“When attendant services are provided under the conventional Medicaid model, state plans contract to provide specified amounts of personal assistance services to disabled people. Recipients have little control over who enters their homes, when they will come, what they will do, and how they will do it.”

The “Iron Triangle” theory of health care policy, which claims that one can achieve, at most, two of the three goals of high quality, universal access, and cost containment, has a number of influential adherents. Bloomberg Government hosted an April 2011 event entitled “The Iron Triangle of Healthcare,”1 which reportedly featured an examination of the “interrelationship between cost, access and quality in the U.S. health care system” and such speakers as the Secretary of Health and Human Services and a number of widely known people from government and business trade associations.

“Iron Triangle” discussions inevitably end up focusing on ways to control the provision of care; on increasing reporting requirements in order to centrally manage quality; and on generally reducing an individual’s ability to manage his own health care. But results from experiments with consumer-directed care in Medicaid show that this view is unduly pessimistic. In fact, acting on this view may add to cost while degrading quality. The experiments suggest that people given the freedom to operate outside of the “Iron Triangle” systems can simultaneously improve quality, provide better access, and reduce expenditures as long as their incentives are properly aligned.

Properly aligning incentives means that the people spending the money must be encouraged to voluntarily economize on their use of health care by being subjected to a spending constraint that requires them to choose between medical care and other goods. They must also be free to negotiate for those services as they see fit. In the Medicaid experiments described below, people are typically given a budget for buying the care that Medicaid officials think their physical condition warrants and then are given the freedom to negotiate the price and bundle of services that best fit their circumstances. Because they have more control over how the money is spent and can get more care if they spend it wisely, they are able to buy services that better fit their health needs and living conditions. Their health improves, thus reducing the cost of providing them with medical care.

Consider one component of Medicaid: the provision of attendant services. Such services include housekeeping, food preparation, bathing, toileting, transportation, administering medications at home, therapeutic services, and shopping. When attendant services are provided under the conventional Medicaid model, state Medicaid plans contract with certified agencies to provide specified amounts of personal assistance services to disabled people. Medicaid caseworkers, in conjunction with agencies in some states, determine the kind of care and the number of hours a Medicaid beneficiary is entitled to receive.

Recipients of agency care, that is, care provided by certified agencies, have little control over who enters their homes, when they will come, what they will do, and how they will do it. Theft is a constant problem. In addition, the recipients cannot reward good performers or dock poor ones because personal assistants work for the agency that hires, trains, schedules, supervises, fires, and bills for their services. Because Medicaid beneficiaries often do not see invoices for their care, state officials have no way of knowing if the services agencies bill for were actually rendered. In some states, reimbursements rates are set so low that good workers refuse to work for the program, and Medicaid recipients cannot get the hours of service that they need.

Regulations governing agency hiring tend to be unnecessarily complex. A 2006 Health and Human Services report counted 301 different sets of requirements for Medicaid attendants, including background checks, literacy, health training, adequate supervision, record-keeping, and education2 In addition to this, agencies must comply with state Nurse Practice Acts. These laws vary from state to state but generally prohibit certain services from being provided without supervision by a licensed nurse. Nurse Practice Acts can increase the cost of caring for people who need both housekeeping services and care requiring nurse supervision.

Workers who fail to show up pose one of the most common problems in agency-directed care. When attendants miss work, incapacitated people may be left waiting in bed all morning, may miss appointments, or may end up spending the night in a wheelchair. No-shows cause such problems that some people have been driven to desperate measures to avoid them. One working parent of an adult with severe disabilities missed so much work covering for no-shows that she quit her job, got a nurse’s aid certification to satisfy certification requirements, and became her son’s full-time caregiver. Another parent started his own nursing agency. He felt it was the only way he could have some control over the quality of attendants hired to provide for his daughter’s needs.3

People dependent on agency care may hesitate to complain if they fear agency retaliation. Poor attendant performance is difficult to prove because home care disputes are generally one person’s word against another’s. Agencies can retaliate against people who make life difficult for them by sending lousy attendants, by classifying someone as a threatening or disruptive patient, or by issuing a finding that the person making the complaint can no longer live on his own and needs to be in a more “protective” environment.

For background information on Medicaid, see Welfare, by Thomas MaCurdy and Jeffrey M. Jones and Health Care, by Michael A. Morrisey in the Concise Encyclopedia of Economics.

When Medicaid personal care services are consumer-directed, however, participants generally control who works for them. They may be able to hire and fire their own service providers and train, schedule, and supervise them. In some programs, they can also set wages and working conditions and meet payrolls with Medicaid personal-assistance funds on deposit with fiscal agents. The fiscal agents help train program participants, help to ensure that all proper employer forms are filed, and provide payroll services.

Although there are hundreds of consumer-directed care programs operating in various states, most are only a few years old. California’s In-Home Supportive Services (IHSS) program has allowed both consumer-directed and agency-directed long-term care since 1979. By 2001, about a dozen California counties were offering consumer-directed programs.4 In 2000, Benjamin et al. conducted a random telephone survey of 1,095 clients. They found that people using the consumer-directed model were more likely to hire people that they knew were compatible, that they invested considerable resources in client-specific, on-the-job training, and that the people they hired were less limited in the work that they undertook. People using the consumer-directed model reported fewer unmet health needs, higher service satisfaction, and less concern about their safety.

Between 1999 and 2003, the Robert Wood Johnson Foundation funded the Cash and Counseling Demonstration in Arkansas, Florida, and New Jersey. Volunteer Medicaid recipients in each state were randomly assigned to either a consumer-directed or an agency-based care option. More than 1,700 people assigned to the Cash and Counseling option in each state received cash allowances that could be used to hire personal-assistance workers, including family members, or, depending on each state’s rules, to purchase other goods and services.

In order to participate, people had to develop plans showing how they would use their budget or designate a representative to do so. More than half of the participants were either elderly or cognitively impaired. Participants submitted time sheets and invoices to a fiscal agent. The agent then wrote checks for the services rendered. Counselors were available to provide help in traditional case management.

Participants used their allowance to hire paid caregivers, modify their homes to better accommodate their disability, and buy equipment, such as microwave ovens, that helped them be more self-sufficient. As in the IHSS program, the people in the Cash and Counseling program were significantly more satisfied with their paid care-givers, had higher overall satisfaction rates, and were significantly less likely to report unmet needs than people who received agency care. People directing their own care also reported fewer adverse outcomes or injuries. Though they tended to receive fewer hours of care, their relatively high satisfaction rates suggested that needed care was provided more efficiently under the consumer-directed model.5

Cash and Counseling was designed to be budget-neutral. People in consumer-directed programs were not allowed to spend more on personal care services than was already allowed for people in the traditional agency-directed programs.

In the first two years, consumer direction increased Medicaid spending for personal care services, as people were finally able to buy the services that they were legally entitled to. In Arkansas, the state with the largest increase, about 28 percent of those in agency-directed groups received no personal care services at all. Those who did receive care generally received only two-thirds of the hours in their care plan.6 In all, Arkansas’ personal care service expenditures were $2,349 for the agency-directed group and $4,605 for the consumer-directed group.

Based on the two-year follow-up that was part of the Arkansas study, Dale and Brown concluded that consumer direction is likely to have a “modest” impact on “total Medicaid costs” in the short run and might generate savings in other long-term care services.7 In Arkansas, costs for other Medicaid services were about 14-percent higher ($1,514) in the first year in consumer-directed plans. They fell to just five-percent higher, a statistically insignificant difference, in the second year. On average, overall Medicaid spending went up by about $500. During the second year, people with consumer-directed care spent fewer days in nursing homes and had fewer home health therapy visits, encouraging speculation that a longer-term program might produce overall savings.

Colorado’s experience with its Consumer Directed Attendant Support Services (CDASS) program suggests that given the right incentives, consumer direction can actually reduce overall Medicaid costs. It also shows that even successful programs can be undone by the reimposition of ill-considered regulations.

The CDASS pilot began in 2002. Designed by a group of disabled people in concert with state officials, it was open to anyone who had received personal services from Medicaid for a year. Those who chose to participate were provided with a personal care budget allocation that was equal to the cost of agency care for the level of service indicated by a client’s condition. The state provided training on the responsibilities of employers, and participants were required to maintain their personal care budget allocation in an account housed with a fiscal intermediary that acted as a bookkeeper and escrow agent. The fiscal intermediary received a 12-percent fee for managing an individual’s account, for training, and for providing payroll services.

The Colorado pilot program was unique in the freedom that it gave participants. Participants paid their attendants as they saw fit and kept half of any savings that they generated, an important freedom given that recipients lived in a variety of urban and rural settings, each of which had unique local labor-market conditions. In the first two years, monthly spending was 21 percent under budget, an average savings of $794 per client per month. Program participants used their half of the savings to buy things that Medicaid did not cover, such as voice-activated telephones for quadriplegics.

The ability to vary wage rates at will was an important factor in realizing CDASS savings. In other states, government or unions often control payments to Medicaid workers. According to a 2010 Pennsylvania Executive Order allowing home care workers to unionize, personal assistants in consumer-directed home care are unionized in at least six other states.89 In 2010, Iowa Republicans complained bitterly when the governor rescinded reimbursement rate cuts for caregivers in the state’s consumer-directed Medicaid program, while keeping them intact in other programs. Iowa Medicaid’s personal care attendants were unionized, and AFSCME, the union representing them, had filed a grievance.10 The union’s refusal to reduce mandated wage rates increased state Medicaid costs by an estimated $1.1 million. This might seem like a small number, but remember that this is for a pilot program with a limited number of participants.

Given that CDASS satisfaction rates approached 100 percent and that its shared savings approach, coupled with hiring flexibility, had generated savings of 20 percent, the pilot was judged a success. Colorado officials sought to make CDASS part of Colorado’s regular Medicaid program. Unfortunately, in the transition from pilot program to Medicaid, CDASS has been compromised by state and federal rule changes. The nature and extent of the rule changes suggest that either state and federal administrators misunderstood the program’s success or that those opposed to it—generally, the agencies providing the attendant care that direct CDASS hiring replaced—had significant influence on the rulemaking process.

The conditions for making Colorado’s CDASS pilot a formal Medicaid program immediately ended the shared savings and eliminated an important incentive for recipients to keep their spending under budget. The federal government required that the fiscal agent fee be reduced to ten percent. In addition, family members were limited to a 40-hour work week. Many family members worked more than 40 hours per week because they were the caregivers preferred by participants, and they might have had to seek other jobs due to the proposed cut in hours. Thus, Colorado’s Medicaid agency recommended that CDASS participants accommodate the 40-hour limit by adjusting the hourly rate that they paid family members. As the state provided the same budget for care, regardless of the hours worked by family members or the amount that they were paid, the 40-hour restriction made little sense within the context of the program.

Once CDASS participants accepted the 40-hour per week limit on family-member labor, the state government moved to further restrict CDASS hiring by applying a $40 per hour cap on the wages that program participants pay. Coupled with limits on hours, setting a maximum wage provides a precedent for putting the state back in control of pay for personal attendants.

State officials also plan to limit CDASS flexibility by requiring participants to submit wage justifications along with their care plans. It is unclear how various levels of the bureaucracy will react to some of the more creative CDASS payment arrangements. For example, a CDASS participant might offer someone $100 a month to be on-call every evening and record the reimbursement as $100 for one hour of work simply because the timesheets approved for the program do not have a category for this kind of work arrangement.

The state also impeded program functioning by failing to provide reliable bookkeeping and training services. CDASS operates on an annual allocation for attendant care rather than on month-by-month payments because labor costs for those who need attendant care can vary throughout the year: holiday coverage costs more; some people’s conditions worsen in cold weather; people may have to pay significantly more for coverage under emergency conditions, such as blizzards; and people work hard to save their allocations so that they can hire attendant care coverage for vacations. As is common practice throughout the private sector, the state’s CDASS fiscal agent is supposed to provide them with a monthly statement of their charges and of the balance of the funds in their account.

Participants claim that the fiscal agent’s performance worsened after the pilot program transitioned to formal Medicaid. Fiscal agent fees come out of the amount budgeted for each client, and matters became critical when the state arbitrarily reduced clients’ attendant care allocations as part of a generalized package of budget cuts at about the same time that the state rebid the fiscal agent contract and, once again, changed the fiscal agent fee. Allocations went down, fiscal agent fees went up, confusion reigned, and some participants went months without receiving accurate account statements. Lawsuits were filed. By December 2009, things were so bad that some caregivers did not receive paychecks.

Groups representing disabled clients say that some state Medicaid officials now claim that CDASS spending is out of control due to client overspending. Spokesmen for CDASS participants maintain that state officials are exaggerating this problem by including cases in which a combination of budget cuts and a lack of monthly statements made it impossible to know what was available to spend.

Some CDASS participants think that opportunistic officials hostile to consumer direction are trying to blame them for state mismanagement. They cite cases of CDASS training CDs being sent to clients who did not have a computer; incompetent fiscal agents; the arbitrary imposition of standardized timesheets ill-suited to attendant hiring; and case managers who threaten to send clients to nursing homes if their worsening health leads them to request an allocation review.11

CDASS, Cash and Counseling, and California’s In-Home Supportive Services programs have demonstrated that voluntary, properly designed, consumer-directed programs can smash the “Iron Triangle” by reducing both health care costs and expenditures while improving quality. Unfortunately, they have also shown that freeing people from the excess regulation that produces the Iron Triangle is no easy task. It requires sustained attention to nuts-and-bolts program operation; a steadfast dedication to protecting participants’ freedom to hire at will and determine reimbursement rates; a commitment to providing timely, accurate information to all parties; and people willing to protect nascent programs from special interests in both the private and public sectors.


Office of Inspector General, Health and Human Services. December 2006. States’ Requirements for Medicaid-Funded Personal Care Service Attendants. OEI-07-05-00250.

Laura Hershey. 2004. Improving Infrastructure: Voices of Attendant Services Users, A Report on Six Focus Groups Between April and September 2004. A report for the Independence Plus Grant, Department of Health Care Policy and Financing, State of Colorado, Denver, Colorado.

Jane Tilly and Joshua M. Wiener. 2001. Consumer-Directed Home and Community Services: Policy Issues. [PDF file.] Occasional Paper No. 44. Assessing the New Federalism, The Urban Institute. Washington, DC.

Barbara Lepidus Carlson, Stacy Dale, Leslie Foster, Randall Brown, Barbara Phillips, and Jennifer Schore. May, 2005. Effect of Consumer Direction on Adults’ Personal Care and Well-Being in Arkansas, New Jersey, and Florida. Mathematica Policy Research, Inc. Princeton, New Jersey.

Stacy Dale, Randall Brown, Barbara Phillips. June 2004. Does Arkansas’ Cash and Counseling Affect Service Use and Public Costs? [PDF file.] Mathematica Policy Research Inc. Princeton, New Jersey.

Stacy B. Dale and Randall S. Brown. February 2007. How Does Cash and Counseling Affect Costs? Health Services Research,42, 1, Part II, 488-509.

Service Employees International Union. Website. Accessed April 15, 2011.

State of Pennsylvania. Governor’s Executive Order 2010-04. Consumer-Directed Home Care Providers.

Iowa House Republicans. September 2010. “Culver Gave Raises for Union Members Only.”

Personal communication from Julie Reiskin, Colorado Cross-Disability Coalition. March 2011.


*Linda Gorman is Director of the Health Care Policy Center at the Independence Institute in Golden, Colorado and a frequent contributor to John Goodman’s Health Policy Blog.

For more articles by Linda Gorman, see the Archive.