BRIEF 39 • MAY 2014 An Introduction to Redesigned Health Insurance Coverage Questions in the 2014 Current Population Survey’s Annual Social and Economic Supplement Introduction Authors The Current Population Survey (CPS), conducted by the U.S. Census Bureau for the Bureau of Labor Statistics, is designed to provide monthly data on labor force participation and unem- Joanna Turner ployment for the civilian non-institutionalized population. Data on health insurance coverage, Michel Boudreaux income, and non-cash employment benefits are collected once each year through the Annual State Health Access Data Assistance Center (SHADAC) Social and Economic Supplement (ASEC), which is administered February through April. The survey provides annual estimates for the nation and states. The reference period for both the income and health insurance questions is the calendar year prior to the survey. Summary The question series has undergone several changes over the years (Davern et al. 2003; SHA- This brief discusses the addition of new DAC 2009; Ziegenfuss and Davern 2011; Boudreaux and Turner 2011). This brief discusses and modified health insurance coverage questions to the 2014 CPS. The question the most recent change: the addition of new and modified health insurance coverage questions series was completely redesigned to: to the 2014 CPS. The health insurance coverage questions were completely redesigned to 1) • Improve the measure of past year improve the measure of past year coverage, 2) add a point in time measure, 3) collect addi- coverage tional information related to health reform (exchange participation and employer-offers of • Add a point in time measure coverage), and 4) provide coverage by month. This brief provides an overview of the new and • Collect additional information modified content. related to health reform (exchange participation and employer-offers of coverage) New and Modified Health Insurance Coverage Questions • Provide coverage by month For decades, the Census Bureau and external researchers have recognized several flaws in the CPS health insurance questions which cause it to over-estimate the number of uninsured (Davern et al. 2007; Pascal 2008; DeNavas-Walt et al. 2013). The new and modified health insurance coverage questions are designed to reduce recall error, reduce respondent fatigue, to improve comprehension, and to provide additional information about employer offers, reasons why eligible workers do not take-up employer sponsored coverage, purchase of coverage on health insurance exchanges, and the receipt of subsidies (U.S. Department of Commerce and U.S. Census Bureau 2013/2014; U.S. Census Bureau 2014). The modifications are described below. To address the recall error, the respondent is first asked about their current coverage and is then asked when that coverage began and if other coverage was held in any month from the start of the previous calendar year through the date of interview. While extensive experience with the CPS’s previous year reference period suggests considerable difficulty in measuring coverage status held in the past, the health insurance estimates from the CPS must align with the previous year income and non-cash benefit measures. In addition to providing a point-in- time estimate, the new questions attempt to improve recall of past coverage by anchoring the respondent to a specific time point before asking about past year coverage. In the household level design of the traditional CPS, respondents were asked if anyone in the | STATE HEALTH ACCESS DATA ASSISTANCE CENTER 1 household had a certain type of coverage and if so any of its employees, if the person is eligible for that which persons were covered. This lead to a tendency coverage, and if they are eligible, why they did not for coverage of certain household members to be take up coverage. missed, but switching to a person-level design where each person is asked the entire question series increas- Evaluations of the New Health es respondent burden (Hess et al. 2002). To improve accuracy without increasing respondent fatigue, a Insurance Questions hybrid of the household and person level design was The new health insurance questions were developed developed. Questions are first asked at the person-lev- and evaluated over the course of several years of el with follow-up questions to determine if others in pilot-testing that involved qualitative interviews, ran- the household also have an identified coverage type. domized survey experiments, and a “content test” in For the additional household members their coverage which the new questions were embedded into the full is verified, if already identified, and they are asked if version of the CPS and fielded on a test sample using The new question they had any additional coverage. The full question the Census Bureau’s regular interviewing procedures series will provide new series is only asked of additional household members (Pascale 2009; Pascale 2012; Pascale et al. 2013; content that can be if they have not had any coverage identified in the Boudreaux et al. 2013; Brault et al. 2014). used to monitor sources first-pass through the question series. While these evaluations are informative they do not of health insurance and Understanding sources of health insurance coverage perfectly mimic the environment of the regular CPS. changes in coverage can be confusing, so to improve comprehension Test samples have different characteristics than the over the course of a the question series is designed to capture coverage regular CPS and test respondents respond at different calendar year. in a more intuitive way, with a series of shorter and rates than regular survey respondents. Therefore, simpler questions. First, a general coverage question results should be interpreted with caution. is asked before asking more detailed questions about The most recent evaluation of the new questions, source. The questions are designed to drill-down conducted in 2013, suggests that the new question as needed, clarify areas of ambiguity, and re-route series produces a lower rate of all-year uninsured respondents to the correct coverage path to capture than the traditional CPS, as was hoped (Brault et people and households with more complicated cov- al. 2014). This finding appears relatively consistent erage. Most respondents should be able to quickly across race and age groups. The additional cover- identify their coverage and avoid many of the detailed age measured by the new instrument appears to be follow-up questions. The entire question series is concentrated in private coverage. The new question not asked of every respondent. The questions are also series estimates a smaller rate of Medicaid coverage targeted in places to better capture underreported and about equal levels of Medicare coverage. Future plans (e.g., respondents age 65 and older or disabled research that merges survey responses from the new are first asked about Medicare coverage while other question series to administrative records of health respondents are asked the general “do you have cover- insurance coverage is needed to more fully measure age now?” question). the validity of the new instrument. In addition to modifications to the general structure The new question series will provide new content that of the health insurance questions, the revised CPS can be used to monitor sources of health insurance asks new questions about whether coverage was and changes in coverage over the course of a calen- purchased on an exchange or marketplace and if the dar year. However, because of these new changes, premiums are subsidized. The questions use a state estimates from the 2013 traditional CPS (which pro- specific name fill so that respondents are asked about vides coverage for calendar year 2012) and the 2014 the specific exchange that operates in their state. redesigned CPS (which provides coverage for calendar There are also new questions targeted at workers year 2013) will not be comparable. The redesigned that do not have coverage through their jobs. They 2014 CPS does capture information from the year determine if a person’s employer offers coverage to 2 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER An Introduction to Redesigned Health Insurance Coverage Questions in the 2014 CPS prior to full implementation of the Affordable Care Bureau plans to split the annual “Income, Poverty, Act (ACA) which can be contrasted with the 2014 and Health Insurance” report into two products. The calendar year coverage from the 2015 CPS. However, CPS Income and Poverty report will be released first trend analysis that incorporates data for years prior to followed by a Health Insurance report that includes 2013 will not be possible. Other surveys such as the both CPS and ACS estimates to provide a broad- American Community Survey, the National Health er picture of coverage. The Census Bureau is still Interview Survey, and the Medical Expenditure Panel finalizing the details of the data release, but it is likely Survey will provide a consistent time series. that all of the new health insurance content (e.g., month-level estimates) won’t be available until a later September 2014 Data Release date. In addition to implementing the new and modi- fied health insurance coverage questions, new and About SHADAC modified income questions were implemented. In the The State Health Access Data Assistance Center is an 2014 CPS, the entire sample received the new and independent health policy research center located at modified health insurance coverage questions, but the University of Minnesota School of Public Health. the sample was split with a phase-in schedule for the SHADAC is a resource for helping states collect and new and modified income questions. Five-eighths of use data for health policy, with a particular focus on the sample received the traditional income questions monitoring rates of health insurance coverage and and three-eighths of the sample received the new and understanding factors associated with uninsurance. modified income questions. The health insurance For more information, please contact us at coverage questions were not implemented with a shadac@umn.edu, or call 612-624-4802. phase-in schedule so that year-to-year comparisons over 2013 and 2014 can be made using the full sam- Suggested Citation ple. The smaller sample size of the phase-in schedule makes it more difficult to detect differences between Turner, J, Boudreaux, M. 2014. “An Introduction to years. Redesigned Health Insurance Coverage Questions in the 2014 CPS.” SHADAC Brief #39. Minneapolis, For the data release in September 2014, the Census MN: State Health Access Data Assistance Center. REFERENCES Boudreaux, M. and J. Turner. 2011. “Modifications to the Imputation Routine for Health Insurance in the CPS ASEC: Description and Evaluation.” SHADAC Working Paper. Minneapolis, MN: University of Minnesota. Available at: http:// www.shadac.org/files/shadac/publications/CPS_Imputation_Dec2011.pdf. Boudreaux, M., B. Fried, J. Turner, and K.T. Call. 2013. “SHADAC Analysis of the Survey of Health Insurance and Pro- gram Participation.” SHADAC Working Paper. Minneapolis, MN: University of Minnesota. Available at: http://shadac. org/files/shadac/publications/SHIPP_final_report.pdf. Brault, M., C. Medalia, B. O’Hara, J. Rodean, and A. Steinweg. 2014. “Changing the CPS Health Insurance Questions And The Implications On The Uninsured Rate: Redesign and Production Estimates.” U.S. Census Bureau SEHSD Working Paper 2014-16. Available at: http://www.census.gov/hhes/www/hlthins/publications/sehsd_wp_2014-16. pdf. Davern, M., T.J. Beebe, L.A. Blewett, and K.T. Call. 2003. “Recent Changes to the Current Population Survey: Sample Expansion, Health Insurance Verification, and State Health Insurance Coverage Estimates.” Public Opinion Quarterly 67 (4): 603-626. Davern, M., G. Davidson, J. Ziegenfuss, S. Jarosek, B. Lee, T. Yu, T.J. Beebe, K.T. Call, and L.A. Blewett. 2007. “A Comparison of the Health Insurance Coverage Estimates from Four National Surveys and Six State Surveys: A Dis- cussion of Measurement Issues and Policy Implications.” Final report for U.S. DHHS Assistant Secretary for Planning MAY 2014 | www.shadac.org 3 and Evaluation, Task 7.2. Minneapolis, MN: University of Minnesota. Available at: http://www.shadac.org/files/sha- dac/publications/ASPE_FinalRpt_Dec2007_Task7_2_rev.pdf. DeNavas-Walt, C., B.D. Proctor, and J.C. Smith. 2013. “Income, Poverty, and Health Insurance Coverage in the United States: 2012.” U.S. Census Bureau Current Population Reports P60-245. Available at: http://www.census.gov/prod/ 2013pubs/p60-245.pdf. Hess, J., J.C. Moore, J. Pascale, J. Rothgeb, and C. Keeley. 2002. “The Effects of Person-Level VS. Household-Level Questionnaire Design on Survey Estimates and Data Quality.” U.S. Census Bureau Research Report Series Survey Methodology #2002-05. Available at: http://www.census.gov/srd/papers/pdf/rsm2002-05.pdf. Pascale, J. 2008. “Measurement Error in Health Insurance Reporting.” Inquiry: Vol. 45, No. 4, pp. 422-437. Pascale, J. 2009. “Findings from a Pretest of a New Approach to Measuring Health Insurance in the Current Popula- tion Survey.” Proceedings of the Federal Committee on Statistical Methodology Conference, November, 2009. Pascale, J. 2012. “Findings from a Split-Ballot Experiment on a New Approach to Measuring Health Insurance in the Current Population Survey.” American Society of Health Economists Conference, Minneapolis, MN, June 2012. Pascale, J., J. Rodean, J. Leeman, C. Cosenza, and A. Schoua-Glusberg. 2013. “Preparing to Measure Health Cover- age in Federal Surveys Post-Reform: Lessons from Massachusetts.” Inquiry: Vol. 50, No. 2, pp. 106-123. State Health Access Data Assistance Center. 2009. “Historical Changes in Current Population Survey Health Insurance Coverage Items for Survey Years 1988 through 2008.” Issue Brief #19. Minneapolis, MN: University of Minnesota. Available at: http://www.shadac.org/files/shadac/publications/IssueBrief19.pdf. U.S. Census Bureau. 2014. “ASEC Questionnaire: Attachment A.” Available at: http://www.reginfo.gov/public/do/ PRAICList?ref_nbr=201311-0607-001. U.S. Department of Commerce and U.S. Census Bureau. 2013/2014. Several ICR Supporting Documents. Available at: http://www.reginfo.gov/public/do/PRAViewDocument?ref_nbr=201110-0935-001. Ziegenfuss, J. and M. Davern. 2011. “Twenty Years of Coverage: An Enhanced Current Population Sur- vey—1989-2008.” Health Services Research 46(1 Pt 1): 199-209. 4 | STATE HEALTH ACCESS DATA ASSISTANCE CENTER An Introduction to Redesigned Health Insurance Coverage Questions in the 2014 CPS MAY 2014 | www.shadac.org 5 Sign up to receive our newsletter and updates at www.shadac.org Bridging the gap between research and policy @ www.shadac.org Funded by a grant from The Robert Wood Johnson Foundation ©2014 Regents of the University of Minnesota. All rights reserved. The University of Minnesota is an Equal Opportunity Employer. 05012014