April 2012 Volume 8 Number 8

Predictive Modeling: The New Frontier in Medicare Claims Review

By Cybil G. Roehrenbeck, American Medical Association, Washington, DC1

AuthorStatutory Basis

Section 4241 of the Small Business Jobs Act of 20102 authorizes the Secretary of the Department of Health and Human Services ("HHS") to use predictive modeling and other analytics technologies to identify improper claims for reimbursement and to prevent the payment of such claims under the Medicare fee-for-service program. The statute provides that the predictive modeling program should capture Medicare provider and beneficiary activities across the Medicare program to provide a comprehensive view of all providers, beneficiaries, and geographies within the program. The goal is to identify and analyze provider networks, provider billing patterns, and beneficiary utilization patterns, and to use that information to detect patterns and networks that represent a high risk of fraudulent activity.

The statute requires HHS to: (1) analyze large data sets for unusual or suspicious patterns or anomalies or contain other factors that are linked to the occurrence of waste, fraud, or abuse and undertake such analysis before payment is made; (2) prioritize such identified transactions for additional review before payment is made in terms of the likelihood of potential waste, fraud, and abuse to more efficiently utilize investigative resources; (3) capture outcome information on adjudicated claims for reimbursement to allow for refinement and enhancement of the predictive analytics technologies on the basis of such outcome information, including post-payment information about the eventual status of a claim; and (4) prevent the payment of claims for reimbursement that have been identified as potentially wasteful, fraudulent, or abusive until such time as the claims have been verified as valid. The statute also requires that the predictive modeling program be integrated into the existing Medicare fee-for-service program claims flow with minimal effort and maximum efficiency.3

Congress appropriated $100 million for the program, from a period beginning January 1, 2011, until expended.4 Further, the statute permits the Secretary of HHS to waive such provisions of titles XI, XVIII, XIX, and XXI of the Social Security Act, including applicable prompt payment requirements under titles XVIII and XIX of the Act, as the Secretary determines to be appropriate to carry out the program.5

Early Stakeholder Concerns

Stakeholders expressed concerns early on regarding the implementation of the predictive modeling program. In particular, the American Medical Association (“AMA”) expressed that the program should be free from false positives or inaccurate results before CMS begins to deny payment based on its findings.6 The AMA also urged CMS not to waive prompt payment in furtherance of the program, noting that numerous errors and glitches were likely to ensue as the program developed, and timely provider reimbursement should not be subject to the program’s growing pains. There is also a concern about the need for clinical input to inform the development of the program’s algorithms, which are unknown to the public. Providers fear that the complexity of claims review may be a challenge for the program, and that instant reviews—such as those that credit card companies conduct to identify fraud—may not be the best fit for complex clinical determinations, as such determinations require specific medical expertise.

While HHS’ Centers for Medicare & Medicaid Services (“CMS”) has made clear that the program is currently limited to fraud identification, the possibility that the program may be used for identification of improper payments unrelated to fraud—or waste—looms large in the provider community. As the Administration embarks on its Campaign to Cut Waste7 and new CMS demonstration programs to identify improper payments are picking up speed,8 the predictive modeling program appears to be a resource that CMS may wish to utilize to identify improper payments unrelated to fraud. However, groups like the AMA have cautioned that many issues that give rise to improper payments, or “vulnerabilities,” are recently-decided matters of policy, and are most productively addressed by provider education and outreach. Irrespective of the program, stakeholders assert, employing the program to identify improper payments unrelated to fraud could pose complicating problems for CMS and providers, including payment denial errors and lengthy, expensive appeals.9

Program Launch

CMS formally launched the predictive modeling program on June 17, 2011.10 Peter Budetti, MD, JD, Director of CMS’ Center for Program Integrity (“CPI”), testified regarding the program before the Senate Committee on Homeland Security and Governmental Affairs, Subcommittee on Federal Financial Management, Federal Services, and International Security, shortly thereafter.11 In CMS’ press release on the program, the agency likened the program to “technology used by credit card companies,” and stated that “predictive modeling helps identify potentially fraudulent Medicare claims on a nationwide basis, and helps stop fraudulent claims before they are paid.”12 CMS further noted that “this initiative builds on the new anti-fraud tools and resources provided by the Affordable Care Act that are helping move CMS beyond its former ‘pay & chase’ recovery operations to an approach that focuses on preventing fraud and abuse before payment is made.”13

On October 19, 2011, CMS issued an educational article on the program entitled Predictive Modeling Analysis of Medicare Claims.14 Importantly, that article verified that “currently, CMS is not denying claims solely based on the alerts generated by predictive models. CMS is focused on developing and refining models that identify unusual behavior without disrupting its claims processing for Medicare providers.”15 CMS also addressed providers’ concerns regarding the clinical complexity of claims review in part by noting that it is “working closely with clinical experts across the country and of every provider specialty…developing and refining algorithms that reflect the complexities of medical treatment and billing.”16 Lastly, CMS made the important provision that “prompt payment of claims is a statutory requirement; only in exceptional and urgent circumstances will CMS leverage its authority to waive prompt payment to conduct further investigation and review.”17 CMS has also indicated when the predictive modeling program indicates fraudulent action, that information will be given to Zone Program Integrity Contractors ("ZPICs") to investigate on the ground—a key component that will be monitored closely by stakeholders as the program proceeds.

Conclusion

To the extent that the predictive modeling program utilizes a targeted, streamlined approach to investigate fraudulent activity, rather than overly burdensome requirements for the majority of providers, it may be a positive development. And, because private insurers are also developing predictive modeling programs for claims review and risk assessment, the utilization of predictive models is likely to grow in significance and impact. As the Medicare predictive modeling program continues to evolve, providers should express their reservations regarding implementation, if any, to CMS and to the provider community at large.


1

The views expressed herein are the author’s own and are not necessarily those of the American Medical Association. Cybil G. Roehrenbeck is Washington Counsel for the American Medical Association. Cybil focuses on legal issues regarding healthcare delivery innovation and fraud and abuse compliance. She received her BA and JD from the University of Virginia and the University of Georgia, respectively. She may be reached at cybil.roehrenbeck@ama-assn.org.

2 42 U.S.C. 1320a-7m. Available at http://www.gpo.gov/fdsys/pkg/USCODE-2010-title42/pdf/USCODE-2010-title42-chap7-subchapXI-partA-sec1320a-7m.pdf.
3

42 U.S.C. 1320a-7m(b)(3).

4

42 U.S.C 1320a-7m(h)(1).

5

These titles of the Social Security Act cover General Provisions, Peer Review, and Administrative Simplification; Health Insurance for the Aged and Disabled; Grants to States for Medical Assistance Programs; State Children's Health Insurance Program; respectively.

6

American Medical Association. August 8, 2011 Letter to Donald Berwick, Administrator, Centers for Medicare & Medicaid Services regarding the predictive modeling program. Available at http://www.ama-assn.org/resources/doc/washington/predictive-modeling-letter-8aug2011.pdf.

7

The White House. Campaign to Cut Waste. Available at http://www.whitehouse.gov/goodgovernment/actions/campaign-cut-waste.

8

The Centers for Medicare & Medicaid Services recently announced three demonstration projects on curbing improper payments. Available at https://www.cms.gov/CERT/02_Demonstrations.asp.

9

American Medical Association. August 8, 2011 Letter to Donald Berwick, Administrator, Centers for Medicare & Medicaid Services regarding the predictive modeling program. Available at http://www.ama-assn.org/resources/doc/washington/predictive-modeling-letter-8aug2011.pdf.

10

Centers for Medicare & Medicaid Services. (June 17, 2011). New Technology to Help Fight Medicare Fraud [press release]. Available at http://www.cms.gov/apps/media/press/release.asp?Counter=3983.

11

Harnessing Technology and Innovation to Cut Waste and Curb Fraud in Federal Health Programs : Hearing before the United States Senate Committee on Homeland Security and Governmental Affairs, Subcommittee on Federal Financial Management, Government Information, Federal Services, and International Security (July 12, 2011) (Statement of Peter Budetti, MD, JD, Deputy Administrator and Director, Center for Program Integrity, CMS). Available at http://hsgac.senate.gov/public/index.cfm?FuseAction=Hearings.Hearing&Hearing_ID=fefef393-0b8f-4fca-9bd9-6298b5fda386

12

Centers for Medicare & Medicaid Services. (June 17, 2011). New Technology to Help Fight Medicare Fraud [press release]. Available at http://www.cms.gov/apps/media/press/release.asp?Counter=3983.

13

Id.

14

Available at http://www.cms.gov/mlnmattersarticles/2011mman/itemdetail.asp?itemid=CMS1253104.

15

Id.

16

Id.

17

Id.


The ABA Health eSource is distributed automatically to members of the ABA Health Law Section . Please feel free to forward it! Non-members may also sign up to receive the ABA Health eSource.