Session 98 L, Medicare Advantage Risk Adjustment: Past, Present and Future. Presenters: Adrian L. Clark, FSA, MAAA David Benjamin Koenig, FSA, MAAA

Transcription

1 Session 98 L, Medicare Advantage Risk Adjustment: Past, Present and Future Presenters: Adrian L. Clark, FSA, MAAA David Benjamin Koenig, FSA, MAAA SOA Antitrust Disclaimer SOA Presentation Disclaimer

2 Medicare Advantage Risk Adjustment: Past, Present and Future Adrian Clark, FSA, MAAA David Koenig, FSA, MAAA Spring 2017 SOA Health Meeting Tuesday June 13, 2017

3 Antitrust Disclaimer Attendees should be mindful of constraints of anti-trust or competition laws when engaging with each other at SOA hosted activities. This meeting is conducted for educational purposes, to provide a forum for the open discussion and free exchange of ideas even amongst companies that may otherwise compete in the marketplace. At any gatherings, members are reminded that it is their responsibility to avoid (and where they see others acting in such a manner, to discourage) any agreement or inappropriate conversations that could have anticompetitive effects or could constitute a restraint on trade. In discussing the pros and cons on a topic of discussion, attendees may express an opinion, but must avoid collective agreements on how any particular issue/matter should be handled 2

4 Additional Disclaimers The views expressed in this presentation are those of the presenters, and not those of Milliman or the SOA. Nothing in this presentation is intended to represent a professional opinion or be an interpretation of actuarial standards of practice. This presentation is intended solely for educational purposes and presents information of a general nature. It is not intended to guide or determine any specific individual situation and persons should consult qualified professionals before taking specific actions. Neither the speakers, the speakers employer, nor the SOA shall have any responsibility or liability to any person or entity with respect to damages alleged to have been caused directly or indirectly by the content of this presentation. 3

5 Agenda History of risk adjustment in Medicare RAPS to EDS transition Ongoing litigation Near term considerations Questions 4

6 History of Risk Adjustment in Medicare 5

7 MA Risk Adjustment Timeline Pre-2000: Adjusted Average per Capita Cost (AAPCC) Payments of 95% of FFS costs by county, adjusted for age, sex, Medicaid status, Inst status, and MSP status No adjustment for morbidity! : Principal Inpatient Diagnostic Cost Group (PIP-DCG) Used IP diagnoses to adjust for morbidity 2004: Original implementation of CMS-HCC model Plans/MedicareAdvtgSpecRateStats/downloads/evaluation_risk_adj_model_2011.pdf 6

8 CMS HCC Risk Adjustment Models Dx G803 Dx B381 Dx R532 Filtering Diagnosis information is filtered into HCCs Demographic info (age, sex, Medicaid status) All morbidity and demographic info is scored using coefficients that predict prospective claim cost Hierarchical Condition Categories (HCCs) Demographics Risk Score 7

9 CMS HCC Risk Adjustment Models Part C and Part D risk adjustment models are periodically updated Can include updates to: Diagnosis -> HCC mapping Coefficients applied to demographic and HCC information Addition of new models for certain populations Model changes produce winners and losers Most recently, CMS announced a new RxHCC model for PY

10 ICD-10 Using ICD-10 diagnosis codes (October 1, 2015 and on), correct specificity is required in many cases to qualify for risk adjustment With thoughtful coding/crosswalking, the ICD-9 to ICD-10 transition should not significantly affect risk adjustment With additional detail available in ICD-10, CMS may create a more predictive HCC model -published/icd-10-impact-provider.pdf 9

11 RAPS to EDS Transition 10

12 RAPS/EDS Comparison Risk adjustment processing system (RAPS) submissions are diagnosis codes that have been filtered for risk adjustment by MAOs Encounter data system (EDS) submissions require full claims detail to be submitted for CMS to filter for risk adjustment RAPS Data Beneficiary identifier (HIC) Date of service Diagnosis code Provider type (IP, OP, phys.) EDS Data Beneficiary identifier (HIC) Date of service Diagnosis code Monetary amounts Quantity Provider NPI Revenue codes Procedure codes Place of service Type of bill A lot more... 11

13 RAPS Data Flow RAPS files are created by MAOs MAOs filter all diagnosis codes based on CMS guidance RAPS files contain narrow set of data diagnosis clusters MAOs submit RAPS files to CMS CMS does basic review of RAPS diagnosis submissions, and provides a return file to the MAO 12

14 EDS Data Flow MAOs regularly submit full claims detail to CMS Number of Medicare enrollees Greater than 100,000 Minimum submission frequency Weekly 50, ,000 Bi-weekly Less than 50,000 Monthly Many data elements Initial edits/tests are applied by CMS (e.g. NCCI) Claims are filtered by CMS for inclusion in risk adjustment 13

15 Transition Timeline PY2015 All diagnosis data from RAPS and EDS was combined PY2016 and on Weighted blend of RAPS and EDS risk scores 100% 75% 50% 25% EDS RAPS 0% Rate Announcement, CMS 2017 Rate Announcement, CMS 14

16 Why Are MAOs Concerned About the EDS Transition? CMS EDS filtering logic differs from MAO s RAPS filtering New submission process causes new errors, rejections, and issues EDS is producing lower risk scores and revenue Limited visibility on EDS based risk scores significant lag in MMR files Delays and revisions to MAO-004 files 15

17 RAPS Vs EDS Scores Milliman study compared RAPS and EDS risk scores Part C risk score difference percentiles (EDS vs RAPS) Plan Type 20 th 40 th 50 th 60 th 80th All plans -7.2% -4.8% -4.0% -3.4% -2.6% SNPs -8.4% -6.6% -5.1% -4.7% -3.1% General enrollment -6.8% -4.2% -3.8% -3.2% -2.5% Methodology notes: i. Payment year 2016 risk scores ii. Phase II MAO-004 files iii. 15 MAOs (small to medium size), 154 plans, 900,000 members iv. Last submission date varied by participating MAO from August to October 2016 v. Consistent submission date for RAPS and EDS vi. 16

18 Final PY2016 Risk Score Settlement PY2016 Diagnosis Submission Deadlines RAPS January 31, 2017 EDS To be determined. Interim May 1, 2017 deadline PY2016 Settlement Interim Settlement October 2017 Based on EDS submissions through May 1, 2017 Final Settlement Date to be determined Based on EDS submissions through final submission deadline 17

19 EDS MAO-004 Files MAO-004 files contain EDS diagnosis information accepted for risk adjustment Phase II version released in October 2016 to address issues with original files Phase III version released in April 2017 to address issues with Phase II files CY15 diagnosis data released in early April 2017 Jan 14 Mar 15 diagnosis data released May 25, 2017 Apr 15 Apr 17 diagnosis data released June 6,

20 What Will an EDS World Look Like? CMS HCC risk score models calibrated on MA data. No more MA coding pattern differences adjustment. Consistent diagnosis code filtering across all MAOs CMS transition from FFS data to MA data for more components of the bid MA county benchmarks Other components? 19

21 Ongoing Litigation 20

22 Making Headlines A Whistle-Blower Tells of Health Insurers Bilking Medicare Is the Tide Turning on Medicare Advantage Fraud Cases? Scheme Tied to UnitedHealth Overbilled Medicare for Years, Suit Says The Risk Adjustment Score Scam A Medicare (Dis) Advantage 21

23 The Details Original complaint by SCAN Health Plan employee (Swoben) in 2009 Alleged that retrospective diagnosis reviews were designed to identify and correct underpayments, but not overpayments Subsequent allegations by Swoben against United Health, HealthCare Partners, Aetna, Wellpoint, and Health Net Based on false claims act allows citizens to sue on behalf of government agencies they believe to have been defrauded United employee (Poehling) turned whistleblower and filed a complaint in Similar allegations to Swoben suit. Against UnitedHealth and 14 other MAOs. Estimates of overpayment in the MA program range from $2 $32 billion 22

24 CMS Guidance 2014 Advance Notice - we are concerned about the high rate of coding of other HCCs by MA organizations, relative to FFS providers, given that the coefficients are calibrated on FFS data Look both ways rule Medical record reviews conducted by an MA organization cannot be designed only to identify diagnoses that would trigger additional payments by CMS to the MA organization; and medical record review methodologies must be designed to identify errors in diagnoses submitted to CMS as risk adjustment data, regardless of whether the data errors would result in positive or negative payment adjustments. CY15 Policy and Technical Changes to the MA and MAPD Benefit Programs, January 10,

25 Developments SCAN settled for $320 million in 2012 In 2016 an appeal court vacated a lower court s decision to throw out Swoben s case DOJ joined the Poehling suit in Feb 2017 DOJ joined the Swoben suit in March 2017 Swoben and Poehling cases are both ongoing DOJ announced it is investigating Aetna, Humana, Health Net, and Cigna s Bravo Health $32 million settlement in DOJ case with Freedom Health in FL in May 2017 Upcoding and network adequacy 24

26 Implications MAOs are reviewing their coding and submission policies and practices Voluntary deletions of non-supported diagnosis codes Possibly more lawsuits to come 25

27 Near Term Considerations 26

28 On the Calendar EDS submissions for PY2016 Mid-year 2017 risk scores RAPS only vs blended RAPS and EDS PY2016 initial risk score settlement in October Accrue for final PY2016 settlement after EDS submission deadline Ongoing assessment of gap between RAPS and EDS based risk scores 27

29 Thank you! Adrian Clark, FSA, MAAA David Koenig, FSA, MAAA 28