DPW says the new reporting method is just as accurate as the old one, merely different. But it will not disclose its new method or recalculate the latest Medicaid data using the old formula.
"The timing is suspicious, and it obscures the impact of the decisions that DPW had made," said Sharon Ward, executive director of the nonpartisan Pennsylvania Budget and Policy Center. "What they've done now is made it really impossible to compare."
Anne Bale, a DPW spokeswoman, said the department made the revisions because its long-standing method of reporting Medicaid enrollment counted recipients more than once if they received benefits from more than one program.
"The hope was to get rid of that duplicate issue so the numbers were more accurate," Bale said.
For consistency, the monthly reports have been revised back to July 2010. A comparison of both methods shows that the number of enrolled adults, just over one million, is between 50,000 and 88,000 lower every month in the revised data. The last month with both sets of numbers was November.
Policy experts who follow the reports said they had expected the downward trend to continue in December, both because duplicates were removed and also because the state was still working through a backlog of controversial eligibility reviews that it says has resulted in 65,000 adults losing benefits over the last few months.
Enrollment data for children, which is still calculated using the old method, showed a drop of 88,000 - half of that in December alone - since August, a period that roughly coincided with reviews that caused 71,000 children to lose benefits. (Bale said the children's data would soon be revised as well.)
The opposite happened for adults.
"I am totally flummoxed by what they are doing," said Richard Weishaupt, senior attorney at Community Legal Services of Philadelphia.
Bale said there was a simple explanation. "The numbers are right," she said. "The program statewide continues to grow even though we are taking people off."
A note that accompanied some of the revised adult tables, on Jan. 13, said that the formula used to calculate medical assistance numbers had changed but gave no details. It added: "This does NOT mean that the previous MA #'s were less accurate. It would be up to the individual which #'s to use when looking @ trends."
Bale said DPW would not provide the data needed to continue the former trend.
Ward, of the budget and policy center, said DPW should report it both ways.
"If you want to give an honest assessment of policy changes and the impact of those changes, you want to compare apples to apples, and this makes that difficult," she said.
DPW officials in July ordered workers to process a backlog of what turned out to be more than 700,000 eligibility reviews, most of them less than a year overdue. Officials said the reviews were not a policy change but enforcement of state law, which requires eligibility checks every six months. People who lost benefits as a result, they said, were ineligible.
Legal advocacy groups believe that many if not most of those cases - about one-fourth of all that were reviewed - were eligible. They describe a "perfect storm" of already overburdened state welfare offices overwhelmed by the additional paperwork as well as applications for disaster aid after summer floods. They say that nearly all the cases they have handled were eligible and were retroactively reinstated.
But they suspect that many recipients still don't know they have lost benefits and may not find out until they go to a doctor or a hospital.
DPW Secretary Gary Alexander has publicized the reviews as an example of the Corbett administration's commitment to ferreting out waste, fraud, and abuse. He has said he plans more sweeping changes, intended to focus safety-net programs more narrowly on emergency aid, in the future.
Health policy analysts argue that targeting beneficiary waste, fraud, and abuse may garner headlines but that there is little evidence of savings as a result. And they questioned why the data was being revised at the very moment when the impact of changes could be measured.
Bale, the DPW spokeswoman, said there was "no particular reason" why the method was revised now.
Nor does it necessarily support Alexander's contention that he is saving money.
Even with tens of thousands of people losing benefits since August, the revised method shows that adult enrollment in Pennsylvania's Medicaid program grew by more than 2 percent. That is nearly the same rate that it expanded from August to December 2010 under then-Gov. Ed Rendell, when unemployment was higher.
Contact staff writer Don Sapatkin at 215-854-2617 or firstname.lastname@example.org.