Decreasing Medicaid eligibility: reducing or shifting costs?

North Carolina House and Senate Republicans are grinding toward a resolution of their differences in how to proceed with Medicaid reform. The latest from the Senate sounds mostly similar to their past offer (reduce Medicaid eligibility for some aged, blind and disabled beneficiairies, and move to bid out the full Medicaid program to private insurance companies, but allowing provider networks to bid). Hereis is a brief review until today of the many twists and turns since complete Republican control of state government commenced in January 2013.

I will get to the ins and the outs of the emerging plan when it emerges some more, and we hear from the House.

A quick point on the continued insistence of Senate Republicans to reduce eligibility for some aged, blind and disabled recipients–Rose Hoban quotes Senator Harry Brown (R-Jacksonville) in a committee meeting yesterday as follows:

“That is a major concession for us. We think that you eventually have to address eligibility requirements in Medicaid if you want to control costs,” Brown told the committee.


Read more of this post

Marsha Gold interview: dual eligibles III

This is the third and final part of my interview with Marsha Gold of Mathematica Policy Research.

Part I of the interview and Part II.

It is my understanding that dual eligible Special Needs Plans (SNPs) do not include capitation for Medicaid costs. Is this correct? Are there likely to be movements in this direction?

 Yes, you are right. The SNP contract is between Medicare and the Medicare Advantage plan operating the SNP. The payments cover only Medicare benefits, though SNPs must account for any savings they may have over and above the payments for providing such benefits and use them to enrich  benefits (or reduce costs) for enrollees. Though some SNPs have contracts with state Medicaid agencies that cover Medicaid benefits for these enrollees, most do not. Instead, state Medicaid agencies pay Medicaid’s share of Medicare benefits and other Medicaid benefits (like long-term care) separately, usually on a fee-for-service basis.[1] Plans seeking to pair a Medicaid capitated contract with an SNP contract face obstacles because the two programs do not necessary apply the same requirements to managed-care entities.[2]

This may change in the future but probably not quickly and not in all states. The Medicare Improvements for Patients and Providers Act of 2008 requires all SNPs to submit models of care but requires only new SNPs to contract with state Medicaid programs. Also, there is no parallel requirement that Medicaid agencies contract with SNPs. Contracting is complicated because state Medicaid agencies have little “skin in the game” now in paying for Medicare’s acute care benefits and face many challenges in contracting for other benefits that are financially relevant to them (like long-term care). The Affordable Care Act requires all dual eligible SNPs to contract with state agencies by 2013, though it does not require states to contract with any or all of them nor do such contracts have to include Medicaid payments.[3] The Affordable Care Act also allows a frailty adjustment to Medicare rates of dual eligible SNPs that are “fully integrated” (that is, have a Medicaid capitation contract that covers benefits including long-term care as well as an SNP contract).

How likely is it that new regulations to enhance coordination between Medicare and Medicaid will work?

 I’m hopeful that the new federal initiatives will support progress by states that have long sought to better integrate care for dual eligibles and by other states that seek to follow their lead. However, I think that coordination can’t be attained by simply legislating it, it must be achieved by hard work on the ground and a commitment to better care for the people who are dually eligible for Medicare and Medicaid. I have some concerns that the current budget climate will lead to unrealistic expectations about the pace of change or the extent of savings to be achieved, thus undercutting the potential to make real improvements for the people covered under these programs.

 Anything else?

 Yes. As these issues are debated, it’s important to avoid thinking of integration strategies as “one size fits all.”  The dual eligible population is highly diverse in ways that are likely to require different models of care and different uses of Medicare and Medicaid benefits. The same strategies may not work for subgroups of dual eligibles who need different kinds of care, particularly in terms of the services that Medicaid covers or that are provided in communities. Care for dual eligibles is likely to require a variety of models and considerable flexibility in adapting them across the states. Creative development, evaluation, and sharing of models will be of great value. The real policy challenge is to encourage such innovation while maintaining appropriate beneficiary protections and national oversight over the use of federal funds in both Medicare and Medicaid.

[1] Medicare Payment Advisory Commission. “Chapter 5. Coordinating Care for Dual Eligibles.”  In Report to the Congress: Medicare and the Health Care Delivery System. Washington, DC: MedPAC, June 2011.

[2] Melanie Bella and Lindsay Palmer. “Encouraging Integrated Care for Dual Eligibles”  Resource Paper.  Princeton, NJ: Center for Health Care Strategies, July 2009.

Marsha Gold interview: dual eligibles II

This is the second part of an interview with Marsha Gold of Mathematica Policy Research on reform options for the care of dual eligibles.

Part I of the interview.

Should one payer be responsible for all care of dual eligibles to encourage such goals (reduce costs and improve quality), and if so, which direction is more likely to achieve success: federalizing the Medicaid costs of dual eligibles or moving them into private plans?

 This is a tough question that can and should be debated. Making one payer responsible for all care for dual eligibles is appealing because it creates a single funding stream in which the incentives can be aligned to encourage coordinated care. But modifying current programs to create such a single stream would be difficult financially and politically and might have significant risks for beneficiaries.

 Currently, Medicare is the main source of payment for acute care services used by dual eligibles and Medicaid is the main source of payment for everything else (including long-term care and supplemental services like transportation) as well as cost sharing for Medicare benefits. Because dual eligibles rely to differing extents on coverage from both programs, it is important to align their incentives and coordinate payments and requirements. The advantage of having Medicare at the helm is that the Medicare program is national, with consistent requirements across states and consistent protections for beneficiaries, whether or not they are dual eligibles. However, it seems unlikely that this or any future Congress would agree to absorb full responsibility for Medicaid’s long-term care services, which now differ substantially across states.

 Medicaid programs have the advantage of being closer to the ground and potentially better positioned to coordinate care systems that differ across communities. Unfortunately, eligibility requirements and benefit levels, as well as the capacity of Medicaid programs, vary from state to state, despite national requirements and standards now in place. States also lack Medicare’s scale and experience in paying for acute care services and gaining provider participation. If the states became responsible for all care for dual eligibles, some states might do better than the federal government but others likely would end up using Medicare financing as a “cash cow” to help them negotiate tough budget constraints, leaving care for dual eligibles the same or potentially even worse.

 Politics aside, it seems to me that the most productive approach might be to require the two programs to work together more effectively, with flexible models that would vary with the state context but would retain a substantial national role in paying for and shaping care for dual eligibles. Under this approach, Medicare would still have final authority for acute care services and states would be responsible for long-term care, but the two would be required to develop operating agreements that support more effective work across programs to enhance care for dual eligibles.

Have dual eligible special needs plans (SNPs) successfully coordinated the Medicare and Medicaid streams of care for these beneficiaries so far?

 Generally no, though it isn’t necessarily the SNPs’ fault because they haven’t had the authority or tools to do so nor have they been selected because of their focus on this objective. Historically, states have had few incentives to make coordinating Medicare and Medicaid a high priority and those who tried faced many constraints (some Medicare related). That may be changing as a result of new initiatives included in the Patient Protection and Affordable Care Act to encourage better systems of care for dual eligibles, including the establishment of the Medicare-Medicaid Coordination Office and the Center for Medicare and Medicaid Innovations.

The interview will conclude this afternoon.

Marsha Gold interview: dual eligibles

Marsha Gold is a Senior Fellow at Mathematica Policy Research who is an expert on Medicare, Medicaid and the care of the dual eligibles. She recently completed work focused on Special Needs Plans (SNPs) that I blogged about.[1] I asked her some questions about care for the dual eligibles via email that she has graciously answered in this, and two subsequent posts to be published later today. Dr. Gold’s responses are printed in full as I received them. Great thanks to Adam Coyne, who is Vice President and Director of Public Affairs for Mathematica, and who facilitated this interview.

What’s your take on why there is interest in dual eligibles?

Beneficiaries who are dually eligible for Medicare and Medicaid have diverse and typically complex health care needs. Compared to the Medicare population overall, dual eligibles are substantially more likely to be in poorer health, suffer from physical disability and/or cognitive or mental impairments, have low incomes and limited education, and reside in long-term care facilities.[2] They account for a disproportionate share of spending in both programs relative to their numbers. In addition, there is widespread concern over the quality of care these beneficiaries receive, because their needs are complex and care is financed through multiple channels.

 Do you think it is possible to reduce the costs of caring for dual eligibles while improving quality?

 Yes, I think it’s possible to do so, but not easily or quickly. We have a fragmented care system that is better designed to provide limited, short-term care in particular settings for particular conditions than to coordinate care over the long term for people who need a variety of medical, mental health, and social services provided by different systems. It’s hard to imagine that care for many dual eligibles cannot be improved or that such improvements would not generate some savings. Reconfiguring delivery to overcome these challenges will not be easy, particularly if it requires the systems and service providers to move out of their “comfort zone” and change the way they operate. Such change does not happen overnight, if it can be done at all. The fact that care for dual eligibles is financed through separate programs makes the challenges even greater, since it complicates the task of creating a consistent set of financial incentives and program requirements to encourage change.

What is the key to generating such change?

 Reaching out to the delivery system and finding ways to encourage and require it to become more focused on patients, families, and caregivers. We currently finance and provide services, instead of caring for people. Anyone with aging parents or an adult child or friend with special needs probably has experienced the frustration of knowing their care could be better, more humane, and less costly, “if only” more providers would talk to one another and focus on the patients, rather than their body parts, and then only “from 9:00 to 5:00.” To generate such change, expectations need to be modified, financial incentives overhauled, and care requirements configured to support new models of delivery. Ultimately, such change probably will be enhanced by capitation or per person payment but changing the payment approach is not a magic bullet for creating change unless the care configuration is improved. In addition, risk adjustment and other features that protect patients are essential to creating successful capitation payment programs that reward good performance rather than ways to game the system.

[1] Marsha Gold has been monitoring the evolution of the managed care marketplace for many years. She has written many seminal studies of Medicaid managed care and is recognized as a leading authority on the evolution of private plans in Medicare. The perspectives expressed here are those of the author only and do not necessarily reflect those of Mathematica or any of the agencies or organizations with whom it works.

Reducing Unnecessary Hospitalizations for persons in NHs

Joesph Ouslander and Robert Berenson have a piece in the NEJM amplifying the problem of unnecessary hospitalizations among persons living in nursing homes (NH). The statistics are sobering: 1.6 Million Americans live in a NH, and nearly 1 in 4 who are first admitted from a hospital to a NH are readmitted to the hospital within 1 month. They provide a face to these statistics: a 90-year old with Alzheimer’s disease, CHF and severe left ventricular dysfunctions who develops a low grade fever and chills. The scenario ends up with a hospitalization in which the patient is treated with IV antibiotics paid for by the Medicare program (at a cost of $10,000) and the fever resolves. However, one night during the hospital stay, the patient becomes disoriented, falls and breaks her hip, returning her to the NH in far worse shape than she left it in.

High cost that results in a worsening of quality of life in this type of population is about as bad as it gets.

The authors paint an alternative scenario in which the patient is evaluated in the NH and receives the same antibiotics in her familiar surroundings with fall prevention controls and her fever has resolved a week later at a cost of around $200 to Medicare, and her hip intact. The paper sketches out the hospitalization experience of the dual eligibles (Medicare and Medicaid) in 2005.

The goal is to reduce the avoidable hospitalizations while maintaining access for the ones that are needed. The authors note several caveats lest you think this is such obviously low hanging fruit that we must definitely be on the cusp of solving these problems.

  • Misaligned financial incentives between Medicare and Medicaid. For example, after a hospitalization, a dual eligible may become eligible for Medicare Part A SNF funding which in some states may pay 3-4 times the Medicaid per diem rate, providing a strong incentive for a NH to send patients to the hospital. A second example is that any savings that accrue to Medicare by reducing avoidable hospitalizations are not shared with Medicaid, reducing the incentive of Medicaid programs to address this issue. There are many others.
  • Not all avoidable hospitalizations can be avoided, so it might be easy to overstate the plausible savings from a program to address these issues. Brad Flansbaum has a nice post on how difficult it is to determine why readmissions occur, who is responsible, and how to fairly and effectively incentivize hospitals to reduce them.
  • Many nursing homes do not have the appropriate staff to safely address acute health problems that could be addressed within the NH.

Ouslander and Berenson note that the incentives must be changed to address the issue and list some possibilities based on passed demonstrations and aspects of the ACA. I have written about two polar opposite approaches: federalizing Medicaid and integrated care provided by private insurance, both of which seek to incentivize one payer to make changes. However, the dual eligible Special Needs Plans that now exist undertake almost no such integration. There are profound political barriers to either approach in addition to many practical difficulties.

We have now well identified that the dual eligibles are a vulnerable group that costs a great deal while having numerous quality issues. We need to move ahead and try and evaluate as many models as possible, but the biggest barrier to doing so is the political stalemate over health reform that our nation is currently in. Until we manage to break through that, it will be lots easier to identify big problems than it will be to take practical steps to address them.

update: my fourth grade spelling teacher wouldn’t have been happy with the original title of the post

Dual Eligible Special Needs Plans: Unrealized Potential

This is the second post highlighting some key findings from the Kaiser Family Foundation’s September 2011 Data Spotlight on Special Needs Plans (a subset of Medicare Advantage plans). The first post focused on overall enrollment.

While enrollment in dual eligible Special Needs Plans (SNPs) has risen, the distribution of such enrollees is unequal, both geographically and across health plans. Around 2% of Medicare beneficiaries are in SNPs. However, there are 8 states in 2011 with no SNPs of any type, while 5 states and the district of Columbia have 15+% of their Medicare beneficiaries enrolled in SNPs. Around 80% of SNP enrollees are in dual eligible plans which are the most common type SNPs in all states where such plans are offered. The 8 states with no SNP enrollees of any type in 2011 are: Alaska, Montana, Wyoming, North and South Dakota, Vermont, New Hampshire and West Virginia.

In addition to geographic variation, a relatively small number of insurance firms are heavily involved in offering SNPs. Virtually all insurance companies had more dual eligible SNPs as compared to institutional or chronic condition plans, with only two insurance providers (see table A5 in the brief) having more patients in other types of plans.

The concentration of dual eligible SNP enrollees into a relatively small number of private insurance companies is consistent with a story of specialization of services to treat the complicated and numerous needs of dual eligibles. However, to this point, most dual eligible SNPs have not lived up to their potential. From the brief:

…coordination between the Medicare and Medicaid programs is more a goal than a reality, posing challenges for SNPs and for enrollees who receive coverage under both programs. While existing dual SNPs will be required to contract with state governments by 2013, it is not clear how this requirement will be implemented and how effective it will be…. Dual SNPs also face enormous challenges in providing appropriate services to enrollees with heterogeneous and often highly‐specialized and diverse needs, ranging from younger beneficiaries with debilitating mental disabilities to older enrollees, with physical and cognitive impairments, living in nursing homes or trying to maintain their independence at home.

The health and social needs of the dual eligibles are complex, and so far SNPs represent mostly unrealized potential in seeking to reduce the cost of their care while improving quality.


Marsha Gold, Gretchen Jacobson, Anthony Damico and Tricia Neuman. Special Needs Plans: Availability and Enrollment, Kaiser Family Foundation Program on Medicare Policy. h/t and thanks to Adam Coyne and Amy Berridge of Mathematica Policy Research for providing me with this brief and other information on SNPs.

Dual Eligible Special Needs Plans-Enrollment

Kaiser Family Foundation has a useful Data Spotlight on Special Needs Plans (Sept 2011), a subset of Medicare Advantage plans that cover dual eligibles (Medicare and Medicaid), persons living in institutions like nursing homes, and individuals suffering from chronic, disabling conditions such as congestive heart failure.

Such plans have provided increased opportunities for high need/high cost Medicare beneficiaries to have private insurance options in lieu of traditional Medicare. I am going to do several posts focusing on dual eligibles and Special Needs Plans (SNP): enrollment, concentration of beneficiaries across plans, evidence that SNPs could reduce costs and/or improve quality, and questions about setting premiums and risk adjustment, following on recent posts on efforts to reduce cost and increase the quality of care for the duals. There are two polar opposite approaches to dealing with the care of the dual eligibles: federalizing Medicaid, and moving them into private plans. Of course a hybrid approach is most likely, and interest is reforming the care of dual eligibles and evaluation of same is likely to remain a high priority.

SNPs were enabled by the Medicare Modernization Act of 2003, and beneficiary enrollment of dual eligibles in SNPs has risen steadily in spite of a decline in the number of plans offered after 2008.

  • After peaking in 2008, the number of SNPs has declined since; SNPs serving dual eligibles have always been the most common plan type, but they too have declined since 2008
  • The number of beneficiaries in dual eligible-SNPs has increased, even as plans have declined
  • The number of beneficiaries in institutional and chronic disease SNPs have declined along with the number of such plans offered
  • Patients in all SNPs are concentrated in a relatively small number of insurance firms/plans; 12% of the firms offering SNPs now enroll 83% of such patients

Around 11% of all dual eligibles are enrolled in dual eligible SNPs, but rates vary tremendously across the nation; no Medicare beneficiaries are enrolled in SNPs of any type in 8 states. There are many questions that need to be answered to determine the potential for such plans to improve quality of care while addressing the very large costs of caring for the dual eligibles and the body of evidence should make us skeptical about the ability of SNPs to reduce costs as compared to FFS Medicare. However, dual eligibles typically have large Medicaid costs as well, and substantial long term care needs that may be unmet, particularly for duals living in the community. A holistic assessment of costs and quality is needed. Answering these inter-related questions is a top priority since over 1 million dual eligibles are now enrolled in dual-SNPs, and cost and quality problems abound in how they are cared for now.


Marsha Gold, Gretchen Jacobson, Anthony Damico and Tricia Neuman. Special Needs Plans: Availability and Enrollment, Kaiser Family Foundation Program on Medicare Policy. h/t and thanks to Adam Coyne and Amy Berrige of Mathematica Policy Research for providing me with this brief and other information on SNPs.

Care Coordination for Dual Eligibles, ctd.

The June 2011 MEDPAC report provides important information about private insurance options to cover dual eligibles (persons covered by Medicare and Medicaid) amplifying yesterday’s post on care coordination.

The Medicare Modernization Act of 2003 created a category of Medicare Advantage plans called Dual-Special Needs Plans (D-SNP) that only cover dual eligibles. D-SNP are a subset of Special Needs Plans (SNPs) that are private insurance options for groups with unique and complex health care needs. SNPs have been created to cover Medicare beneficiaries living in a nursing home, or having chronic disabling conditions such as End Stage Renal Disease, HIV, and congestive heart failure. Table 5-2 (p. 128) from the MEDPAC report provides SNP enrollment:

There were around 1.05 million dual eligibles in D-SNP plans in February 2011 (11.4% of all duals), and nearly 1.3 million persons enrolled in all SNPs. SNPs differ from other types of Medicare Advantage plans in several ways (e.g. they must cover part D; you can enroll at any time and not only during open enrollment windows if such a plan is available in an area). Several quick thoughts:

  • The problem being addressed is the high cost and potentially non-optimal quality of the care provided to the dual eligibles.
  • The goal of proposals to federalize the Medicaid costs of the duals, as well as to move duals into private plans are to incentivize and empower one payer to decrease costs while improving quality for dual eligibles.
  • Austin and Aaron have pulled together lots of evidence showing that public payers have done a better job at restraining costs over time than private ones. Most of this literature predates D-SNPs. Studies showing slower cost inflation in Medicare than private insurance were obtained with the dual eligibles (a very expensive group) being cared for by FFS Medicare.
  • MEDPAC says that most of the D-SNP plans (p. 128) do not provide fully integrated and coordinated care to the degree believed to be optimal to duals, typically because they do not coordinate Medicaid benefits. Standards and regulation for these plans will expand due to aspects of the Affordable Care Act that will require quality assessments of plans (2012), and a requirement that D-SNPs have state contracts with Medicaid for the coordination of and provision of services (mostly LTC, by 2013).
  • There are other types of coordinated activities for dual eligibles described in the MEDPAC report that are implemented outside of the D-SNP approach (PACE, state carve outs, other demonstrations p.124-26 ). Overall, MEDPAC says 2% (p. 124) of dual eligibles are in fully integrated plans. This is not a subset of the D-SNPs but includes other options as well.
  • Finally, Austin’s point from last week that lowering spending may not be the appropriate question to ask about whether expanding private insurance options in Medicare is a good one to remember. It would be great to improve quality and reduce costs, but doing both will be hard. The current default for how care is financed for dual eligibles leaves room for improvement on both counts and we need to try and evaluate different options.


Care Coordination for Dual Eligibles

Ken Thorpe has a new study funded by AHIP out in white paper form that suggests that moving all dual eligibles (covered by Medicare and Medicaid) into private plans using care coordination could save the federal government and states $125 billion over 10 years. Thorpe notes that CBO projects that combined Medicare and Medicaid spending on the duals will be $3.7 trillion over 10 years by default, owing to their multiple health care needs (over half have 5+ chronic conditions and many reside in nursing homes). The 9 million dual eligibles are among the most sick and vulnerable members of society.

Thorpe proposes defaulting dual eligibles into private plans who will have strong incentive to reduce costs since they would be responsible for all care; similar incentives for Medicare and Medicaid are now reduced owing to the ways they share the cost of caring for the duals. A few details:

  • All dual eligibles would be enrolled in a health plan with an opportunity to opt out. All covered Medicaid and Medicare services would be provided through the health plans.
  • States can innovate and design their own approaches, but must have an evidence based design.
  •  Health plans would be responsible for coordinating the multiple needs of their patients, including contracting with other entities such as community health teams (defined in section 3502 of the Affordable Care Act) or medical homes.

The paper includes a rich literature review that suggests costs can be reduced while improving quality, and he includes a simulation of potential 10 year savings if all duals were in private plans: $80.9 billion for Medicare and $44.6 billion for Medicaid, around $35 billion of which would be savings to the states (clearly upper bound estimates):

I have suggested that dual eligible costs should be federalized (essentially making Medicare responsible for their total cost), with the main point being to put one payer in charge of all their care in order to incentivize cost reduction while increasing quality. Thorpe’s idea would achieve the same general goal using private insurance. There are many relevant details to discuss about Thorpe’s proposal (how would premiums for private plans be set?, what are realistic uptake estimates?, etc.), but it is clear that the current approach to caring for the dual eligibles is wanting both in terms of cost and quality. The only question is what changes should be undertaken to better care for them?

Update: There has been lots done on TIE about private v. public plans and cost control and the evidence tilts toward public doing better than private at controlling costs [FAQ and podcast]. Most Dual Eligibles have not been in private Medicare plans, so the balance of the evidence does not include them. However, the special needs plans do cover duals. There was some interesting tweeting yesterday about distribution of dual eligibles in such plans; I will try and track down a linkable source and blog about these issues more over the next week. Finally, by noting in the first sentence of the post that this is a ‘white paper’ that means it has not been subjected to peer review.

Health Costs of Working Age Disabled Adults

There are around 19 million disabled persons between the ages of 18 and 64 in the U.S. (6.4% of all persons in this age range), and the federal government and states spent around $425 billion on assistance to them in 2008 according to a new paper in Health Affairs.* This paper focuses on the proportion of this assistance that goes for health care as compared to the other aid components. Health care costs for this group are increasing rapidly, while spending on education, training and employment is declining.

Around half of the disabled working age adults in the U.S. have their health care costs paid for by three programs: the Veterans Administration, Medicare and Medicaid (some are dual eligibles, covered by Medicare and Medicaid due to their disability/poverty status). Health care expenditures were both the largest and fastest growing component of assistance provided to working age disabled persons from 2002 to 2008. A limitation noted by the authors is their inability to provide per capita expenditures for persons by program, eligibility for multiple programs, and reason for disability. This would be especially useful for the dual-eligible patients. Such persons are a part of the roughly 6 million persons nationally who are eligible for Medicare and Medicaid who are not age 65 or over and are a part of what I call “the third Medicaid program (with low income non disabled persons being program 1, and elderly dual eligibles being program 2).

This “third program” in Medicaid–non elderly dual eligibles is the most diverse portion of the Medicaid population due to the many reasons for their disability. More clarity on the reasons for the qualifying disability could be useful to identify strategies to address patient needs in the most efficient means possible, as well as to inform any potential reform strategies for how to best care for such persons.

The need to better understand the needs of these groups is heightened by the looming expansion of Medicaid as part of the Affordable Care Act. Expanded access to Medicaid under the ACA will likely serve to increase the differential eligibility (lag between Medicaid eligibility and Social Security and Medicare disability)  for the variety of programs for which disabled working age adults may be eligible. There are likely issues of duplication and of unmet need, and certainly of timing. Similarly, persons with disabilities that prevent them from working but who are now ineligible for programs due to the degree of their disability will presumably be added to the ranks of the insured via income based subsidies (if incomes are above 133% of poverty).

Care coordination issues abound for disabled persons of working age and these issues require more attention.

*Gina Livermore, David C. Stapleton, Meghan O’Toole. Health Care Costs Are a Key Driver of Growth in Federal And State Assistance to Working-Age People with Disabilities. Health Affairs 2011;30:9:1664-72.

update: added the link to the paper