State-Level Impact of Health Reform: What Can Existing Data Sources Tell Us?

State Health Research and Policy Interest GroupAcademyHealth Annual Research Meeting

June 11, 2011

Julie Sonier, Deputy DirectorState Health Access Data Assistance CenterUniversity of Minnesota

SHADAC is supported by a grant from the Robert Wood Johnson Foundation.

www.shadac.org

Goals

• How can state-level data be used to inform planning, implementation, and evaluation of the ACA?

• Focus on four types of data:– Population surveys– Medicaid administrative data– All-payer claims databases– Health plan data reported to insurance

regulators

2

www.shadac.org

Overview

• Background on data and how they are collected

• Opportunities and challenges for using existing data

• What changes would make the data more useful for policymakers and/or researchers?

• What important gaps need to be filled?

3

www.shadac.org

Panelists

• Sharon Long, Ph.D., University of Minnesota

• Chris Peterson, MPP, Medicaid and CHIP Payment and Access Commission (MACPAC)

• Sean Kolmer, MPH, Office of Oregon Governor John Kitzhaber

• Deborah Chollet, Ph.D., Mathematica Policy Research

4

Using Federal and State Survey Data to Inform State Health Reform

Sharon K. LongUniversity of Minnesota

State Health Research Policy Interest GroupAcademyHealth Annual Research MeetingSeattle, WAJune 11, 2011

www.shadac.org

Key Federal Surveys

• General household survey– ACS: American Community Survey

• Employment/Income surveys– CPS: Current Population Survey– SIPP: Survey of Income and Program Participation

• Health surveys– NHIS: National Health Interview Survey– MEPS-HC: Medical Expenditure Panel Survey-

Household Component– BRFSS: Behavioral Risk Factor Surveillance System

6

www.shadac.org

Feasibility of State-Level Estimates from Federal Surveys

ACS SIPP CPS NHIS MEPS HC

BRFSS

State-level estimates possible?

All states *

Some states ~20 ~20 ~35

State identifiers available on public use files?

Sample size by state

High 353K 9K 20K 13K 5K 20K

Median 44K 1K 3K 1K 400 7K

Low 6K 160 2K 110 <100 2K

7* Two-year averages recommended.

www.shadac.org

Potential Outcomes of Interest

• Health insurance coverage

• Affordability and scope of insurance coverage

• Access to care/barriers to care

• Health care use

• Affordability of care

• Quality of care

8

www.shadac.org

Availability of Outcome Measures in Federal Surveys

ACS CPS SIPP NHIS MEPS HC

BRFSS

Insurance coverage

Point in time

Prior calendar year -

Affordability of coverage/comprehensiveness of coverage

- -

Access to care/barriers to care - - -

Health care use + -

Affordability of care - -

Quality of care

9

www.shadac.org

Timeliness of Data Release for Federal Surveys

• Data files available now:– ACS 2009– SIPP 2008 – CPS 2010 (data for 2009)– NHIS 2010– MEPS-HC 2008– BRFSS 2010

10

www.shadac.org

Strategies to Increase the Value of Federal Surveys for State Studies

• Expand state-level estimates– Increase state sample sizes– Expand use of small area estimation methods

• Expand survey content

• Improve data availability – More timely data release– Make state identifiers available outside of

Research Data Center settings

11

www.shadac.org

Changes in NHIS and BRFSS to Address Needs Under Health Reform• NHIS

– State-level estimates: Added sample size & exploring use of small area estimation methods

– Content:• Added questions• Expanding links with administrative data

– Availability: More timely release of data

• BRFSS – State-level estimates: States can add sample size– Content:

• Core questions include emerging issues• States can add their own questions

12

www.shadac.org

The Bottom Line on Federal Surveys for State-Level Analyses of Health Reform

• No single data source is “best”– CPS, ACS, and BRFSS have accessible 50-

state estimates but are limited in scope– NHIS and MEPS-HC have greater scope but

can’t provide estimates for all states and access to state identifiers is restricted

– MEPS-HC on slow track for data release

• No good alternatives available for some states and for some outcome measures

13

www.shadac.org

State Survey: One Strategy for Addressing Gaps in Federal Surveys

• Larger state sample sizes

• Potential to oversample key subgroups

• Information on state-specific insurance and health care programs

• More outcomes: health care access & use, costs, quality, barriers to care, awareness of reform, attitudes toward reform

• More timely access to data to inform policy and program design

14

www.shadac.org

State Surveys Addressing Health Insurance Coverage and Other Issues

15

Last 5 years

Considering

Long ago/no survey

www.shadac.org

Federal and State surveys will play an important role under health reform, but…

• Sample sizes often can’t support analyses of important subgroups, including Medicaid/CHIP enrollees

• Surveys rely on self-reported information, including type of insurance coverage

• Some outcomes are difficult to capture in surveys—such as costs, diagnoses, quality of care

16

www.shadac.org

References

• SHADAC. 2011. “Monitoring the Impacts of Health Reform at the State Level: Using Federal Survey Data.” Brief #24. Minneapolis, MN: University of Minnesota.

• Sonier, J. and E. Lukanen. Forthcoming. “A Framework for Tracking the Impacts of the Affordable Care Act in California.” Minneapolis, MN: University of Minnesota.

17

Federal Medicaid Federal Medicaid Administrative DataAdministrative Data

for State Usefor State Use

Chris L. PetersonDirector—Eligibility, Enrollment &

Benefits

June 11, 2011Seattle, WA

Presentation Overview

• Brief MACPAC background

• A few key sources of federal Medicaid administrative data and how to use them– Form CMS-64 and Form CMS-21

– MSIS

– Form CMS-416

• Policy issues regarding federal Medicaid/CHIP administrative data

19

MACPAC Statutory ChargeLegislative History

• Established in February 2009 (Children’s Health Insurance Program Reauthorization Act, CHIPRA)

• Expanded and funded in March 2010 (Patient Protection and Affordable Care Act, PPACA)

Commission

• Appointed by the Comptroller General of the United States to 3-year terms

• 17 Commissioners represent broad spectrum of interests and expertise on Medicaid and CHIP

Goals

• Serve as a federal non-partisan and analytic resource on Medicaid and CHIP policy for the Congress

• Review federal and state Medicaid and CHIP policies and data sources

• Regularly consult with states

20

MACPAC Duties• Review Medicaid and CHIP policy issues listed in

statute:

– Payment policies and payment methodologies – Access– Eligibility and enrollment– Quality– Interactions between Medicaid and Medicare – Data for policy analysis and program accountability

• Review and comment on Secretarial reports and regulations that relate to policies under Medicaid and CHIP

• Submit two annual reports to the Congress—March 15 and June 15– March 2011 report Chapter 6 on federal administrative data– June 2011 report on managed care– MACStats in every report

21

MACStats• State-level and national data on Medicaid and CHIP

• March 2011 MACStats state-by-state tables include:

– Medicaid enrollment by eligibility group

– CHIP enrollment by FPL (e.g., 90% of CHIP kids <200% FPL)

– Medicaid and CHIP income eligibility levels

– Medicaid spending by service, federal and state dollars

– Medicaid as a share of states budgets (total and state-funded)

• Preview of MACStats in June 2011 Report

– Trends in Medicaid enrollment and spending

– Health characteristics, enrollment and spending among Medicaid and CHIP populations

– State-by-state info on Medicaid managed care enrollment & spending 22

Key Sources of Federal Medicaid Administrative Data

23

Form CMS-64 & Form-CMS 21• Overview/purpose

– Form used by states to report actual Medicaid and CHIP spending

– Basis of federal payment• Available on CMS website as Excel file thru

FY09• Policy application—Examine each state’s

Medicaid and CHIP aggregate spending– Medicaid services: 40+ categories, $360B– Medicaid administration: 30+ categories, $18B– Medicaid-expansion CHIP: 30+ service and

premium-collection categories, $2.8B– Separate CHIP: 30+ service, premium-

collection, and administration categories: $7.9B– §2105(g) spending in up to 11 states, $0.02B

24

Form CMS-64 and CMS-21, cont.

• Limitations– “As submitted by states”

• States may report spending in different places• A line of spending may not actually include

everything– Only about spending; no enrollment numbers

• Availability and usage tips– Annual summary spreadsheets:https://www.cms.gov/MedicaidBudgetExpendSystem/

Downloads/FY1998throughFY2009MedicaidCHIPTotals.zip

– Detailed Financial Management Report (FMR) spreadsheets from CMS-64 and CMS-21 data:

https://www.cms.gov/MedicaidBudgetExpendSystem/Downloads/NetExpendituresFY2002throughFY2009a.zip

25

Medicaid Statistical Information System (MSIS)

• Overview/purpose of MSIS: Provide person-level eligibility and claims info on every Medicaid enrollee in every state

• Available in a variety of forms, including:– Online state summary “Datamart” (FY)– Annual person summary file (FY)– MAX (cleaned up, by CY, but long lag)

• Policy application—Examine each state’s Medicaid enrollment and spending by individuals’ characteristics (basis of eligibility, BOE)

26

MSIS, cont.• Limitations

– 9% of records have “unknown” BOE—usually prior-year enrollees with a claim paid this year

– MSIS spending does not align with CMS-64• MSIS 13% lower nationally• By state, ranges from 40% lower to 19% higher

– About 25% of spending is for managed care payments, but little is known about enrollees’ encounters

• Availability and usage tips– State summary Datamart available (Internet

Explorer):

http://msis.cms.hhs.gov/ 27

MSIS Datamart Cheat Sheet• For “annual,” click on “quarterly” (FY08=all states;

FY09=80%)• 3rd icon at bottom: Get data later• Take out CHIP enrollees

– Drag CHIP folder onto line under “2009 annual”– Right-click to “insert calculation” to create a new No-CHIP

variable, based on adding: NON-CHIP, CHIP UNKNOWN– Right click to “Hide/show” and hide everything but new No-CHIP

• Break by Basis of Eligibility (BOE)– Drag BOE folder onto line under “Uniq Elgbls Cnt”– Right-click to “insert calculation” to create a new NondisabKids

variable, based on adding: CHILDREN, CHILDREN (UNEMPLOYED PARENT), FOSTER CARE CHILDREN

– Right-click to “insert calculation” to create a new NondisabAdults (under 65) variable, based on adding: ADULTS, UNEMPLOYED ADULTS, BCCA WOMEN

– Right click to “Hide/show” and hide everything used to build new variables (i.e., capped names in last two bullets)

• Click “Get Data”• Once numbers are shown, click penultimate button to

“export to CSV” 28

Form CMS-416• Overview/purpose of Form CMS-416

– State Medicaid programs must provide children <21 with early and periodic screening, diagnostic, and treatment services (EPSDT)

– CMS-416 is the 2-page state-required report on EPSDT, by state, age group, and categorically needy (CN) vs. medically needy (MN)

• Reports available online as pdfs• Policy application: Assess Medicaid

children’s utilization of a handful of services (screenings, whether referred for treatment, dental services, blood lead test)

29

CMS Form-416, cont.• Limitations

– Many concerns about accuracy of data (see GAO and OIG reports)

– However, may be better than MSIS because of broader inclusion of managed care

• Availability and usage tips– https://www.cms.gov/

MedicaidEarlyPeriodicScrn/03_StateAgencyResponsibilities.asp

30

Policy issues regarding Medicaid/CHIP

administrative data

31

Selected Policy Issues on Data

• Competing desires on data quality– If data aren’t being used by policymakers and

researchers, states won’t (can’t) make data quality a priority

– If data quality isn’t a state priority, policymakers and researchers won’t use the data

– Use the data• Competing needs—states and federal

government– Example 1: Voluntary initial core quality measures for

Medicaid/CHIP children could inform policymakers about what works

– Example 2: Encounter data in MSIS could be used to reduce federal reporting redundancies

– Requires state and federal investment32

Selected Data Issues, cont.

• Federal Medicaid and CHIP administrative data only tell you so much about Medicaid and CHIP. Other sources of data:– Survey data

• Surveys of households• Surveys of providers

– States’ own Medicaid and CHIP administrative data, state all-payer claims databases (APCDs)

– All are being used in MACPAC analyses33

34

Health reform in OregonUsing data to drive reform

Sean Kolmer, MPHDeputy Health Policy AdvisorOffice of Governor Kitzhaber

The challenge before us

• Health care has become increasingly unaffordable for businesses large and small, families and individuals and state and local governments, due in large part to inefficiencies in the system that waste dollars and drive people to more expensive care.  

• Public and private dollars are being spent in an inefficient system at a time when there is no public dollar to spare.

• Inefficiencies in care drive people to getting care in the most expensive settings such as hospitalization and institutionalization.

• Between 30-40 percent of health care costs are attributed to inefficiencies and waste.

35

36

The challenge before Oregon

• There are too many uninsured in Oregon: 16.5 percent uninsured, or approximately 614,000 Oregonians, are without health insurance.

• Costs are unsustainable for individuals and families in Oregon:– Average family health insurance premiums in Oregon

increased by 44 percent between 2003 and 2009.– Average premiums accounted for 19 percent of the

median income in Oregon in 2009.– The average deductible for family coverage increased

by 94 percent between 2003 and 2009.

Source: Uninsured: 2008 American Community Survey. Premium data: The Commonwealth Fund, December 2010.

The imperative and the opportunity

• Medicaid is eating up an increasing portion of Oregon’s budget – unnecessary health care costs that are depleting dollars that could be used for education and other important public services. T

• The state can no longer afford to throw dollars into an inefficient system that drives people to the most expensive care and services.

• Today, the state’s fiscal reality gives Oregon an opportunity to remake the Medicaid service and health delivery system to help address the immediate shortfall and reap savings in the years to come.

• If we don’t take action now, we will pay the price later.

37

38

The imperative and the opportunity

Comparing the rate of increase in Medicaid and PEBB health care expenditures vs rate of increase in state General Fund revenue

100

150

200

250

300

350

400

2001-2003 2003-2005 2005-2007 2007-2009 2009-2011 2011-2013(proj)

2013-2015(proj)

2015-2017(proj)

2017-2019(proj)

Perc

ent C

hang

e (In

dex=

100)

Medicaid (TF) PEBB (TF) Statewide General Fund Revenue

Health system transformation principles

• Local coordination and accountability of all benefits, including physical health, mental health and addiction services, oral health and long-term care services.

• Coordination of social supports that promote health and keep people out of high cost medical care.

• Oregonians who are qualified for both Medicaid and Medicare should be included in improvements. Blend funding to create more efficient use of resources, care management, aligned incentives.

• Set global budgets at levels that are sustainable, affordable, and sufficient for best-practices, incentives for prevention efforts, and lower costs.

• Build on best practices in local communities where health entities partner with each other and consumers and are accountable for improved health.

39

Coordination Care Organization Principles

• An organization that serves as a single-point of accountability for the cost of health care within a global budget and for access to and quality of a coordinated system of physical health, behavioral health and oral health care services delivered to the specific population of patients enrolled with the organization.

• Global budget should be configured to hold CCOs accountable for outcomes.

• Responsible for the full integration of physical, behavioral and oral health care services for the Medicaid enrollees, including their members who are dually eligible for Medicaid and Medicare.

• Governance reflect the responsibility for risk, the major components of the health care delivery system, and the community at large.

40

41

Oregon Health Authority spending projected with and without system transformation

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

2009-11 Actual 2011-13 2013-15 2015-17

Biennium

Billi

ons

of $

Gen

eral

Fun

d

With system transformation Without system transformation

We don’t know, what we don’t know

• Oregonians pay for health care without comparable information about cost and quality across the health care system settings.

• Currently, Oregon has fragmented, inconsistent and incomplete information about how our health care system is performing

• An all payer, all claims data database is a tool for better understanding of cost, quality, and utilization across Oregon’s health care system.

What is all payer, all claims (APAC) data?

• Data includes– Claims (medical, pharmacy, and/or dental) – Health plan member demographic information– Health plan provider information– Pharmacy, and/or dental files from:

• Data from– Commercial health insurers– Third party administrators – Public payers (Medicaid, Medicare)– Pharmacy benefit managers

APAC data includes

• Plan payments• Member financial responsibility (co-pay,

coinsurance, deductible) • Patient demographics (date of birth, gender,

geography, race/ethnicity)• Diagnoses• Procedures performed • Type of bill (i.e. inpatient, outpatient,

emergency department)

45

Statutory framework

• Oregon Law Chapter 595, Section 1200-1202

• OHA has authority to:– Require collection

• Commercial insurers (all lines of business)• Third party administrators• Medicaid managed care organizations• Pharmacy benefit managers

– Write administrative rules– Penalties for non-compliance

46

Statutory framework of uses

• Determining the maximum capacity and distribution of existing resources allocated to health care.

• Identifying the demands for health care.• Allowing health care policymakers to make

informed choices.• Evaluating the effectiveness of intervention

programs in improving health outcomes.• Comparing the costs and effectiveness of

various treatment settings and approaches.

47

Statutory framework of uses

• Providing information to consumers and purchasers of health care.

• Improving the quality and affordability of health care and health care coverage.

• Assisting the administrator in furthering the health policies expressed by the Legislative Assembly in ORS 442.025.

• Evaluating health disparities, including but not limited to disparities related to race and ethnicity.

APAC data value

• APAC data provides information necessary to describe health care utilization, patterns of care, disease prevalence, and cost of care.

• Bringing in all payers eliminates biases created when assessing only Medicare, commercial or Medicaid in isolation—much greater precision.

• Community level assessment and comparisons– Comparison of the current world– Comparison of a future world (i.e. accountable care

organizations)

49

Variation in payments by geography

50

State B average rate = $4,064

State A average rate = $4,614If State A avg Pmt = State B avg pmt:Potential Payment savings = $1.5M

Variation in asthma by payers

10% 10%

9%

7%8%

9%

11%

13%

15%

17% 17%

16%

17%

5%

7% 7%

6%

6%

4% 4%

5%5% 5% 5% 5% 5%

0%

2%

4%

6%

8%

10%

12%

14%

16%

18%

All Ages 0-4 5-9 10-14 15-18 19-20 21-24 25-34 35-44 45-49 50-54 55-59 60-64

Medicaid-only CHIS Commercial

Essential element for reform

• Metrics are critical to accountability• Accountability for health outcomes requires

robust, dynamic information systems• System integration across public and private

payers requires across payer information– Active purchasing (850,000 OHA lives)

52

53

Health reform in Oregon

Sean Kolmer, MPHDeputy Health Policy AdvisorOffice of Governor Kitzhabersean.p.kolmer@state.or.us

503-798-2208

What Do States Need to Know aboutTheir Health Insurance Markets?

Deborah Chollet

AcademyHealth State Health Research and Policy

Interest Group Meeting

June 11, 2011

Minimum medical loss ratios, rate oversight

Health insurance exchanges for individuals

SHOP exchanges for small groups– Group size redefined from 2-50 to 2-100

Qualified health plans– Minimum essential benefits– Quality improvement– Tiered products in the Exchange

Navigators + brokers/agents

State risk adjustment - individuals and small groups

ACA is business not as usual

55

Annual statement data

– Company-wide financial/solvency data

– “State page” data, aggregated to line of business

(individual, all group, large and small)

• Premiums (earned, written)

• Losses (paid, incurred)

• Enrollees (lives, member months)

In ~15 states, supplemental filing of small- group data

What do states know now?

56

Rate filing

– Highly variable among states

– Some, not all companies: nonprofits/HMOs only

– Some, not all lines: individual, not group

Product filing

– Original + amendments, but not updated original

What else do states know now?

57

What does the ACA small group market (2-50 and 51-100) look like?

– Number of lives; number of groups

– Demographics

– Disease characteristics

– Burden of illness

– Products offered and taken (minimum essential coverage)

– Quality improvement activities (activities to improve health outcomes, prevent hospital readmissions, improve patient safety, reduce, medical errors, wellness and health promotion, HIT)

What does the association market look like?

What do self-insured small (stop loss) groups look like?

What do states need to know to manage their markets?

58

Market dynamics– Individual exit to/from Medicaid and the exchange

– Small group exit to/from self-insured status

– Small group/individual exit to/from insured and self-insured association plans

– Navigators/licensure

Risk adjustment– Distribution of lives and burden of illness among products and

carriers, individuals and small groups

Market efficiency– Product variation inside/outside the exchange/SHOP– Meaningful choice: fewer or more products?

What do states need to know to evaluate impacts?

59

National Association of Insurance Commissioners http://www.naic.org/index_health_reform_section.htm

The Center for Consumer Information and Insurance Oversight www.cciio.cms.gov

States’ insurance department web sites

More information

60

Sign up to receive our newsletter and updates at www.shadac.org

@shadac

www.facebook.com/shadac4states

State Health Access Data Assistance Center University of Minnesota, Minneapolis, MN

612-624-4802