By Lambert Strether of Corrente.
Spoiler alert in one word: Privatization. Here is a brief description of the privatized Kansas Medicaid program, “KanCare,” from the Kansas state legislature’s Legislative Division of Post Audit (DPA). Page 7 of the PDF:
Launched in January 2013, KanCare is the program through which the State of Kansas administers Medicaid. KanCare offers health care for people with limited income, which may include pregnant women, children, and low-income families with children. The Kansas Department of Health and Environment (KDHE) and the Kansas Department for Aging and Disability Services (KDADS) jointly administer KanCare. KDHE maintains financial management and contract oversight of the KanCare program. KDADS administers the Medicaid waiver programs for people with disabilities, mental health conditions, and substance abuse problems, and oversees the state hospitals and institutions.
As the state’s Medicaid program, KanCare focuses on providing person-centered care coordinated through contracts with three private managed care organizations [MCOs]: Amerigroup of Kansas Inc., Sunflower Health Plan, and United Healthcare Community Plan of Kansas. The state also contracts with Maximus [hmm…], a company that processes the state’s Medicaid applications and provides support services during the eligibility process.
Or in short form:
[KanCare is depends on] a federal waiver that allows the state to contract with three private companies to operate KanCare, a $3 billion program that covers more than 420,000 low-income, elderly and disabled Kansans.
And let’s not forget Maximus, the fourth company!
I’m sure the KanCare story is a lot more complicated and horrid than this superficial survey will reveal — Kansas readers please chime in! — but even so things look pretty bad. (Note that KanCare is for existing Medicaid, not Medicaid expansion.) I’m going to start with the story on KanCare’s data debacle, move through their eligibility determination debacle, and close, for a change, with a bit of good news.
KanCare’s Bad Data and Phishing Equilibria
Here’s the story that set me off. Governing:
Legislators directed their independent auditors to find out whether KanCare is working.
After a year of work, those auditors recently released their determination: the state’s data is so bad [How bad was it?], there’s no way to know.
“These data issues limited our ability to conclude with certainty on KanCare’s effect on service use and limited our ability to interpret cost trends,” the auditors wrote. “More significantly, data reliability issues entirely prevented us from evaluating KanCare’s effect on beneficiaries’ health outcomes.”
Holy moly. (The article says “the state’s data,” which really means the data managed for the state by KanCare’s four vendors.) So I went looking for the report, ” “Medicaid: Evaluating KanCare’s Effect on the State’s Medicaid Program” (PDF). Here’s the section on data:
- Data reliability issues and state policy changes limited our ability to fully conclude on KanCare’s effect on claims costs and service use. Although the total claims dataset was reliable, 6 of the 12 individual service datasets were unreliable due to inaccurate or inconsistent data. This limited our ability to fully conclude on KanCare’s effect on service use, which also limited our ability to interpret post-KanCare cost trends.
- Data reliability issues prevented us from evaluating KanCare’s effect on beneficiaries’ health outcomes. Five of the seven health outcome datasets were unreliable because of inaccurate data. These inaccuracies were likely the result of difficulties KDHE had in correctly pulling data from its complex Medicaid systems. We did not have sufficient time to correct the errors we identified and ultimately decided we lacked sufficient evidence to conclude on KanCare’s effect on beneficiary health outcomes.
Note that the service data was bad because the medical coding (covered at NC here and here) was bad: “For example, prescription drug costs fluctuated between $24 million and $1.4 million between the first and third quarter of 2013. KDHE officials told us these differences were due to inconsistent coding and reporting by the MCOs” (that is, by Amerigroup, Sunflower, and United HealthCare). That’s exactly what you’d expect from three separate vendors in a system with no quality assurance. Here is the DPA report on quality assurance for claims data (page 28):
Although KDHE [Kansas Department of Health & Environment] has a process to ensure MCOs claims are allowable, they lack a process to ensure they are accurate.
Holy moly. More:
KDHE does not have a process to ensure reported costs accurately reflect MCOs actual costs. We expected KDHE to have a process to check a sample of MCO reported costs against original provider claims to verify the accuracy of reported costs. KDHE and DXC officials told us no such process exists. KDHE officials told us it was unlikely MCOs could overreport their costs because it would require them underpaying providers. KDHE officials told us providers would alert them to these issues.
So, KDHE processes claims on the honor system. Good to know. Not necessarily a debacle. Still on page 28:
We compared a judgmental sample of 19 provider claims from two Medicaid providers against MCO reported costs and found no significant discrepancies. These 19 claims showed how much the MCOs actually paid the providers. We worked with DXC [another contractor, not Maximus] staff to compare the actual paid amount in each claim to what the MCOs reported to KDHE. In all but one case, the paid amounts matched MCO reported costs. In the one exception, the paid amount exceeded what the MCO reported to KDHE. Although inaccurate, this type of error would not inappropriately increase the MCO’s capitated payment rates.
Although our review of 19 claims did not uncover instances of improper reporting, this was a very small, non-representative sample of claims. As such, there is still a risk that MCOs could improperly report costs to KDHE, resulting in incorrect state payments to the MCOs. KDHE could compare a larger, more representative sample of provider claims to MCO reported costs to help reduce this risk.
At this point, let’s remember the notion of a phishing equilibrium, which we might crudely summarize as “If there can be fraud, there is already fraud.” And here we note that LPA is defining accuracy in a curious fashion: the equality of provider billing and MCO payments. But what if the providers are playing games with the coding to maximize their billing, as we already know they do? KanCare doesn’t seem to be checking for that. What we do know is that KanCare’s honor sytem has allowed what we might categorize as, if not fraud, fraud-like behavior:
To encourage prompt payment of claims, K.S.A. 39-709(f) requires MCOs pay interest to providers on valid claims not paid within 30 days. However, MCOs should not include these interest payments in the data they submit to KDHE to avoid inappropriately inflating claims costs and therefore future state payments to the MCOs.
We verified one MCO, Sunflower, improperly included interest payments in the claims data it submitted to KDHE. We interviewed officials from all three KanCare MCOs to determine how they handled interest payments. Officials from Amerigroup and United told us they did not include interest paid in claims submitted to KDHE. However, officials from Sunflower told us they did report interest paid in their claims data. To confirm this information, we asked each MCO to submit five late claims on which they paid provider interest. We compared the fifteen total claims against data in the state’s system. The claims data did not include interest amounts for Amerigroup or United but did include interest amounts for Sunflower. It is important to note that although we did not find interest included in the claims for Amerigroup or United, our data was based on a non-representative sample of self-reported data.
MCO officials we spoke to had different understandings of whether or how to report interest payments to KDHE. It is possible clear guidance from KDHE, along with continued monitoring of MCO claims, could help prevent this issue from occurring in the future.
(“[O]ur data was based on a non-representative sample of self-reported data.” So maybe Sunflower was just stupid?) I dunno. Getting reimbursed for interest paid on a late claim doesn’t seem an awful lot like claims data to me. To start with, how would you have coded it? I have to work awfully hard to see what Sunflower did as an honest mistake, though I’m not familiar enough with Kansan to know if “continued monitoring of MCO claims” in “Kansas nice” translates to “fraud detection.”
KanCare’s “Circus Maximus”
Maximus is a publicly traded multinational based in — wouldn’t you know it — Reston, VA. They have quite a reputation in the UK (here, here), where they’re doing their little bit to help gut the NHS, and in the United States (Wisconsin). And they don’t seem to be doing very well by Kansas, either. First, there’s the $11 million Kansas is on the hook for from the Feds. Wichita Eagle March 30, 2018:
Kansas is suing over $11 million that it says a federal agency is improperly trying to take through an “egregiously hardline position.”
(Pause to note that the current Kansas administration is Republican.)
The lawsuit stems from a dispute over fees the Kansas Department for Children and Families paid into a state debt collection program run by the Kansas Department of Administration between 2003 and 2010. DCF typically seeks reimbursement from the federal government for a portion of the fees it pays into the program.
The Department of Administration each year files documents with the federal government about the federal funding the state plans to claim. Until 2013, the agency never included the revenue it retained from administering the debt collection program in the filings.
“During that prior time period, the KDOA had relied in good faith on a consultant” to help prepare the filings and “never once” did it include the debt collection revenue in its draft filings, the lawsuit says. The federal government asked about the debt collection revenue in 2012, the lawsuit says, and an audit began. Kansas worked with HHS to provide documentation, but the federal agency refused to negotiate a compromise, the lawsuit says.
There’s that honor system again.
The consultant, Maximus Consulting Services, is an arm of Maximus, which has had other contracts with Kansas throughout the years.
Oopsie. Then there are the problems with the “KanCare Clearinghouse.” Kansas City Star, February 22, 2018:
In 2015, Kansas moved to a new computer system [uh-oh] for applying for Kansas Medicaid, or KanCare. Then it funneled applications and annual reviews that used to be handled in regional offices into a single “KanCare Clearinghouse” in Topeka. It contracted with a company called Maximus to staff the Clearinghouse starting in 2016.
Since then the number of seniors covered by KanCare for in-home nursing help has gone down and so has the number being covered for nursing home beds. Meanwhile, the state’s population has been aging. Dan Goodman, the director of the Johnson County Area Agency on Aging, said it just doesn’t add up unless there’s a problem with the system.
“Some seniors are really having a tough time getting onto Medicaid,” Goodman said. “They get frustrated, are in poor or declining health, become defeated by the process and give up.”
Instead, Richardson said the past several months have been a nightmare, as they play phone tag with a revolving door of Maximus employees in Topeka, none of whom know her husband’s case and one of whom suggested she just go on welfare.
But the computer system roll-out was rocky and complaints about the Clearinghouse began almost immediately. A backlog of Medicaid applications pending past the 45-day federal limit quickly formed and has never been fully resolved. Nursing home administrators have repeatedly said it’s hurting them financially and the delays have caused some homes to limit the number of people with pending Medicaid applications they will take.
Funny how these bugs never give too many people health care, isn’t it? Here are the horrid numbers. Wichita Eagle February 16, 2018:
[Maximus] is so out of compliance with its contract that it could face fines of more than $250,000 a day.
Maximus is operating with 40 percent accuracy on financial payments; 98 percent is required. The company is also falling behind on handling applications and cases, the state of Kansas says.
Kansas has given the company until June 1 to shape up. Otherwise, Maximus will face retroactive fines back to January that could total tens of millions of dollars.
And an interesting quote from a Republican appointee, unintentionally explaining one reason #MedicareForAll is a superior solution to a body shop like Maximus:
Kansas’ Medicaid director on Friday described a sense of despair among workers during a recent visit to the Topeka [Clearinghouse] facility. “It’s night and day. It’s a completely different work environment on our side of the house and their side of the house and I think that’s because we at the state value the employees that work for us,” Kansas Medicaid Director Jon Hamdorf said.
As I keep saying, holy moly!
The Good News: Proposal for a Medicaid Lifetime Cap Rejected by CMS
Finally, good news from Kansas in Modern Health Care, Recall that KanCare depends on a Federal waiver, so Kansas has to get a CMS OK for changes:
The CMS will reject Kansas’ request to impose lifetime limits on Medicaid coverage, following an outcry from clinicians and others in the state. “We’re determined to make sure that Medicaid remains the safety net for those that need it most,” Verma said at the American Hospital Association’s annual meeting in Washington. “To that end, we have determined we will not approve Kansas’ recent request to place a lifetime limit on Medicaid benefits.”
Verma’s announcement also likely spells doom for Arizona’s similar pending request, which sought a five-year benefits limit. It also indicates that the Trump administration has a hard line on which conservative policies it will support.
Vox has an explainer:
The CMS decision to reject Kansas’s waiver is a very big deal: Lifetime limits have not gotten the same amount of attention as work requirements, but they would have signaled an equally important shift in Medicaid, away from an entitlement for all eligible Americans toward a shrinking program actively culling people from the rolls.
Even if work requirements are allowed, it matters that the administration is not going to let states to kick people off Medicaid simply because they have been on the program for a long time.
“I do think it is significant that CMS has decided to say no to one of the pending requests that cut people off of Medicaid,” Joan Alker, executive director at Georgetown University’s Center for Children and Families, told me.
Of course, Medicaid work requirements are truly crazed, but a lifetime limit would have gone beyond crazed to outright evil.
#MedicareForAll would make Amerigroup of Kansas Inc., Sunflower Health Plan, and United Healthcare go away, and also take away a big chunk of business from Maximus. That is all good. Amazingly, Kansas Medicaid Director Jon Hamdorf gives a very good reason why replacing privatized body shops with state employees would make for a better experience for citizens interacting with the heatlh care system. Kansas, as big as you think!
 They are not “officials.” MCOs are private entities. They are “employees.” Neoliberal brain damage at the word level!