Monitoring Health Access using

the American Community Survey

Kathleen Thiede Call

Workshop on the Benefits (and Burdens) of the

American Community Survey

June 14, 2012

Washington, DC

Funded by the Robert Wood Johnson Foundation and Federal and State Agencies

About SHADAC

• We help states collect and analyze data to

inform state health policy decisions relating to

health insurance coverage and access to care.

• Our goal: To help states bridge the gap

between health data and the policy-making

process.

• Based at the University of Minnesota

2

States’ needs for monitoring coverage

• Consistent estimates

• Trends over time

– Monitor impacts of health reform

• Comparisons across states

• Subpopulation analysis

– Race/ethnicity, poverty, age

– Counties/sub-state areas

• Access to microdata

3

Key federal survey data sources

• General household survey

– ACS: American Community Survey

• Employment/Income survey

– CPS: Current Population Survey (ASEC)

• Health surveys

– NHIS: National Health Interview Survey

– MEPS-HC: Medical Expenditure Panel Survey-Household

Component

– BRFSS: Behavioral Risk Factor Surveillance System

4

CPS: the good, the bad and the ugly

5

Good • Historic trends

• State-level estimates

• Several control

variables available

• State-specific public

health insurance

program names

• Timely data release

Bad • Low sample in

smaller states

• 10% of respondents

have entire

supplement imputed Ugly • Concerns about the

coverage questions

ACS to the rescue

• SAMPLE SIZE!!!!!!!!!!!!!!

• Sub-state estimates

• Robust subpopulation analysis

• Representativeness

• Current coverage measured

• WISH LIST:

– Data updates during the year, like NHIS

– Self reported health status

6

Questions we help states answer

• How many uninsured are in my state and

where do they live? What is their

demographic profile?

• How many people in my state will be eligible

for Medicaid under ACA.

• How many kids in each county are eligible for

CHIP or Medicaid but not enrolled?

7

Where should we allocate funds for

community clinics?

8

West Virginia Sub-State Uninsurance Estimates, Age 0-64

Source: 2009 American Community Survey

All income levels ≤200% of poverty

How many Minnesotans will be NOT be

eligible for Medicaid under ACA?

Eligible, 89% Eligible, 92%

Not Eligible, 11% Not Eligible, 8%

0%

20%

40%

60%

80%

100%

United States Minnesota

462,000

9

38,000 4.0 Million

32.7 Million

How many uninsured kids in Colorado are

eligible for CHIP but not enrolled?

10 Source: Colorado Health Institute Analysis of 2008-2010 American Community Surveys

What percent of poor kids are uninsured?

11

Uninsurance Rates, Age 0-18, ≤200% FPG

Source: 2010 American Community Survey

Unmet needs with American FactFinder

• Does not provide all policy-relevant measures

– Health insurance unit rather than Census family

– Federal poverty guidelines (HHS) rather than

thresholds

– FPG cuts at ACA policy levels (138%, 200%)

• Not user-friendly

12

SHADAC’s technical assistance for states

• SHADAC’s Data Center

– Online table and chart generator of policy-relevant tables

of health insurance coverage estimates from the ACS and

CPS.

• 12 years of rigorous investigation

• Technical assistance to help states use data

• Education and capacity building

• FPG at 138% and 200%

• Health Insurance Unit

– Family relationships assigned according to health plan

eligibility

13

SHADAC’s Data Center

14

SHADAC’s Health Insurance Unit (HIU)

Married couple

with two children

15

Married

Kids

Grandparents

Child Spouse

Grandchild

• Census family has 1 unit

• HIU family has 1 unit

• Census family has 1 unit

• HIU family has 2 units

Grandparents, married

child, and grandchild

Conclusions

• The ACS fills a gap in state-level data to inform

policy decisions

– Statewide and locally

– Subpopulations; small minority, age or income groups

• The ACS is widely accessible

– Don’t need to be a data programmer to access this

information

• The ACS is a great tool for modeling/linking with

other data to develop state-level estimates

• Benchmarking and sub-population analysis will be

compromised if ACS becomes voluntary

16

Sign up to receive our

newsletter and updates at

www.shadac.org

@shadac