Washington ? State Medicaid programs, which provide health care to some of the nation’s poor, vary wildly in their eligibility criteria, the scope and quality of their care, and the amount they reimburse physicians providing it, according to an independent assessment published last week.

Overall, the programs are “severely challenged,” with the best scoring the equivalent of a low D and the worst way below an F.

“This evaluation demonstrates a bleak picture for millions of people in many states,” wrote the authors of the 143-page evaluation, produced by Public Citizen’s Health Research Group.

The top five programs, in order of rank, were in Massachusetts, Nebraska, Vermont, Alaska and Wisconsin. The bottom five, with lowest-ranked last, were in South Dakota, Oklahoma, Texas, Idaho and Mississippi.

Medicaid is paid for by both state and federal tax revenue. The federal government specifies the minimum services that must be offered, which states can broaden but not narrow.

For example, states must cover children on welfare and poor pregnant women. All states and the District of Columbia also choose to cover uninsured poor women needing care for breast or cervical cancer, even though this is not a requirement. Thirteen states and the District of Columbia also cover uninsured people with tuberculosis. (There are many other eligible groups, as well.) States also may apply for waivers to try experimental strategies for delivering services.

The program covers 55 million Americans and accounts for about 20 percent of U.S. health-care spending. It is the largest source of federal grants to states.

The anecdotal differences among states can be dramatic.

For example, a pregnant woman in a family of three must have a household income of less than $22,128 to qualify for Medicaid in Wyoming. In Minnesota, she could qualify with an income of $45,650, according to the report.

“We know that the differences between programs reflect both differences in priorities and resources, but nobody knows the extent of them. The programs haven’t been subjected to a uniform scoring scheme,” said Annette B. Ramirez de Arellano, a health-policy expert who headed the project.

The authors used published data to measure Medicaid performance in 55 areas. In calculating a final score out of a possible 1,000 points, they weighted issues of eligibility and reimbursement more heavily than breadth of services and quality of care.

The highest- and lowest-ranked states differed in their scores by a factor of two – 646 for Massachusetts vs. 318 for Mississippi. There was even more variation in the components that went into the total scores.