On September 16, the Census Bureau and the National Center for Health Statistics (NCHS) will release the first federal statistics on health insurance coverage during the period after the implementation of the Affordable Care Act (ACA). The agencies held a technical meeting on August 18 to go over what will be released later this month and provide background on the methods used in the respective household surveys. Census Bureau Director John Thompson and NCHS Director Charles Rothwell introduced the meeting. A webcast is available here.
Two sources of data will be used together to assess the impact of the ACA on health insurance coverage. The Census’ Annual Social and Economic Supplement (ASEC) to the Current Population Survey (CPS) asks participants about their health insurance status in the previous calendar year, providing a baseline estimate of health insurance coverage before the ACA’s coverage expansion went into effect. NCHS’ National Health Interview Survey (NHIS) collects data on participants’ current health insurance coverage. The upcoming data release will cover the first quarter of 2014 (coverage expansion began in January 2014).
Jennifer Cheeseman Day, Census Bureau, explained how Census collects insurance coverage data. There are three main sources within Census for this type of data: the CPS ASEC (this is what will be released on September 16), the American Community Survey, and the Survey of Income and Program Participation. Day described the years’ worth of research and testing that were undertaken to ensure that the insurance coverage questions produce the most accurate estimates of coverage possible. Stephen J. Blumberg, NCHS, discussed the evolution of the NHIS questions on insurance coverage since collection of this data began in 1959. Since 1997, NIHS (which is an in-person survey conducted in participants’ homes) has collected data on the names of participants’ insurance plans and, when possible, verified this information with participants’ insurance cards. For 2014, NHIS is also collecting information on if coverage was obtained through the federal health insurance marketplace or state-based exchanges.
Details on the Releases
Victoria Velkoff, Census Bureau, said that the Census’ estimates of uninsured coverage will be based on a sample of 68,000 households, rather than the full 98,000 sample, because the Bureau is testing redesigned questions on income in the remaining 30,000 households. Only responses to the standard questions will be included in the 2013 Income & Poverty and Health Insurance reports to ensure comparability. The report, Health Insurance in the United States: 2013, will include the same cross-sectional characteristics as the 2012 report (family status, race and Hispanic origin, age, nativity, region, metropolitan residence, work experience, and disability status). Trend estimates from 2008-2013 will be drawn from the American Community Survey, which will release its own one-year estimates of health insurance coverage on September 18.
Jennifer Madans, NCHS, explained that the September release of NHIS data is part of the NHIS Early Release Program, which produces quarterly reports on 15 key health indicators. The report will include information about persons’ uninsured status and types of insurance (none, private, or public), sorted by factors including age, poverty status, race/ethnicity, sex, education, employment status, and marital status. The report will also include estimates of the population with high-deductible health plans and, for the first time, estimates of the population with exchange-based coverage.
Perspectives on Producing Reliable Coverage Statistics
Two discussants gave their perspectives on producing reliable estimates of insurance coverage and on some of the challenges the agencies may face going forward. Michael O’Grady, Principal at O’Grady Health Policy, LLC and former Assistant Secretary for Planning and Evaluation (ASPE), shared his experience trying to produce good data on the uninsured during the George W. Bush administration. Different surveys across the federal government produced widely divergent estimates of the uninsured population, so ASPE convened meetings of experts from the various agencies and also looked to outside organizations to provide an objective perspective on the best way to determine the right number. O’Grady observed that while arriving at an accurate estimate of the uninsured population is critically important to making good policy, those engaged in the work should be prepared to face the criticism of “defining away the problem.”
Gary Claxton, Kaiser Family Foundation, described some key issues to be aware of after the data is released. He predicted that the September release will give a good sense of the transition taking place in the U.S. health insurance landscape, but may not give the full picture right away. Because there was a late surge in enrollment (which was extended through the first months of 2014), the surveys may not “catch” those who signed up late or whose coverage was not yet effective when the survey was conducted. Until now, observers have been relying on privately conducted surveys, which give an idea of the insurance rate, but do not go to the lengths the NHIS does to verify the type of coverage people have. But, Claxton suggested, even the NHIS may have trouble accurately categorizing all insurance plans. The different subsidies and tax breaks offered by the federal and state governments may confuse peoples’ understanding of whether they have a public or private health plan, and some insurance companies offer identical plans on and off the state exchanges, further complicating the issue.