS notifies them that their income exceeds the limit.
21 See the Appendix for a fuller discussion of factors that affect the accuracy of our estimates. 22 Insurance is subsidized if the employer pays at least 50 percent of the cost of coverage. Minn. Stat. (2002), §256L. 07, subd. 2( c).
Insurance-related eligibility is largely self-declared. The health care application and renewal forms ask questions, listed in Table 2.8, related to these insurance criteria. MinnesotaCare workers are supposed to follow up only when an applicant gives a positive response to one of these questions or when the worker has conflicting information regarding insurance (for example, the worker sees a health insurance deduction on a pay stub, but the applicant did not report health insurance coverage on the application). If more information is required, the worker can contact the applicant by phone, mail a form requesting health insurance information, or send a form about employer-subsidized insurance that applicants are to give their employers.
In our review, we examined (1) the accuracy of self-reported information about insurance, (2) workers' efforts to pursue potential insurance barriers and their ability to determine whether insurance available to applicants is employer-subsidized, and (3) the extent to which DHS policy is consistent with statutory requirements for insurance-related eligibility criteria.
Accuracy of Self-Reporting Because applicants self-report insurance information, enrollment workers' ability to accurately assess insurance-related eligibility depends upon the extent to which applicants flag possible problems on their applications. In some cases, self-reported information from applicants accurately identifies insurance-related eligibility barriers. DHS has analyzed data on the reasons applicants are denied MinnesotaCare coverage. According to DHS data for fiscal year 2002, 46 percent of individuals denied coverage were denied for insurance-related reasons. In these cases, applicants informed workers about insurance-related information necessary to determine their eligibility.
Table 2.8: Insurance Questions on the Application and Renewal Forms Application Form a Is anyone covered under health insurance, Medicare, or prescription drug coverage?
Has anyone had health insurance or prescription drug coverage in the past four months?
Is anyone working for an employer who offers health insurance or has offered it in the past?
Renewal Form Does the employer of any family member offer health insurance?
Tell us if you or anyone in your family had changes occur in the past year or if changes are expected to occur for health insurance.
Tell us if you or anyone in your family had changes occur in the past year or if changes are expected to occur for Medicare benefits.
a The application form directs applicants to complete an attached form with more detailed insurance information if they answer yes to any of the three questions listed.
SOURCES: Department of Human Services, Minnesota Health Care Programs Application( July 2002); Department of Human Services, Minnesota Health Care Programs Renewal Form( April 2002).
DHS relies primarily on applicants to report whether they have access to private insurance.
But, our review of case files and survey of employers of MinnesotaCare enrollees suggest that more insurance-related eligibility issues may be going undetected. Our analysis showed that:
Self-reports regarding insurance-related eligibility are not sufficiently reliable to ensure that workers have information needed to follow up on potential insurance barriers. Based on our file review, households submitting new or renewal applications answered "yes" to an insurance question in an estimated 32 percent of cases. As stated above, workers are supposed to rely on insurance information declared on the application unless conflicting evidence is present. In an additional 12 percent of our file review cases, workers used a source other than the application and found information leading them to believe that an applicant might have access to other insurance. 23 For example, the worker may have checked computer systems for other human services programs for information on private insurance or may have found conflicting information in the application materials, such as a payroll deduction for health insurance. In these cases, applicants should have responded positively to one or more of the insurance questions but did not.
In addition to reviewing case files, we independently tested the accuracy of enrollees' responses to questions about insurance available from employers by surveying MinnesotaCare enrollees and their employers. 24 We asked both enrollees and their employers the following three questions regarding the enrollee's insurance status: (1) whether the employer offered insurance to any of its employees, (2) whether the enrollee was eligible for insurance benefits, and (3) whether the enrollee or any family members were actually enrolled in the employer's health insurance plan. According to the director of MinnesotaCare Operations, 25 DHS wants applicants to report potential insurance issues in the broadest sense, as in the first question. Workers will then probe for more detailed information on eligibility and enrollment.
The majority of employers responding (74 percent) reported that they offered health insurance benefits to their employees. Thirty-four percent stated that the MinnesotaCare enrollee we were asking about was eligible for those benefits, and 18 percent of employers indicated that at least one person in the enrollee's household was enrolled in a health care plan. The majority of employers (85 percent) offering employee-only coverage provided coverage that was employer-subsidized. In contrast, only 48 percent of employers offering family coverage provided coverage that was subsidized. 26
Some enrollees misreported information related to their insurance status on our survey. Twenty-two percent of the time, employers surveyed reported that they offered health insurance benefits, but the enrollees reported that no health insurance benefits were offered. Of employers who offered health insurance as a benefit, 9 percent reported that the enrollee was eligible for benefits when the enrollee indicated that he or she was ineligible. Of employers who offered health insurance benefits for which the enrollee was eligible, 12 percent stated that someone in the enrollee's household was enrolled in the employer's health plan when the enrollee indicated that no one in the household was enrolled. 27
Applicants often misreport their access to private insurance.
23 The 95 percent confidence intervals are 28 to 36 percent for the 32 percent estimate and 9 to 15 percent for the 12 percent estimate. 24 Our survey methodology is discussed in more detail in the Appendix. 25 MinnesotaCare Operations is the DHS unit responsible for processing the vast majority of MinnesotaCare cases.
26 The 95 percent confidence intervals are 68 to 80 percent for the 74 percent estimate, 27 to 42 percent for the 34 percent estimate, 12 to 24 percent for the 18 percent estimate, 71 to 92 percent for the 85 percent estimate, and 33 to 63 percent for the 48 percent estimate.
There was a greater degree of mismatch between employer and employee when we compared employers' responses on the survey to information provided by enrollees on their MinnesotaCare new or renewal application forms. 28 Fifty-two percent of the time, employers reported in our survey that they offered insurance benefits, but the enrollee did not flag possible employer insurance on the MinnesotaCare application or renewal form. Applicants may believe that they only have to answer yes on application insurance questions if they are actually eligible for the employer's health insurance. Even using this more narrow interpretation, 24 percent of employers responding to the survey indicated that an enrollee was eligible for health insurance when the enrollee did not answer yes to any insurance questions on the application or renewal forms. 29 The larger degree of mismatch between employer reports and application information may be due to timing problems. Because of waiting periods and other employer eligibility rules, employees may not have been eligible for employer health insurance benefits in January of 2002 but were eligible at the time of our survey in the fall of 2002.
Many applicants misreport information on insurance available from their employers, making it more difficult for MinnesotaCare staff to detect insurance-related eligibility barriers. 27 The 95 percent confidence intervals are 16 to 29 percent for the 22 percent estimate, 5 to 15 percent for the 9 percent estimate, and 5 to 26 percent for the 12 percent estimate. 28 We restricted this analysis to cases in which at least one household member worked for the same employer at the time of their 2002 application (sometime between January and March) and at the time of the survey (September and October 2002). 29 The 95 percent confidence intervals are 42 to 61 percent for the 52 percent estimate and 16 to 35 percent for the 24 percent estimate.
Enrollees may also have been more accurate on our survey than on their application or renewal form because our questions were worded more clearly and were linked to specific employers. It is also possible that employers misreported insurance information on our survey.
We did not have sufficient information to determine whether enrollees misreporting insurance information would have been ineligible on the basis of insurance criteria. However, the application questions clearly do not reliably identify applicants who might face insurance eligibility barriers. While the survey responses indicate that some MinnesotaCare enrollees do not accurately report their access to employer-based health insurance, the survey does not provide insight on enrollees' intentsome may simply not have been aware of the benefits offered while others may have intentionally misreported their insurance status. Regardless of intent, problems with the accuracy of self-reports undermine a key program goaltargeting of MinnesotaCare to those who do not have access to other coverage.
DHS has considered various options for verifying insurance information. For example, it could require employer verification of insurance status for all employed applicants. This option would improve the accuracy of insurance information received by DHS but would also impose an additional burden on workers and employers. This burden could be reduced by targeting mandatory verification to cases in which applicants are more likely to have access to insurance, such as cases in which individuals have relatively high wage income. However, this method may raise equity concerns.
In the future, computerized verification could increase compliance with insurance requirements with less administrative burden on applicants and workers. One option, according to DHS, is to modify a child support enforcement program that requires all employers in the state to report to DHS when they hire a new employee. Employers could also be required to report whether they offer health insurance benefits to any employees. These data could be used to build a statewide database of employers offering health insurance benefits that workers could use as the basis of follow-up regarding individual MinnesotaCare applicants. Such a system is not currently feasible because the computer system used for MinnesotaCare is not linked with the system containing the information on new employment. A second option is to contract with a private firm to build and maintain a database of Minnesota employers offering health insurance benefits.
On a smaller scale, DHS could improve the questions about insurance asked on the application and renewal forms. As discussed above, enrollees answered questions about their insurance status on our survey more accurately than they did on their application or renewal forms. Our survey questions may have been worded more clearly. In addition, our survey asked questions about eligibility and enrollment specific to a particular employer, rather than a general question about insurance offered by employers. 30 Similar changes to application and renewal forms may result in more accurate self-reported insurance information. DHS could also add more insurance questions to the forms. This choice may provide workers with more accurate insurance information; however, it would lengthen the application form, perhaps deterring some from applying for the program.
DHS needs to find better ways to verify insurance availability.
30 DHS audited the accuracy of self-reported insurance information in 1995 and also found that applicants were more accurate when asked about insurance available from a specific employer. Minnesota Department of Human Services, MinnesotaCare Quality Assurance Status Report (St. Paul, 1995), p. 3.
Problems Assessing Insurance Eligibility Although we reviewed insurance-related information for each of our file review cases, we were not able to systematically assess the accuracy of workers' decisions regarding insurance eligibility. Our review was limited because verifications of insurance are generally not required, and because workers only have to document in case files a yes or no decision on whether applicants meet insurance eligibility criteria. They are not required to document the evidence used to support their decisions. Still, we observed that:
Workers pursue possible insurance issues with different levels of effort, and when applicants report that an employer may offer insurance, information on whether that insurance is employer-subsidized is difficult for enrollment staff to obtain and interpret. Workers vary in the extent to which they probe for potential insurance eligibility barriers. DHS policy directs workers to rely on insurance information provided on the application unless the information conflicts with (1) other information in the case file or (2) information commonly known to workers in the agency (e. g., that it is commonly known that an employer offers health insurance benefits). 31 During our file review, we found that some workers sought mandatory verification even when no conflicting evidence was present. For example, some workers required verification when an individual reported changing employers even if the application did not indicate any potential insurance barriers. This proactive probing for insurance verifications violates DHS policy. Additionally, the use of information "commonly known to workers" is not consistent among DHS and counties. DHS policy states that information must be available to all workers in the agency and recommends that agencies develop lists of employers known to offer employer-subsidized insurance. Some, but not all, counties keep such lists. Contrary to its own policy, DHS does not provide its workers with a list of employers offering insurance primarily out of concern that workers will deny eligibility based solely on the list rather than using it to prompt additional follow-up specific to the applicant. Inconsistency in the extent to which DHS and county workers probe on potential insurance issues raises equity concerns because some applicants will undergo a more rigorous eligibility determination process than others.
Workers find that determining whether employer insurance is subsidized is not always easy. MinnesotaCare applicants are ineligible for having access to insurance from an employer only when that insurance is subsidizedthat is, when the employer pays 50 percent or more of the cost of the coverage. 32 For employee-only coverage the calculation is fairly easyinsurance is subsidized if the employer pays 50 percent or more of the premium. The calculation is more complex for spouse and dependent coverage. This coverage is considered to be employer-subsidized if the employer's share of the marginal, or added, cost of coverage for the spouse or dependents is 50 percent or more. 33 According to the director of MinnesotaCare Operations, both enrollees and employers have difficulty understanding what cost information a worker needs to determine if coverage is employer-subsidized. While new applicants are asked to provide premium allocation information on a form attached to the application, workers often have to request the information directly from the employer. Even information received from an employer can be difficult to interpret. For example, we observed that some employers sent benefit booklets describing a wide variety of coverage options without distinguishing which options were available to the MinnesotaCare applicant. We also observed numerous cases in which it was unclear whether the cost of spouse or dependent coverage reported by the employer included or excluded the cost of the employee's coverage. Revising the form to more clearly distinguish between the marginal and total cost may help to improve the accuracy of cost information provided by employers but may not resolve problems understanding complex coverage options.
Enrollment staff are inconsistent in how vigorously they verify self-reported insurance information.
31 DHS, Health Care Programs Manual, §§ 0904.13, 0910.11. 32 Minn. Stat. (2002), §256L. 07, subd. 2.
The interpretation of cost information as marginal or total can affect whether insurance for a spouse or dependent is employer-subsidized, and consequently, whether a spouse or dependent is eligible for MinnesotaCare. In our survey, we asked employers who reported that the MinnesotaCare enrollee was eligible for health insurance benefits to provide the cost of the employee and employer share of premiums. In collecting these data, we provided instructions clearly stating that we wanted the employer to provide the cost of employee-only coverage and the full cost for family enrollment options, not the marginal cost. We then calculated the employer's share of the premium using both the total cost and marginal cost methods. The method used often affected whether the insurance is determined to be employer-subsidized. For example, using the marginal cost method, for 44 percent of employers offering an employee plus spouse coverage option, coverage for the spouse was employer-subsidized. However, when we used the full cost method, 73 percent of the employee plus spouse options would be considered subsidized. 34 Thus, misinterpretation of the premium costs may affect eligibility decisions for a spouse or a dependent.
Consistency Between DHS Policy and Statute DHS has also had difficulty implementing the statutory requirement that applicants not have had access to employer-subsidized insurance in the last 18 months. In reviewing statute and DHS policy, we found that:
Some aspects of DHS policy implementing insurance eligibility criteria contradict statutory language.
Determining whether employer insurance is subsidized is often complex.
33 Ibid. 34 The 95 percent confidence intervals are 27 to 64 percent for the 44 percent estimate and 56 to 86 percent for the 73 percent estimate.
The statute states that: To be eligible, a family or individual must not have access to subsidized health coverage through an employer and must not have had access to employer-subsidized coverage through a current employer for 18 months prior to application or reapplication. A family or individual whose employer-subsidized coverage is lost due to an employer terminating health care coverage as an employee benefit during the previous 18 months is not eligible." 35
The statute exempts children in households with income at or below 150 percent of the federal poverty guidelines from this requirement. 36
The DHS policy manual section implementing this law states: In addition to the limitations on current coverage or access to ESI [employer-subsidized insurance], adults who lost coverage or access to ESI because the employer chose to drop coverage in any of the 18 months prior to the month of application are ineligible for MinnesotaCare . Adults who lost coverage or access to ESI for any other reason are not subject to this restriction. Children under 21, regardless of group status, are exempt from this restriction even if ESI was lost because the employer dropped coverage. 37
DHS policy exempting all children rather than just those with low household income from the 18-month insurance rule contradicts statutory language. Even for adults, the policy narrowly applies the rule to only those cases in which an applicant lost access because the employer dropped coverage. DHS does not apply the rule when an applicant had access to employer-subsidized insurance and lost it for other reasons, such as working fewer hours.
PROGRAM OVERSIGHT The Department of Human Services not only processes MinnesotaCare applications, but also has oversight authority over the program. Based on our review and in light of the problems we found with income and insurance eligibility:
DHS' oversight of MinnesotaCare does not provide sufficient assurance that applicants are accurately self-reporting eligibility or that enrollment staff are making correct eligibility determinations.
35 Minn. Stat. (2002), §256L. 07, subd. 2. 36 Minn. Stat. (2002), §256L. 07, subd. 1. 37 DHS, Health Care Programs Manual, §0910.11.03.
We found weaknesses in a number of policies and procedures that DHS uses to ensure day-to-day integrity of case processing, including computer support, application forms, staff training, and supervisory review. We also found deficiencies in DHS' efforts to monitor compliance with eligibility requirements through audits and fraud prevention efforts.
Computer systems are one tool that agencies can use to control eligibility for government programs. The MinnesotaCare eligibility determination process lacks reliability, in part, because the process is not automated. Currently, enrollment workers calculate total household income and evaluate other eligibility criteria manually and then enter the results into the MinnesotaCare computer system. MinnesotaCare automated records are kept in MMIS, a system originally designed for recording payments to medical providers. MMIS basically keeps records; it does not provide computer-aided eligibility determinations. In contrast, MAXIS, the computer system used for MA and other social service programs, was revised in 2002 to provide automated support for workers determining eligibility for MA and other health care programs (other than MinnesotaCare). Workers enter information from the application and documentation provided, and the system informs the worker, for which, if any, of the various programs the applicant is eligible. MAXIS is also linked to other systems containing, for example, data on child support and federal benefits.
DHS is currently planning to develop one automated eligibility system, known as HealthMatch, for MinnesotaCare, MA, and GAMC. If implemented as planned, workers will use this system to determine eligibility for all three health care programs. DHS is in the process of choosing a vendor to develop the system and expects the first phase to be in place about 18 months after a contract is awarded. As part of the HealthMatch project, DHS is also developing an online application for all of the health care programs. 38 A contract to develop this application began in November of 2002, and a prototype is expected in the spring or summer of 2003.
Another planned component of HealthMatch is the ability to electronically verify income by accessing third-party income data. 39 Under state law enacted in 2001, MinnesotaCare is to use electronic verification as the primary method of verifying income at the time of application. Federal law applying to MinnesotaCare and the MAprogram also requires the state to use electronic verification, but it does not specify that it be the primary means of verifying income. Currently, the MA program complies with federal requirements by using electronic verification in conjunction with paper documentation.
DHS does not use electronic sources to verify income for MinnesotaCare applicants and enrollees. According to DHS, there are two interrelated reasons why. First, MMIS is not linked to the computer system containing third-party income data used for MA and other programs. Second, according to DHS, the third-party data are not sufficient to serve as the primary means of income verification. Based on a pilot test using wage data reported to the unemployment system, DHS found that electronic verification may be of limited use as a primary method of verification because many applicants did not have matching unemployment wage data and because these data were not current. 40 Because the HealthMatch vendor is to integrate third-party income matching into the new automated eligibility system, DHS did not want to invest additional resources in an interim solution based on MMIS. As a result, DHS relies solely on earned income documentation provided by applicants.
While eligibility is now determined manually, DHS plans to develop an automated eligibility system.
Current systems do not support electronic income verification.
38 The online form will allow applicants to apply for any of the programs and will be designed to ask only questions relevant to that applicant's eligibility determination. While the online application will provide a more efficient means of applying for many applicants, DHS will still accept paper applications.
39 Electronic verification involves matching income data collected by DHS (based on self-report or paper documentation) with other income data such as the unemployment system wage data collected by the Department of Economic Security or tax return data collected by the Department of Revenue.
DHS also ensures accuracy in eligibility determinations by developing forms that gather the appropriate information from applicants needed to determine eligibility. All three of Minnesota's public health care programs use the same application and annual renewal forms. Workers processing MinnesotaCare may also rely on many other forms to request additional information in certain cases, and sometimes use forms specific to the MA and GAMC programs to establish MinnesotaCare eligibility when individuals are no longer eligible for MA or GAMC. 41 Several years ago, DHS, with the support of interested legislators, shortened the health care application form from 22 to 4 pages to simplify the application process.
According to DHS and county officials, the application and renewal forms are not as helpful as they could be. While shortening the application form may have made MinnesotaCare more accessible, officials said that the new form has lengthened the determination process because workers need to do more follow-up with applicants to obtain information not included in the application. In addition, some of the questions on the forms are poorly phrased or ask for the wrong information. Based on our interviews and our own analysis of the forms, we identified some questions that could be corrected or clarified.
DHS is counting on computer modernization projects to help staff make more consistent and accurate eligibility decisions.
40 We did not evaluate DHS' pilot test or conduct our own evaluation of the feasibility of using electronic verification as the primary means of verifying income. 41 Policy requires that, in these cases, the last form in the file be used to determine eligibility. When reviewing case files, we often observed workers using forms from other programs.
While some forms give applicants clear choices for specifying how frequently they receive wage income (weekly, every two weeks, twice a month, etc.), the health care application form asks only whether pay is received monthly or hourly. Pay frequency is an essential element in determining annual income, and we found many income determination errors related to using the wrong pay frequency. The application should help clarify information on pay frequency found on the pay stubs.
The health care application includes a question about the 18-month insurance rule that reflects obsolete policy. If applicants indicate a possible insurance issue on the main application, they are asked to provide more detailed insurance information on an attached form. On this attached form, applicants are asked to indicate whether any employer, rather than a current employer, offered health insurance in the past 18 months. The inaccuracy of this question may result in unnecessary follow-up by workers or an improper denial.
Determining household size can be difficult because the health care application and renewal forms do not clearly ask for a list of all individuals living in the household and their relationships to the applicant. Most of the county MinnesotaCare directors we interviewed stated that applicants often failed to list all household members on the application. Missing individuals in the household can result in incorrect eligibility decisions or premium amounts.
The form used to request information from employers about employer-subsidized insurance does not clearly state how the costs of each coverage option are to be presented. Workers could misinterpret total cost information as marginal and make an incorrect determination as to whether spouse or dependent coverage is employer-subsidized.
The six-month income renewal form for the MA program, sometimes used as an application form for MinnesotaCare, does not include any information about access to other insurance. Workers using this form as a MinnesotaCare application would need to follow up with an applicant to determine eligibility.
According to DHS managers, many of these difficulties are a direct result of making the application shorter. They said that maintaining a short, accessible application form while collecting the best eligibility information possible is an ongoing struggle. The department is working with health literacy professionals to improve the structure and wording of its applications and forms.
Some questions on application and renewal forms are inaccurate or unclear.
DHS also trains workers on the eligibility requirements for MinnesotaCare to ensure accuracy in eligibility determinations. While DHS provides extensive entry-level training for MinnesotaCare staff, it offers little training for experienced workers. Workers at both MinnesotaCare Operations and the counties undergo four weeks of training on the eligibility determination process, much of which is offered in a web-based format. Following this training, workers receive updates about changes in policy through written policy bulletins, but DHS does not offer any refresher training on eligibility for experienced workers. In the past, two DHS staff members traveled across the state to discuss policy changes and address issues workers may have. In response to recent budget cuts, DHS reduced this service and may eliminate it entirely. 42 According to the director of MinnesotaCare Operations, because some experienced workers incorrectly apply obsolete policy, increased training for experienced workers may reduce inconsistency and error.
In addition to training, DHS and counties also ensure accuracy by directly reviewing whether workers apply DHS policy in determining eligibility. Supervisors in MinnesotaCare Operations and in some county agencies review workers' eligibility determinations, but on a limited basis. In MinnesotaCare Operations, supervisors are supposed to review all cases processed by a new worker until the supervisor is comfortable with the worker's performance. For experienced workers, supervisors are required to review one case per month. If supervisors identify a problem in one of these reviews, they are supposed to review a few additional cases. DHS does not systematically summarize the results of supervisors' reviews to get an aggregate picture of case quality problems. Based on interviews with six county MinnesotaCare directors, frequency of supervisory review in counties administering MinnesotaCare varies widely. Supervisors in two counties did not conduct reviews, and in another county, supervisors reviewed 15 cases per year. In light of the error rate we found in reviewing case files, the current level of supervisory review is insufficient to identify problems in eligibility determination.
DHS offers extensive training for new MinnesotaCare staff, but little training for experienced workers. To improve accuracy, MinnesotaCare staff need more training.
42 One of the two staff members responsible for providing information on policy updates will now be working on the HealthMatch project.
While the correct tools and proper training aid in ensuring program compliance, monitoring mechanisms such as random audits and fraud detection are means of directly testing whether households enrolled in the program are eligible. According to statute, DHS " shall perform random audits to verify reported income and eligibility," though the law does not further define the nature or frequency of such audits. 43 DHS' last compliance audit specific to MinnesotaCare was done in 1995. Since then, DHS has completed several broader evaluations related to other aspects of MinnesotaCare such as health care access for welfare leavers, reasons for disenrollment from MinnesotaCare, and assets held by MinnesotaCare enrollees. However, none of these studies directly addressed the accuracy of information provided by applicants. DHS initiated a MinnesotaCare evaluation in the fall of 2002, several months after we began our study.
DHS also does not have a formal mechanism in place to detect or investigate allegations of applicant fraud. While county agencies assign staff to investigate fraud, no DHS staff are responsible for investigating fraud by MinnesotaCare enrollees whose cases are processed by DHS. Workers handle the few allegations that they receive in-house or transfer them to county fraud divisions. According to DHS managers, the agency is planning to introduce a fraud prevention program for MinnesotaCare Operations. While the specifics of this program have not been decided, the agency plans to develop a means of identifying cases at risk for fraud and a system to assign these cases to specially-trained workers. DHS added that, although the law allows the department to pursue fraud in MinnesotaCare, it provides fewer enforcement mechanisms, such as a disqualification period, than are available for other social service programs.
CONCLUSIONS In designing MinnesotaCare, the legislature faced trade-offs between careful targeting of coverage and encouraging program participation. One challenge for DHS and counties in administering MinnesotaCare, particularly in tight fiscal times, is to strike a balance between providing reasonably easy access to the program through an efficient, understandable application process and ensuring that workers have sufficiently detailed information with which to make accurate eligibility decisions. Deciding where this balance lies involves philosophical judgments about the proper level of scrutiny to be placed on applicants' eligibility. Too much, and the application process will be longer, more costly, and more intrusive; as a result, eligible Minnesotans may be reluctant to apply. Too little, and the state may waste resources providing coverage for people for whom it was not intended.
Although required by law, DHS has not done a random audit specific to MinnesotaCare since 1995.
43 Minn. Stat. (2002), §256L. 05, subd. 2.
While we discuss the efficiency of MinnesotaCare's application process in Chapter 3, the work presented in this chapter shows that MinnesotaCare policies and oversight do not provide reasonable assurance that applicants are accurately self-reporting eligibility information or that enrollment staff are making correct eligibility and premium decisions. We identified several weaknesses that point to a need for improving the accuracy of information collected directly from applicants, clarifying policy, ensuring that workers adhere to policy, and strengthening oversight. While no method of determining eligibility will eliminate all errors, effective, equitable targeting of MinnesotaCare requires that eligibility criteria be applied as consistently and accurately as possible.
DHS is counting on its computer modernization projectsan online application and automated eligibility determinationto help the agency resolve many of these issues. In DHS' view, and we agree, automating eligibility decisions is an important part of a long-term solution to problems with consistency and accuracy. But, the new system will not be implemented for at least 18 months, and some problems require additional steps, such as clarifying when to make exceptions to standard policies. In the meantime, we think DHS can take several steps to address problems raised in our report and to strengthen its controls over the eligibility determination process, including clarifying its policies and requiring workers to take refresher training. As DHS health care managers acknowledge, the department also needs to regularly monitor the accuracy of information provided by applicants and to have procedures for identifying and dealing with applicant fraud and abuse. DHS recently initiated a MinnesotaCare performance review, but it is the first such study to be done since 1995. Similarly, DHS does not have a formal fraud-prevention program, but has plans for one.
Any eligibility decision, whether automated or manual, will only be as good as the information that goes into it. We identified two key issues in this regard. First, the Legislature felt it was important to restrict MinnesotaCare eligibility to those who do not have access to affordable private health insurance to prevent individuals and employers from dropping private health coverage in favor of MinnesotaCare. Yet, insurance-related eligibility is largely self-reported. Self-reported eligibility is inherently risky from a compliance standpoint, and our research shows that applicants' statements regarding their insurance status are not sufficiently reliable to ensure proper eligibility decisions. Second, for a variety of reasons, MinnesotaCare estimates of annual income often do not accurately reflect actual income. Much of this problem occurs because MinnesotaCare's point-in-time method of estimating wage income does not adequately reflect income for an entire year. This is less of a problem when MinnesotaCare income estimates are too high because applicants are allowed mid-year corrections when their income declines. When MinnesotaCare income is too low, however, DHS cannot make mid-year premium adjustments. Both issuesreliance on self-report and current income determination policyneed to be addressed at the policy level, considering trade-offs between improved compliance, cost, and administrative ease.
RECOMMENDATIONS Improve the Accuracy of Eligibility Decisions
RECOMMENDATIONS To make the eligibility determination process less vulnerable to inconsistency and error, DHS should:
expedite development and use of an automated eligibility computer system.
In the meantime, to improve the accuracy of eligibility decisions, DHS should:
Clarify its policy on documenting wage income by requiring pay stubs and by more clearly defining those circumstances in which a worker may deviate from the standard method of estimating annual wage income;
Revise its health care policy manual to ensure that policies and procedures implementing the 18-month insurance eligibility rule are consistent with statutory language; Use more frequent, targeted refresher training to help ensure that workers understand current policy; and
Revise applications and other forms to more closely link insurance and income questions to specific eligibility criteria.
Our recommendation to develop an automated eligibility system supports a process that DHS has already begun. But, because of its potential to address a number of weaknesses in the eligibility determination process, we feel it is important to emphasize that the project should be completed as expeditiously as possible. DHS has started its search for a project contractor and expects the new system to be ready in about 18 months at a total cost of $13-15 million dollars. According to DHS officials, federal funds will cover half the project cost, and the agency has funds available to cover the state share. While DHS does not expect to request additional funds for the project, it will be reallocating staff from other areas. For example, some individuals previously responsible for providing health care policy updates will now be working solely on the automated eligibility project. Later, when the new system is implemented, all health care training staff will be focused on the new automated system.
Standardizing its wage verification policy will require changes to DHS forms, publications, training materials, and policy documents. The department has standard procedures in place to update these items, but, according to DHS, it would need to devote additional resources to implement our recommendations.
DHS should take interim steps to tighten eligibility procedures while developing the new automated eligibility system.
DHS currently allows applicants to submit tax returns or pay stubs to document wage income. Our recommendation to eliminate use of tax returns to verify wage income should not impose a significant administrative burden on applicants since most are currently providing pay stubs rather than tax returns, and DHS could continue to accept an employer statement of current earnings if pay stubs are unavailable. DHS can also change its policy applying the 18-month insurance rule through its ongoing policy update process. The change should not affect case processing time as workers are to evaluate insurance-related eligibility for every case. DHS should be able to revise applications and other forms without undermining its preference for a short application form. In most cases, questions do not need to be added to the form; rather current questions need clarification. For the most part, the changes suggested also do not limit the use of the application form for the other two health care programs. Instead of adding questions unrelated to MA to the six-month MA renewal form, DHS may decide that that form should not be used as an application form for the MinnesotaCare program.
DHS managers agree with the need to improve training for staff handling MinnesotaCare cases, but said that doing so may be difficult in light of recent budget cuts and the agency's need to reallocate training staff to the automated eligibility projectwhich we agree is a high priority. Additional training also reduces the time enrollment staff spend processing applications and renewals. Still, our work demonstrates that more training is needed to improve the accuracy of eligibility determinations. DHS has a web-based training curriculum already in place for new MinnesotaCare workers that, while not as ideal as training materials developed specifically for experienced workers, could be used for refresher training as well.
Clarify Procedures for Verifying Insurance-Related Eligibility
RECOMMENDATIONS To improve the accuracy of insurance-related eligibility decisions, DHS, in consultation with the Legislature, should reconsider the method to be used for enforcing insurance-related eligibility criteria. If DHS continues to accept self-reported insurance information, it should clarify in policy when workers need to verify information provided by applicants and the extent to which lists of employers offering insurance should be kept and used by state and county agencies.
DHS should reconsider the method by which insurance-related criteria should be verified. A range of options is available from continued reliance on self-reported information to development of an electronic verification system using a database of Minnesota employers. Each of these options involves tradeoffs between administrative ease and compliance.
Providing more training will be difficult with current resources, but is needed to improve the accuracy of eligibility decisions.
If DHS chooses to accept self-reported insurance information in any circumstance, DHS should clarify its policy regarding when self-reported insurance information should be verified and how workers should use lists of employers offering insurance. The resources required to maintain a list of employers offering insurance will vary depending upon the method used to obtain the list. If DHS and agencies develop the list based on self-reported information from applicants, the resources should be minimal, but the list may be less accurate. Obtaining the information from private sector sources would be more accurate, but would have added costs.
Ensure That Premiums Reflect Actual Income RECOMMENDATIONS To help ensure that MinnesotaCare premiums reflect actual income throughout the year:
DHS should incorporate third-party income sources in its automated HealthMatch system to help determine income at application, renewal, and periodic mid-year reviews. DHS should examine the costs and benefits of alternative ways of reviewing income throughout the year.
The Legislature should change the law to allow DHS and counties to adjust premiums between annual renewals when income increases as it now allows for changes when income decreases.
DHS could use third-party income sources such as quarterly unemployment wage reports and income tax data to help determine income not only at application and annual renewal, but also when enrollees' income changes during the year. As discussed earlier, MinnesotaCare's computer systems currently lack the capacity to integrate these third-party income sources with its eligibility system, but DHS is planning to include this capacity in its new automated HealthMatch system. We support this effort and think that it should include the ability to identify income changes between annual renewals. Data from these sources would provide workers with additional information, such as indicators of unreported income, to guide eligibility decisions. How DHS will obtain electronic data sufficient to serve as the primary means of verifying current income is still an open question.
DHS should determine the best means of reviewing income throughout the year. Periodic reviews could range from more frequent full-scale renewals for all enrollees to targeted income (and possibly insurance) reviews when third-party income sources show a significant change in income. In fact, DHS has considered conducting full-scale renewals every six months instead of every 12 months. It estimated that the additional cost of conducting six-month renewals would be about $2.7 million per year with a first year start-up cost of $0.6 million. This cost is less than the loss in premiums that we estimate is attributable to differences between MinnesotaCare income determinations and actual income. However, the margin of error for our estimate is large, and it is not clear what proportion of this loss in premiums would be captured by six-month renewals. Thus, we recommend that DHS study this approach as well as other less expensive approaches that target mid-year reviews at cases with large changes in income. DHS could target cases by using the new automated system to monitor changes in income. It could also review cases in which income dropped at the time of application. For example, if applicants reported that they quit or lost their jobs, DHS could check back later in the year to determine whether their income changed.
Many alternatives exist for reviewing income between renewals, each with different costs and benefits.
As noted earlier, state law prevents DHS from raising premiums between annual renewals because of a change in income though it allows DHS to lower premiums if income declines. To help ensure that premiums reflect actual income, we think the Legislature should change the law to allow mid-year corrections for higher income. As discussed above, how DHS uses this authority to make mid-year income adjustments will depend on judgments regarding tradeoffs between administrative cost, improved targeting of program benefits, and enrollment stability during the year.
Strengthen Oversight RECOMMENDATIONS To heighten oversight of MinnesotaCare eligibility, DHS should:
Increase supervisory review of staff's eligibility decisions to the extent possible; and
Do more frequent compliance reviews to check the accuracy of information self-reported by applicants.
According to MinnesotaCare managers, increasing the level of supervisory case review would be difficult in light of current staff resources and duties. We agree that more frequent reviews will place additional demands on supervisors, but believe that the need to tighten oversight of eligibility determinations merits the extra effort. Targeting reviews to certain problem issues related to income and insurance may help reduce the additional burden on supervisors' time.
DHS has a division responsible for internal program evaluations, and it was this group that completed the last MinnesotaCare audit in 1995. Since then, the group has concentrated its efforts on evaluating the Minnesota Family Investment Program. Turning to MinnesotaCare will limit other types of evaluations that this group may perform, but regular MinnesotaCare compliance reviews are an essential element of program oversight. As is DHS' practice, these compliance reviews may be targeted to certain aspects of the eligibility determination process.
A change in state law would be required for DHS to raise premiums between renewals.
Case Processing SUMMARY Over the past four years, Minnesotans sending a MinnesotaCare application to the Department of Human Services (DHS) often had to wait more than 20 days for the department to begin processing their applications and 60 to 90 days in total for health coverage to begin. Recently, however, DHS reduced the time to begin processing applications to less than one week, although application processing remains vulnerable to large backlogs. MinnesotaCare workloads have increased rapidly, and DHS has responded by hiring more staff, reassigning staff to handle peaks in new applications, and improving productivity. Still, staffing issues and heavy reliance on manual operations have hampered DHS efforts to better manage caseloads. The agency is investing in a new electronic case management system that it hopes will provide quicker access to case files, more flexibility in assigning work, and more detailed performance data. We make several recommendations regarding collection and use of performance data to help DHS better manage the program.
As discussed in Chapter 1, MinnesotaCare enrollment has been steadily increasing and is projected to grow through 2005. Understanding how DHS handles this workload is an important aspect of evaluating the program. Inefficiencies in its work processes can delay health insurance coverage and impose unnecessary burdens on applicants, enrollees, and the state. Accordingly, this chapter addresses the following question:
How efficiently does DHS process MinnesotaCare cases? Our study focused on how promptly DHS processes MinnesotaCare cases and the factors that influence processing time. Prompt processing of new applications enables eligible individuals to be covered by health insurance in a reasonable amount of time. Similarly, prompt processing of renewal applications enables enrollees to maintain coverage without interruption. We limited the scope of our review to MinnesotaCare Operationsthe DHS unit responsible for processing and maintaining the vast majority of MinnesotaCare cases. To meet our objective, we collected and analyzed DHS data on processing times over the past four years. We also obtained available data regarding factors that influence processing times, such as workload and staffing trends, and interviewed DHS officials on agency efforts to improve case processing.
WORKLOAD AND PRODUCTIVITY In fiscal year 2002, DHS received about 58,000 new applications, 45,000 annual renewal applications, and about 500,000 telephone inquiries. It employed about 116 enrollment staff and 13 supervisors to process these applications, manage ongoing cases, and answer telephone calls. DHS divides these staff into teams of about ten enrollment workers and one supervisor. On each team, about half of the enrollment workers process new applications, while the other half maintain ongoing cases and handle annual renewals. All workers spend time covering the telephone assistance lines, including four hours per week on the general MinnesotaCare information line plus any case-specific calls on their direct lines.
Obtaining health insurance coverage through MinnesotaCare can be a lengthy process, as outlined in Figure 3.1. Applicants must complete an application form and submit it to DHS or their county for processing. The application form often does not provide information sufficient to determine eligibility, primarily because of the complex criteria involving insurance and income. In these cases, enrollment workers request additional information from applicants or their employers. 1 After they obtain sufficient information, workers determine eligibility, and the department's computer system calculates the household premium. Insurance coverage begins on the first day of the month following the date DHS receives the first monthly premium payment.
To maintain MinnesotaCare coverage, enrollees must annually submit renewal applications, which they can do beginning three months before their health insurance renewal date. If the renewal is not submitted and processed prior to that date, MinnesotaCare coverage ends, and individuals must reapply to reinstate coverage. The renewal application is shorter than the new application (two pages rather than four), but requires the same documentation of income as the new application.
Minnesota law requires DHS and counties to "determine an applicant's eligibility for MinnesotaCare no more than 30 days" after receiving a new application. 2 DHS measures compliance with this standard by tracking backlog timethe number of days between receipt of the application and the date a worker first looks at it. Currently, it is more practical to track backlog time because the department's computer system does not track the date eligibility decisions are made. Application backlog time approximates the time it takes to determine eligibility if the worker can make a decision based on the application and, perhaps, a few quick phone calls. When applications are incomplete, however, it may take several days or weeks before the eligibility decision is actually made, depending on how long it takes the applicant, and in some cases the employer, to return the required information.
MinnesotaCare enrollment staff at DHS work in teams to process new applications, manage ongoing cases, and answer telephone inquiries.
By law, eligibility is supposed to be determined within 30 days.
1 If the applicant did not send required proof of income but the application form contains enough information for the worker to reasonably estimate annual income, the worker can grant eligibility to the household for one month. Eligibility will continue after that month if the household provides the required proof of income. We found that workers rarely used this option.
2 Minn. Stat. (2002), §256L. 05, subd. 4.Application Received and Date-Stamped Application Received and Date-Stamped Case Worker Takes Application Case Worker Takes Application Case Worker Makes Initial Determination Case Worker Makes Initial Determination Applicant Pays First Premium Applicant Pays First Premium Health Insurance Coverage Begins Health Insurance Coverage Begins Eligibility Determination Eligibility Determination Backlog Time Has Ranged From 2 to 46 Days a Initial Processing Time Typically 1 Day Time to Return Requested Information Depends on Applicant or Employer Time to Receive Payment Depends on Applicant Wait Time to First of Month From1 to 30 Days Need Data Approve Deny Need Data Approve Deny
To examine how long DHS takes to process cases, we analyzed DHS data on new application backlog times. We were not able to measure the entire time it takes to determine eligibility because DHS does not track the eligibility decision date. Our analysis shows that:
DHS recently reduced backlog times to less than one week, but long backlog times have been common during the past four years.
As Figure 3.2 shows, the average monthly time in backlog has fluctuated greatly during the past four years, ranging from 5 to 44 days with an average of about 22 days. Average monthly backlog time remained over 20 days for extended time intervals on three occasions, including 15 consecutive months from October 1999 through December 2000. The most recent peak occurred after the Minnesota state employee strike of October 2001. After each of these three peaks, DHS reduced
CASE PROCESSING 51 Figure 3.1: Processing a New Application
a Based on DHS data for longest and shortest backlog times from July 1998 through October 2002. SOURCE: Office of the Legislative Auditor analysis.
DHS tracks backlog timethe time from an application's receipt to a worker's first look at it. the backlog time to less than ten days. As of November 2002, however, it has not been able to keep the backlog time to less than ten days for more than five consecutive months.
Rapid growth in MinnesotaCare workloads is an important factor behind long backlog times. Between fiscal years 1999 and 2002, the number of new applications increased by about 64 percent, and the number of renewals increased by 36 percent. 3 In addition, the number of telephone inquiries increased by about 50 percent between fiscal years 2000 and 2002, reaching about 8,800 calls per week during fiscal year 2002. DHS enrollment estimates for the next biennium suggest that these trends will continue through 2005.
The total time it takes an applicant to obtain health coverage fluctuates with the backlog time, the individual's response times to data requests, and the timing of the first premium payment. While we do not have systematic data on these times, our file review and interviews with DHS staff indicate that the total time for coverage to begin has often been between 60 and 90 days. When backlog times are short and the initial application is complete, applicants can be enrolled within a month of applying. But if backlog times are long, the application is incomplete, or the premium payment is not promptly received, it can take over 90 days to become enrolled in MinnesotaCare.
0 10 20 30 40 50
7/98 1/99 7/99 1/00 7/00 1/001 7/01 1/02 7/02 SOURCE: Office of the Legislative Auditor analysis of Department of Human Services data.
Figure 3. 2: Application Backlog Time, MinnesotaCare Operations, July 1998-October 2002
Number of Days
DHS has not been able to hold backlog times to under ten days.
Rapid growth in MinnesotaCare workloads is projected to continue.
3 Through the first four months of fiscal year 2003 (July through October 2002), new applications increased by 7 percent over the same time period the previous year.
Our review of how DHS managed the rising caseload over the past four years shows that:
In response to increasing MinnesotaCare workloads, DHS hired more staff, reassigned staff to handle peaks in new applications, and improved productivity. With additional funding from the Legislature, DHS increased its MinnesotaCare staff each year between 1999 and 2002, but these increases did not keep pace with the increase in workload. The number of hours worked by enrollment staff increased by 25 percent between fiscal years 1999 and 2002. This percentage increase is lower than the increase in workload, particularly compared with the 64 percent increase in new applications.
DHS was able to close the gap between workload increases and staff increases by improving productivity. Between fiscal years 1999 and 2002, the number of new and renewal applications processed per work week increased by 17 percent. 4 Much of this gain was achieved in short bursts of heightened productivity. When backlog times were unusually long, DHS management assigned new applications to all enrollment workers, including those who normally just process renewals. Management also pushed the entire staff to work at a faster pace. For example, after the backlog time reached 44 days in November 2000, DHS focused its resources on reducing the backlog time for new applications. In December 2000, DHS processed considerably more new applications than normal (6,322 compared with a range of 2,372 to 4,565 for the previous 12 months), reducing its backlog time to 19 days in January 2001. According to DHS management, processing this many applications meant cutting back on customer service for MinnesotaCare enrollees, though they said the department was able to maintain essential services.
CASE PROCESSING MinnesotaCare enrollment staff responded to about 8,800 general inquiry calls per week in fiscal year 2002. Application backlogs grew when staff increases and productivity gains lagged behind increases in workload.
4 We defined a work week as 40 hours of work plus overtime and compensatory time worked. We excluded vacation, sick leave, and other leaves of absence.
Reassigning staff and pushing productivity helped DHS catch up with its workload, but there are limits on how much it can rely on this strategy. DHS gives priority to processing renewals in time to maintain enrollees' coverage. During the past 20 months, DHS processed about 35 percent of its renewals within 30 days of the expiration date, 58 percent between 30 and 60 days, and 6 percent between 60 and 90 days of the expiration date. 5 DHS could delay processing many of these renewals, but not for long. If DHS does not allow enough time for enrollees to respond to requests for additional information, DHS delays could cause a break in coverage if the renewal deadline is missed.
While DHS has recently managed to catch up with the rising workload, long backlog times have been common over the past several years because the gains in productivity and staff increases have lagged behind increases in workload. For example, between fiscal years 1999 and 2000, the workload increased by 17 percent, but the number of staff work hours increased by 11 percent and productivity remained the same. One reason that staff hours did not increase as fast as applications was that DHS greatly reduced its use of overtime, going from 7 percent of staff hours in fiscal year 1999 to less than 1 percent thereafter. 6 As a result, in fiscal year 2000, DHS processed 5,000 fewer new applications than it received. The time in backlog remained at a high level until DHS improved its productivity in late 2000.
FACTORS AFFECTING PERFORMANCE The large fluctuations in backlog times during the past four years raise the concern that DHS will not be able to avoid long backlog times if MinnesotaCare caseloads continue to rise as expected. In this section, we examine staffing issues and work processes that affect how well DHS keeps up with the MinnesotaCare caseload. In general:
Staffing issues and heavy reliance on manual operations have hampered DHS efforts to manage workload peaks and keep case processing timely.
Staffing DHS management points to various staffing issues that have hampered its efforts to keep up with its case processing workload. First, DHS argues that the number of ongoing cases assigned to each enrollment worker is higher than it should be, placing the program at risk for falling behind if caseloads increase and jeopardizing service to MinnesotaCare enrollees. Managers point out that the average caseload for workers at MinnesotaCare Operations was about 850 to 900 cases in October 2002, much higher than caseloads for county workers who manage health care cases. While they acknowledge that county workers may have more complex caseloads than DHS workers, DHS managers believe that caseloads for their workers are too large. In fact, according to DHS managers, the agency has requested budgets based on a ratio of 600 cases per employee in past budget requests but has not been successful in reaching this goal as costs and caseloads have grown.
5 Percentages are based on renewals processed from March 2001, when DHS began collecting these data, through October 2002. 6 DHS management stopped relying on overtime because it was more expensive than hiring more employees, and it hurt employee morale.
We reviewed health care caseloads in six counties, but concluded that several important differences between counties and DHS made meaningful comparisons impractical without further study. 7 The number of cases per worker ranged from about 150 to 400 in these six counties. County workers with caseloads at the low end of this range typically manage cases for a variety of programs, including health programs (Medical Assistance, General Assistance Medical Care, or MinnesotaCare) and one or more non-health programs (Food Stamps, Minnesota Family Investment Program (MFIP), Emergency Assistance, or child care). This makes comparisons with DHS caseloads difficult because individual cases may involve several programs, and each program has different eligibility requirements. For example, several of these programs (including Medical Assistance, MFIP, and Food Stamps) require income to be reviewed more frequently than MinnesotaCare. County workers with caseloads between 300 and 400 worked mostly on MinnesotaCare cases. Even in these instances, comparisons are difficult because, unlike DHS, county workers who manage ongoing MinnesotaCare cases are also usually responsible for processing new applications.
A second factor affecting case processing times is high turnover among enrollment staff. As shown in Table 3.1, annual turnover rates during the past four years have ranged from 16 to 29 percent. According to DHS managers, the enrollment representative position is an entry-level job category, and many enrollment staff are promoted to other positions within the agency. Turnover hampers MinnesotaCare Operations because it takes considerable time and resources to hire and train new staff. According to the director of MinnesotaCare Operations, it typically takes four to six months for new employees to be able to process applications on their own. Since turnover varies from month to month, unusually high turnover rates can disrupt operations for several months.
Table 3.1: MinnesotaCare Operations Staffing, FY 1999-2002 Fiscal Year 1999 2000 2001 2002 Number of Supervisors 10 11 12 13 Average Number of Enrollment Workers on Staff 85 99 107 116 Turnover Rate For Enrollment Workers 29% 18% 16% 19%
SOURCE: Office of the Legislative Auditor analysis of Department of Human Services data.
The length of time required to hire and train new staff has hampered efforts to keep pace with case processing.
7 The six counties were Anoka, Carlton, Dakota, Itasca, Kandiyohi, and Morrison. These counties processed about half of the new and renewal applications received by counties between January and March of 2002.
A third factor is that DHS has not been able to maintain staff at authorized levels. 8 For example, DHS had 10 to 15 vacancies at several times during the past two years. DHS hired staff on nine occasions during the past two years, but several times could not fill as many vacancies as needed. For instance, in February 2002, it had 15 vacancies but hired only 4 employees. DHS management cited several factors that have made it difficult to attract applicants and compete with other job opportunities, including the tight job market of the late 1990s, the talk of the state shutdown in June 2001, and the state employee strike of October 2001.
Work Processes As with the eligibility determination process discussed in Chapter 2, DHS uses manual, paper-driven processes to receive, assign, and manage MinnesotaCare cases. DHS receives new applications through the mail. Clerical staff date stamp them upon receipt and place them in a central location. Workers assigned to process new applications go to this central location to sign out applications (a few at a time, based on the order received) and take the files to their desks for processing. After making a final eligibility decision, the case files go back to a clerical worker. Files for denied cases are stored in central file cabinets, and approved case files are assigned and transferred to a case maintenance worker based on current workloads. The case maintenance worker takes a look at the file to become familiar with the case, then stores the file with other active cases in cabinets, shown in the picture above, located near his or her work area. Renewal forms are also received through the mail and date stamped and then are routed to the case maintenance worker assigned to that case.
Because of the sheer volume of case files and physical layout of MinnesotaCare Operations' office space, reliance on paper case files affects how DHS manages the MinnesotaCare workload. MinnesotaCare work teams are physically located on five floors in a downtown St. Paul office building. Cases maintained by a particular team are stored on the floor where that team is located. Because it is inefficient for workers to travel to different floors to retrieve case files, these physical storage limitations affect DHS' ability to quickly reallocate work among teams or to have workers on other teams respond to case-specific customer service calls. The paper system also increases the risk that files will be lost or misplaced, as we discovered during our file review when DHS staff were not able to immediately locate several files that we requested. According to DHS staff, files on occasion need to be entirely recreated.
DHS uses manual, paper-driven processes to receive, assign, and manage MinnesotaCare cases. Important records are kept in thousands of hard-copy case files.
8 As of October 2002, MinnesotaCare Operations employed 138 enrollment workers.
Use of hard-copy case files also affects the quality of customer service. When an enrollee calls with a question about his or her case that requires access to the file, workers may not be able to answer the question during the call. Rather, the worker may have to retrieve the file and call the enrollee back. Because each file must be copied and mailed, reliance on paper files also slows transfer of cases between DHS and counties.
DHS has some evidence that current application procedures may be discouraging some applicants from completing the enrollment process. According to DHS data for fiscal year 2002, about 24 percent of individuals denied coverage were denied because they did not complete the application process. According to DHS, many of these households reapply at a later date, creating added burden for applicants and DHS. Because these data imply that some applicants find the process too difficult to successfully navigate and because of the drain on agency resources, DHS is doing more research on these cases to understand why applicants are not completing the process.
DHS recognized that it could not continue to handle growing workloads through staff increases, and that opportunities to make significant productivity gains were hampered by its existing work processes. Accordingly, DHS contracted with a consulting firm in 2000 to study its case processing system and make recommendations for its redesign. Based on its study, the consultant recommended that DHS invest in an electronic case management system with two primary componentselectronic document storage and automated workflow. Under the recommended system, all elements of the traditional, hard-copy case files (the application form, pay stubs, tax returns, correspondence, etc.) would be scanned and stored as electronic images that would be indexed to a case identification number. The workflow component of the system would control access to case documents, prompt required work actions, and monitor the status of each case. For example, the system would automatically distribute new applications to available staff. DHS accepted the consultant's recommendation and started its search for a project contractor in December 2002. According to DHS, the new system is to be completed in the spring of 2003.
Based on data from the consultant, DHS expects the new system to reduce administrative costs associated with handling and storing hard copy case files. According to estimates prepared in 2001, DHS annually spends about $650,000 to store files, $180,000 to retrieve and maintain files, and $41,000 to copy and mail files being transferred to counties. 9 The estimated cost to develop the new document imaging system is between $818,000 and $982,000 with ongoing annual maintenance costs of roughly $100,000. While all of the expenses associated with hard-copy files will not be eliminated, particularly in the short term, it is reasonable to expect some savings from electronic document management.
DHS is counting on a new automated system to improve management of its workload.
9 We did not evaluate in detail the consultant's methodology for making these estimates or its underlying assumptions.
It is not clear yet how much the electronic document management system will improve productivity or streamline case management. As initially implemented, the system will not fundamentally change the basic procedures that guide how applicants apply for or renew their coverage; for the most part, the new system takes the existing work steps and translates them to an electronic format. Applicants will continue to submit hard copy applications and verifications, and workers will continue to make eligibility decisions and enter the results into MMIS. Still, DHS should see some immediate benefits from quicker access to case files and more flexibility in assigning work.
The new system is also expected to provide more comprehensive data on workload and performance because it is supposed to be capable of logging every action on a case with a time stamp. These data can then be used to track, for example, the number of cases worked by each individual or aggregate data on the time it takes to make a final eligibility decision. Once workers become more experienced working with the new system, DHS should be able to use this performance data to reevaluate its productivity and staffing assumptions. MinnesotaCare applicants and enrollees should see some direct benefit from DHS' use of electronic case files. As mentioned above, workers will have immediate, desk-top access to case information and should be able to answer questions more quickly. To the extent that the system speeds case processing times, applicants may benefit by getting faster coverage.
Improvements on these issues are related to other modernization projects discussed in Chapter 2the automated eligibility and online application systems. A key issue for DHS in the future will be ensuring that the electronic case management system is fully integrated with the other two projects.
CONCLUSIONS Over the past four years, DHS has had trouble keeping pace with growing workloads. For long stretches of time, applicants had to wait more than 20 days for DHS to start processing applications, and many had 60-to 90-day waits for coverage to begin. Recently, however, DHS has improved the way it manages processing of new applications. The circumstances leading to the most recent peak in application backlog timethe October 2001 state employee strikewere largely outside of DHS' control, and the agency recovered faster than it had following previous peaks in backlog time. Still, the frequency with which backlog times have risen and stayed above 20 days indicates that application processing is vulnerable to unacceptably large backlogs.
DHS has used a variety of techniques to manage its growing workload, including staffing increases, temporary reallocation of staff, and productivity pushes. DHS has recognized, though, that significant improvements in case processing will require new tools, and the electronic case management system is a step in that direction. If implemented as planned, DHS should benefit from quicker access to case files and more flexibility in assigning work. The extent to which this system will succeed in streamlining the application process and improve service to applicants and enrollees is unclear, particularly since its full impact on work processes will not be seen until DHS implements the automated eligibility and online application systems discussed in Chapter 2.
When integrated with other new computer systems, automated case management should help streamline the application process.
As it implements major automation projects, DHS has an opportunity to improve its collection and use of performance data to manage MinnesotaCare Operations. For example, DHS could design this system to track time from application receipt to eligibility decision in order to more accurately measure compliance with the 30-day standard. Tracking application backlog time is still important, but DHS needs a backlog time standard well under 30 days so that applicants have a reasonable amount of time to provide additional information and still have their eligibility determined within 30 days. In addition, managers need to understand how automation will affect workflow and productivity and, in turn, how these factors will influence overall staffing needs.
RECOMMENDATIONS To better understand and manage processing of application and renewal cases, DHS should:
Ensure that the electronic case management system is capable of tracking key performance data, such as application backlog time, time to make eligibility decisions, and total time from application receipt to coverage, and use these data to monitor performance;
Develop standard productivity measures, track changes in these measures as various automation projects are implemented, and use the data to modify staff performance expectations and staffing estimates; and
Set performance goals for application backlog time that will generally allow eligibility decisions to be made within 30 days.
Throughout our review, DHS managers emphasized the importance of having specific performance and workload data and how difficult it has been for them to produce such data with existing systems. As it begins modernizing, DHS has an opportunity to build into system designs the ability to collect these data. Adding these features should entail little additional cost.
In the future, new automation projects should help DHS process cases within the statutory 30-day limit. In the meantime, DHS will need to use its current techniques, such as filling existing staff vacancies, productivity pushes, and reallocation of staff from renewals to new applications. To keep backlog times low, DHS may need to apply these techniques sooner than in the past when backlog times begin increasing.
CASE PROCESSING DHS should monitor how computer modernization will affect productivity and overall staffing needs.
Summary of Recommendations Improve the Accuracy of Eligibility Decisions (p. 45) To make the eligibility determination process less vulnerable to inconsistency and error, the Department of Human Services (DHS) should:
Expedite development and use of an automated eligibility computer system.
In the meantime, DHS should: Clarify its policy on documenting wage income by requiring pay stubs and by more clearly defining those circumstances in which a worker may deviate from the standard method of estimating annual wage income;
Revise its health care policy manual to ensure that policies and procedures implementing the 18-month insurance eligibility rule are consistent with statutory language;
Use more frequent, targeted refresher training to help ensure that workers understand current policy; and
Revise applications and other forms to more closely link insurance and income questions to specific eligibility criteria.
Clarify Procedures for Verifying Insurance-Related Eligibility (p. 46) To improve the accuracy of insurance-related eligibility decisions, DHS, in consultation with the Legislature, should reconsider the method to be used for enforcing insurance-related eligibility criteria.
If DHS continues to accept self-reported insurance information, it should clarify in policy when workers need to verify information provided by applicants and the extent to which lists of employers offering insurance should be kept and used by state and county agencies.
Ensure that Premiums Reflect Actual Income (p. 47) To help ensure that MinnesotaCare premiums reflect actual income throughout the year:
DHS should incorporate third-party income sources in its automated HealthMatch system to help determine income at application, renewal, and periodic mid-year reviews.
DHS should examine the costs and benefits of alternative ways of reviewing income throughout the year.
The Legislature should change the law to allow DHS and counties to adjust premiums between annual renewals when income increases as it now allows for changes when income decreases.
Strengthen Oversight (p. 48) To heighten oversight of MinnesotaCare eligibility, DHS should: Increase supervisory review of staff's eligibility decisions to the extent possible; and
Do more frequent compliance reviews to check the accuracy of information self-reported by applicants.
Improve Data Used to Manage Case Processing (p. 59) To better understand and manage processing of application and renewal cases, DHS should:
Ensure that the electronic case management system is capable of tracking key performance data, such as application backlog time, time to make eligibility decisions, and total time from application receipt to coverage, and use these data to monitor performance;
Develop standard productivity measures, track changes in these measures as various automation projects are implemented, and use the data to modify staff performance expectations and staffing estimates; and
Set performance goals for application backlog time that will generally allow eligibility decisions to be made within 30 days.
Methodology and Sample Design APPENDIX
This appendix describes three methods we used to assess the accuracy of MinnesotaCare eligibility decisions, as discussed in Chapter 2. These include: (1) reviewing a sample of 594 new and renewal application case files; (2) matching income reported to MinnesotaCare with income reported on tax returns and to the unemployment system; and (3) surveying enrollees and their employers regarding access to employer-based health insurance. FILE REVIEW To evaluate how accurately DHS and counties determine eligibility, we reviewed a stratified, random sample of 594 new and renewal applications submitted from January through March 2002. To identify applications and renewals received during this period, we requested and used an extract of MinnesotaCare's computerized records that we obtained in May 2002. As shown in Table A. 1, the sample was stratified according to (1) where the application was processed (DHS or county) and (2) the type of case (new application or renewal application). We over-sampled cases processed by counties to ensure that we had enough county cases to compare DHS and county processing. Within the DHS and county strata, we selected approximately equal numbers of new applications and renewal applications. Our final sample included 398 cases processed centrally by the MinnesotaCare Operations unit at DHS and 196 cases processed by county social service offices.
For each case in the sample, we used a structured format to collect data from case file documents, including application forms, workers' case notes, follow-up data request forms, verifications provided by applicants, and other relevant documents. For each member of the household, we collected detailed data on countable earned income directly from income verifications in the file, such as pay stubs and tax forms. We also collected data on unearned income and income amounts that DHS used to determine eligibility. We compared the method the worker used to determine income to DHS policy, and if there was a discrepancy between this method and policy, we recorded the type of error. For renewal applications in the sample, we also recorded income information used to determine eligibility in 2001. In addition to detailed income data, we collected information on insurance-related eligibility, including applicants' responses on relevant questions from the application and renewal forms and the methods workers used to follow up on potential insurance-related eligibility barriers. We also determined whether the case file contained documentation supporting other eligibility decisions, such as duration of Minnesota residency for childless adults and verifications of pregnancy and immigration status.
Table A. 1: File Review Sample New Applications Renewal Applications State County State County Total Number of Applications Submitted January-March 2002 14,931 2,163 12,661 851 30,606
Number Selected for Sample 197 99 201 97 594 Proportion Selected 1.3% 4.6% 1.6% 11.4% 1.9%
SOURCE: Office of the Legislative Auditor sample of Department of Human Services data.
Most of the data reported in Chapter 2 are estimates to the population of new and renewal applications submitted from January through March 2002. To ensure that our estimates reflect the actual proportion of cases processed by DHS and counties and the actual proportion of new and renewal applications, we weighted each case according to the strata to which it belongs. 1 To reflect the precision of each estimate, we also calculated and present in the report a confidence intervala range of values within which we expect the actual value to usually fall. By convention, we used 95 percent confidence intervals. Over time, we would expect 95 percent of such confidence intervals to contain the actual value. To determine the confidence interval for each estimate, we used STATA, a computer software package designed to calculate statistics for stratified, random samples that use weights.
INCOME MATCH To assess how well MinnesotaCare's method of estimating income reflects actual income during the time people are enrolled in the program, we compared MinnesotaCare income determinations made in early 2001 with enrollees' actual household income for 2001. Since MinnesotaCare determines income annually, its income estimate in early 2001 determined eligibility and premiums until the next renewal in early 2002. Thus, we sought measures of income for the entire year, not just income at the time of application. To approximate actual income for 2001, we used income reported on 2001 state and federal tax returns from the Department of Revenue and 2001 wage data reported by employers to the Department of Economic Security's unemployment system.
Our income comparisons are based on a subset of our file review sample. We restricted our analysis to cases (1) for which we were able to obtain income data from 2001 income tax returns or unemployment wage reports, and (2) that were on MinnesotaCare from early 2001 through the end of the year. After applying these criteria to our file review sample, we obtained 215 cases out of 298 renewal applications. Our analysis did not include any cases from our file review sample of 2002 new applications because few of these cases were on MinnesotaCare during 2001.
1 We weighted each case by the inverse of its probability of being selected in our sample.
We also restricted our analysis to individuals within households for whom we had comparable data and whose income was counted, per MinnesotaCare rules, in total household income for 2001. For example, if MinnesotaCare's estimated income included income from two household members but we could only obtain comparable tax or unemployment system data for one member, our income comparisons were based only on the income for that one household member. Finally, we excluded income earned by children under age 19 since their income is generally excluded from household income.
Our income comparisons are based on earned income (including wage income and self-employment income) and unemployment benefits. We did not include other types of unearned income such as social security income, child support, interest, and dividends. For some individuals, we were able to obtain wage data but not self-employment income for 2001. In these cases, we made the comparison based on wage income only, and we excluded any self-employment income that was included in MinnesotaCare's income determination.
While we believe that our comparisons are generally reliable, our results may understate the extent to which MinnesotaCare income is less than actual income because neither the wage data from the unemployment system nor the tax data are complete. For example, wage data from the unemployment system do not include self-employment income, payments to independent contractors, or wages paid by the federal government (such as wages for postal employees). Also, income tax data are incomplete because some workers do not file income tax returns. In addition, at the time we obtained our data in August 2002, the Department of Revenue had not yet processed about 14 percent of 2001 tax returns, partly because of a backlog and partly because of late filings. Finally, income reported under each of these sources does not include income from the underground economy.
Because the data are incomplete, we used the higher of these two sources as our estimate of actual earned income. This reduces but does not eliminate the problem. For example, consider an individual who had two jobs, but only one employer reported wages to the unemployment system. In this case, we would have wage data for the individual but could not tell it was too low unless we were also able to obtain tax return data.
To assess how income discrepancies affected aggregate collection of MinnesotaCare premiums, we compared premiums set by DHS in early 2001 with premiums calculated on the basis of actual earned income for 2001. To estimate the aggregate annual impact on MinnesotaCare premiums for the entire caseload, we assumed that the average impact for cases in our sample was the same for other MinnesotaCare cases. These estimates of the impact on premiums have large margins of error due to various factors, including our limited sample size and the high degree of variability in premium impacts. For example, the margin of error for our $14 million net premium impact estimate was about $8 million, and the 95 percent confidence interval was $5-22 million.
There are various other sources of uncertainty in our estimates, including some conservative assumptions that tend to make our estimates too low and other assumptions that may tend to make them too high. Although we cannot measure the magnitude of these effects, they go in both directions and, thus, may offset each other.
One way in which these estimates may understate the net premium loss to the program is that they do not reflect the fact that enrollees may adjust their premium when their income declines. As we discussed in Chapter 2, some households whose income declined actually had their premiums reduced before the next renewal. However, it is difficult to estimate how much this affects our estimated premium impact because the number of these cases (27 cases would have been eligible for a premium reduction of at least $10 per month) is a very small sample size from which to generalize.
Our estimates of the impact on premiums are also conservative because they are based only on cases for which we were able to obtain comparable income data. These cases represent 75 percent of the renewal cases in our sample. The remaining cases include households without earned income as well as households with earned income but for which we were not able to obtain income data from tax returns or the unemployment system. We assumed that the premium impact for these remaining cases was zero.
Another source of uncertainty is our assumption that households that applied for MinnesotaCare in the first quarter of 2001 had the same income discrepancies as households that applied during the rest of the year. While it is reasonable to assume that the problems we found for cases that applied for renewal during the first quarter exist in other cases, it is not clear that they occur to the same extent. Also, the accuracy of income estimates may vary from year to year because of changes in the economy. Income estimates are more likely to understate future income during periods of economic expansion than during economic declines.
Finally, our estimates may be affected by how enrollees would respond if premiums were increased based on higher income. It is not clear how many enrollees would pay the higher premium, transfer to Medical Assistance (MA), switch to private insurance, or become uninsured. Our estimates assume that enrollees would remain in the program at the higher premium level, even if the actual income places the enrollee over the income limit. 2 To the extent that enrollees switch to MA, the state would lose premium revenue. DHS contends that, because many MinnesotaCare enrollees are eligible for MA, the apparent gains from raising premiums may often not be realized. However, the enrollees included in our analysis who would be charged higher premiums also had incomes substantially higher than the average MinnesotaCare household. As a result, they would be less likely to meet MA income limits. Also, if enrollees reacted to higher premiums by obtaining private insurance or becoming uninsured, the program savings may be higher than estimated.
2 As discussed in Chapter 1, some enrollees may remain on the program paying an unsubsidized premium for 18 months after DHS notifies them that their income exceeds the limit.
INSURANCE SURVEYS In addition to reviewing case files, we independently tested the accuracy of enrollees' responses to questions about insurance available from employers by surveying MinnesotaCare enrollees and their employers. To take advantage of data we had already collected in our file review, we surveyed a subset of enrollees from our file review sample. Specifically, we surveyed all cases from our file review sample that (1) had at least one household member who was active on MinnesotaCare at the time of our computer data extract in May 2002 and (2) had at least one member of the household who reported wage income or unemployment compensation in the current or prior year. In total, 335 cases met these criteria.
We used two surveys to obtain insurance information. First, we surveyed enrolled households to collect the names and addresses of employers for all members of the household. For each employer named, we asked three questions regarding insurance status: (1) whether the employer offered insurance to any of its employees, (2) whether the household member working for that employer was eligible for insurance benefits, and (3) whether any household members were actually enrolled in the employer's health insurance plan. Second, we surveyed the employers named by enrollees on the same issues. For those employers offering health insurance benefits for which the enrollee was eligible, the survey also asked for employer and employee shares of premium costs for various types of coverage.
As shown in Table A. 2, we received responses from 245 of the 335 households surveyed, a response rate of 73 percent. 3 We did not include 51 of these responses in our analysis because, at the time of our survey in August 2002, no one in the household had a wage job, no one was enrolled in MinnesotaCare, or the survey was returned too late. 4 For the remaining 195 cases, we mailed employer surveys to each of the 247 employers identified (some households had more than one employer). Eighty-nine percent of these employers responded, giving us a total of 219 employer surveys that we used in our analysis.
Table A. 2: Enrollee and Employer Survey Response Rates Number Percentage of of Surveys Surveys Mailed Enrollee Survey Enrollee Surveys Mailed 335 Enrollee Surveys Returned 246 73% Returned, but No Employer Survey Mailed 51 15 Returned After Our Deadline 15 4
Currently Unemployed 14 4 Self-Employment Income Only 13 4 Not Currently on MinnesotaCare 7 2 Other Problem 2 1
Surveys Returned With Employers Named 195 58 Employer Survey Employer Surveys Mailed 247
Employer Surveys Returned 219 89
SOURCE: Office of the Legislative Auditor survey data.
3 We mailed the 335 enrollee surveys on August 28, 2002, and sent 182 follow-up surveys on September 17, 2002.
4 Because of the time involved in mailing and getting back employer surveys, our cut-off date for using enrollee responses was October 3, 2002.
Further Reading DeFrancesco, Leanne, Christina Folz, Madeliein Konig, and Carole Lee. State of the States. Washington, DC: Academy Health for State Coverage Initiatives, 2002.
Holahan, John, and Mary Beth Pohl. States as Innovators in Low-Income Health Coverage. Washington, DC: The Urban Institute, 2002.
Kaye, Neva, and Lynda Flowers. How States Have Expanded Medicaid and SCHIP Eligibility. Portland, ME: National Academy for State Health Policy, 2002.
Long, Sharon K., and Stephanie J. Kendall. Recent Changes in Health Policy for Low-Income People in Minnesota. Washington, DC: The Urban Institute, 2002.
Minnesota Department of Health. Employer-based Health Insurance: Family Decisions to Enroll. St. Paul, 2002.
Minnesota's Uninsured: Findings from the 2001 Health Access Survey. St. Paul, 2002.
MinnesotaCare Disenrollee Survey Report. St. Paul, 2002.
Minnesota Department of Human Services. A Profile of the MinnesotaCare Program: The First Six Years. St. Paul, 1998.
Minnesota House of Representatives Research Department. Minnesota Family Assistance: A Guide to Public Programs Providing Assistance to Minnesota Families. St. Paul, 2002.
MinnesotaCare Information Brief. St. Paul, 2002.
Oliver, Leah N., and Martha P. King. SCHIP: Money Matters. Denver, CO: National Conference of State Legislatures, 2002.
Socks, Heather, Todd Kutyla, and Sharon Silow-Carroll. Toward Comprehensive Health Coverage for All: Summaries of 20 State Planning Grants From the U. S. Health Resources and Services Administration. New York: The Commonwealth Fund, 2002.
State Coverage Initiatives. State Coverage Matrix. Washington, DC: Academy Health, 2002; http://www.statecoverage.net/matrix.htm.
Lafayette Road North Saint Paul, Minnesota 55155 An Equal Opportunity Employer January 9, 2003 James R. Nobles, Legislative Auditor Office of the Legislative Auditor Centennial Office Building 658 Cedar St. St. Paul, MN 55155
Dear Mr. Nobles: The Minnesota Department of Human Services appreciates the opportunity to respond to the January, 2003 program evaluation report on MinnesotaCare. The report will help guide us as we continue to improve the program's operations. We appreciate the acknowledgement of our improvement in worker productivity and in the timeliness of processing new applications. It confirms our assessment that a significant investment in automation is necessary to support the proper administration of MinnesotaCare to assure maximum accuracy and efficiency. We will apply knowledge gained from this report as we continue to improve our operations.
The Department's goal is to provide efficient access to MinnesotaCare enrollment while achieving a high level of program integrity. The report identifies the challenges we face as we pursue our goal. Steady growth in enrollment, fluctuations in the availability of qualified staff and continuous changes in program policy has affected the Department's ability to monitor and address administrative performance (see Attachment 1 for a legislative history of MinnesotaCare policy changes). Combining large caseloads with manual processing of complicated eligibility rules has resulted in an unacceptable level of inaccuracy. Your review, by design, cast a broad net to look both at eligibility and processing errors. That approach is useful in highlighting opportunities for improvement. While we agree with many of your findings and see the opportunity for improvement, not all the documentation errors in your first finding resulted in inappropriate eligibility results or premium calculations.
We would like to clarify your second finding. State law requires the 12-month projection used currently. Your reviewers found that savings could by achieved if that law was changed. The report summary (p. XI) states that "income estimates for MinnesotaCare frequently did not match actual income, often because income changed after eligibility was determined . Had premiums been based on actual income, enrollees would have paid an estimated $5-22 million more in annual premiums." When DHS and county workers review applications, they look at a snapshot of annual income based on the point in time in which the application is submitted. The report considered estimate determinations made in early 2001, then compared the data to actual income earned throughout the calendar year. The estimates were not necessarily erroneous; rather, they were calculations based on information required at the time of enrollment. Increasing the frequency of income reviews may more accurately reflect enrollee income, but because income is fluid in many cases, it would be difficult to ever capture total accuracy.
We are proposing to review eligibility more frequently and will build the capacity to perform third party income matches into HealthMatch, so we take your recommendations very seriously. Naturally, implementing more frequent reviews comes with a cost. In the long term, the Department will resolve a majority of the issues and concerns raised in this report through our commitment to develop a fully automated eligibility system called HealthMatch. Access to health care program enrollment and program integrity will be addressed with this new technology. HealthMatch will assure consistent and accurate data processing and allow applicants and enrollees to access program information and preliminary eligibility results via the Internet. Interfaces with data from other government agencies and vendors will be used to confirm information provided by the enrollees and improve the accuracy of eligibility determinations. An electronic document management system, to be implemented this summer, will eliminate the inefficiencies related to paper files and improve workflow. We are pleased that the report recognizes and supports the value of this long-term solution. While this new system is at least 18 months from implementation, there are a number of short-term solutions that are being put in place: DHS is updating its MinnesotaCare policy manual so that our policies and procedures strictly follow statute and are understandable for the workers who enroll applicants;
DHS is close to implementing an electronic case management system that will allow quicker access to case files and more flexibility in assigning work; DHS also will launch two pilot projects later this month that will focus on applicants who declare zero income and those who indicate no employer-subsidized insurance, both regardless of employment status and for those employed 30 hours per week. We expect that this information will alert us to cases in which: 1) applicants may have had a change in income since applying (e. g., got a new job, were approved for unemployment insurance) or may have misunderstood the question; and, 2) applicants have access to employer-subsidized insurance but were unaware, misunderstood coverage, or failed to provide an accurate answer; and,
DHS has assigned an employee with fraud investigation experience to examine allegations of MinnesotaCare fraud and implement an auditing process to verify targeted information. It is important to note that focus on a short-term response will require a reevaluation of priorities. Some of the recommendations are likely to require additional staff in order to achieve both access to enrollment and program integrity. For example, amending Minnesota Statute to allow for premium increases when enrollee income rises will result in additional premium revenue. However, the savings should be offset with an appropriation for additional staff to process the added workload and detect unreported income increases. The alternative, more work without additional staff, will likely result in slower processing and declining service for eligible applicants and enrollees.
Additional training and oversight may also have added costs. An analysis of the costs vs. benefits will be completed as we proceed with our interim actions. Attachment 2 outlines current and proposed activities to address immediate needs. Conclusion The report recognizes that administering MinnesotaCare requires that there be a proper balance between providing reasonably easy access to the program through an efficient, comprehensible application process and ensuring that workers have sufficiently detailed information to make accurate decisions regarding eligibility. On the one hand, if the level of scrutiny is too high, where the application process is lengthier and more intrusive, eligible Minnesotans may be reluctant to apply and eligible people will wait unnecessarily for health care coverage. Conversely, DHS is responsible for providing a thorough review of each case to ensure that only qualified applicants are enrolled in the program.
MinnesotaCare is a critical tool for Minnesotans who don't have health insurance. When it comes to health care, this program can be credited with helping Minnesota achieve one of the lowest rates of uninsured in the country. DHS remains committed to targeting MinnesotaCare to lower-income individuals who do not have access to affordable health insurance while controlling program costs.
In closing, we compliment your staff on its professionalism and thoroughness. As you know, MinnesotaCare is a complicated program. DHS supports revisions resulting from this report that ultimately contribute to the program's continued success. Thank you, again, for the opportunity to respond to this report.
Sincerely, /s/ Kevin Goodno Kevin Goodno Commissioner
Attachment #1 Minnesota Department of Human Services Health Care Eligibility and Access January 14, 2003 1
DATE EVENT SOURCE 1987 Legislature passed the Children's Health Plan (CHP). The program provided coverage to pregnant women and children under six years of age who had gross incomes up to 185% FPG, and were ineligible for MA or GAMC, and otherwise uninsured.
1987 Ch. 403, Art. 2, Sec. 63
1988 Federal government amended the Medicaid program to allow coverage of pregnant women and infants under age one in families with income up to 185% FPG. As a result, the Legislature amended CHP to exclude pregnant women and infants and expanded CHP eligibility to include children ages one through eight. The annual enrollment fee per child was $25, not to exceed $150 per family. Funding for the program came from enrollment fees and a one-cent increase in the cigarette tax.
1988 Ch. 689, Art. 2, Sec. 137, 144
July 1, 1988 Enrollment in the CHP began. A Profile of the MinnesotaCare Program: The First Six Years 1992-1996, p. 1 (Hereafter, "Profile") 1989 Creation of the Health Care Access Commission (HCAC) Group charged with gathering data and recommending a plan to insure the uninsured. Profile, p. i. July 1, 1989 Legislature expanded CHP eligibility to include children 1 to 18. This expansion was to become effective in 1991. 1989 Ch. 282, Art. 3, Sec. 33 January 1991 HCAC submits final report to Legislature. Recommends sweeping reform measures, emphasizing universal access to health care, helping lower-income individuals and families, ending discrimination in underwriting and other insurance practices, and providing equitable benefits.
Profile, p. 3.
May 1991 Health care legislation recommended the Health Care Access Commission was passed by the Legislature but vetoed by Governor Arne Carlson. Profile, p. i.
1992 The HealthRight Act (now known as the MinnesotaCare Act) which was drafted by a bipartisan group of seven legislators was passed by the Legislature and signed by Governor Carlson. The Act extended coverage to parents and dependent siblings of children already eligible for CHP. These individuals had to be MN residents, have household incomes less than 185% FPG, be underinsured, and not have access to ESI within the previous 18 months. Enrollees charged premium based on a sliding-scale.
1992 Ch. 549, Art. 4 MinnesotaCare SONAR
Attachment #1 Minnesota Department of Human Services Health Care Eligibility and Access January 14, 2003 2
DATE EVENT SOURCE The act established a two-year phase-in period of expanding eligibility: October 1992 families of children currently covered under CHP; January 1993 all families with children; July 1, 1994 single adults and couples without children. July 1, 1992 A new 5 cents per-pack tax on cigarettes is effective. 1992 Ch. 549, Art. 4, Sec. 16. Oct. 1, 1992 MinnesotaCare program begins. Children 1 to 18 with gross family incomes less than 185% FPG and who are not eligible for MA or otherwise insured are eligible. Parents and dependent siblings residing in the same household as an eligible child are also eligible.
1992 Ch. 549, Art. 4, Sec. 5
Jan. 1, 1993 Families and children eligible at 275% FPG. 1992 Ch. 549, Art. 4, Sec. 5 Jan. 1, 1993 A 2% tax on gross patient revenues of hospitals and surgical centers became effective. 1992 Ch. 549, Art. 9, Sec. 7 July 1, 1993 CHP enrollees were transferred into MinnesotaCare. Children in households with income below 150% FPG were charged a premium of $4 per month. 1993 Ch. 345, Art. 9, Sec. 7
July 1, 1993 Pregnant women and infants with incomes greater than or equal to 185% FPG are required to pay a premium. 1993 Ch. 345, Art. 9, Sec. 12 July 1, 1993 Health care services were expanded to include inpatient hospital benefits. 1993 Ch. 345, Art. 9, Sec. 3 July 1, 1993 No asset test for children or parents (under previous law, only children were exempted). 1993, Ch. 345, Art. 9, Sec. 13 Jan. 1, 1994 A 2% health care provider tax on gross revenues or licensed health care providers including doctors, dentists, chiropractors, wholesale drug distributors, pharmacies. . . etc. became effective. 1992 Ch. 549, Art. 9, Sec. 7
Recent Program Evaluations: Residential Facilities for Juvenile Offenders, February 1995 95-01 Health Care Administrative Costs, February 1995 95-02 Guardians Ad Litem, February 1995 95-03 Early Retirement Incentives, March 1995 95-04 State Employee Training: A Best Practices Review, April 1995 95-05 Snow and Ice Control: A Best Practices Review, May 1995 95-06 Pollution Control Agency's Use of Administrative Penalty Orders, Update July 1995 95-07 Development and Use of the 1994 Agency Performance Reports, July 1995 PR95-22 State Agency Use of Customer Satisfaction Surveys, October 1995 PR95-23 Funding for Probation Services, January 1996 96-01 Department of Human Rights, January 1996 96-02 Trends in State and Local Government Spending, February 1996 96-03 State Grant and Loan Programs for Businesses February 1996 96-04 Post-Secondary Enrollment Options Program, March 1996 96-05 Tax Increment Financing, March 1996 96-06 Property Assessments: Structure and Appeals, A Best Practices Review, May 1996 96-07 Recidivism of Adult Felons, January 1997 97-01 Nursing Home Rates in the Upper Midwest, January 1997 97-02 Special Education, January 1997 97-03 Ethanol Programs, February 1997 97-04 Statewide Systems Project, February 1997 97-05 Highway Spending, March 1997 97-06 Non-Felony Prosecution, A Best Practices Review, April 1997 97-07 Social Service Mandates Reform, July 1997 97-08 Child Protective Services, January 1998 98-01 Remedial Education, January 1998 98-02 Transit Services, February 1998 98-03 State Building Maintenance, February 1998 98-04 School Trust Land, March 1998 98-05 9-1-1 Dispatching: A Best Practices Review, March 1998 98-06 Minnesota State High School League, June 1998 98-07 State Building Code, January 1999 99-01 Juvenile Out-of-Home Placement, January 1999 99-02 Metropolitan Mosquito Control District, January 1999 99-03 Animal Feedlot Regulation, January 1999 99-04 Occupational Regulation, February 1999 99-05 Directory of Regulated Occupations in Minnesota, February 1999 99-05b Counties' Use of Administrative Penalties for Violations of Solid and Hazardous Waste Ordinances, February 1999 99-06 Fire Services: A Best Practices Review, April 1999 99-07 State Mandates on Local Governments, January 2000 00-01 State Park Management, January 2000 00-02 Welfare Reform, January 2000 00-03 School District Finances, February 2000 00-04 State Employee Compensation, February 2000 00-05 Preventive Maintenance for Local Government Buildings: A Best Practices Review, April 2000 00-06 The MnSCU Merger, August 2000 00-07 Early Childhood Education Programs, January 2001 01-01 District Courts, January 2001 01-02 Affordable Housing, January 2001 01-03 Insurance for Behavioral Health Care, February 2001 01-04 Chronic Offenders, February 2001 01-05 State Archaeologist, April 2001 01-06 Recycling and Waste Reduction, January 2002 02-01 Minnesota Pollution Control Agency Funding, January 2002 02-02 Water Quality: Permitting and Compliance Monitoring, January 2002 02-03 Financing Unemployment Insurance, January 2002 02-04 Economic Status of Welfare Recipients, January 2002 02-05 State Employee Health Insurance, February 2002 02-06 Teacher Recruitment and Retention: Summary of Major Studies, March 2002 02-07 Local E-Government: A Best Practices Review, April 2002 02-08 Managing Local Government Computer Systems: A Best Practices Review, April 2002 02-09 State-Funded Trails for Motorized Recreation, January 2003 03-01 Professional/Technical Contracting, January 2003 03-02 MinnesotaCare, January 2003 03-03 Metropolitan Airports Commission, January 2003 03-04 Preserving Housing: A Best Practices Review, April 2003 03-05 Medicaid Fraud and Abuse, forthcoming Charter School Financial Accountability, forthcoming Higher Education Tuition Reciprocity, forthcoming