Thinking About Medicaid Buy-In Enrollment Projections for Ohio

Lessons from Other States

August 2004

Table of Contents

Chapter 1 Executive Summary

Chapter 2 Introduction

Chapter 3 Enrollment Lessons from Other States

Chapter 4 Cost Projections

Chapter 5 Conclusions

About the Author

Acknowledgements

Chapter 1

Executive Summary

Thinking about Medicaid Buy-In Enrollment Projections for Ohio

Introduction

American attitudes towards persons with severe disabilities have changed dramatically over the last few decades. No longer is it customarily assumed that disabling conditions render a person unsuitable to do work, or that persons with severe disabilities necessarily require on-going cash assistance. Nevertheless, only 18% of Ohioans 18-64 years old with severe disabilities work at all during the course of a year.

While only one of several important federal work incentives, Medicaid Buy-In is widely viewed as one of the most critically important reforms to help people with severe disabilities gain and keep employment that will enable them to contribute financially to their well-being. Under Medicaid Buy-In, participants can earn wages while still maintaining health insurance through the Medicaid program.

Aside from the Massachusetts demonstration project, Medicaid Buy-In programs are only a few years old. They have now been implemented in about half the states. Ohio seriously considered the implementation of a program in 2001. This report was prepared to inform current efforts to design and gain legislative support for an Ohio program.

Enrollment Lessons from Other States

• The degree to which people with disabilities are integrated into the workforce in the state prior to the implementation of Medicaid Buy-In. Ohio scores about average in this respect.
• The degree to which program policies encourage higher or lower enrollments. Ohio’s advocates are proposing policies that can largely be characterized as moderate in terms of implications for enrollment.
• The degree to which advocacy organizations help mobilize the community of persons with disabilities to take advantage of Medicaid Buy-In. The advocates of an Ohio program believe themselves to be well organized but working relationships with the state Medicaid agency need strengthening.
• The degree to which existing state Medicaid programs are structured in such a way as to make Medicaid Buy-In more or less attractive to persons with disabilities who are interested in working. Ohio’s income eligibility standard for its existing Medicaid program should cause Medicaid Buy-In to be viewed as an attractive option. This will operate to inflate enrollment.
• How long the program has been operating.

The net effect of all of these factors should be to produce an enrollment level in Ohio in line with the recent projection of 7,073 people that was produced by The Lewin Group. However, there are two important caveats. First, this research underscores the fact that it will take several years before that level of enrollment is reached. Second, this research projects a different mix of participants new to the system than did Lewin, with the effect of lowering the net cost of the program.

Cost Projections

The cost of a Medicaid Buy-In program equals participant costs plus overhead expenses minus revenue enhancements. This report does not attempt to estimate overhead expenses associated with program operation and marketing, nor does it discuss revenue enhancements.

The Lewin Group projected participant costs to Ohio of $22.2 million, based on a cost of $906/month for participants new to the Medicaid system and $118/month net cost for participants who already receive Medicaid (but who would no longer have to pay some of their medical expenses out of pocket). The present analysis suggests that the Lewin cost estimate is too high, not because of their estimated cost per participant, but because they assumed a mix of existing and new Medicaid recipients remarkably different from that found in other states. Assuming Ohio’s mix is similar to the average in other states, the participant costs could be only half of what Lewin estimated.

Chapter 2

Introduction

Background and Objectives

Preliminary Considerations

About 425,000 Ohio adults between the ages of 18 and 64 have severe disabilities. Throughout this report, the phrase severe disability refers to a level of physical or mental impairment sufficiently disabling that the person would qualify for Social Security Disability Income (SSDI) or Supplemental Security Income (SSI). These two income assistance programs are operated by the Social Security Administration (SSA). Individuals must be found eligible through a disability determination performed by the state (in Ohio, the evaluations are done by the Bureau of Disability Determination, part of the Ohio Rehabilitation Services Commission).

• SSDI is an insurance program funded by monthly FICA taxes withheld from the wages of workers covered by Social Security. Workers are eligible for SSDI after 20 quarters of work. Monthly payments are graduated to reflect the level of earnings prior to the disability.
• In contrast, SSI is an entitlement program for persons with disabling conditions who have not accumulated 20 quarters of work history. Monthly SSI payments for an Ohio individual living alone in 2002 averaged $412.

A person might receive both SSI and SSDI if his or her earnings history was such that the SSDI income benefits are less than the SSI benefit amount. Without some reversal of the disabling condition, individuals may expect to continue receiving income assistance until the age of 65, whereupon Social Security Retirement benefits begin.

It is important to emphasize that not every person with a severe disability receives either form of federal income assistance. Not enough is known about why individuals opt out of such coverage or why they might be judged ineligible for it. Some workers (e.g., state employees) are not covered by Social Security but do have other disability insurance. Some people with a disability probably choose not to apply for income assistance. No doubt, some who do apply are incorrectly denied. Finally, of course, some individuals are able to work and achieve self-sufficiency in spite of having physical or mental disabilities that would render another person eligible for income assistance.

American attitudes towards persons with severely disabling conditions have undergone a remarkable change over the past 30 years. Until relatively recently, persons with disabilities were consigned to a lifetime of severely diminished expectations by others; we did not expect them to be self-sufficient, or even to contribute to their ability to live in the community. Today, there is widespread recognition that people with severe disabilities can contribute to their ability to live in the community and, further, that they, their families, and their communities fare better if their efforts to achieve at least partial self-sufficiency are nurtured and reinforced. Some legislative milestones that mark this evolution in thought include the Americans with Disabilities Act of 1990, the Balanced Budget Act of 1997 and the Ticket to Work and Work Incentives Improvement Act of 1999.

The Logic of Disability Programs - Then and Now

Logic modeling is an important new tool in public administration and program evaluation. Upon first exposure, logic models strike some people as trivial. A logic model strips a program down to its essence. What is this program trying to do? How is it designed to work? Bereft of details about requirements and procedures, logic models starkly portray what programs are about, and the results are sometimes disconcerting.

What was the logic of programming for persons with severe disabilities in some comfortably distant enough past - say the 1950s - that it will elicit no defensiveness from the reader? The following figure suggests that starting from the assumptions that disabilities are permanent and that adaptability is low, the best solution to caring for persons with disabilities is to institutionalize them, isolate them, and enforce that isolation from the mainstream of community life. Keep them in homes or at home, send them to special schools, create sheltered workshops for them. Maintain low expectations. Certainly do not accommodate such people in normal schools or the workplace. Disabled persons (the condition defining the individual) were treated comparatively better than they had been before the movement to adopt a medical model of disabilities began in the 19th century, and reasonably well from a narrow perspective of physical safety and health. As a result, the general community felt its treatment of such people was progressive.

Many consumers, their relatives and the professionals who serve them worked tirelessly for many years to undermine the conventional wisdom embodied in the above logic model. Mental health consumer advocates argued that perhaps the focus should be more on adaptation than cure. Civil rights attorneys fought involuntary hospitalization practices. Inventors developed improved assistive technology. Educators studied the salutary effects of mainstreaming children with special needs. In addition, legislators listened to their constituents and proposed measures to help advance all of these ideas. A more modern attitude has gradually developed, although it has not fully permeated program design or the way persons with disabilities think about themselves. This new look is built on three assumptions.

• Some physical and mental health problems may be permanent but people are resilient and adaptive. With assistance, people can at least partly rise above such impairments so as not to be disabled by them.
• People with disabilities do better when they can make an economic contribution to their well-being, even if they cannot be wholly self-sufficient economically, and when they are subject to the same kinds of expectations about self-sufficiency that are applied to people in general, if not perhaps the same level of expectations.
• Our children, employees and neighbors do not have to be protected from people with disabilities.

Given the striking discrepancies between these new assumptions about people with disabilities and our older notions that still echo in discourse about people with disabilities, it should not be surprising to learn that the implications of this new look have not everywhere been incorporated fully into the way people think about themselves or the way programs operate.

Employment for Ohioans with Severe Disabilities

There are about 425,000 Ohioans who are between the ages of 18 and 64 and who have severe disabilities, or 6.1% of the population of 6,957,044 people ages 18-64 in Ohio. These statistics, like all of the ones to follow in this section of the report, are based on a pooled data file from the Current Population Survey for Ohio for the period 1996-2002 using methodology previously reported.

Of these persons with severe disabilities, 47% receive SSDI, 39% receive SSI, and 79% receive one or the other, or both. Thus, SSA policies exert a powerful influence on persons with severe disabilities. Since Medicare is a concomitant benefit of SSDI and since Medicaid often accompanies SSI benefits in Ohio , it is not surprising that relatively few persons with severe disabilities lack health insurance.

One-third of persons with severe disabilities live in families with incomes below 100% of the Federal Poverty Level (FPL). While representing just 6% of all persons 18-64, families containing persons with severe disabilities represent 12% of families below 50% of the FPL, 30% of families between 50% and 99% of the FPL and 16% of families between 100% and 149% of poverty. Thus, an important segment of the population of poor people in Ohio is the group with severe disabilities. If it were possible to increase the self-sufficiency of people with severe disabilities, a concomitant benefit would be to decrease the poverty rate and the cost of poverty. Households that contain persons with severe disabilities represent 30% of households with adults 18-64 who receive food stamps and 29% of the households who have subsidized rents. Almost 20% of people who receive public assistance (i.e., welfare, or more properly, Temporary Assistance for Needy Families) have a severe disability.

Severe disabilities increase with age. Only 3% of Ohioans 18-29 years old have severe disabilities, but that number increases to 11% of Ohioans 50-65 years old. Age is a risk factor because older individuals have had a longer period of exposure to accidents and because the risk of some chronic diseases and conditions increases with age.

Only 18% of Ohioans 18-64 with severe disabilities worked at all during the previous calendar year (compared to 86% of other Ohioans 18-64). Another 2% looked for work. In other words, 80% of Ohioans 18-64 with severe disabilities had no standing in the labor market during the previous calendar year. When asked why they were not working or looking for work, 87% cited their illness or disability. The point in time unemployment rate for persons with severe disabilities is nearly 18% (meaning that 18% of individuals who are working or are looking for work are unemployed at the time interviewed).

On the one hand, the new look in thinking about people with disabilities holds that everyone should at least contribute to their economic well-being. On the other hand, the low labor force participation rate and high unemployment rate for people with disabilities suggest that most persons with severe disabilities have little earned income. The reasons for this discrepancy between expectations and outcomes are the numerous and substantial barriers to labor force participation faced by persons with severe disabilities.

This is a report designed to inform planning for a Medicaid Buy-In program in Ohio. In greatly over-simplified terms, Buy-In programs allow persons with severe disabilities to escape from the dilemma of having to choose health insurance or work. While people with severe disabilities are often interested in working, they are dependent upon health insurance to a critical degree. If an employer does not provide adequate health insurance, either because it is prohibitively expensive or merely to control labor costs by hiring people part-time instead of full-time, people with severe disabilities may feel trapped into maintaining their SSDI or SSI eligibility status to ensure the continuation of the health insurance. In other words, without Buy-In, the disincentives to work outweigh the incentives.

Implementing the New Logic of Disability Programs

It is useful first to place Medicaid Buy-In within the context of more general efforts to change the logic model for programs aimed at persons with disabilities. There has been a broadly supported movement in America over the past 14 years to increase the opportunities for persons with disabilities to live independently and with a greater degree of self-sufficiency. Some of the more important landmarks along the way have been the Americans with Disabilities Act of 1990, signed by the first President Bush, the Ticket to Work and Work Incentives Improvement Act of 1999 (TWWIIA), signed by President Clinton, and the New Freedom Initiative of President George W. Bush.

To appreciate just how far-reaching this movement is, consider the aims of the New Freedom Initiative, which seeks to ensure that persons with disabilities have improved educational opportunities, improved employment opportunities, and improved opportunities to live where they choose and to participate in the life of their communities. Among the focal areas of the initiative are employment, education, housing, transportation, health, income supports, assistive technology, independent living and civil rights. Later sections of this report will allude to some equivocal results of Medicaid Buy-In programs over the first several years of their existence. Thus, it is worthwhile stating explicitly why the New Freedom Initiative casts so wide a net: Implementing the new thinking about persons with severe disabilities is not as simple as making work pay, and in particular, it is not as simple as implementing Medicaid Buy-In.

Even within the domain of making work pay, Buy-In is only one tool among many. Following are descriptions of some of the work incentives being encouraged and supported by the Social Security Administration:

• Persons with disabilities can deduct the expense of impairment related work expenses from their earnings in order to keep their earned income below the level of Substantial Gainful Activity (SGA), beyond which their cash assistance benefits are reduced. A person who earns $900 a month but must pay $200 from their earnings for specialized transportation to work can still stay below the SGA level of $780.
• Persons who make attempts to work that are unsuccessful because the disabling condition could not be overcome are not penalized for these efforts in the calculation of SSDI or SSI benefit amounts.
• Under certain conditions, including the person participating in a program of vocational rehabilitation, SSDI and SSI benefit amounts can be continued after a person’s medical condition improves to the point that he or she no longer meets criteria for having a disability.
• Full income benefits may be continued during trial work periods (up to nine months) even if earnings exceed the SGA limit.
• Persons on SSDI can be reinstated without a new disability determination following the conclusion of a trial work period if their earnings fall below the SGA level, assuming that the reason for leaving the job was the disabling medical condition.

While all of these policy incentives to make work pay are important, the idea that loss of health insurance is the critical disincentive to work has had currency. As long as ten years prior to the formal legislative authorization for Medicaid Buy-In programs, SSI recipients could, under the Section 1619(b) provision, maintain their Medicaid coverage even if their earnings exceeded SGA levels. The number of SSI recipients who took advantage of 1619(b) was always relatively small, and Buy-In may properly be understood as a liberalization of the 1619(b) eligibility criteria. Still, in retrospect, the 1619(b) enrollments raise questions about the adequacy of better health insurance alone as a tool to help persons with disabilities gain independence through work. This report will also document that there is remarkable variability among states with Buy-In programs in the percentage of the population of persons with severe disabilities who enroll. Two or three years ago, this variability would have been attributed by most Buy-In enthusiasts to the relative newness of the programs. This report instead concludes that as important as Buy-In programs are, they are merely one more brick in the foundation of supports that has to be constructed to make the vision of the New Freedom Initiative a reality.

Medicaid Buy-In

The federal Balanced Budget Act of 1997 (BBA) first permitted states to create Medicaid Buy-In programs. However, the more important piece of enabling legislation was the Ticket to Work and Work Incentives Improvement Act of 1999 (TWWIIA). Not only did TWWIIA give states greater flexibility in the design of a Buy-In program than did the BBA, Buy-In was also better placed amidst an assortment of tools designed to encourage persons with severe disabilities to make efforts to secure and maintain employment.

TWWIIA created two new Medicaid eligibility groups: a Basic Coverage Group and a Medical Improvement Group. The second of these - the Medical Improvement Group (MIG) - is less relevant to the present analysis than the first, if for no other reason than that states which offer a MIG must also provide a Basic Coverage Group and Ohio has not yet created a Basic Coverage Group (Medicaid expansion under TWWIIA requires state enabling legislation). Thus, for the remainder of this report, Medicaid Buy-In will refer, minimally, to a Basic Coverage Group.

In creating a Buy-In program, states have the opportunity to offer Medicaid coverage to working individuals 16-64 years old who would meet standards of disability under SSI or SSDI even though they exceed income or resource limitations that would disqualify them for SSI. It is important to note that individuals do not have to actually be receiving SSDI or SSI at the time of their enrollment in Buy-In, but merely qualify based on their level of disability. States have great flexibility in designing eligibility standards for a Buy-In program.

• Under TWWIIA, states may set income eligibility criteria for a Buy-In program as high as 450% of the Federal Poverty Level guidelines (FPL) . The 2004 FPL for a person living alone is $9,310, so using the 450% standard would mean that a state could declare that anyone living alone with an income less than $41,895 would meet the income standard for Buy-In eligibility.
• States have the further option of disregarding kinds or amounts of either earned or unearned income. States vary widely in their policies. Minnesota chose to disregard all unearned or earned income, thus effectively declaring that there were no income restrictions for its Buy-In program. Some states assess income before taxes and others after taxes. States may count the earnings of spouses and other family members or not.
• Similarly, states differ greatly in terms of their policies regarding other resources. All states disregard the value of the primary residence and one automobile in determining eligibility. Most disregard savings in qualified retirement programs. Some states have policies that disregard Medical Savings Accounts. All states have explicit limits on cash resources, such as savings accounts and non-qualified investments, but these limits range from as little as $2,000 to as high as $40,000.

States also have wide latitude in designing premium policies and co-pay policies and, as with eligibility criteria, states have enacted strikingly different rules. While in principle higher premiums should and probably do limit enrollment, other things being equal, in reality Buy-In is still so new that those "other things" are seldom equal enough that the impact of premium policies can be understood fully.

Once a person begins working, states may require continuous employment or they may allow a worker to have some gaps in employment. Further, these requirements may be implemented along with strict monitoring requirements or with reasonably lax oversight that effectively softens the impact of any requirement that a person be continuously employed.

As compared to ordinary Medicaid, Buy-In has three important advantages:

• People are offered an incentive to work because earnings beyond the level of Substantial Gainful Activity may be retained and used to improve quality of life.
• As compared to Social Security Section 1619(a) and (b) provisions, which permit some earnings to be retained, Buy-In may allow people greater opportunity to accumulate assets.
• Medicaid recipients with a spend-down (that is, who have income in excess of the Medicaid income standard that needs to be expended each month on medical expenses in order to maintain eligibility for Medicaid) often find that managing their spend-down is difficult. Buy-In with premium payments promises to be easier for people to manage.

As of mid 2002, 25 states had passed enabling legislation for a Buy-In program and had started enrolling participants. Several other states have subsequently initiated programs for which little if any data are available.

Previous Relevant Ohio Research

Medicaid Buy-In programs show enormous variation. Advocates in Ohio have urged the legislature to implement enabling legislation for an Ohio Buy-In program and, naturally, legislators and administrators in the Oho Department of Job and Family Services (ODJFS) wish to extract and digest whatever lessons can be learned from research before committing themselves to a design. Following is a description of the three pieces of previous work that bear on the design of an Ohio program and that strongly influenced the direction of the current research.

Projecting Enrollment in a Medicaid Buy-In Program for Ohio

With funding from the Ohio Developmental Disabilities Council, Howe (2001) developed a series of projections for the number of people who might eventually enroll in an Ohio Medicaid Buy-In program under a variety of assumptions concerning eligibility criteria and premium collections. The Ticket to Work Study Group of the Ohio Legislature (chaired by Senator Bill Harris) recommended the following set of policy options:

• Family income would be capped at 250% FPL (or, for a person living alone in 2004 dollar terms, $23,275) but persons with a severe disability who were working would be given an earned income disregard of $20,000/year, meaning that someone could earn as much as $43,275 and still be eligible for Buy-In.
• Adults living with their parents would be treated as one-person households for eligibility determination, thus making their parents’ income and resources irrelevant.
• Resources would be capped at $10,000 disregarding a primary residence, an automobile, qualified retirement accounts and medical savings accounts.
• Premiums would be set to 10% of income in excess of 150% FPL. Thus, in 2004, an individual living alone would pay nothing in premiums as long as his or her income was less than $13,965. If the person earned $17,500, his or her annual premium would be 10% of the difference between $17,500 and $13,965.

The projection methodology necessarily made many other assumptions about the percentage of the population of persons with severe disabilities who would work, the percentage who would want to take advantage of Buy-In, etc. Rather than reviewing these projections in detail here, the more important point to make is that the projections were done just two years after the first Buy-In programs had started program operations. Because Minnesota had, in just two years, enrolled over 5,500 people (in a state with less than half the population of Ohio), the assumptions underlying the projections were nearly certainly too skewed in the direction of higher enrollments. Howe projected that 12,542 Ohioans might eventually enroll in a Medicaid Buy-In program.

Administrative and Systems Changes for Implementation of a Medicaid Buy-In Program

The Lewin Group was retained by ODJFS to study options for designing and implementing the administrative systems necessary to operate an Ohio Buy-In program. The final report was published in August 2003.

Based on data collected from 15 other states with Buy-In programs, Lewin reviewed options for 1) eligibility determination, 2) disability determination, 3) premium collection, 4) administrative systems changes, 5) staffing, 6) training and 7) outreach and marketing. The report makes no recommendations in these areas but does discuss the advantages and disadvantages of various options and often attempts to estimate the cost of implementing each of the options.

More relevant to the current analysis, the Lewin report draws several conclusions about the experiences other states have had with Buy-In programs. Those that are most important to mention because they provide a basis for some of the conclusions of this report are as follows:

• Enrollments in other states vary dramatically, but in nearly all states, most Buy-In participants were previously enrolled in other Medicaid programs. This is extremely important because in Howe’s original Ohio projections it was estimated that the effect of a Buy-In program would be to draw very significant numbers of people into the Medicaid system who were not previously recipients of Medicaid. Indeed, Howe had projected that eventually as many as 82% of Buy-In enrollees might be new to Medicaid.
• Lewin also found that Buy-In program participants are believed by some key informants to be suppressing their earnings to keep their incomes below the SGA (beyond which they might lose their eligibility for SSDI or SSI benefits). While not ignoring the fact that relatively few Buy-In participants put their SSDI and SSI benefits at risk, this report concludes it is too early to know whether people are increasing their earnings under Buy-In, and if not, why. For example, some advocates believe strongly that employers control their labor costs by offering persons with disabilities part-time, but not full-time work. Buy-In is almost certainly a necessary ingredient in the transition from dependency to self-sufficiency, though it is possible that it will not be a sufficient reason.
• Premium collection systems are expensive. States collect relatively little in premiums, which probably means that premiums are less a source of revenue than a means of enrollment management.
• Lewin states, "The advocacy community is an important partner in policy development, the legislative process, and outreach."
• While framing the issue quite differently than other experts have, the Lewin report does raise the important issue that a Buy-In program may serve people other than those committed to engaging in substantial work activities. In contrast to the tone of the Lewin conclusion, Jenson, et al note that a state’s Buy-In policy may legitimately have one of two quite different policy objectives: either to serve people committed to substantial work or to serve to increase the disposal income of people on SSDI and SSI.
• Lastly, among those Lewin conclusions most relevant to this analysis, workers should understand the goals of the program. As strange as it may sound, it appears as though one of the difficulties in assessing the impact of Buy-In programs nationwide is that neither workers nor enrollees are necessarily aware that a Buy-In program has over-arching goals that go beyond those of other Medicaid programs.

Study of Medicaid Eligibility Options

Lewin also contracted with ODJFS to prepare a projection of the participant-based costs of a Medicaid Buy-In program in Ohio (as opposed to the administrative costs of operating the program). Participant-based costs are a function of the number of enrollees and the average cost of coverage/enrollee. The final report was issued in November 2003.

In preparing their report, Lewin took note of the fact that Ohio’s status as a 209(b) state means that it uses different standards for SSI and Medicaid eligibility determination (in other words, in non-209(b) states, eligibility for SSI confers automatic eligibility for Medicaid). In Ohio, a person with a severe disability who is eligible for SSI is only eligible for Medicaid if their income is less than 64% of the FPL ($479/month in 2003). With an income above that amount, Ohioans who desire Medicaid cover must be given an opportunity to "spend down" to the eligibility standard in order to receive coverage. If a person has an income of $600/month, they would have to spend over $100 out of pocket on medical expenses before being eligible for Medicaid. This Ohio Medicaid need standard is among the lowest in the country. Regardless of the state of residence, an SSI recipient becomes eligible for coverage under the 1619(a) and 1619(b) provisions of Social Security once their earned income reaches the level of Substantial Gainful Activity (SGA), or $780/month.

Lewin conducted its own micro-simulation of enrollment in a Medicaid Buy-In program using methodology similar to that developed by Howe (2001). They examined five different Buy-In scenarios, one of which incorporated the recommendations of the Ohio Ticket to Work Study Group (recall that Howe had projected 12,542 eventual enrollees under that set of assumptions). Lewin projected enrollments from 3,451 participants under the most restrictive set of assumptions to 9,056 participants under the most liberal set of assumptions. For the set of assumptions endorsed by the Study Group, Lewin projected enrollment of 7,073.

What produced this dramatic difference in projected enrollment (12,542 versus 7,073) under identical program design assumptions? Recall that Howe’s 2001 work was done when there was very little data on Buy-In enrollments nationwide. He made employment and take-up rate assumptions that were consistent with an enrollment level similar to that experienced by Minnesota. Based on later work for the state of Louisiana (unpublished), Howe determined through an analysis of more recently available enrollment data for other states that the assumptions about employment and take-up rates he had used were almost certainly too high. Lewin similarly had concluded that some of the simulation parameters used by Howe needed to be scaled back.

In projecting the costs of a Buy-In program, Howe (2001) had used a per member per month (PMPM) cost of $437 given to him by ODJFS staff. This cost represented the cost of caring for an adult 18-64 years old living in the community (as opposed, for example, to living in a nursing home). At first blush, it might seem entirely inappropriate to assume that severely disabled individuals would have the same PMPM as non-disabled individuals. However, that lower figure could be justified, in the absence of any better one, using the following argument - the population of persons 18-64 already receiving Medicaid in Ohio includes some number who are severely disabled, and under the assumption that the number of new Buy-In enrollees would be small relative to the total adult Medicaid population, they would probably not inflate the $437 PMPM cost significantly.

Lewin was able to develop a better estimate of the PMPM cost of providing Medicaid to persons with severe disabilities. They used an estimated PMPM cost of $906, which proved to be an excellent estimate as subsequently published research by Mathematica Policy Research found that the average PMPM cost in states with Buy-In programs was $916.

With respect to the Study Group’s recommended design for a Buy-In program, Lewin estimated the net cost to the state would be $22.2 million a year, not substantially different from Howe’s estimate, although the similarity is coincidental as Howe projected a much greater enrollment but at a lower PMPM. The overall cost of the program, according to Lewin, would be $54.1 million, of which the state’s share is 41% (or $22.2 million). This total cost takes into account the loss of spend-down and the receipt of income from premiums, but it does not take account of any increased costs associated with the design or on-going administration of the program, costs which were the focus of the previous Lewin report.

The only critically relevant point to make about the Lewin projections is that, like the Howe projections of 2001, Lewin assumed that a Buy-In program would be attractive to individuals who are not currently on Medicaid. Howe had predicted that perhaps 82% of the eventual enrollees in a Buy-In program in Ohio might be new to the Medicaid system. Lewin’s projection for the Ticket to Work Study Group assumptions was that 66% of enrollees might eventually be people not currently receiving Medicaid. Later sections of this report will call both of these results into question.

Objectives

This work had two objectives:

• To investigate what might be learned from enrollments in other states. Based on a combination of data analyses and reviews of other research, six conclusions with implications for Ohio were developed about Buy-In enrollments.
• To comment on how the mix of recipients enrolled in an Ohio Buy-In program might influence program participant costs.

Chapter 3

Enrollment Lessons from Other States

Predicting how many enrollees a state will have

Introduction

A key factor in the cost of a Medicaid Buy-In program is enrollment. Examination of data on enrollments in other states, along with a review of other research and interviews with key informants, suggests that the following factors largely account for variability in enrollments in Medicaid Buy-In programs:

• The employment climate for people with severe disabilities prior to the implementation of Medicaid Buy-In,
• Whether the program’s policies encourage or limit enrollments,
• Outreach and advocacy efforts,
• To what degree, if at all, Buy-In is advantageous to enrollees compared to existing state Medicaid programs,
• To what degree, if at all, Buy-In is advantageous compared to non-Medicaid-financed programs, and
• How long the program has been operating.

Disability Climate Pre-Medicaid Buy-In

The purpose of the analyses presented in this section is to demonstrate the extent to which the climate in a state for persons with disabilities prior to the start of a Medicaid Buy-In program helps to predict Buy-In enrollment.

Perhaps the most useful compendium of data on state experiences with Buy-In is a recent report from Mathematica Policy Research, funded by the Centers for Medicare and Medicaid Services. The researchers compiled data on Buy-In enrollments from 21 of the 25 states with programs that were operating in 2002. However, somewhat more recent estimates of Buy-In enrollment for these states is available on the www.medicaidbuyin.org web site maintained by Allen Jenson, and it was those enrollments as of June 2003 that were used in the following analysis.

For these 21 states and for Ohio, Table 1 shows the population of persons 18-64 years old based on the 2000 Census of Population and Housing and the number of Buy-In enrollees at year-end 2002. The table also shows the year the Buy-In program was started and the number of persons ages 18-64 in the state who received SSDI or SSI in the year immediately prior to the start of the Buy-In program. The final column of Table 1 shows the estimated number of persons in the state with severe disabilities. These estimates were developed for this study using the methodology developed by Howe for the 2001 Ohio projections. The technique involved examining data from the Current Population Survey for all years between 1996 and 2002. However, only those years prior to the introduction of the Buy-In program were used in developing estimates for a given state. The data are difficult to interpret because of the huge differences in population size, but it is clear that Buy-In enrollment varies widely.

Table 1: State Data on Population and Program Participation

Table 2 clarifies the situation by expressing the Buy-In enrollments as of June 2003 as a percentage of the number of people with severe disabilities in the state. Similarly, the numbers of people 18-64 on SSDI or SSI are expressed as percentages of people with severe disabilities. Table 2 introduces one additional column of information, the percentage of people on SSI who were working in the year prior to the introduction of Buy-In.

Table 2: Pre-Buy-In State Employment Climate Measures that Predict Buy-In Enrollment

As a percentage of the population of people with severe disabilities, it can be seen that Iowa and Missouri have the highest Buy-In participation rates (each over 7%) and that California and Washington have the lowest (less than 1/10th of 1%). Such discrepancies in participation rates demand some sort of explanation, and indeed are the focus of this entire chapter. However, at present, the issue is on how well the first three columns of information in Table 2 can predict the rate of enrollment. Each of these three predictors - % of those disabled on SSDI, % of those disabled on SSI, and % of those on SSI who are working - tells us something different about the state’s climate for persons with disabilities prior to the start of the Buy-In program.

• Perhaps the easiest case can be made for the first predictor: the percentage of people on SSI who are working. While the relationship with Buy-In enrollment is only modest, it is positive, meaning that as the percentage of people on SSI who are working increases, the percentage of people with severe disabilities who enroll in Buy-In increases. A state in which more SSI participants work is a state with a more favorable climate for people with disabilities.
• States with high Buy-In enrollments (Iowa and Missouri) also have higher percentages of their populations of people with severe disabilities enrolled in SSDI. States with low enrollments (California and Washington) have low percentages on SSDI. In what way may this predictor be taken as an indicator of the state’s climate for persons with severe disabilities? SSDI eligibility requires 20 quarters of employment (with a few exceptions). It would be untenable to assume that these individuals earned their 20 quarters while working with a disability, although that may be true of some people. More likely, this predictor measures the ease with which workers who become disabled are able to secure income supports through the SSDI disability determination process. In theory, this should have nothing to do with the medical examination process, although in practice it might. However, it almost certainly reflects the extent to which the state’s government, employers and advocates have created a comfortable pathway for workers who have become disabled to enter the SSDI system.
• States with a higher percentage of people with severe disabilities who are on SSI is a negative indicator of a good climate for people with disabilities. People with disabilities who are not eligible for SSDI but who are unable to support themselves with employment need the support of the SSI program. The simple correlation between SSI enrollment and Buy-In enrollment is very modest, but negative.

Before using these three predictors to model Buy-In enrollment in other states, consider the crudest sort of model for Ohio’s Buy-In enrollment. From the last column of Table 2, it can be seen that Buy-In enrollment ranges from 0.06% to 7.28% of the population of people with severe disabilities. Table 3 shows what enrollments these percentages would correspond to if they were applied to Ohio. The experiences of other states, with no refinement, suggest that a Buy-In program in Ohio might enroll anywhere from 268 people to over 31,000. It may further be concluded that there are approximately two chances in three that Ohio’s enrollment would be between 893 and 14,577 persons, again based on nothing beyond the unadjusted enrollment rates in other states.

Table 3: Unrefined Estimates of Ohio’s Buy-In Enrollment

A multiple regression model was developed to predict Buy-In enrollment based on the three predictors in Table 2. The model explained 42% of the variability in enrollment (corresponding to a correlation of .65). The most important predictor was the state SSI employment rate and the least important was the percentage of the state’s population of people with severe disabilities on SSI. By conventional standards of statistical inference, this last predictor was not statistically significant but it was retained because its negative sign highlights an important difference between SSDI enrollments (positively correlated with Buy-In enrollment) and SSI enrollments (negatively correlated).

Using the values in Table 2 for Ohio, the regression model produces a predicted value of 2,158 people for Buy-In enrollment in Ohio, with 68% confidence interval of enrollment ranging from 865 to 5,382. Without considering the three predictors, Ohio was predicted to have an enrollment of 3,741 with a range of 893 to 14,577. Based on the three predictors, the range of the estimate is narrowed considerably. The appropriate conclusion is as follows: Based solely on the climate in the state for people with severe disabilities prior to the introduction of a Buy-In program – climate comprising such factors as employment opportunities for people with disabilities and pathways to SSDI participation – Ohio could expect an enrollment lower than any that has been projected to date. In other words, Ohio’s pre-Buy-In ground is not particularly fertile in terms of supporting high Buy-In enrollments.

It must be acknowledged that this model for predicting Buy-In enrollment is far from perfect, accounting as it does for only 42% of the variability in enrollments. Following are the eight states where the model most seriously over-estimated Buy-In enrollment as of June 30, 2003. Table 5 shows the eight states where the model most seriously under-estimates enrollment.

Table 4: States for Which Model Over-predicts Buy-In Enrollment

Table 5: States for Which Model Under-predicts Buy-In Enrollment

Thus far, only one set of influences on Buy-In enrollment have been considered. Attention will now turn to others. As these are introduced, occasional reference will be made back to Tables 4 and 5, for presumably what pre-Buy-In conditions in a state cannot explain about Buy-In enrollments must be explainable by other considerations.

Program Policies

The most extensively discussed influences on Buy-In enrollments are the policy parameters a state selects in designing the program. The following discussion owes much to the excellent summaries prepared by Allen Jensen and his colleagues (see note 6) and the work of Mathematica Policy Research (see note 7). With respect to each policy parameter, some conclusions are offered as to how the recommendations of the Ohio Ticket to Work Study Group might encourage either higher or lower Buy-In enrollments as compared to other states.

Limits on Unearned Income

The effect of a limit on unearned income is complex. One the one hand, it reduces program costs by increasing the percentage of the Buy-In caseload that is already receiving Medicaid, merely switching them from one eligibility group to another but having little impact on state costs. On the other hand, by excluding former workers with a history of higher wages, such a policy reduces the likelihood that Buy-In will be a path to eventual self-sufficiency and increases the likelihood that Buy-In will be a path to increasing by a small amount the disposable income of SSI recipients. The current Buy-In recommendations by the Ohio Study Group impose no limits on unearned income, which allows for the possibility that more SSDI and private insurance recipients might benefit from the Buy-In program.

Minimum Required Earnings

Nebraska’s minimum earning’s level has the remarkable effect of helping to restrict Buy-In almost entirely to SSDI recipients (according to Mathematica Policy Research, 98% of Nebraska’s enrollees were receiving SSDI at the time of Buy-In enrollment). It also helps Nebraska realize higher average earnings/recipient and lower PMPM costs of Medicaid coverage (perhaps because of dual enrollment with Medicare, Medicaid costs are lower). It probably helps to explain why the climate model discussed in the previous section more seriously over-estimated enrollment in Nebraska than in any other state. Such a policy as Nebraska’s regarding minimum earnings will restrict enrollments and skew them toward people with a greater chance of achieving self-sufficiency. The Ohio Study Group made no recommendation regarding minimum earnings.

Premiums

It is worth noting exactly why people with very limited incomes, many of whom already have Medicaid coverage, would be willing to pay premiums. People who have unearned income in excess of the Medicaid eligibility standard have to “spend-down” that excess unearned income on health care expenses in order to qualify for continuing Medicaid coverage. These costs can easily exceed monthly Buy-In premium amounts and, further, are hard to manage. This issue is especially relevant to Ohio, where the Medicaid eligibility standard is only $495 (much lower compared to many other states), suggesting that many SSDI and SSI recipients might find Buy-In attractive. The Ohio Study Group recommended that premiums be set at 10% of income in excess of $150% of poverty, a policy that is superficially neither liberal nor conservative compared to other states. However, in conjunction with Ohio’s very low eligibility standard, the premium policy may be more liberal than it would otherwise appear.

Income Eligibility and Resource Standards

Treatment of Taxes and Spousal Income

Job Loss Protection

Summary

Table 6: Summary of Policy Parameters

Outreach and Advocacy Efforts

The Lewin Group, in its report to ODJFS entitled, Administrative and Systems Changes for Implementation of a Medicaid Buy-In Program, collected key informant interview data with government staff members in a number of states. Lewin inquired about marketing efforts in five states with higher than expected Medicaid enrollments, based on the climate model presented earlier. In every case, there was evidence of strong marketing efforts and collaborations. New Mexico contracts with the Department of Vocational Rehabilitation to handle outreach and marketing. With a budget of over $100,000 a year, the Vocational Rehabilitation staff is able to run public service announcements, operate job fairs, and work directly with employers. Missouri described encouraging stakeholder involvement, creating a strong university partnership and creating strong ties to the advocates for the community of people with disabilities. Connecticut described some advocacy-run outreach efforts but also noted that the department had sent an informational mailing to all prospective Buy-In enrollees. Indiana’s effort was strongly advocate-driven, as was Minnesota’s effort.

While it appears that effective marketing efforts are involved when states have higher Buy-In enrollments than expected, the data Lewin collected only partially support the idea that lower enrollments than are predicted by state climate are associated with weak marketing. Nebraska state officials frankly acknowledged that marketing may help to explain their lower than predicted enrollment. In addition, while Illinois talked about its marketing efforts, it appears that those were largely in-house efforts; only one advocacy group was mentioned, and it appears to be rather narrowly focused on one segment of the population of people with disabilities (AIDS). Pennsylvania officials described printing 25,000 brochures but nothing was said about how they were distributed or whether they were effective. More importantly, it was the only outreach effort described. Finally, Lewin reported that New Jersey reported spending $500,000 on outreach (although few specifics were given), a sum that seems sufficiently large to support a robust outreach and marketing effort.

In addition to the data reported by Lewin, interviews with several key informants from across the country highlighted the importance of marketing and the involvement of advocates. There is a uniform perception that weak marketing efforts contribute to California’s lower than expected Buy-In enrollment. State governments in Iowa, Minnesota and Wisconsin, were all acknowledged to have nurtured their relationships with advocates for the population of people with disabilities for years, and not just in preparation to marketing Buy-In. Minnesota and Wisconsin both have higher than predicted Buy-In enrollments, and while the climate model did not under-estimate Iowa’s enrollment, it is the second highest in the country on a per capita basis (see Table 2).

Occasionally key informants from outside a state attributed some of a state’s success to advocacy outreach and marketing efforts even though key informants within the state were so exceedingly modest about their efforts that one had to conclude other factors might have played a role in producing the high enrollments. That is a weakness to key informant interviewing.

After these interviews were completed, the results seemed unsatisfying, as if something had been overlooked. Then it became apparent. All of the informants had talked about the importance of ties between state government and advocacy groups, or about the importance of collaborations between state Medicaid agencies and other groups, such as the one-stops or departments of vocational rehabilitation. However, nobody had talked about employers, the organizations that have the jobs and that will actually be hiring people with disabilities. Some web-based research revealed that some states are sensitive to this issue. Virginia is a state with a Buy-In program too new to have been included in the Mathematica study but it has a page aimed at employers.

Each of the two preceding sections has closed with a comment about the implications of the findings for Buy-In enrollments in Ohio. The Lewin report to ODJFS clearly highlighted the importance of marketing and noted the effectiveness of ties to advocacy groups, saying, "Working with community advocates is a key component of state outreach efforts" (page 26). Curiously, however, later in the report, when presenting four options for ODJFS to consider in how to accomplish marketing and outreach, advocates are mentioned only to the extent that they might object to certain options as being limited. Nowhere is there a recommendation that ODJFS enter into collaboration with advocacy groups.

As prepared as the advocacy community in Ohio appears to be to market and endorse a Medicaid Buy-In program, the efforts of advocates would be enhanced if there were closer tie between ODJFS and the advocacy community.

Interaction of Existing Medicaid Policies and Buy-In Policies

The Mathematica Policy Research report demonstrates conclusively what some observers have long suspected: most people (74%) enrolled in Buy-In programs nationwide were already enrolled in Medicaid under a different eligibility group. (It also appears that an approximately equal percentage of people are dually enrolled in Medicare.) This is important for several reasons.

• First, it strongly suggests that the models developed by Howe and the Lewin Group over-estimated the attractiveness of Buy-In programs for people who are not currently enrolled in Medicaid. Howe, in particular had estimated that eventually Buy-In might be attractive enough to such individuals that perhaps only 17% of Buy-In enrollees would have been on Medicaid at the time of enrollment. Only one state in the Mathematica report showed a lower percentage of current Medicaid participants (New Jersey) and the report implied that there were some data quality issues with the New Jersey data.
• Second, by extension, it raises the possibility that the cost projections of the Lewin Group are too high. This does not mean that a state can provide Medicaid to Buy-In enrollees at the same cost as Medicaid under different programs. Buy-In enrollees who had previously had incomes in excess of the Medicaid eligibility standard might have been "spending down" to maintain eligibility, and while it is surely an oversimplification to assume that one spend-down dollar is one Medicaid dollar saved, consumers who save themselves spend-down dollars are costing the state something in the way of increased Medicaid expenditures. Lewin projected that the net cost to the state of 2,411 Medicaid participants moving into a Buy-In program would be $3.4 million, or $118 PMPM.
• Thirdly, and most immediately relevant to the present discussion, it prompts speculation as to whether Ohio might have a higher or lower percentage of Buy-In participants who had already been Medicaid participants at the time of Buy-In enrollment. That in turn has implications for whether an Ohio Buy-In program will have the effect of operating primarily to produce small increases in the disposable income of SSI participants or the effect of moving people out of SSI and SSDI status and into wage-based self-sufficiency.

It is questionable if Ohio’s status as a 209(b) state has a direct implication for Buy-In enrollments. Some 209(b) states - including Indiana, Minnesota and Missouri - have higher Buy-In enrollments than can be explained by the pre-Buy-In climate. However, at least one other 209(b) state, Illinois, has a lower than expected enrollment. As a 209(b) state, Ohio is not required to provide Medicaid to SSI recipients, or more properly to SSI recipients who do not qualify for Aged, Blind and Disabled (ABD) Medicaid. Under ABD Medicaid, a person with a severe disability is eligible for Medicaid if their income (as an individual) is less than 64% of the Federal Poverty Level ($490/month). However, even as a 209(b) state, Ohio is required to make Medicaid available to individuals with higher monthly incomes who spend down to the Medicaid income standard by paying for medical expenses or even by transferring income to their county department of Job and Family Services.

Jensen et al (see note 6) have clearly explained how the Medicaid income standard in a state affects Buy-In enrollment. In states where the Medicaid income standard is higher, producing a small gap between SSI Medicaid eligibility and 1619(b) eligibility, Buy-In programs only materially benefit those current Medicaid beneficiaries who fall into the gap. In contrast, in states such as Ohio, where the gap is much wider – and Ohio may have one of the widest gaps in the country – many more current Medicaid beneficiaries will find Buy-In attractive because it will replace their spend-down amount with a monthly premium, which will nearly always be smaller than the spend-down and in any case will be easier to manage.

Ohio’s income standard will likely have the following effects:

• The wide income range in which people with severe disabilities can obtain Medicaid only with a spend-down will tend to drive Buy-In enrollment up. However, the direct fiscal effect of this will be about $118 PMPM, to use the Lewin estimate, and not the cost of a new Medicaid beneficiary.
• Perhaps more importantly, it may create a Buy-In climate in Ohio where the program implicitly is assumed to operate for the purpose of increasing the monthly disposal income of SSI participants, as opposed to moving people off of SSI/SSDI cash benefits and into self-sufficiency.

Interaction of Other State Policies and Buy-In Policies

Time in Operation

Unlike many of the other variables that have been considered, which are difficult to quantify, months in operation can easily be used to improve the climate model presented earlier . Initially, there was little reason to believe it would be useful; months from program start to June 30, 2003 did not appear to be correlated with the percentage of the population of persons with severe disabilities who were enrolled in Medicaid as of that time. (Recall, for example, that Indiana’s enrollment rocketed upon program inception whereas California’s has languished.) But a peculiar statistical situation developed, known as suppression, in which a predictor that does not initially seem useful actually improves a model because it suppresses some of the variability in the other predictors that is unrelated to the thing being predicted.

A new multiple regression analyses revealed that just three predictors, all of them making significant contributions to the model, help to predict enrollment as of year-end 2002: % of the population on SSDI at program inception, % of the population on SSI at program inception, and months of operation as of year end 2002. The effects of SSDI and months of operation were positive, meaning that as the pre-Buy-In % of people on SSDI increased and as months of operation increased, enrollments increased. As before, the effects of SSI enrollment was negative, meaning that Buy-In enrollment went up as the % of the population of people with disabilities who receive SSI went down. Overall, the model accounted for 48% of the variance in enrollment (even better than the climate model).

In examining the errors of prediction for this model, it was found that the states that were over-predicted and under-predicted were largely the same as for the climate model. Nebraska, New Jersey, Washington, Illinois, California and Utah were over-predicted by both models. Kansas and Pennsylvania were over-predicted by the climate model but not by the current model, which makes perfect sense as Kansas and Pennsylvania were both cited earlier as states with relatively liberal policies but new in terms of operations. They were replaced by Maine and Oregon, which had been neither over nor under-predicted by the climate model. As for errors such that the model under-predicted enrollment, both models identified Wisconsin, Minnesota, New Mexico, Missouri, Alaska, Massachusetts and Indiana. Connecticut moved from among the worst eight states for under-prediction to position nine and Iowa took its place.

The new prediction model allows us to project how Ohio’s enrollment might increase over the first several years of the program. The projections shown below are not inconsistent with the climate model, which by ignoring time in operation implicitly assumed Ohio would have operated about as long as the other programs. In other words, the climate model prediction of 2,158 should be compared to 2,251, which is the value the current model gives if Ohio were operating for 33 months (the average of all the other states considered). It would be a mistake to assume that in out-years the program will continue growing faster and faster. There is too little data to model long-term growth in Buy-In enrollments. However, the data available to date do seem to suggest that the rate of enrollment accelerates for a while early in program operations.

Table 7: Projected Ramp-up for an Ohio Buy-In Program

Summary

• Most indications suggest Ohio will have a moderate Buy-In enrollment compared to other states, as expressed as a percentage of the population of people with severe disabilities who will enroll.
• Previous projections for Ohio have probably overestimated the percentage of Buy-In participants who will be new to Medicaid.
• The projected ramp-up model (projecting 3,350 participants after four years) is not inconsistent with pre-Buy-In climate model. By ignoring time of operation, the climate model essentially assumed Ohio would have been in operation the same length of time as the other programs (average 33 months). It can be seen from Table 7 that the climate projection of 2,158 people falls neatly between 24 and 36 months, and closer to the 36-month mark.

Table 8: Influences on Buy-In Enrollments

Chapter 4

Cost Projections

Total Cost of Medicaid Buy-In

Introduction

where T = the total cost to the state of operating the program, ENew = number of new enrollees not currently enrolled in Medicaid, CNew = cost of Medicaid coverage to the state per new enrollee, EExisting = number of enrollees already enrolled in Medicaid, CExisting = cost of Medicaid coverage to the state per existing enrollee, O = overhead expenses associated with running the program and R = revenue enhancements the state might expect as a result of operating the program. The first four variables to the right of the equal sign comprise enrollee costs.

Enrollee Costs

The present analysis offers a quite different perspective than that afforded by earlier Ohio projections. Howe (2001) originally projected that more than 80% of Buy-In enrollees might be new to the Medicaid system. The Lewin Group projected that 66% of Buy-In enrollees would be new to the system. Both of these figures are well-outside the middle range established by the research by Mathematica, which found that an average of 74% of Buy-In enrollees had been receiving Medicaid at the time of their enrollment. The issue matters enormously because of the two PMPM parameters of the cost equation. The Lewin Group estimated that the PMPM cost of providing Medicaid to a Buy-In participant would be $906. For a current beneficiary who moved from another eligibility group into the Buy-In program, Lewin estimated the net change in Medicaid expenses would be $118 (taking account of spend-down loss and premium revenue).

Table 9 attempts to project the net cost to the state of a Buy-In program that would begin in 2007. The Ohio Department of Job and Family Services estimates that the PMPM cost for a Buy-In participant in 2007 will be $1,198.95. Using the same inflation factor but applied to The Lewin Group’s estimate of the net cost of a Buy-In participant who was already enrolled in Medicaid, it will be assumed that each such participant will be associated with an incremental PMPM of $155.91. The table demonstrates that the proportion of enrollees who are new to the system has enormous implications for program cost. (The analysis assumes that the state share of Medicaid in 2007 will be 40.32%.)

Table 9: Program Cost in 2007 Assuming Enrollment of 7,073 People Under Two Mixes(Disregarding Administrative Costs and Revenue Enhancements)

Comparison between Lewin's Assumptions (34/66) and Assume Average Mix (74/26)

Total State Enrollee Expenses - $28,863,011 (Lewin's Assumptions) $14,616,332 (Assume Average Mix)

It is, of course, impossible to know if Ohio’s mixture will be average compared to other states. However, the low Medicaid income standard in Ohio (which should make Buy-In attractive to SSI recipients) along with the absence of unearned income limits and minimum earnings requirements, all seem to suggest that Ohio will be more like other states, on average, than like one of the few states with an unusually low percentage of people on Medicaid.

Table 9 is based on the expected eventual enrollment in an Ohio Buy-In program. Recall from Table 7 that the expected first year enrollment is projected to be in the neighborhood of 770 enrollees. Assuming a 74/26 percent mix of existing enrollees to new enrollees, the projected liability of the state in 2007 will be $1.6 million (excluding overhead expenses, which are not included in Table 9). The Ohio Department of Job and Family Services, using a more sophisticated methodology, which assumes these first year enrollees will come into the system gradually, projects a first year cost of less than $1 million. However, the state projection shows program costs increasing rapidly to just over $11 million a year by 2010.

Chapter 5

Conclusions

Ohio Compared to Other States

• Based solely on its pre-Buy-In climate for people with severe disabilities, Ohio’s climate is not particularly conducive to a high Buy-In enrollment. Of course, any program’s enrollment will be based on many other factors, but based solely on climate, we might expect an enrollment in the range of 3,700 people.
• The Ohio Ticket to Work Study Group in 2001 endorsed a set of program policies that would be consistent with mid-range enrollment compared to other states (based on percentage of persons with disabilities who might enroll). Limits on unearned income and a minimum earnings requirement can impose severe limits on enrollments, but Ohio proposes neither of these policies. Ohio’s premium and resource limit policies are moderate. While its income policy - with a $20,000 earned income disregard - is supportive of high enrollments, this is offset by the fact that income will be assessed before taxes and a spouse’s income will be considered.
• With regard to outreach and advocacy, the evidence suggests that close ties between the state Medicaid agency and the advocacy community help promote higher enrollments. Advocates in Ohio would like to see such ties strengthened.
• Ohio’s relatively low income standard for Medicaid will probably drive SSI and SSDI recipients to enroll in a Buy-In program in order to avoid their spend-downs. While this may promote higher enrollments, it will also produce a less expensive program if more enrollees in an Ohio Buy-In program were previous Medicaid recipients.
• A model based on some climate conditions and on months of program operation results in a projection that Ohio’s enrollment will increase from under 1,000 people after one year to about 3,350 people after four years. This is not inconsistent with The Lewin Group’s projection of 7,073 people, but it underscores the fact that the Lewin estimate is a projection of where the program will plateau.

Program Costs

Beyond number of enrollees and PMPM costs, each of which has been projected by The Lewin Group, the single most important determinant of total cost is the mix of participants according to whether they were receiving Medicaid at the time of Buy-In enrollment. Previous projections for Ohio – completed prior to the publication of the Mathematica Policy Research survey of state programs - assumed that most Buy-In enrollees in Ohio would be new to Medicaid. Not only would this be discrepant with the experiences of most other states, it is not consistent with the conclusion that Ohio’s low income standard should attract people anxious to avoid spend-down. The cost implications of this point are enormous. The state’s share of a Buy-In program’s cost under the assumption of an average percentage of people already on Medicaid (74%) is only half of what was projected by The Lewin Group.

Program Logic

• It must be acknowledged that Medicaid Buy-In is not sufficient to support people’s movement to self-sufficiency. There is too much evidence that health insurance is necessary for people with severe disabilities who want to work for there to be any question about the importance of Buy-In. However, Buy-In in and of itself does not seem capable of generating a wholesale movement of people with severe disabilities into self-sufficiency.
• The conventional wisdom in America about work for people with severe disabilities has undergone fundamental change over the past few decades. However, that is not to say that everyone with a severe disability is psychologically prepared to think about serious efforts to re-enter the world of work, and it is certainly not to say that employers and other organizations that interact with people with severe disabilities are fully equipped to help support such serious efforts.
• For example, according to Jensen & Folkemer, the higher the level of SSDI cash benefits, the less likely enrollment in Buy-In is. For example, in Connecticut, 25% of SSDI participants have cash benefits in excess of $1,000/month but only 12% of Buy-In enrollees have unearned income in excess of $1,000/month. (Bear in mind that 82% of Connecticut enrollees have SSDI and Medicare.) This seems to suggest that the people who have had longer and more rewarding experiences in the work world are less likely to enroll in Medicaid Buy-In. Why? The answer, of course is unknown. Perhaps they are older and do not require Medicaid as a wrap-around to their Medicare coverage. But another obvious possibility is that not everyone with a severe disability wants to step onto a path that might lead them back into the world of work.
• under the assumption that the population of people interested in Buy-In may not include those people with severe disabilities who are most able to expand the amount of work they do for pay, the inescapable conclusion appears to be that Buy-In needs to be well integrated with an array of clinical and vocational services to support the movement to self-sufficiency.
• The issue of whether and how much earnings will increase post-Buy-In enrollment is one on which key informants are divided. Some presume that people are managing their earnings to stay below the level of Substantial Gainful Activity at which they would begin losing cash benefits under SSDI. A standard explanation for this finding is that participants are conservative - it is hard to forego a reliable stream of income, they have concerns about the future of the Buy-In program, etc. In other words, participants do not yet believe in Buy-In in the way that people really understand and believe that, for example, going to college will result in increased earning power. Another possibility, however, voiced by other advocates, is that people may not even know they are enrolled in a Buy-In program. Especially in states where counties or the state Medicaid agency automatically moved people from another eligibility group into a Buy-In program, advocates feel convinced that enrollees are ignorant of the possibilities for earnings increases. Some advocates believe that the economy over the last three years has been one in which persons with severe disabilities would find it hard to increase their earnings. Their concern are captured in the saying, "Last hired, first fired." Another school of thought is that employers are managing their health insurance expenses by providing workers with disabilities only a limited number of hours. Still other informants would argue that we simply do not yet know why some Buy-In participants have not substantially increased their earnings, or even if it is true that many have not.
• Therefore, should Ohio implement a Buy-In program, it is extremely important that the advocacy community broadly defined be drawn into all aspects of program operations, not just asked to advise on marketing and recruitment. These stakeholders have the potential to ensure that Buy-In will function as a pathway into self-sufficiency. Some participants may enroll to increase their disposable income. Others may have dreams of self-sufficiency through work. But with the integrated assistance of the advocacy community, the participants and their employers will know not just that they are on a path, but that there is a map showing how that path can deliver them to a more ambitious destination when they are ready.

About the Author

Beginning in 1980, he worked for 13 years at the UC Institute for Policy Research (IPR), an interdisciplinary social science research organization, where he served as Director of the Southwest Ohio Regional Data Center.

In 1993, Dr. Howe assumed a faculty position in the UC Department of Psychology. He teaches courses in statistics and methodology and conducts externally funded evaluation and policy studies.

He can be contacted at (513) 253-8718 or by email at howe@srhassociates.com. Examples of work by Steven R. Howe and Associates, LLC may be found at www.srhassociates.com.

Acknowledgements

Doug DeVoe, of the Ohio Advocates for Mental Health, and Beverly Johnson and Janet Miller, of the Cerebral Palsy Association, made many useful and important suggestions on earlier drafts of the report.

David Ellis and George Weiner of the Federation for Community Planning contributed some of their substantial expertise on state budgeting to this project.

Lorin Ranbom of the Ohio Department of Job and Family Services performed some projections of monthly costs of Medicaid that were used in the cost projections presented in this report.

Allen Jensen of George Washington University, who must surely know Medicaid Buy-In policy better than almost anyone, read an earlier draft of this report and made useful suggestions. More importantly, Allen very generously spent time with me on the phone and via email patiently explaining certain subtleties of Buy-In policy.

And to my various key informants who spoke frankly about the experiences in their states, thank you.

| Home | About Us | Calendar of Events | Grants and NOFAs | Links | Publications and Products | Site Map | What's New