AHLA W. Responding to CMS Overpayment Demands: Legal, Statistical, and Clinical Defense Strategies Christine N. Bachrach Vice President and Chief Compliance Officer University of Maryland Medical System Baltimore, MD Christopher Haney Duff & Phelps Chicago, IL Mary C. Malone Hancock Daniel Johnson & Nagle PC Richmond, VA Fraud and Compliance Forum October 6-7, 2014
Responding to CMS Overpayment Demands: Legal, Statistical, and Clinical Defense Strategies Christine Bachrach, VP & Chief Compliance Officer, University of Maryland Medical Center Baltimore, MD cbachrach@umm.edu Christopher Haney, CPA, CFE, Director, Dispute Consulting, Duff & Phelps, LLC Washington, DC Christopher.Haney@duffandphelps.com Mary Malone, Shareholder, Hancock, Daniel, Johnson & Nagle, PC Richmond, VA mmalone@hdjn.com Enforcement Environment, Recovery Programs, and Appeal Process 1
CMS Recovery Audit Program Recovery Auditors or Recovery Audit Contractors (RACs) Mission The Recovery Audit Program s mission is to identify and correct Medicare improper payments through the efficient detection and collection of overpayments made on claims of health care services provided to Medicare beneficiaries, and the identification of underpayments to providers so that the CMS can implement actions that will prevent future improper payments in all 50 states Medicare Fee For Service 3 Parameters for RACs Same Medicare policies as Medicare Administrative Contractors (MACs) National and Local Coverage Determinations and CMS Manuals Review Types Automated (no medical record needed) Semi Automated (claims review using data and potential human review of a medical record or other documentation) Complex (medical record required) 4 2
Parameters for RACs (cont d) Limits on number of medical records requests 3 year look back period based on date claim was paid Some demonstration projects on pre pay review IP medical necessity reviews on hold during implementation of two midnight rule CMS was awarding new contracts so all RAC activity ceased June 2014, but restarting in late August while court decides contract award dispute 5 CMS Recovery Statistics Overall Recoveries FY2013 $3.7 billion recovered Q1 3 2014 $2.2 billion year to date Appeals FY2012 (most recent) 26% of the claims appealed Success rate of appeals 47% per CMS for FY2011 claims 67% per American Hospital Association survey 6 3
Appeal Process Discussion Period from the time that organization receives letter until recoupment at 40 days Opportunity for provider to give RAC additional information AHA survey report 1/3 hospitals reversed some decisions in this period Redetermination by day 120 to MAC Reconsideration within 60 days of redetermination decision to Qualified Independent Contractor (QIC) Administrative Law Judge within 60 days of reconsideration decision Further to Medicare Appeals Council, US District Court 7 Legal Strategies for Combating Appeals 4
Legal Strategies Authority to Audit Limitation of Liability Provider without Fault Validity of Extrapolation 9 Authority to Audit Medicare reopening regulations allow for reopening in the following timeframes: (1) Within one year from the date of the initial determination or redetermination for any reason; (2) Within four years from the date of the initial determination or redetermination for good cause as defined in Section 405.986; or (3) At any time if there exists reliable evidence as defined in Section 405.902 that the initial determination was procured by fraud or similar fault as defined in Section 405.902. Generally, CMS contractors use the four year time period to reopen claims based on good cause. 10 5
Authority to Audit Good cause for reopening when: there is new and material evidence that (1) [w]as not available or known at the time of the determination or decision; and [m]ay result in a different conclusion; or (2) [t]he evidence that was considered in making the determination or decision clearly shows on its face that an obvious error was made at the time of the determination or decision. 42 CFR 405.986. CMS contractor s decision to reopen a claim not subject to appeal, but can be discussed during rebuttal or discussion period. 11 Authority to Audit RACs must comply with CMS Statement of Work. The RAC Statement of Work forbids RACs from attempting to identify overpayments more than three years past the date of the initial determination made on the claim. The first step in responding to a RAC records request should be to confirm that all of the claims under review are within the three year window allowed by the Statement of Work. If the claims are more than three years old, the provider can challenge (in writing) the RAC s authority to select the claims. 12 6
Limitation of Liability Section 1879 of the Social Security Act ( SSA ) states: Where (1) a determination is made that, by reason of Section 1862(a) (1) or (9)... payment may not be made under Part A or Part B of this title for any expenses incurred for items or services furnished an individual by a provider of services or by another person pursuant to an assignment under section 1842 (b) (3) (B) (ii); and (2) both such individual and such provider of services or such other person, as the case may be, did not know, and could not reasonably have expected to know, that payment would not be made for such items or services under Part A or B, then to the extent permitted by this title, payment shall, notwithstanding such determination, be made for such items or services... as though section 1862(a) (1) and section 1862 (a) (9) did not apply... 42 U.S.C. 1395pp. 13 Limitation of Liability What does this mean? The limitation of liability provision allows a provider subject to post payment review to argue that even though the contractor may have identified an overpayment, the provider should retain its payment because the provider did not know, and could not reasonably be expected to know, that the payment would later be denied. The limitation on liability defense is best used in cases where there is little or no specific guidance on the Medicare coverage or documentation requirements for a specific service. 14 7
Limitation of Liability Case Study Baptist Health Care, 2009 WL 5764270 (H.H.S. June 26, 2009) Medicare Appeals Council case Hospital challenged a RAC s determination that the hospital was inappropriately paid for inpatient stays related to automatic implantable cardioverter defibrillator ( AICD ) procedures in 2003, 2005, and 2006. During these years, the AICD procedures were included on Medicare s inpatient only list. 15 Limitation of Liability Case Study (cont.) The Florida QIO, Florida Medical Quality Assurance, Inc., ( FMQAI ) issued a policy in 2006 that stated it would no longer uniformly allow AICD procedures to be billed as inpatient services, but that it would consider denying routine cases with an anticipated discharge within 24 hours. For services provided before FQMAI issued its policy, however, the Council agreed with the hospital s limitation on liability defense. The Council overturned the RAC s overpayment determination, ruling that until FMQAI issued its policy, the hospital did not know, and did not have reason to know, that Medicare would not cover automatic implantable cardioverter defibrillators services provided in an inpatient hospital setting. 16 8
Limitation of Liability For purposes of determining the provider s knowledge actual or constructive an ALJ or the Council look for the following evidence: A Medicare contractor s prior denial of payment for similar or reasonably comparable services; Medicare s general notices to the medical community of Medicare payment denial of services under all or certain circumstances, including manual instructions, bulletins, contractor s written guides and directives; Provision of services inconsistent with acceptable standards of practice in the local medical community; The provider s utilization review committee informed the provider in writing that the services were not covered; or A Medicare contractor previously issued a written notice to the provider that Medicare payment for a particular service or item is denied. This also includes notification of Quality Improvement Organization ( QIO ) screening criteria specific to the condition of the beneficiary for whom the furnished services are at issue and of medical procedures subject to preadmission review by the QIO. 17 Provider without Fault Section 1870(b) of the SSA states that a provider is deemed to be without fault (absent evidence to the contrary) with respect to an overpayment if the overpayment determination is made subsequent to the fifth year following the year of the initial determination. If a provider was paid for services in 2008 (or earlier) and has retained the payment until 2014, the provider is now deemed to be without fault with respect to any overpayments, unless there is evidence to the contrary. 18 9
Provider without Fault Medicare manual guidance states that a provider is considered to be without fault if it exercised reasonable care in billing for, and accepting, the payment, i.e.: It made full disclosure of all material facts; and On the basis of the information available to it, including, but not limited to, the Medicare instructions and regulations, it had a reasonable basis for assuming that the payment was correct, or, if it had reason to question the payment; it promptly brought the question to the Fiscal Intermediary or carrier s attention. 19 Provider without Fault Case Study Comprehensive Decubitis Therapy, Inc., 2009 WL 5788648 (H.H.S. Mar. 13, 2009) Medicare Appeals Council case Facts: A supplier argued that it was without fault with respect to payments it received for supplies that should have been bundled into the beneficiary s home health benefit. 20 10
Provider without Fault Case Study (cont.) Comprehensive Decubitis Therapy, Inc., 2009 WL 5788648 (H.H.S. Mar. 13, 2009) In finding that the supplier was without fault, the Council determined that the following facts demonstrated that the supplier exercised reasonable care in billing for the services: the physician s order for the supplies indicated that the beneficiary was not receiving home health; a progress note stated that the beneficiary had discontinued home health; and, the supplier obtained pre authorization from a Medicare contractor that did not indicate that the beneficiary was receiving home health. Importantly, in this case, the presumptive without fault time period had not elapsed, but the provider was able to show it was without fault through strong documentation. 21 Challenging Use of Extrapolation Section 1893(f)(3) of the SSA states that a Medicare contractor may not use extrapolation to determine overpayment amounts to be recovered... unless the Secretary determines there is a sustained or high level of payment error or documented educational intervention has failed to correct the payment error. In order for a provider to successfully challenge an extrapolated overpayment, it must challenge the statistical analysis supporting the extrapolation, and disprove the auditor s determination on underlying sample claims. 22 11
Challenging Use of Extrapolation Use the facts and circumstances of the audit to argue that extrapolation is unfairly punitive. For example, for documentation issues where a physician can demonstrate when a fix was made, an auditor may agree to drop the extrapolation and recover only the actual overpayment amount. 23 Statistical Defense Strategies 12
Statistical Defenses: Background Complex audits typically involve sampling and extrapolation, and they necessitate a high degree of statistical competency to ensure claims are properly estimated. Medicare Program Integrity Manual ( MPIM ) includes guidelines on data analysis, statistical sampling, extrapolation, and estimation of overpayments. Successful appeals often highlight a contractor s nonconformance with MPIM, including the following: Generally Accepted Standards Sample Size Precision and Confidence Representativeness and Randomness 25 Sampling Terminology Extrapolation: Projecting the results of your sample onto the entire population. Observed ratios: o Failure rate of an audit o Proportion of red M&Ms Observed descriptive statistics: o Mean overpayment per claim o Mean household income Results of extrapolation within a specified level of significance. Different sample sizes yield results with different levels of significance Confidence level (i.e. 95%, 99%, etc.) Margin of error or precision level (i.e. ±3 percentage points) e.g. Candidate X is expected to receive 47% of votes, ±2 percentage points, at a 90% confidence level 26 13
Statistically Valid Random Sample MPIM Guidance: If a particular probability sample design is properly executed, i.e., defining the universe, the frame, and the sampling units; using proper randomization; accurately measuring the variables of interest; and using the correct formulas for estimation, then assertions that the sample and its resulting estimates are not statistically valid cannot legitimately be made. 27 Burden of Proof Relevant CMS Requirement: The burden of proof is on the appellant to prove a contractor s statistical sampling methodology invalid, not on the contractor to establish that it chose the most precise methodology. Significance Federal courts are reluctant to overturn a contractor s methodology without explicit evidence of errors. Relevant Case: Border Ambulance Service, LLC v. TrailBlazer Health Enterprises [a]ppellant s challenges to the sample are not based on demonstrable errors in the sample or reference to specific supporting evidence in the record The appellant s speculative assertions do not satisfy its burden of proving that the statistical sampling methodology at issue is invalid. Disputing a RAC s methodology is not enough for successful appeal. Appellants must prove that a contractor s statistical analysis is invalid. 28 14
Generally Accepted Standards Relevant MPIM Requirement: Any sampling methodology must be reviewed by a statistician or a person with equivalent expertise in probability sampling and estimation methods. Significance A great deal of subjectivity exists when implementing statistical procedures. Several appeal decisions have concluded that no generally accepted standards exist for the application of sampling. Relevant Case: Michael King, M.D. and Kinston Medical Specialists v. Cigna [w]hile there may well be theories on the right way to conduct a sample, there is no formal recognition of generally accepted statistical principles and procedures. No recognized statistical principles exist. When appealing statistical methodology, show how the method impacts conclusions. 29 Sample Size Relevant Requirements: There is no CMS standard for minimum sample size. However, OIG s Provider Self Disclosure Protocol ( SDP ) requires a sample size of at least 100 claims. This does not apply to CMS contractors for recovery audits, but may serve as an indication of reasonable sample sizes. Significance Federal courts are reluctant to overturn a contractor s methodology based on insufficient sample size since no standard exists. Relevant Cases: Ratanasen v. State of California, Webb v. Shalala, and Pruchniewski v. Leavitt Courts have ruled that no minimum sample size exists. No generally accepted minimum sample size exists. Appeals should demonstrate how a larger sample size would impact conclusions. 30 15
Precision and Confidence Relevant Requirements: No CMS standard exists for minimum precision levels of extrapolation. However, per MPIM, In most situations the lower limit of a one sided 90 percent confidence interval shall be used as the amount of overpayment to be demanded for recovery. Significance Federal courts are reluctant to overturn a contractor s methodology based on insufficient levels of precision since no standard exists. Relevant Case: Pruchniewski v. Leavitt Because there is no established standard of precision providers must go further and establish that the degree of imprecision is such that the extrapolation does not reasonably approach the actual overpayment No generally accepted minimum precision or confidence level exists. Appeals should demonstrate how a more precise analysis would impact conclusions. 31 Random and Representative Relevant Requirements: Valid samples must be selected using proper randomization to minimize selection bias, and should be representative of the total universe. Significance Analysis may improperly sample only certain subsets of a population and attempt to extrapolate conclusions across the entire universe. i.e. disproportionate samples of high dollar claims or a focus on one particular facility or provider. Software such as RAT STATS are considered a reliable means of selecting a sample, however these programs are only as effective as their operator. Relevant Case: John Sander, M.D. v. CIGNA either samples themselves were not drawn correctly or the claims were not correctly assigned to the correct stratum Sample design and selection are imperative to reliable results. RAT STATS can be effective, but users need to understand the broader strategy. 32 16
RAT STATS Statistical Software RAT STATS is statistical software developed by the government Free and available online along with 394 page user guide and 245 page companion manual. Practically, RAT STATS is a calculator with three main functions: Calculating sample size Generating random numbers to aid sample selection Extrapolating (projecting) results of the sample to a broader population RAT STATS is used in conjunction with a broader strategy * * * Sampling Plan * These steps involve using RAT-STATS 33 Parting Tips for Statistical Appeals Evaluate a contractor s sampling plan for soundness and completeness. Ensure sufficient documentation exists. Assess sample size and levels of confidence and precision. Evaluate sample design and weigh the benefit of other designs. Consider underlying causes of bias within the sample design. Evaluate the methods of randomization and sample selection. If you identify areas of weakness in the contractor s analysis: Identify issues that may render the analysis invalid (sample too small, etc.) Conduct your own audit with improved inputs (larger sample size, etc.) Demonstrate contractor s demand does not equal actual overpayments Meet your burden of proof by demonstrating their analysis is invalid 34 17
Elements of a Sampling Plan Population of Interest (POI) This can help you prepare your request for data Sampling Unit Population of interest is composed of all possible sampling units Sampling Frame Population from which sample is drawn (explain if not POI) Sample Size Minimum or any other procedural requirements/thresholds Required Precision and Confidence possibly 95% confidence ±2% precision Sample Design Simple, Stratified, Clustered, etc. Specify strata or cluster criteria Source of Random Numbers often RAT STATS Method of Selecting Sampling Units Ensure randomization without bias Procedures for Missing Data Typically failures, however may use spares Estimation Methodology Also referred to as extrapolation methodology 35 QUESTIONS Christine Bachrach, VP & Chief Compliance Officer, University of Maryland Medical Center Baltimore, MD cbachrach@umm.edu Christopher Haney, CPA, CFE, Director, Dispute Consulting, Duff & Phelps, LLC Washington, DC Christopher.Haney@duffandphelps.com Mary Malone, Shareholder, Hancock, Daniel, Johnson & Nagle, PC Richmond, VA mmalone@hdjn.com 18