Providing Meaningful Oversight of Risk Adjustment Programs. February 1, 2016

Transcription

1 Providing Meaningful Oversight of Risk Adjustment Programs February 1, 2016 Presented by: Richard Lieberman Chief Data Scientist TODAY S AGENDA A brief history of Medicare-Advantage rate-setting and risk adjustment The process of revenue management- medical record review, in-home assessments, etc. The role compliance officers should play in overseeing both internal operations and vendors The False Claims Act and several open cases headed to trials on the merits 1

2 WHO ARE WE? Mile High Healthcare Analytics is a recently founded population health analytics enterprise. Richard Lieberman is the Chief Data Scientist who has been actively involved in the development of risk adjustment and quality measurement systems for over 20 years Duke Owen, an expert in the NCQA and PQA quality measure certification process, among other talents! Eric Olmsted is a Senior Consultant with 15 years of experience designing, implementing and evaluating total population health management programs HOW DID WE GET HERE? Prior to the Balanced Budget Act of 1997, Medicare paid plans 95% of average traditional Medicare costs in each county HMOs were thought to be able to provide care more efficiently than could be provided in traditional Medicare. These payments were not adjusted for health status, and HMOs typically enrolled beneficiaries who were healthier than average. 2

3 THE BALANCED BUDGET ACT (BBA) OF 1997 The BBA of 1997 established a payment floor, applicable almost exclusively to rural counties. The Benefits Improvement and Protection Act (BIPA) of 2000 created payment floors for urban areas and increased the floor for rural areas. BBA-97 set in motion what would have become the destruction of the Medicare managed care industry It limited year-over-year increases in the county base rates (the maximum amount a plan could be paid) to 2 percent in most counties THE BALANCED BUDGET ACT (BBA) OF 1997 Health plans perceived they could not earn sufficient returns, with many exiting the market In 1997, there were over 7 million Medicare beneficiaries enrolled in plans; by 2003 the industry had shrunk to 5.3 million enrollees 3

4 THE MEDICARE MODERNIZATION ACT OF 2003 In December 2003, President Bush and Congressional Republicans passed the Medicare Modernization Act (MMA) Not a single Democrat voted yes for the legislation The House of Representatives employed parliamentary shenanigans to get the bill passed The MMA created the Part D prescription drug benefit (which everyone, including Democrats, now love!) MMA also changed the way Medicare-Advantage (MA) plans had to justify their capitation rates each year Beginning in 2006, MA plans had to submit competitive bids COMPETITIVE BIDDING COMES TO MEDICARE-ADVANTAGE Since 2006, MA Plans have bid to participate in Medicare- Advantage Plans prepare a set of utilization and cost estimates by type of service that resolve to a pmpm rate Cost estimation uses a combination of experience and manual rates, trended to the subsequent contract year The cost estimates are adjusted for each county the Plan is authorized to service The cost estimates comprise what is known as the Standardized Part A/B bid Standardized refers to a risk score 1.0 4

5 WHAT IS THE BENCHMARK? Each year (the first Monday in April), CMS publishes a rate book The rate book contains the maximum amount that it will pay an MA Plan in each county of the United States These county-specific rates are known as Standardized Part A/B benchmark rates Each county rates assumes that the population has a risk score of 1.0 Since 2012, star ratings are integrated into the rate book BID VS. BENCHMARK Bids are submitted by MA Plans to CMS on the first Monday in June If the bid is less than the benchmark, the MA Plan does not have to charge a supplemental premium If bid exceeds the benchmark, a supplemental premium must be charged Impact of Star-ratings: Contracts at the 4-Star level or higher bid against contracts with similar Star ratings; contracts with less than 3- Stars bid against lower-quality contracts 5

6 IMPACT OF THE AFFORDABLE CARE ACT The ACA of 2010 revised the methodology for paying plans and reduced the benchmarks For 2011, benchmarks were frozen at 2010 levels Reductions in benchmarks will be phased-in over 2 to 6 years between 2012 and 2016 By 2017, when the new benchmarks are fully phased-in, the benchmarks will range from 95% of traditional Medicare costs in the top quartile of counties with relatively high per capita Medicare costs (e.g., Miami-Dade), to 115% of traditional Medicare costs in the bottom quartile of counties with relatively low Medicare costs (e.g., Boise). The ACA also reduced rebates for all plans, but allowed plans with higher quality ratings to keep a larger share of the rebate than plans with lower quality ratings. 6

7 COMPREHENSIVE RISK ADJUSTMENT The Medicare Modernization Act of 2003 (MMA) ushered in comprehensive risk adjustment for Medicare-Advantage plans 4-year phase-in began in 2004 Medicare was not first to the party with risk adjustment For most of the first decade of the 21 st century, many plans did not pay that much attention to risk adjustment Many other bigger fish to fry: competitive bidding, Part D rollout, new plans entering the market, private-fee-for-service, etc. There was more than enough money flowing into Medicare- Advantage plans by way of the county level benchmarks RISK ADJUSTMENT GOES MAINSTREAM Limiting the annual growth of MA benchmarks by the ACA has led to: Plans must rely on risk adjustment and quality measurement to cover the cost of medical price inflation and to maintain profit margins The advent of risk adjustment in the Exchanges has dramatically raised awareness (and mass confusion) Approximately 24 states risk adjust their Medicaid managed care programs 7

8 AREAS OF RISK FOR MA PLANS Most vendor-supplied services are black-box MA Plans typically accept this arrangement Most vendors have very limited analytical capabilities Inside of health plans, there is very limited knowledge of the nuances of risk adjustment Most risk adjustment managers have segued from finance positions and have learned risk adjustment on the job. You don t know what you don t know! Vendors typically know very little about risk adjustment process and its complex calculations WHAT SHOULD COMPLIANCE OFFICERS DO? Medical record review, in-home assessments, and other revenue management programs are a necessary component of MA Plan functions Risk adjustment entitles plans to be paid for the illness burden of the members they enroll Demand that vendors have the capacity to calculate member-level risk scores at the beginning, during, and after a revenue management intervention Most vendors just deliver the data and the plans typically accept this as sufficient 8

9 GOLD STANDARD REVIEW OF CODING Most MA Plan contracts with coding vendors require certified medical record coders Coding certification requirements are a very low bar AHIMA certifications are more comprehensive then AAPC certifications What counts is the coders experience and the amount of oversight the coder has received Plans need to assess the degree to which coding vendors: Train their coders Perform quality assurance review of coding on an ongoing basis Inter-rater reliability testing GOLD STANDARD (CONT D) Plans need to conduct their own quality assurance coding reviews Use a small sample of charts Hire in-house coders or contract with an independent firm Is the data coming back from the vendor subjected to consistency and quality checks? 9

10 SOURCE OF DIAGNOSIS CODES Since the inception of comprehensive risk adjustment in 2004, the sole source of diagnosis codes were RAPS files RAPS are a very limited extract from claims data: five fields, with filtering the responsibility of the plan (or its vendor) CMS has threatened to sunset the RAPS submission process for several years EDPS data has been collected since 2012, but is fraught with concerns about its completeness For 2015, diagnosis codes from EDPS were added to RAPS data, using the two sources in tandem (with no preference for one over the other) 10

11 SOURCE OF DIAGNOSIS CODES: EDPS VS. RAPS Beginning in 2016, CMS proposes to calculate risk scores by blending 90 percent of the score from RAPS with 10 percent of the score from EDPS Deficiencies in the EDPS data stream will, for the first time, alter a contract s aggregate risk score Phasing out RAPS transfers the responsibility for diagnosis code filtering from plans and vendors to CMS FILTERING OF DIAGNOSIS CODES IN RAPS AND EDPS! RAPS diagnosis codes filtering has always been problematic There are frequent occurrences of mass deletes (or large number of additions) right before the final reconciliation deadline in January of each year CMS specifications say that diagnosis codes incorporated into risk scores are supposed to be from face-to-face visits only But the primary guidance published by CMS for Medicare-Advantage uses provider specialty codes to identify face-to-face visits Many filtering algorithms are black-box, either because they are vendor solutions or within plans they are handled by IT departments 22 11

12 RAPS DATA SUBMISSION RULES ARE MURKY The RAPS data submission process requires the plans (or vendors) to filter out diagnosis codes that did not emanate from a face-to-face visit CMS guidance uses provider specialty as the primary filtering tool This approach works poorly and results in many inappropriately filtered diagnoses The process is even worse for institutional claims The best way to filter is to use procedure codes Mimic published CMS logic for either MA Plans or Exchange issuers FILTERING FOR SUBMISSION TO THE EDGE SERVERS Conceptually, the rules for submitting diagnosis codes for commercial risk adjustment are identical to the concepts used by Medicare-Advantage Face-to-Face visits Service directly rendered by a physician or mid-level practitioner Diagnosis codes from inpatient episodes of care enjoy blanket inclusion Outpatient facility encounters rely on both bill type codes and the CPT/HCPCS codes CMS has published the comprehensive filtering rules for commercial risk adjustment since

13 MEDICARE-ADVANTAGE FILTERING FOR RAPS SUBMISSION In contrast to the specifications that exist for the EDGE server submission process, no such comparable detailed specifications existed for Medicare- Advantage Until now! The filtering specifications are similar, but not identical to the EDGE server specifications 25 DIAGNOSIS CODE FILTERING FOR EDPS On July 21, 2015, CMS issued a, Draft Encounter Data Diagnosis Filtering Logic memo Industry comments were sought through August 21, 2015 For the most part, the draft filtering specifications for EDPS are conceptually very similar to the commercial exchange paradigm But it is not identical! There are some Medicare-specific bill types (e.g. Part B only) because of the high frequency use of observation stays Provider specialty is no longer part of the filtering paradigm 26 13

14 WHAT SHOULD MA PLANS BE FOCUSING ON ICD-10 should be a concern, but there are bigger financial risks that can have immediate negative impacts on CY2015 revenue Audit your diagnosis code filtering algorithm! Use the July 2015 draft guidance on EDPS filtering to determine if RAPS filtering is being done correctly This can impact CY2015 revenue, even though EDPS will only impact CY2016 risk scores It is essential for plans to start using EDPS data for operational purposes The MAO-004 reports can be used to validate the filtering algorithm 27 PREPARING FOR THE MIGRATION TO EDPS Use the published specifications (even in draft) to audit how diagnosis codes are being submitted to RAPS Remember RAPS still accounts for 90 percent of CY2016 payments Suboptimal filtering even impacts CY2015 payments 10 percent of risk-adjusted payments will be determined in 2016 by EDPS 28 14

15 Ensuring you are being paid accurately by both RAPS and EDPS data streams requires more than just comparing diagnosis code streams Is every rendered service (even capitated encounters) incorporated into EDPS? If not, then when CMS uses EDPS to recalibrate the CMS- HCC models, the industry won t be happy! MIGRATING TO EDPS 29 FALSE CLAIMS ACT The False Claims Act, which has been dramatically expanded under the ACA Overpayments now have to be reported to HHS within sixty days of detection Elements of a False Claims Act violation: defendant makes a false statement or engages in a fraudulent course of conduct do so with the required scienter (intent or knowledge of wrongdoing) the statement or course of conduct is material the statement or course of conduct caused the government to pay out money of forfeit moneys due 15

16 TWO ACTIVE COURT CASES U.S. v. Isaac Kojo Anakwah Thompson U.S. District Court for the Southern District of Florida ( CR-ZLOCK/HUNT) Criminal fraud case Olivia Graves, on behalf of herself and the U.S. Humana is defendant in this case U.S. District Court for the Southern District of Florida ( CIV-MORENO) False Claims Act case DEPARTMENT OF JUSTICE IS TRYING OUT A FRAUD THEORY AGAINST ALLEGED RISK ADJUSTMENT VIOLATIONS 16

17 TRY TYPING MEDICARE ADVANTAGE WHISTLEBLOWER INTO GOOGLE WHAT SHOULD MA PLANS DO? Many of the alleged whistleblower lawsuits will be bogus. But not all of them! Get Your Head Out of the Sand Providers need real training on optimal clinical documentation Not coding! Documentation! Providers need oversight Just because the contract transfers risk, doesn t mean it transfers responsibility Risk adjustment is hard and has lots of nuances Medicaid and exchange risk adjustment are not just like Medicare! 17

18 FREE MONTHLY WEBINAR SERIES Mile High Healthcare Analytics will continue to present key risk adjustment and performance improvement topics to health plans and provider groups on a monthly basis throughout the year. Our next webinar will be held on Thursday February 11, 2016 at 2 p.m. Eastern time. This webinar will take a deeper dive into CMS recent proposal for Medicare-Advantage RACs and other oversight topics Register at: CONTACT INFORMATION Richard Lieberman rlieberman@healthcareanalytics.expert or rlieberman@milehighhealthcareanalytics.com (voice) 18

19 THANK YOU FOR JOINING US!! Our website continues to evolve. Please visit us at: 19