Figuring out which approach to managing care for people who are eligible for both Medicare and Medicaid is difficult.

Tamara Hayford and Andrea Noda, analysts at the Congressional Budget Office (CBO), make that observation in a report on “dual eligibles” — people are eligible for both Medicare and Medicaid.

Those people are generally both very poor and sick enough to be disabled, or they are people who are old enough to be in Medicare who have made themselves poor on purpose, in an effort to qualify for Medicaid nursing home benefits.

Dual eligibles make up only 13 percent of  the population enrolled in either Medicare or Medicaid but account for 34 percent of total Medicare and Medicaid spending.

Because caring for dual eligibles is so expensive, policymakers are extremely interested in finding ways to reduce the cost of care for dual eligibles while maintaining the quality,

One table in the CBO report shows that ordinary managed care has the lowest per-enrollee cost for dual eligibles, and that another type of all-inclusive managed care, PACE, has the highest per-enrollee cost.

But simply comparing per-enrollee costs may be misleading, because officials resist any efforts to force enrollees and their caregivers to choose between one type of program and another, the CBO analysts said.

In most cases, analysts said, enrollees or caregivers have made a voluntary decision to choose a particular test program or regular program option.

Because the decisions are usually voluntary, researchers know that antiselection may have affected the nature of the enrollees in a particular program.

The enrollees in one program could be sicker, healthier, richer, needier, more motivated, less motivated or different in other ways from the enrollees in the competing programs, the analysts said.

Some evidence suggests that bundling as many types of care, such as long-term care and behavioral care, into a program as possible might help control costs, by giving the providers a stronger incentive to coordinate care, the analysts said.

See also: