Blog & News
Small Area Health Insurance Estimates (SAHIE) Allow First Look at Post-ACA Uninsurance Rates in All U.S. Counties
March 18, 2019:UPDATE: WEBINAR DETAILS
SHADAC will host a coffee-break webinar on June 9, 2016, featuring experts from the U.S. Census Bureau, who will discuss findings from the 2014 SAHIE data along with the recent Medicaid data changes used in SAHIE production.
The webinar will take place from 12:00 - 12:30 a.m. ET / 11:00 - 11:30 CT.
REGISTER HERE.
May 16, 2016:
Although All States Saw Declines in Overall Uninsurance, Changes in Uninsurance Vary at the Sub-State Level
The U.S. Census Bureau released the 2014 Small Area Health Insurance Estimates (SAHIE) on May 12, 2016. The Census Bureau updated the SAHIE models for 2014, incorporating more current Medicaid data sources in order to better capture Medicaid expansion under the Affordable Care Act (ACA). The Census Bureau also re-released 2013 estimates using this updated methodology to facilitate comparisons between 2013 and 2014.
The current SAHIE release provides researchers and policy analysts with the first opportunity to compare uninsurance rates for all U.S counties before and after full implementation of the coverage provisions of the ACA.
Highlights from the 2014 Data
State-level estimates from the 2014 ACS show decreases in the uninsured rates for all 50 states and the District of Columbia, but the new SAHIE results indicate that coverage changes vary at the sub-state level:
- • According to the new SAHIE analysis, 2,325 counties, or 74.1 percent of all counties, saw a decrease in the uninsurance rate.
- • In Medicaid expansion states, 96.1 percent of counties saw a decrease in their estimated uninsured; in non-expansion states, on the other hand, only 36.8 percent of counties saw a decrease in uninsurance.
Table 1 displays the 10 counties with the highest number of nonelderly (0-64) uninsured for 2013 and 2014.
- • Approximately 17% percent of the nation’s 2014 uninsured lived in these 10 counties.
- • All the 2014 top counties saw year-over-year decreases in the estimated number of uninsured.
- • Nine of the top ten 2014 counties were also among the top ten counties in 2013.
- • The top 4 counties with the highest number of uninsured remained the same from 2013 to 2014.
SHADAC's full analysis of the geographic concentration of the uninsured in 2013 and 2014 (updating past work) will be available soon.
Learn More about the New SAHIE Data
2014 SAHIE Data Release Highlights can be viewed here, and Interactive estimates can be viewed here.
SAHIE: BackgroundThe SAHIE program uses statistical modeling to create estimates of health insurance coverage for counties and states. SAHIE data can be used to analyze geographic variation in health insurance coverage as well as coverage disparities by age, sex, race, and income. Why Are the SAHIE Estimates Important? The SAHIE are the only source of single-year health insurance coverage estimates for all U.S. counties. The American Community Survey (ACS) also provides county-level estimates of the uninsured, but only provides estimates for counties with smaller populations (i.e., below 65,000) in the five-year ACS file. (These pooled estimates for 2010 to 2014 are accessible via clickable map on SHADAC’s website.) SAHIE Methodology SAHIE data products are based on statistical modeling and combine data from a number of sources:
SAHIE estimates are more precise than the both the ACS one-year and ACS five-year estimates for most counties because the SAHIE incorporates other data sources in addition to the ACS. Estimates are based on an area-level model that uses survey estimates for domains of interest, rather than individual responses, and the estimates are “enhanced” with administrative data. Additional Resources SHADAC conducted a formal evaluation of the SAHIE experimental estimates and methodology in 2005. The review can be viewed here. SHADAC Issue Brief #26 highlights what was new for the release in 2008 and 2009, provides an overview of how the SAHIE estimates were developed, and compares the SAHIE model based methodology to the ACS. |
||