The Medicare Master Beneficiary Summary File THE OLD AND NEW - PowerPoint PPT Presentation

The Medicare Master Beneficiary Summary File THE OLD AND NEW INFORMATION IT CONTAINS Marshall McBean, M.D., M.Sc. Principle Investigator and Director, ResDAC Professor, University of Minnesota

Beneficiary-Level Part D Information  Enrollment or “denominator” information – this presentation  Utilization or numerator information – later presentations ˗ Prescription Drug Event (PDE) information ˗ “Characteristics” files 2

Acronyms in this Presentation  BSF – Beneficiary Summary File  CCW – Chronic Condition Warehouse  EDB – Enrollment Database  RTI – Research Triangle Institute  CAHPS - Consumer Assessment of Healthcare Providers and System  PDE – Prescription Drug Event  QMB – Qualified Medicare Beneficiary 3

Acronyms in this Presentation  SLMB – Specified Low-income Medicare Beneficiary  QI – Qualifying Individuals  QDWI – Qualified Disabled and Working Individuals  PACE – Program of All-Inclusive Care for the Elderly  LIS – Low Income Subsidy  PDP – Prescription Drug Plan  MA-PD – Medicare Advantage Prescription Drug Plan 4

History of Enrollment or “Denominator” Information for use in Part D studies  Until recently there was a single Denominator File created originally to work with Part A and Part B data; thin (80 columns), and we suggested that everyone get a copy of it for every study.  Now different ‘denominator files’ may be mentioned: ˗ The ‘old’ Denominator File is no longer available ˗ The Part D Denominator – short lived ˗ Beneficiary Summary File (BSF) – the next one ˗ Master Beneficiary Summary File (MBSF) -- this is it! (for now) 5

Master Beneficiary Summary File Segments  There are 4 segments to the Master Beneficiary Summary File ˗ Beneficiary Summary File or Medicare Enrollment (A/B/C/D information) ˗ Chronic Conditions ˗ Cost & Utilization ˗ NDI Death Information (includes ICD-10 Cause of Death) ; through 2008.  This presentation is about the Beneficiary Summary File (BSF) segment; sometimes shown as Part A/B or Part A/B/D 6

Beneficiary Summary File (BSF) Segment of the MBSF  Similar to the traditional Medicare Denominator File ˗ 100% of all Medicare beneficiaries enrolled in that year ˗ Annual, calendar year file ˗ one record per beneficiary  Information is what is known by CMS as of 12/31/20XX  Traditional denominator information available by end of summer of 20XX+1 – in what is now called the Beneficiary Summary File (BSF) segment of the MBSF  “Part D denominator” information added in January or February/March 20XX+2 , also, to the BSF segment of the MBSF  So, you could have received the BSF segment for one year in 1 or 2 pieces depending on when you request data. May not be true in 2013 and forward. 7

Traditional Demographic and Enrollment Information in BSF Segment of MBSF  Unique beneficiary identification number  Demographics  Medicare Part A, Part B, and Medicare Advantage Plan enrollment information by month and annually  Information about death ˗ Died during that year (yes/no) ˗ Date the person died  Whether the beneficiary is in the 5% CCW random sample or the enhanced 5% CCW random sample – Note: Part D data, unless otherwise specified in the Data Use Agreement (DUA) will be on the 5% samples 8

Identification Number and Demographic Information  A beneficiary identification number: encrypted BENE_ID is unique to each Data Use Agreement (DUA); created by CCW/Buccaneer  Date of birth and Age as of 12/31/20XX  Gender  Race/ethnicity – a one column field  Place of residence – state, county and zip code 9

Agreement Between Demographic Information in BSF and in Claims Files  All demographic information in BSF and claims files comes from the Enrollment Database (EDB) maintained at CMS Data Center  As claims are processed, the demographic information known to CMS at that time overwrites any demographic information in the claim  Therefore, there is agreement between BSF and claims unless change in status during the year – essentially talking about change in residency  PDE File gender and date of birth variables from a different source; therefore, recommend linking PDE and BSF for all demographic information. 10

Residency in BSF  State, county and ZIP code of residence are the mailing address for official correspondence  Some persons have their mail sent to another person (e.g., son, daughter, guardian)  Because residency in BSF is based on the information available at the end of the calendar year, there may be some mismatch with residence on a claim from earlier in year.  No residency variables in PDE or other Part D data files. 11

Traditional Enrollment Information A variable for each month and one indicting the total number of months enrolled in that year  Enrolled in Medicare Part A  Enrolled in Medicare Part B  “State buy - in” – did a state pay Medicare for beneficiary’s Part B coverage in traditional Medicaid Program or in a Medicare Savings Program (QMB, SLMB, QI)  Enrolled in Medicare Advantage (Part C); a.k.a. managed care, HMO 12

Medicare Savings Programs http://www.medicare.gov/publications/pubs/pdf/10126.pdf  http://www.medicareinteractive.org/page2.php?t opic=counselor&page=script&slide_id=390 and link to the individual state requirements  New York state (see handout in this tab) - http://www.health.ny.gov/health_care/medicaid/program /update/savingsprogram/medicaresavingsprogram.htm  QMB – Qualified Medicare Beneficiary  SLIMB – Specified Low-income Medicare Beneficiary  QI – Qualifying Individuals  QDWI – Qualified Disabled and Working Individuals 13

State Buy-in Variable Values  0 = NOT ENTITLED  1 = PART A ONLY  2 = PART B ONLY  3 = PART A AND PART B  A = PART A, STATE BUY-IN  B = PART B, STATE BUY-IN  C = PARTS A AND B, STATE BUY-IN 14

State Buy-In Indicator Values, June 2008 5% Sample, all ages State Buy-In Frequency Percent 0 = Not entitled (Why?) 126,068 4.82 1 = Part A only 164,463 6.28 2 = Part B only 3,826 0.15 3 = Part A and Part B 1,913,210 73.07 A = Part A, State Buy-in 58 0.00 B = Part B, State Buy-in 13,603 0.52 C = Parts A & B, State Buy-in 396,942 15.16 15

How ResDAC Recommends You Use the State Buy-In Variable  1. Use it to select those with both Part A and Part B coverage ˗ State Buy- in Variable values of “3” and “C” ˗ Do “all” of your studies on those who have both Part A and Part B coverage. ˗ Why? – see next slide » Those with Part A only have lower rates of utilization (of hospitalization, e.g. » Not many beneficiaries with only Part B, and probably have different health utilization patterns  2. To identify those with less resources (will discuss in Segments C and D of this workshop) 16

Hospitalization rates by type of Medicare coverage (per 100 enrollees) supports the conclusion that persons with A-only coverage probably have incomplete claims, even for Part A services 35 30 25 20 A+B 15 A only 10 5 0 65-74 75-84 85+ 17

Part D-related Denominator and Enrollment Information in the BSF Segment of the MBSF  Types of information ˗ RTI race variable (RTI = Research Triangle Institute) ˗ Additional enrollment information ˗ Information for linking with the Part D Characteristics Files  Enrollment information is presented in 2 different ways ˗ For each month of the year; therefore, 12 variables ˗ Number of months (0-12), 1 variable 18

Race – A One Column Variable Hispanic ethnicity not asked nor coded separately  Originally, race coded as: ˗ white, black, other, unknown  Effective 1994, race codes were expanded to: ˗ white, black, Asian, Hispanic, Native American, other, unknown – the “old” race variable.  New “RTI Race Code” variable is available in the Beneficiary Summary File segment of the Master Beneficiary Summary File (The old race code information is available, also.) Was in the Part D sub-segment, but now in the Part A/B sub- segment. 19

Traditional Medicare Race Variable Accuracy and agreement between race variable in the Enrollment Database (EDB) and “SELFRACE”, self -reported race, from Medicare CAHPS Surveys of 2000 - 02 Positive RACE/ Sensitivity Specificity predictive Kappa Ethnicity value White 99.3 61.7 91.7 0.71 Black 97.4 98.8 86.3 0.91 Hispanic 29.5 99.9 92.7 0.43 Asian/P.I. 54.7 99.8 84.5 0.66 Am. Indian/ Ak.Native 35.7 99.9 59.9 0.45 Other/Unk. 1.7 98.8 4.9 0.01 20

Information Used in Algorithm Improving the Identification of Hispanics and Asian/Pacific Islanders  Language preference for materials requested from CMS and Social Security Administration  Source of the race/ethnicity code (survey, Indian Health Service, other)  State of residence (PR or HI)  Hispanic and Asian/Pacific Islander surname lists from U.S. Census Bureau 21

Accuracy and Agreement Measures for RTI Race and EDB Race Compared with SELFRACE in the CAHPS Surveys Positive Race/ Ethnicity Sensitivity Specificity predictive Kappa value Hispanic 79.2 29.5 99.7 99.9 81.5 92.7 0.80 0.43 Asian/PI 76.6 54.7 99.2 99.8 84.5 84.5 0.79 0.66 22