OMB to agencies: Bump up IT use to get control of payments

UPFRONT CHECK: To rein in Medicare overpayments, CMS does automated data scans, Melanie Combs says.

Rick Steele

On any given workday, common payment processing mistakes cost the government a fortune in improper payments'adding up to $45.1 billion last year alone.

Complicated rules, relentless volume and layers of contractors using aging systems result in agencies distributing benefits payments to ineligible applicants, overpaying beneficiaries and sending duplicate payments, said Melanie Combs, technical adviser for the Program Integrity Group at the Centers for Medicare and Medicaid Services.

And it's a growing problem.

Improper payments governmentwide grew $10.1 billion last year, up from $35 billion in fiscal 2003, according to the Office of Management and Budget. The problem is at the heart of Combs' work at CMS because the Medicare program accounts for almost half of annual improper payments.

With pressure from the Bush administration to staunch the flow of errant money from federal coffers, agencies are turning to IT, particularly matching and analysis applications that can help agencies put a halt to the problem before a check is ever disbursed.

'Utilizing IT is the core of our strategy to reduce improper payments,' said Fritz Streckewald, assistant deputy commissioner of the Social Security Administration for program and policy. 'If we can invest a dollar in systems development, we may be able to save $10 in improper payments.'

At Social Security, the agency has found that it pays to keep tabs on beneficiaries. To avoid wrongful payments of its supplemental security income, a means-tested benefit paid to low-income elderly and disabled persons, SSA shares applicant data with the Veterans Affairs Department and the Health and Human Services Department's Office of Child Support Enforcement.

The agency reduced improper payments for the benefits program by more than $100 million in the last year, according to OMB, which has been pushing agencies to find ways to tackle the problems through its new Eliminating Improper Payments Program.

At SSA, 'efforts that began in January 2001 are yielding results, as evidenced by using online queries to access the Office of Child Support Enforcement's quarterly wage data and new hires directory,' OMB noted in a recent report, Improving the Accuracy and Integrity of Federal Payments.

Based on reporting required for the first time last year under the Improper Payments Information Act of 2002, OMB found that 92 percent of improper payments are overpayments.

The biggest culprit: common mistakes by state and federal workers, contractors and medical providers. To a lesser degree, fraud adds to the problem, Combs said.

'Based on the volume of disbursements that are made, you just can't do this manually, so IT is vital,' said McCoy Williams, director of financial management and assurance for the Government Accountability Office.

Complex search

For Medicare, CMS' use of contractors adds another layer of complexity. It relies on its claims processing contractors, such as Blue Cross/Blue Shield of Maryland, to handle claims submitted by doctors, hospitals and other health care providers.

'They primarily are on the front lines of looking at claims and trying to identify which ones to pay and which not to pay,' Combs said.

CMS owns the legacy claims processing systems the contractors use and which run on IBM Corp. mainframes. Most of the code is written in Cobol and use IBM's CICS transaction processing applications.

To try to rein in the improper payments, CMS has turned to its contractors for help. By adding on apps, the contractors can set up automatic scans that compare the contents of fields on the claim forms submitted by doctors and hospitals.

By comparing the information from two different fields on a claim, for example, the automated scans can catch errors and fraud attempts. For instance, a scan would flag a form seeking payment for a male patient receiving a hysterectomy. In that simple example, the provider obviously made a data entry mistake, so the system would automatically deny the claim and return to the doctor for correction.

Other applications look for problems across a series of claims, Combs said. These apps set up parameters for a time period and establish what number of claims would be reasonable. Claims exceeding the parameter are flagged for further review.

But IT alone can't solve all of CMS' errant payment problems, Combs said. Some claims fall in a gray area that require a nurse or clinician to review the medical record and then decide if a claim is correctly coded and whether it is covered in accordance with all the Medicare rules, she said.

CMS' goal is to reduce the dollars it pays out incorrectly to 4.7 percent of all payments by 2008; last year, that figure was 10.1 percent of Medicare payments.

Using IT helps immensely, Combs said, but creating software that can account for Medicare's complex rules is no easy task. Every medical service has codes and variants, plus national and local coverage rules add another layer of complexity.

To help with the coding, AdminaStar Federal of Indianapolis, which manages the Medicare contractors for CMS, developed the Correct Coding Initiative.

AdminaStar evaluates all coding instructions and other Medicare rules quarterly and updates its application with the most appropriate automatic edits to help contractors determine if a claim is covered.

'It is a daunting task and always changing,' Combs said.

To back up its contractors, CMS next plans to pilot the use of recovery audit contractors to identify and collect Medicare overpayments that claims processors missed. CMS last week awarded contracts to conduct demonstrations in three states.

GAO's Williams said other agencies, such as CMS and SSA, are increasingly using automated checks to ferret out mistakes in benefits processing and sharing data to validate information about program participants.

'If you qualify for one program, such as the Housing and Urban Development's Section 8 housing, you might qualify for another, such as Food Stamps'and vice versa,' he said.

Data exchange

At Social Security, data exchanges are the chief method used to spot problems, Streckewald said.

'Paying benefits correctly and avoiding improper payments is all about having access to good information on a timely basis,' he said.

SSA sets up written agreements with agencies that it wants to swap records with. The automatic data exchanges generally occur on a monthly or quarterly basis.

The agencies for the most part use computer-to-computer interfaces for immediate matching of records, al-though agencies also still exchange data on magnetic tape sometimes, Streckewald said.

Social Security for instance receives data from VA about payments to SSI participants.

If a particular recipient's information does not jibe with SSA records, the system generates an alert that is sent to a Social Security field office for follow-up.

Social Security also has crafted software that automatically reviews files in its SSI Master Record and then assigns scores to beneficiaries' cases based on unverified income and other data deemed questionable.

For some agencies, the programs cut across levels of government.

The Labor Department working with the Health and Human Services Department and three states has established a National Directory of New Hires database. The pilot system lets states check if an individual with a job is still collecting unemployment.

Through the unemployment insurance program, Labor provides benefits to eligible workers who become unemployed. The department reported a 10.3 percent improper payment rate last year for the unemployment insurance program.

Labor is testing its new hire database with Texas, Utah and Virginia. The department expects that its use will generate reductions in improper payments by October, Labor CFO Samuel Mok said. Labor anticipates reducing mispayments by $74 million this year and $259 million by 2007.

To further identify problems before unemployment payments are made, rather than after, the department also has a data exchange agreement with Social Security that lets states verify each claimant's identity and Social Security numbers when someone applies for benefits, Mok said. It has long been a problem that these checks were done after states had already paid out benefits erroneously, he said.

Finding the problems up front is key, OMB noted. Obviously, the chief target is Medicare. And on the priority list this year for the program:

'Developing new data analysis procedures to help identify payment aberrancies and using that information in order to stop improper payments before they occur.'

Reader Comments

Please post your comments here. Comments are moderated, so they may not appear immediately after submitting. We will not post comments that we consider abusive or off-topic.

Please type the letters/numbers you see above