ICRC Extended Study Hall Call Series: An Update on Using Medicare Data to Integrate Care for Medicare-Medicaid Enrollees

Transcription

1 ICRC Extended Study Hall Call Series: An Update on Using Medicare Data to Integrate Care for Medicare-Medicaid Enrollees December 3, 2012 For audio, dial: ; Passcode The Integrated Care Resource Center, a joint initiative of the Centers for Medicare & Medicaid Services Medicare-Medicaid Coordination Office and the Center for Medicaid and CHIP Services, provides technical assistance for states coordinated by Mathematica Policy Research and the Center for Health Care Strategies.

2 Agenda Welcome and Roll Call MMCO/CMS Update New York State Update Washington State Update Questions from States 2

3 MMCO/CMS Update Wendy Alexander, Program Alignment Group, MMCO, CMS 3

4 4 New York State s Experience in Accessing and Using Medicare Data Patrick J. Roohan, Director Office of Quality and Patient Safety New York State Department of Health December 3, 2012

5 5 New York s Data Use Agreement Data is being used for program design and analysis. New York State does not re-release Part A and B claims data for care coordination and other quality improvement activities. New York was recently granted approval by CMS to share with an outside Vendor for purposes of data analysis (that work has not yet begun).

6 6 Program Design and Analysis New York Medicaid Redesign Fully Integrated Duals Advantage Program New York State Data Request for the Full Benefit Dual Enrollee Population Actuarial analysis on membership and expenditures. Fully Integrated Dual Advantage (FIDA) Health Homes for Community Well

7 7 New York s Dual Eligible Data Files Medicare Data File Years Identifier Crosswalk Buccaneer BENE ID to HIC Identifier Crosswalk Buccaneer BENE ID to SSN Master Beneficiary Summary Files (Base, Chronic Conditions, Cost and Utilization) 2011 Beneficiary Annual Summary Files (Demographics/Enrollment) Part A Inpatient Part A Outpatient Part A SNF Part A Hospice Part A Home Health Part B Carrier Part B DME Part D Drug Data COBA Claim Files (Part A, Part B and NCPDP) 2007 forward 2011 forward

8 8 Difficult Issues in Using Data: Beneficiary Identification Not all Beneficiary IDs present on Parts A/B claims data (BENE_CLM_ACNT_NUM) are listed on the Crosswalk File (HIC). A Medicare HIC when matched to Medicaid enrollment and eligibility information, may be a one to many, or many to many match (NY 1-2%). Medicare HICs may be more susceptible to change over time than Medicaid identification numbers (NY ~30%).

9 9 Difficult Issues in Using Data: Claim Level Linkages Linkages with Medicaid payment data is critical to understand the relationship between payers. Without claims linkages in place, utilization patterns cannot be fully analyzed. Understanding payment dynamics on an episode of care. For example, the relationship of the Medicaid/Medicare dynamic on home health care coverage; Complex interplay for nursing facility stays and inpatient admissions: both short term and long term.

10 10 Difficult Issues in Using Data: Mapping / Interpreting Data Elements Creating comparable categories of service across Medicaid and Medicare. State s should prepare their own Metadata for interpreting the data elements received. ResDAC has a Medicare Data Documentation link: The Annual Summary Files received from Buccaneer come with documentation, code reference sets and tips on getting started with the data.

11 11 Difficult Issues in Using Data: Format Conversions COBA Claim Data Need to be Translated Transmitted in an X12 837I (Institutional) and 837P (Professional) data format. New York is still pending translation to usable data formats. Needed to procure software vendor.

12 12 Difficult Issues in Using Data: Expenditures: Part C and Part D Part C Expenditure Data is Not Available To estimate, NY used a CY 2011 weighted benchmark PMPM on the Kaiser Family Foundation statehealthfacts.org; applied a 2% reduction for 2010 calculations; and then multiplied by MA member months (e.g., $ % = $ for CY 2010; $ * 2,210,580 MM = $2.1 billion). Part D: Financial and health plan information are not made available as part of the data feed. To estimate, NY had a 2013 estimated PMPM from actuarial analyses and multiplied by FBDE member months ($ * 7,458,022 MM= $3.6 billion).

13 13 Difficult Issues in Using Data: Devoted Resources A significant investment of program management, information technology and analytical staff is needed for converting, storing interrogating, linking and analyzing Part A and B claim data. Part D data feeds are received monthly. COBA Claim data is sent bi-weekly. Data Use Agreement monitoring.

14 14 Risk Profile New York State uses 3M Clinical Risk Groups (CRGs) to assess the severity of illness of its Medicaid enrollees. The CRG software is currently used for risk-adjusted payment for managed care enrollees, to determine thresholds for utilization review and to stratify the Medicaid population for potential programs including care management and health homes.

15 15 Base Health Status and Severity of Illness (Unique Beneficiaries and Percent of Total Community Based LTC and DD Cohort, CY2010) Severity of Illness Level Base Health Status Grand Total Pct Healthy/Acute 2,888 2,888 2% 100% 100% Minor Condition ,691 1% 51% 23% 16% 11% 100% Single Chronic 7,660 3,112 1, ,906 7% 59% 24% 10% 2% 4% 0% 100% Pairs Chronic 18,824 18,263 19,832 21,043 18,492 2,657 99,111 52% 19% 18% 20% 21% 19% 3% 100% Triples Chronic 2,644 5,270 19,155 9,624 11,209 4,727 52,629 28% 5% 10% 36% 18% 21% 9% 100% Malignancies ,612 4,082 1,833 8,010 4% 1% 5% 20% 51% 23% 100% Catastrophic 218 1,978 2,223 1, ,831 11,095 6% 2% 18% 20% 17% 8% 35% 100% HIV / AIDS ,355 1% 17% 23% 37% 23% 100% Grand Total 2,888 30,250 29,855 44,923 37,985 33,544 11, , % Pct 2% 16% 16% 24% 20% 18% 6% 100% Note: Medicaid, Medicare and Part D claims included in classification.

16 16 Prevalence of Chronic Health Conditions (Top 20): Community Based LTC and DD Cohort Chronic Physical Only, 48% Chronic Behavioral Only <1% Chronic Behavioral & Physical, 52% Percent Episode Disease Condition Unique Beneficiary Total Cohort Hypertension 144, Hyperlipidemia 116, Chronic Joint and Musculoskeletal Diagnoses - Minor 78, Diabetes 71, Osteoarthritis 62, Depression 57, Coronary Atherosclerosis 50, Chronic Gastrointestinal Diagnoses - Minor 42, Angina and Ischemic Heart Disease 41, Congestive Heart Failure 40, Peripheral Vascular Disease 39, Chronic Thyroid Disease 38, Osteoporosis 38, Schizophrenia 38, Chronic Endocrine, Nutritional, Fluid, Electrolyte and Immune Diagnoses - Moderate 35, Alzheimer's Disease and Other Dementias 35, Chronic Stress and Anxiety Diagnoses 33, Mild / Moderate Mental Retardation 33, Asthma 33, Chronic Genitourinary Diagnoses 32,

17 17 Top 20 Co-Occurring Chronic Health Conditions: Community Based Developmentally Disabled Chronic Behavioral Only <1% Chronic Physical Only, 48% n =44,190 Chronic Behavioral & Physical, 52% Percent Episode Disease Condition Unique Beneficiary DD Cohort Hyperlipidemia 17, % Hypertension 16, % Schizophrenia 15, % Chronic Joint and Musculoskeletal Diagnoses - Minor 15, % Epilepsy 15, % Depression 14, % Chronic Thyroid Disease 9, % Chronic Stress and Anxiety Diagnoses 8, % Chronic Gastrointestinal Diagnoses - Minor 6, % Osteoporosis 6, % Chronic Endocrine, Nutritional, Fluid, Electrolyte and Immune Diagnoses - Moderate 5, % Diabetes 5, % Conduct, Impulse Control, and Other Disruptive Behavior Disorders 4, % Depressive and Other Psychoses 4, % Chronic Genitourinary Diagnoses 4, % Asthma 4, % Chronic Mental Health Diagnoses - Minor 3, % Chronic Hearing Loss 3, % Obesity 3, % Extrapyramidal Diagnoses 3, %

18 18 Full-Benefit Dual Eligible Recipients Population Cohort (700,000 approx.)* Population Member Months Medicaid $ Medicare $ Total Total PMPM Institutional NH 1,006,147 $5,695,115,759 $1,835,235,425 $7,530,351,185 $7,484 Community Based LTC 1,639,374 $5,683,607,363 $2,661,299,331 $8,344,906,694 $5,090 OPWDD 517,506 $4,521,383,716 $272,818,618 $4,794,202,335 $9,264 Community Well 4,141,923 $1,104,714,346 $2,919,032,042 $4,023,746,388 $972 Total 7,304,923 $17,004,821,184 $7,688,385,416 $24,693,206,602 $3,380 *Reflects Medicare Part A and B only.

19 Community-Based LTC Cohort 19 Dual Eligible Recipients by Category of Service COS Recipients Medicaid $ Medicare $ Total $ PMPM $ Inpatient 57, ,070,134 1,243,565,704 1,358,635, SNF 13,850 43,925, ,617, ,542, Hospice 2,731 4,274,284 26,146,362 30,420, Non-ER HOPD 74,817 51,354, ,836, ,190, ER (HOPD) 39,515 2,963,721 21,899,290 24,863, FS Clinic 22,867 45,976,937 88,482, ,459, Home Health Care 71,739 1,777,985, ,648,732 1,998,633,859 1,219 Physician/Specialist 126,325 42,298, ,808, ,106, DME 96,675 66,211,854 82,380, ,591, Pharmacy 116,302 56,594, ,594, Capitation 41,425 1,182,771, ,182,771, Personal Care 54,350 1,682,541, ,682,541,484 1,026 Waiver Services 2,645 94,165, ,165, ALP/Adult Day Care 15, ,780, ,780, Case Mgmt. 2,084 8,637, ,637,716 5 Other Services 125, ,057, ,913, ,971, Total 146,287 $ 5,683,607,363 $ 2,661,299,331 $ 8,344,906,694 $ 5,090

20 20 Technical Assistance Buccaneer: Beneficiary ID Crosswalk; interpreting Annual Summary files. Mathematica: Estimating Part C costs; Understanding netting Part D data; Understanding HIC numbers are issued, etc. Acumen: Help with linking questions and understanding payment dynamics.

21 21 Contact Information Patrick Roohan: Mary Beth Conroy: Office of Quality and Patient Safety New York State Department of Health Albany, New York Phone: Fax:

22 Patterns of Hospital Readmissions and Nursing Facility Utilization among Washington State Dual Eligibles David Mancuso, PhD December 3, 2012 DSHS Planning, Performance and Accountability Research and Data Analysis Division DECEMBER

23 Hospitalizations frequently restart Medicare payments for nursing facility stays Discharge Status of Dual Eligibles Admitted to a Hospital from a Medicaid-Paid Nursing Facility Stay Dual Eligible Elders Dual Eligible Disabled Directly discharged to NF paid by Medicaid 34% Directly discharged to NF paid by Medicare 43% Directly discharged to NF paid by Medicaid 45% Directly discharged to NF paid by Medicare 31% Discharged to other setting 23% Discharged to other setting 24% Number of discharges in SFY 2010 = 3,135 Average age = 80.5 Hospital Discharge Status Analysis Timeline Medicaid Paid Nursing Facility Stay HOSPITAL STAY ADMISSION Number of discharges in SFY 2010 = 957 Average age = 54.3 Discharged to... Medicare Paid Nursing Facility Medicaid Paid Nursing Facility Other (home, other facility, death) DSHS Planning, Performance and Accountability Research and Data Analysis Division DECEMBER

24 Dual eligibles admitted to the hospital from nursing facility stays have relatively high rates of subsequent hospital readmissions Readmission within 90 Days of Discharge Dual Eligible Elders Dual Eligible Disabled Admitted to hospital from nursing facility paid by Medicaid 26%... paid by Medicare 38%... paid by Medicaid 47%... paid by Medicare 54% 0% 0 Number of discharges in SFY 2010 = 3,135 Average age = , Number of discharges in SFY 2010 = 957 Average age = Readmission Analysis Timeline Medicare Or Medicaid Paid Nursing Facility HOSPITAL Stay STAY HOSPITAL HOSPITAL ADMISSIO DISCHARG N E DSHS Planning, Performance and Accountability Research and Data Analysis Division DECEMBER Day Readmission Rate Analysis 24

25 Case-mix adjusted measurement of rehospitalization risk appears feasible Facility average readmission rates by average patient PRISM risk score for dual eligible patients, by nursing facility receiving the patient following initial discharge 70% 60% 50% 40% 30% 20% 10% NOTE: Unit of observation is a nursing facility receiving at least 25 dual eligible hospital discharges in SFY 2010 LOWER HHIGHER 90-Day Inpatient Readmission Rate LOWER RISK PRISM Risk Score HIGHER RISK 0% SOURCE: Washington State Department of Social and Health Services, Research and Data Analysis Division, Integrated Client Database, May DSHS Planning, Performance and Accountability Research and Data Analysis Division DECEMBER

26 Policy Implications Many nursing facilities serve relative low acuity patients who may be appropriate to consider for transition to community-based care Identifying clients with high risk of hospital readmission and nursing facilities with persistently high case-mix adjusted hospital readmission rates appears feasible Case-mix adjustment of hospital readmission rates is essential to accurately measure facility performance The performance payment terms of our Duals Demonstration MOU with CMS provides the state with an incentive to invest in strategies to reduce the rate of hospitalization of dual eligibles emanating from nursing facilities DSHS Planning, Performance and Accountability Research and Data Analysis Division DECEMBER

27 Using Risk Models To Identify Patients Most at Risk of Potentially Avoidable Adverse Health Outcomes Medicare $ PMPM... Elders HIGH RISK PRISM Score > 1.5 Other $ 875 Disabled HIGH RISK PRISM Score > 1.5 Other $ 1,258 LOW RISK PRISM Score < 1.5 SNF $ 339 Inpatient $ 809 LOW RISK PRISM Score < 1.5 SNF $ 153 Inpatient $ TOTAL = $ 334 TOTAL = $ 2,023 SFY 2010 Excludes Medicaid Expenditures Low Risk PRISM Score < 1.5 TOTAL = $ 357 TOTAL = $ 2,371 Medicare Costs Per Member Per Month (PMPM) ELDERS DISABLED High Risk Low Risk PRISM Score > 1.5 PRISM Score < 1.5 High Risk PRISM Score > 1.5 Total PMPM $334 $2,023 $357 $2,371 Inpatient PMPM $59 $809 $70 $960 SNF PMPM $27 $339 $7 $153 Covered Lives 46,241 28,703 39,560 20,117 DSHS Planning, Performance and Accountability Research and Data Analysis Division DECEMBER

28 Acknowledgements Chad and Summer cracked mainframe Medicare historical data files and built a 5-year Medicare claims analytical data repository, in addition to integrating Medicare data into PRISM Dan calibrated the PRISM risk model used to illustrate the potential for case-mix adjustment of hospital readmission rates Elizabeth developed the multidimensional day array analytical programming processes that these analyses required Barb designed the layout and presentation of complex information Bev co-authored the paper and made essential contributions to improve the narrative and better capture the policy environment Thanks! DSHS Planning, Performance and Accountability Research and Data Analysis Division DECEMBER

29 Questions? 29

30 About ICRC Established by CMS to advance integrated care models for Medicaid beneficiaries with high costs and high needs Provides technical assistance (TA) to help states integrate care for: (1) individuals who are dually eligible for Medicare and Medicaid; and (2) high-need, high-cost Medicaid populations via health homes as well as other emerging models TA coordinated by Mathematica Policy Research and the Center for Health Care Strategies Visit to submit a TA request and/or download resources, including briefs and practical tools to help address implementation, design, and policy challenges 30