By Rose Hoban
Like cats and dogs, state officials and auditors often don’t get along.
When State Auditor Beth Wood came out last week with an audit of the state Department of Health and Human Services, it temporarily set fur flying in both agencies.
Initially, DHHS officials said the auditor had overstepped when she determined there were deficiencies in DHHS’ processing of Medicaid claims that could mean that as much as $853 million in payments were “questionable.”
Dave Richard, the state Medicaid director, pointed out that the total calculated by Wood was an extrapolation based on a sample of 396 claims her auditing staff chose at random. Out of that sample, 50 claims had a total of $4,288 in overpayments to Medicaid providers, payments to physicians and others who had not submitted qualifying paperwork and payments to providers who were ineligible to provide Medicaid services.
Scaled up, Wood argued, that meant as much as $835 million could have been misspent.
But Richard said that extrapolation was too high and misleading.
“We don’t want the public believing that there’s this amount of money out there that the Medicaid agency can go and recoup back into the program,” he said.
But Wood retorted that Richard was not a statistician, and that she had the evidence to support her conclusion.
She pointed out that this kind of sampling is adequate for the census, adequate for medical research, adequate for opinion polling.
“Our work was statistically valid, our work was statistically sound and we had a renowned statistician working with us through this complex project,” Wood said.
North Carolina Health News asked some folks who have “no dog in this fight” to take a look at the documents and answer a few broader questions about audits.
Was the sample big enough?
Jeff Ovington has worked in Medicaid finance at the Texas Health Authority for a decade. He’s got North Carolina connections (he went to Duke) and follows what’s happening here. He took the time to read the 54-page audit and wrote to N.C. Health News to state his disagreement with the auditor’s findings.
“Now, that 396 claims may actually be a statistically-valid sample size, and if so, extrapolation is a valid tool,” Ovington wrote. “But everyone needs to be clear about this: The auditors did not find millions of actual specific dollars, they found four grand.”
But Lynn Dikolli, who teaches accounting and auditing at the UNC Kenan-Flagler Business School, said the sampling approach is an industry standard.
Auditors have to sample, she said, “because they need to draw a conclusion about a population without testing every single item in that population; it’s just not efficient or effective. And when you have millions of transactions, it’s just not possible.”
Dikolli said there is a risk that the sample might lead to the wrong conclusion. Auditors use statistical sampling to overcome that sampling risk.
Wood used a large enough set of claims data to create a “90 percent confidence interval.” In other words, if Wood were to repeat the whole process with a completely different set of randomly chosen files, she would replicate the results nine out of 10 times. At Wood’s sample size, the result falls in a range that stretches from about $492 million to more than $1.2 billion.
That’s a pretty big range, Dikolli said. The only thing she might have done differently is look at more records to get a tighter range of possibilities.
Signed, sealed, delivered?
Wood found that many of the providers who billed DHHS for services didn’t have the correct paperwork filed with the state. Technically, this lack of documentation means these providers could have been improperly paid.
“You can talk about the logic of it: ‘The provider wasn’t enrolled,’” Ovington said in a phone interview. “But does that mean the person who needed the services never got the services? No.”
During a review, auditors backtrack to providers and ask them to show their paperwork. And if after several months the providers haven’t offered it, an error is recorded. Ovington pointed to those paperwork requests as a reason doctors hate being part of a state Medicaid program.
He also argued that there’s a difference between saying that paperwork transactions were incorrect and that all the money was squandered.
“So what you’ve got here is some issues of tightness of control and paperwork. It means that there’s some looseness in the system,” he said. “You’re going to find some degree of looseness in all of these systems because they’re so complicated.”
Dikolli said when Wood pointed out the lack of documentation, she was really pointing to those kinds of control issues put in place precisely because the program is so complicated.
And if you find too many lapses in documentation, it could point to a bigger problem.
“Controls are put in place to address risk,” Dikolli explained. “One of the big risks is that we make payments to someone that are unsupported. It is the wrong provider. Or it’s too much money. And is the service unsupported?”
She said the auditor’s office is trying to confirm that the billions of dollars being spent providing these services are actually valid. But the projection is still shockingly large.
“So you say, ‘How did we go from $4,000 to a billion?’ and that’s how I can see how DHHS is saying, ‘Wait a minute,’” she said, especially after DHHS officials showed many of the issues were paperwork problems.
It adds up
That gets us back to the question of whether $853 million was actually missing or misspent.
And that was at the heart of a discussion during Tuesday’s meeting of the Joint Legislative Oversight Committee on Medicaid and NC Health Choice, where Wood presented her findings.
On this occasion, Wood and DHHS Sec. Rick Brajer noted the respect they have for one another’s work.
“My objective for the audit of any of these federal programs is to meet the federal requirements, which is to go in and identify errors,” Wood said. “When we find errors, we identify them as questioned costs.”
She explained that many of the errors she found were overpayments to providers. But when DHHS realizes it’s overpaid someone by less than $150, they take no action, because it could cost more than that to recoup the money.
Wood calculated those overages at about $200 million, something DHHS officials didn’t dispute.
“Think about it: $150 spread out over a $11 billion program, 127 million transactions … stopping it from going out the door,” she told the committee.
According to DHHS, North Carolina has about 80,000 providers. If each provider has 30 claims per year that are $100 over the accepted amount, that totals $240 million.
And that’s just in the Medicaid system. Wood found similar problems in the children’s health insurance program, NC Health Choice. But since that’s a smaller program, the total was likely only several million.
“It’s important if there’s a way that we can find, economically, to stop these types of payments from going out the door in the first place,” Wood told the committee.
She said she and Brajer would work together to see which overages are bigger costs to the state in order to fix them.
Which, Dikolli said, is the best possible outcome of an audit.
In the end, what in the past would have been a tense encounter between DHHS officials and Wood was tempered by news that Medicaid is running more than $200 million under budget this year. That change in atmosphere was noted by committee co-chair Sen. Ralph Hise (R-Spruce Pine).
“Given what we’ve been in over the past few years, I think it’s most exciting that there’s a working relationship now between the department and your office,” Hise said to Wood.
Republish our articles for free, online or in print, under a Creative Commons license.
I don’t disagree with sampling, and as I said, if 396 claims was a statistically-valid sample, then extrapolation is certainly a valid tool.
My biggest issue is the characterization of the entire amount ($835 million) as being wasted or misspent. If indeed there is something like $200 million that is being lost due to thousands of small over-payments, then I agree with the statement that it would be valuable to develop a methodology to “economically … stop these types of over-payments from going out the door in the first place.”
But the small over-payments were a minority of the audit findings. Most of the rest of the dollars were for issues that likely would have resulted in eventual similar payments anyway, after the paperwork or enrollment problems were resolved. The majority of the dollars in the findings do not point definitively to a situation that, had everything been perfect, the state would never have had to spend the money. That is a critical distinction.
To suggest that DHHS wasted $835 million is not only not warranted by the specifics that were contained within the audit report, but it is misleading to the taxpayers, and it is an affront to the people that work at DHHS.